diff options
| author | Christian Schneppe <christian@pix-art.de> | 2016-10-16 12:01:48 +0200 |
|---|---|---|
| committer | Christian Schneppe <christian@pix-art.de> | 2016-10-16 12:01:48 +0200 |
| commit | 94a45d8458ac8e02c994a16023ecf026210a2bad (patch) | |
| tree | 30edb2d9ba528414c943cc6f6e561af5f9ae9b96 | |
| parent | 0af5c054c386dcafeedded2c467b22b53afc1d0f (diff) | |
delete unused libs
167 files changed, 0 insertions, 76582 deletions
diff --git a/src/main/jni/Android.mk b/src/main/jni/Android.mk deleted file mode 100644 index 29f9f9c14..000000000 --- a/src/main/jni/Android.mk +++ /dev/null @@ -1,120 +0,0 @@ -LOCAL_PATH := $(call my-dir) - -LOCAL_MODULE := avutil - -ifeq ($(TARGET_ARCH_ABI),armeabi-v7a) - LOCAL_SRC_FILES := ./ffmpeg/armv7-a/libavutil.a -else - ifeq ($(TARGET_ARCH_ABI),armeabi) - LOCAL_SRC_FILES := ./ffmpeg/armv5te/libavutil.a - else - ifeq ($(TARGET_ARCH_ABI),x86) - LOCAL_SRC_FILES := ./ffmpeg/i686/libavutil.a - endif - endif -endif - -include $(PREBUILT_STATIC_LIBRARY) - -include $(CLEAR_VARS) - -LOCAL_MODULE := avformat - -ifeq ($(TARGET_ARCH_ABI),armeabi-v7a) - LOCAL_SRC_FILES := ./ffmpeg/armv7-a/libavformat.a -else - ifeq ($(TARGET_ARCH_ABI),armeabi) - LOCAL_SRC_FILES := ./ffmpeg/armv5te/libavformat.a - else - ifeq ($(TARGET_ARCH_ABI),x86) - LOCAL_SRC_FILES := ./ffmpeg/i686/libavformat.a - endif - endif -endif - -include $(PREBUILT_STATIC_LIBRARY) - -include $(CLEAR_VARS) - -LOCAL_MODULE := avcodec - -ifeq ($(TARGET_ARCH_ABI),armeabi-v7a) - LOCAL_SRC_FILES := ./ffmpeg/armv7-a/libavcodec.a -else - ifeq ($(TARGET_ARCH_ABI),armeabi) - LOCAL_SRC_FILES := ./ffmpeg/armv5te/libavcodec.a - else - ifeq ($(TARGET_ARCH_ABI),x86) - LOCAL_SRC_FILES := ./ffmpeg/i686/libavcodec.a - endif - endif -endif - -include $(PREBUILT_STATIC_LIBRARY) - -include $(CLEAR_VARS) - -LOCAL_CPP_EXTENSION := .cc -LOCAL_ARM_MODE := arm -LOCAL_MODULE := breakpad -LOCAL_CPPFLAGS := -Wall -std=c++11 -DANDROID -finline-functions -ffast-math -Os -fno-strict-aliasing - -include $(CLEAR_VARS) - -LOCAL_SRC_FILES += \ -./libyuv/source/compare_common.cc \ -./libyuv/source/compare_gcc.cc \ -./libyuv/source/compare_neon64.cc \ -./libyuv/source/compare_win.cc \ -./libyuv/source/compare.cc \ -./libyuv/source/convert_argb.cc \ -./libyuv/source/convert_from_argb.cc \ -./libyuv/source/convert_from.cc \ -./libyuv/source/convert_jpeg.cc \ -./libyuv/source/convert_to_argb.cc \ -./libyuv/source/convert_to_i420.cc \ -./libyuv/source/convert.cc \ -./libyuv/source/cpu_id.cc \ -./libyuv/source/mjpeg_decoder.cc \ -./libyuv/source/mjpeg_validate.cc \ -./libyuv/source/planar_functions.cc \ -./libyuv/source/rotate_any.cc \ -./libyuv/source/rotate_argb.cc \ -./libyuv/source/rotate_common.cc \ -./libyuv/source/rotate_gcc.cc \ -./libyuv/source/rotate_mips.cc \ -./libyuv/source/rotate_neon64.cc \ -./libyuv/source/rotate_win.cc \ -./libyuv/source/rotate.cc \ -./libyuv/source/row_any.cc \ -./libyuv/source/row_common.cc \ -./libyuv/source/row_gcc.cc \ -./libyuv/source/row_mips.cc \ -./libyuv/source/row_neon64.cc \ -./libyuv/source/row_win.cc \ -./libyuv/source/scale_any.cc \ -./libyuv/source/scale_argb.cc \ -./libyuv/source/scale_common.cc \ -./libyuv/source/scale_gcc.cc \ -./libyuv/source/scale_mips.cc \ -./libyuv/source/scale_neon64.cc \ -./libyuv/source/scale_win.cc \ -./libyuv/source/scale.cc \ -./libyuv/source/video_common.cc - -ifeq ($(TARGET_ARCH_ABI),armeabi-v7a) - LOCAL_CFLAGS += -DLIBYUV_NEON - LOCAL_SRC_FILES += \ - ./libyuv/source/compare_neon.cc.neon \ - ./libyuv/source/rotate_neon.cc.neon \ - ./libyuv/source/row_neon.cc.neon \ - ./libyuv/source/scale_neon.cc.neon -endif - -LOCAL_SRC_FILES += \ -./jni.c \ -./video.c \ - -include $(BUILD_SHARED_LIBRARY) - -$(call import-module,android/cpufeatures)
\ No newline at end of file diff --git a/src/main/jni/Application.mk b/src/main/jni/Application.mk deleted file mode 100644 index bd546a41a..000000000 --- a/src/main/jni/Application.mk +++ /dev/null @@ -1,4 +0,0 @@ -APP_PLATFORM := android-9 -APP_ABI := armeabi armeabi-v7a x86 -NDK_TOOLCHAIN_VERSION := 4.9 -APP_STL := gnustl_static
\ No newline at end of file diff --git a/src/main/jni/ffmpeg/COPYING.GPLv2 b/src/main/jni/ffmpeg/COPYING.GPLv2 deleted file mode 100644 index d159169d1..000000000 --- a/src/main/jni/ffmpeg/COPYING.GPLv2 +++ /dev/null @@ -1,339 +0,0 @@ - GNU GENERAL PUBLIC LICENSE - Version 2, June 1991 - - Copyright (C) 1989, 1991 Free Software Foundation, Inc., - 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - Everyone is permitted to copy and distribute verbatim copies - of this license document, but changing it is not allowed. - - Preamble - - The licenses for most software are designed to take away your -freedom to share and change it. By contrast, the GNU General Public -License is intended to guarantee your freedom to share and change free -software--to make sure the software is free for all its users. This -General Public License applies to most of the Free Software -Foundation's software and to any other program whose authors commit to -using it. (Some other Free Software Foundation software is covered by -the GNU Lesser General Public License instead.) You can apply it to -your programs, too. - - When we speak of free software, we are referring to freedom, not -price. Our General Public Licenses are designed to make sure that you -have the freedom to distribute copies of free software (and charge for -this service if you wish), that you receive source code or can get it -if you want it, that you can change the software or use pieces of it -in new free programs; and that you know you can do these things. - - To protect your rights, we need to make restrictions that forbid -anyone to deny you these rights or to ask you to surrender the rights. -These restrictions translate to certain responsibilities for you if you -distribute copies of the software, or if you modify it. - - For example, if you distribute copies of such a program, whether -gratis or for a fee, you must give the recipients all the rights that -you have. You must make sure that they, too, receive or can get the -source code. And you must show them these terms so they know their -rights. - - We protect your rights with two steps: (1) copyright the software, and -(2) offer you this license which gives you legal permission to copy, -distribute and/or modify the software. - - Also, for each author's protection and ours, we want to make certain -that everyone understands that there is no warranty for this free -software. If the software is modified by someone else and passed on, we -want its recipients to know that what they have is not the original, so -that any problems introduced by others will not reflect on the original -authors' reputations. - - Finally, any free program is threatened constantly by software -patents. We wish to avoid the danger that redistributors of a free -program will individually obtain patent licenses, in effect making the -program proprietary. To prevent this, we have made it clear that any -patent must be licensed for everyone's free use or not licensed at all. - - The precise terms and conditions for copying, distribution and -modification follow. - - GNU GENERAL PUBLIC LICENSE - TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION - - 0. This License applies to any program or other work which contains -a notice placed by the copyright holder saying it may be distributed -under the terms of this General Public License. The "Program", below, -refers to any such program or work, and a "work based on the Program" -means either the Program or any derivative work under copyright law: -that is to say, a work containing the Program or a portion of it, -either verbatim or with modifications and/or translated into another -language. (Hereinafter, translation is included without limitation in -the term "modification".) Each licensee is addressed as "you". - -Activities other than copying, distribution and modification are not -covered by this License; they are outside its scope. The act of -running the Program is not restricted, and the output from the Program -is covered only if its contents constitute a work based on the -Program (independent of having been made by running the Program). -Whether that is true depends on what the Program does. - - 1. You may copy and distribute verbatim copies of the Program's -source code as you receive it, in any medium, provided that you -conspicuously and appropriately publish on each copy an appropriate -copyright notice and disclaimer of warranty; keep intact all the -notices that refer to this License and to the absence of any warranty; -and give any other recipients of the Program a copy of this License -along with the Program. - -You may charge a fee for the physical act of transferring a copy, and -you may at your option offer warranty protection in exchange for a fee. - - 2. You may modify your copy or copies of the Program or any portion -of it, thus forming a work based on the Program, and copy and -distribute such modifications or work under the terms of Section 1 -above, provided that you also meet all of these conditions: - - a) You must cause the modified files to carry prominent notices - stating that you changed the files and the date of any change. - - b) You must cause any work that you distribute or publish, that in - whole or in part contains or is derived from the Program or any - part thereof, to be licensed as a whole at no charge to all third - parties under the terms of this License. - - c) If the modified program normally reads commands interactively - when run, you must cause it, when started running for such - interactive use in the most ordinary way, to print or display an - announcement including an appropriate copyright notice and a - notice that there is no warranty (or else, saying that you provide - a warranty) and that users may redistribute the program under - these conditions, and telling the user how to view a copy of this - License. (Exception: if the Program itself is interactive but - does not normally print such an announcement, your work based on - the Program is not required to print an announcement.) - -These requirements apply to the modified work as a whole. If -identifiable sections of that work are not derived from the Program, -and can be reasonably considered independent and separate works in -themselves, then this License, and its terms, do not apply to those -sections when you distribute them as separate works. But when you -distribute the same sections as part of a whole which is a work based -on the Program, the distribution of the whole must be on the terms of -this License, whose permissions for other licensees extend to the -entire whole, and thus to each and every part regardless of who wrote it. - -Thus, it is not the intent of this section to claim rights or contest -your rights to work written entirely by you; rather, the intent is to -exercise the right to control the distribution of derivative or -collective works based on the Program. - -In addition, mere aggregation of another work not based on the Program -with the Program (or with a work based on the Program) on a volume of -a storage or distribution medium does not bring the other work under -the scope of this License. - - 3. You may copy and distribute the Program (or a work based on it, -under Section 2) in object code or executable form under the terms of -Sections 1 and 2 above provided that you also do one of the following: - - a) Accompany it with the complete corresponding machine-readable - source code, which must be distributed under the terms of Sections - 1 and 2 above on a medium customarily used for software interchange; or, - - b) Accompany it with a written offer, valid for at least three - years, to give any third party, for a charge no more than your - cost of physically performing source distribution, a complete - machine-readable copy of the corresponding source code, to be - distributed under the terms of Sections 1 and 2 above on a medium - customarily used for software interchange; or, - - c) Accompany it with the information you received as to the offer - to distribute corresponding source code. (This alternative is - allowed only for noncommercial distribution and only if you - received the program in object code or executable form with such - an offer, in accord with Subsection b above.) - -The source code for a work means the preferred form of the work for -making modifications to it. For an executable work, complete source -code means all the source code for all modules it contains, plus any -associated interface definition files, plus the scripts used to -control compilation and installation of the executable. However, as a -special exception, the source code distributed need not include -anything that is normally distributed (in either source or binary -form) with the major components (compiler, kernel, and so on) of the -operating system on which the executable runs, unless that component -itself accompanies the executable. - -If distribution of executable or object code is made by offering -access to copy from a designated place, then offering equivalent -access to copy the source code from the same place counts as -distribution of the source code, even though third parties are not -compelled to copy the source along with the object code. - - 4. You may not copy, modify, sublicense, or distribute the Program -except as expressly provided under this License. Any attempt -otherwise to copy, modify, sublicense or distribute the Program is -void, and will automatically terminate your rights under this License. -However, parties who have received copies, or rights, from you under -this License will not have their licenses terminated so long as such -parties remain in full compliance. - - 5. You are not required to accept this License, since you have not -signed it. However, nothing else grants you permission to modify or -distribute the Program or its derivative works. These actions are -prohibited by law if you do not accept this License. Therefore, by -modifying or distributing the Program (or any work based on the -Program), you indicate your acceptance of this License to do so, and -all its terms and conditions for copying, distributing or modifying -the Program or works based on it. - - 6. Each time you redistribute the Program (or any work based on the -Program), the recipient automatically receives a license from the -original licensor to copy, distribute or modify the Program subject to -these terms and conditions. You may not impose any further -restrictions on the recipients' exercise of the rights granted herein. -You are not responsible for enforcing compliance by third parties to -this License. - - 7. If, as a consequence of a court judgment or allegation of patent -infringement or for any other reason (not limited to patent issues), -conditions are imposed on you (whether by court order, agreement or -otherwise) that contradict the conditions of this License, they do not -excuse you from the conditions of this License. If you cannot -distribute so as to satisfy simultaneously your obligations under this -License and any other pertinent obligations, then as a consequence you -may not distribute the Program at all. For example, if a patent -license would not permit royalty-free redistribution of the Program by -all those who receive copies directly or indirectly through you, then -the only way you could satisfy both it and this License would be to -refrain entirely from distribution of the Program. - -If any portion of this section is held invalid or unenforceable under -any particular circumstance, the balance of the section is intended to -apply and the section as a whole is intended to apply in other -circumstances. - -It is not the purpose of this section to induce you to infringe any -patents or other property right claims or to contest validity of any -such claims; this section has the sole purpose of protecting the -integrity of the free software distribution system, which is -implemented by public license practices. Many people have made -generous contributions to the wide range of software distributed -through that system in reliance on consistent application of that -system; it is up to the author/donor to decide if he or she is willing -to distribute software through any other system and a licensee cannot -impose that choice. - -This section is intended to make thoroughly clear what is believed to -be a consequence of the rest of this License. - - 8. If the distribution and/or use of the Program is restricted in -certain countries either by patents or by copyrighted interfaces, the -original copyright holder who places the Program under this License -may add an explicit geographical distribution limitation excluding -those countries, so that distribution is permitted only in or among -countries not thus excluded. In such case, this License incorporates -the limitation as if written in the body of this License. - - 9. The Free Software Foundation may publish revised and/or new versions -of the General Public License from time to time. Such new versions will -be similar in spirit to the present version, but may differ in detail to -address new problems or concerns. - -Each version is given a distinguishing version number. If the Program -specifies a version number of this License which applies to it and "any -later version", you have the option of following the terms and conditions -either of that version or of any later version published by the Free -Software Foundation. If the Program does not specify a version number of -this License, you may choose any version ever published by the Free Software -Foundation. - - 10. If you wish to incorporate parts of the Program into other free -programs whose distribution conditions are different, write to the author -to ask for permission. For software which is copyrighted by the Free -Software Foundation, write to the Free Software Foundation; we sometimes -make exceptions for this. Our decision will be guided by the two goals -of preserving the free status of all derivatives of our free software and -of promoting the sharing and reuse of software generally. - - NO WARRANTY - - 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY -FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN -OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES -PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED -OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF -MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS -TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE -PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, -REPAIR OR CORRECTION. - - 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING -WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR -REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, -INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING -OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED -TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY -YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER -PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE -POSSIBILITY OF SUCH DAMAGES. - - END OF TERMS AND CONDITIONS - - How to Apply These Terms to Your New Programs - - If you develop a new program, and you want it to be of the greatest -possible use to the public, the best way to achieve this is to make it -free software which everyone can redistribute and change under these terms. - - To do so, attach the following notices to the program. It is safest -to attach them to the start of each source file to most effectively -convey the exclusion of warranty; and each file should have at least -the "copyright" line and a pointer to where the full notice is found. - - <one line to give the program's name and a brief idea of what it does.> - Copyright (C) <year> <name of author> - - This program is free software; you can redistribute it and/or modify - it under the terms of the GNU General Public License as published by - the Free Software Foundation; either version 2 of the License, or - (at your option) any later version. - - This program is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. - - You should have received a copy of the GNU General Public License along - with this program; if not, write to the Free Software Foundation, Inc., - 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. - -Also add information on how to contact you by electronic and paper mail. - -If the program is interactive, make it output a short notice like this -when it starts in an interactive mode: - - Gnomovision version 69, Copyright (C) year name of author - Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. - This is free software, and you are welcome to redistribute it - under certain conditions; type `show c' for details. - -The hypothetical commands `show w' and `show c' should show the appropriate -parts of the General Public License. Of course, the commands you use may -be called something other than `show w' and `show c'; they could even be -mouse-clicks or menu items--whatever suits your program. - -You should also get your employer (if you work as a programmer) or your -school, if any, to sign a "copyright disclaimer" for the program, if -necessary. Here is a sample; alter the names: - - Yoyodyne, Inc., hereby disclaims all copyright interest in the program - `Gnomovision' (which makes passes at compilers) written by James Hacker. - - <signature of Ty Coon>, 1 April 1989 - Ty Coon, President of Vice - -This General Public License does not permit incorporating your program into -proprietary programs. If your program is a subroutine library, you may -consider it more useful to permit linking proprietary applications with the -library. If this is what you want to do, use the GNU Lesser General -Public License instead of this License. diff --git a/src/main/jni/ffmpeg/COPYING.GPLv3 b/src/main/jni/ffmpeg/COPYING.GPLv3 deleted file mode 100644 index 94a9ed024..000000000 --- a/src/main/jni/ffmpeg/COPYING.GPLv3 +++ /dev/null @@ -1,674 +0,0 @@ - GNU GENERAL PUBLIC LICENSE - Version 3, 29 June 2007 - - Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/> - Everyone is permitted to copy and distribute verbatim copies - of this license document, but changing it is not allowed. - - Preamble - - The GNU General Public License is a free, copyleft license for -software and other kinds of works. - - The licenses for most software and other practical works are designed -to take away your freedom to share and change the works. By contrast, -the GNU General Public License is intended to guarantee your freedom to -share and change all versions of a program--to make sure it remains free -software for all its users. We, the Free Software Foundation, use the -GNU General Public License for most of our software; it applies also to -any other work released this way by its authors. You can apply it to -your programs, too. - - When we speak of free software, we are referring to freedom, not -price. Our General Public Licenses are designed to make sure that you -have the freedom to distribute copies of free software (and charge for -them if you wish), that you receive source code or can get it if you -want it, that you can change the software or use pieces of it in new -free programs, and that you know you can do these things. - - To protect your rights, we need to prevent others from denying you -these rights or asking you to surrender the rights. Therefore, you have -certain responsibilities if you distribute copies of the software, or if -you modify it: responsibilities to respect the freedom of others. - - For example, if you distribute copies of such a program, whether -gratis or for a fee, you must pass on to the recipients the same -freedoms that you received. You must make sure that they, too, receive -or can get the source code. And you must show them these terms so they -know their rights. - - Developers that use the GNU GPL protect your rights with two steps: -(1) assert copyright on the software, and (2) offer you this License -giving you legal permission to copy, distribute and/or modify it. - - For the developers' and authors' protection, the GPL clearly explains -that there is no warranty for this free software. For both users' and -authors' sake, the GPL requires that modified versions be marked as -changed, so that their problems will not be attributed erroneously to -authors of previous versions. - - Some devices are designed to deny users access to install or run -modified versions of the software inside them, although the manufacturer -can do so. This is fundamentally incompatible with the aim of -protecting users' freedom to change the software. The systematic -pattern of such abuse occurs in the area of products for individuals to -use, which is precisely where it is most unacceptable. Therefore, we -have designed this version of the GPL to prohibit the practice for those -products. If such problems arise substantially in other domains, we -stand ready to extend this provision to those domains in future versions -of the GPL, as needed to protect the freedom of users. - - Finally, every program is threatened constantly by software patents. -States should not allow patents to restrict development and use of -software on general-purpose computers, but in those that do, we wish to -avoid the special danger that patents applied to a free program could -make it effectively proprietary. To prevent this, the GPL assures that -patents cannot be used to render the program non-free. - - The precise terms and conditions for copying, distribution and -modification follow. - - TERMS AND CONDITIONS - - 0. Definitions. - - "This License" refers to version 3 of the GNU General Public License. - - "Copyright" also means copyright-like laws that apply to other kinds of -works, such as semiconductor masks. - - "The Program" refers to any copyrightable work licensed under this -License. Each licensee is addressed as "you". "Licensees" and -"recipients" may be individuals or organizations. - - To "modify" a work means to copy from or adapt all or part of the work -in a fashion requiring copyright permission, other than the making of an -exact copy. The resulting work is called a "modified version" of the -earlier work or a work "based on" the earlier work. - - A "covered work" means either the unmodified Program or a work based -on the Program. - - To "propagate" a work means to do anything with it that, without -permission, would make you directly or secondarily liable for -infringement under applicable copyright law, except executing it on a -computer or modifying a private copy. Propagation includes copying, -distribution (with or without modification), making available to the -public, and in some countries other activities as well. - - To "convey" a work means any kind of propagation that enables other -parties to make or receive copies. Mere interaction with a user through -a computer network, with no transfer of a copy, is not conveying. - - An interactive user interface displays "Appropriate Legal Notices" -to the extent that it includes a convenient and prominently visible -feature that (1) displays an appropriate copyright notice, and (2) -tells the user that there is no warranty for the work (except to the -extent that warranties are provided), that licensees may convey the -work under this License, and how to view a copy of this License. If -the interface presents a list of user commands or options, such as a -menu, a prominent item in the list meets this criterion. - - 1. Source Code. - - The "source code" for a work means the preferred form of the work -for making modifications to it. "Object code" means any non-source -form of a work. - - A "Standard Interface" means an interface that either is an official -standard defined by a recognized standards body, or, in the case of -interfaces specified for a particular programming language, one that -is widely used among developers working in that language. - - The "System Libraries" of an executable work include anything, other -than the work as a whole, that (a) is included in the normal form of -packaging a Major Component, but which is not part of that Major -Component, and (b) serves only to enable use of the work with that -Major Component, or to implement a Standard Interface for which an -implementation is available to the public in source code form. A -"Major Component", in this context, means a major essential component -(kernel, window system, and so on) of the specific operating system -(if any) on which the executable work runs, or a compiler used to -produce the work, or an object code interpreter used to run it. - - The "Corresponding Source" for a work in object code form means all -the source code needed to generate, install, and (for an executable -work) run the object code and to modify the work, including scripts to -control those activities. However, it does not include the work's -System Libraries, or general-purpose tools or generally available free -programs which are used unmodified in performing those activities but -which are not part of the work. For example, Corresponding Source -includes interface definition files associated with source files for -the work, and the source code for shared libraries and dynamically -linked subprograms that the work is specifically designed to require, -such as by intimate data communication or control flow between those -subprograms and other parts of the work. - - The Corresponding Source need not include anything that users -can regenerate automatically from other parts of the Corresponding -Source. - - The Corresponding Source for a work in source code form is that -same work. - - 2. Basic Permissions. - - All rights granted under this License are granted for the term of -copyright on the Program, and are irrevocable provided the stated -conditions are met. This License explicitly affirms your unlimited -permission to run the unmodified Program. The output from running a -covered work is covered by this License only if the output, given its -content, constitutes a covered work. This License acknowledges your -rights of fair use or other equivalent, as provided by copyright law. - - You may make, run and propagate covered works that you do not -convey, without conditions so long as your license otherwise remains -in force. You may convey covered works to others for the sole purpose -of having them make modifications exclusively for you, or provide you -with facilities for running those works, provided that you comply with -the terms of this License in conveying all material for which you do -not control copyright. Those thus making or running the covered works -for you must do so exclusively on your behalf, under your direction -and control, on terms that prohibit them from making any copies of -your copyrighted material outside their relationship with you. - - Conveying under any other circumstances is permitted solely under -the conditions stated below. Sublicensing is not allowed; section 10 -makes it unnecessary. - - 3. Protecting Users' Legal Rights From Anti-Circumvention Law. - - No covered work shall be deemed part of an effective technological -measure under any applicable law fulfilling obligations under article -11 of the WIPO copyright treaty adopted on 20 December 1996, or -similar laws prohibiting or restricting circumvention of such -measures. - - When you convey a covered work, you waive any legal power to forbid -circumvention of technological measures to the extent such circumvention -is effected by exercising rights under this License with respect to -the covered work, and you disclaim any intention to limit operation or -modification of the work as a means of enforcing, against the work's -users, your or third parties' legal rights to forbid circumvention of -technological measures. - - 4. Conveying Verbatim Copies. - - You may convey verbatim copies of the Program's source code as you -receive it, in any medium, provided that you conspicuously and -appropriately publish on each copy an appropriate copyright notice; -keep intact all notices stating that this License and any -non-permissive terms added in accord with section 7 apply to the code; -keep intact all notices of the absence of any warranty; and give all -recipients a copy of this License along with the Program. - - You may charge any price or no price for each copy that you convey, -and you may offer support or warranty protection for a fee. - - 5. Conveying Modified Source Versions. - - You may convey a work based on the Program, or the modifications to -produce it from the Program, in the form of source code under the -terms of section 4, provided that you also meet all of these conditions: - - a) The work must carry prominent notices stating that you modified - it, and giving a relevant date. - - b) The work must carry prominent notices stating that it is - released under this License and any conditions added under section - 7. This requirement modifies the requirement in section 4 to - "keep intact all notices". - - c) You must license the entire work, as a whole, under this - License to anyone who comes into possession of a copy. This - License will therefore apply, along with any applicable section 7 - additional terms, to the whole of the work, and all its parts, - regardless of how they are packaged. This License gives no - permission to license the work in any other way, but it does not - invalidate such permission if you have separately received it. - - d) If the work has interactive user interfaces, each must display - Appropriate Legal Notices; however, if the Program has interactive - interfaces that do not display Appropriate Legal Notices, your - work need not make them do so. - - A compilation of a covered work with other separate and independent -works, which are not by their nature extensions of the covered work, -and which are not combined with it such as to form a larger program, -in or on a volume of a storage or distribution medium, is called an -"aggregate" if the compilation and its resulting copyright are not -used to limit the access or legal rights of the compilation's users -beyond what the individual works permit. Inclusion of a covered work -in an aggregate does not cause this License to apply to the other -parts of the aggregate. - - 6. Conveying Non-Source Forms. - - You may convey a covered work in object code form under the terms -of sections 4 and 5, provided that you also convey the -machine-readable Corresponding Source under the terms of this License, -in one of these ways: - - a) Convey the object code in, or embodied in, a physical product - (including a physical distribution medium), accompanied by the - Corresponding Source fixed on a durable physical medium - customarily used for software interchange. - - b) Convey the object code in, or embodied in, a physical product - (including a physical distribution medium), accompanied by a - written offer, valid for at least three years and valid for as - long as you offer spare parts or customer support for that product - model, to give anyone who possesses the object code either (1) a - copy of the Corresponding Source for all the software in the - product that is covered by this License, on a durable physical - medium customarily used for software interchange, for a price no - more than your reasonable cost of physically performing this - conveying of source, or (2) access to copy the - Corresponding Source from a network server at no charge. - - c) Convey individual copies of the object code with a copy of the - written offer to provide the Corresponding Source. This - alternative is allowed only occasionally and noncommercially, and - only if you received the object code with such an offer, in accord - with subsection 6b. - - d) Convey the object code by offering access from a designated - place (gratis or for a charge), and offer equivalent access to the - Corresponding Source in the same way through the same place at no - further charge. You need not require recipients to copy the - Corresponding Source along with the object code. If the place to - copy the object code is a network server, the Corresponding Source - may be on a different server (operated by you or a third party) - that supports equivalent copying facilities, provided you maintain - clear directions next to the object code saying where to find the - Corresponding Source. Regardless of what server hosts the - Corresponding Source, you remain obligated to ensure that it is - available for as long as needed to satisfy these requirements. - - e) Convey the object code using peer-to-peer transmission, provided - you inform other peers where the object code and Corresponding - Source of the work are being offered to the general public at no - charge under subsection 6d. - - A separable portion of the object code, whose source code is excluded -from the Corresponding Source as a System Library, need not be -included in conveying the object code work. - - A "User Product" is either (1) a "consumer product", which means any -tangible personal property which is normally used for personal, family, -or household purposes, or (2) anything designed or sold for incorporation -into a dwelling. In determining whether a product is a consumer product, -doubtful cases shall be resolved in favor of coverage. For a particular -product received by a particular user, "normally used" refers to a -typical or common use of that class of product, regardless of the status -of the particular user or of the way in which the particular user -actually uses, or expects or is expected to use, the product. A product -is a consumer product regardless of whether the product has substantial -commercial, industrial or non-consumer uses, unless such uses represent -the only significant mode of use of the product. - - "Installation Information" for a User Product means any methods, -procedures, authorization keys, or other information required to install -and execute modified versions of a covered work in that User Product from -a modified version of its Corresponding Source. The information must -suffice to ensure that the continued functioning of the modified object -code is in no case prevented or interfered with solely because -modification has been made. - - If you convey an object code work under this section in, or with, or -specifically for use in, a User Product, and the conveying occurs as -part of a transaction in which the right of possession and use of the -User Product is transferred to the recipient in perpetuity or for a -fixed term (regardless of how the transaction is characterized), the -Corresponding Source conveyed under this section must be accompanied -by the Installation Information. But this requirement does not apply -if neither you nor any third party retains the ability to install -modified object code on the User Product (for example, the work has -been installed in ROM). - - The requirement to provide Installation Information does not include a -requirement to continue to provide support service, warranty, or updates -for a work that has been modified or installed by the recipient, or for -the User Product in which it has been modified or installed. Access to a -network may be denied when the modification itself materially and -adversely affects the operation of the network or violates the rules and -protocols for communication across the network. - - Corresponding Source conveyed, and Installation Information provided, -in accord with this section must be in a format that is publicly -documented (and with an implementation available to the public in -source code form), and must require no special password or key for -unpacking, reading or copying. - - 7. Additional Terms. - - "Additional permissions" are terms that supplement the terms of this -License by making exceptions from one or more of its conditions. -Additional permissions that are applicable to the entire Program shall -be treated as though they were included in this License, to the extent -that they are valid under applicable law. If additional permissions -apply only to part of the Program, that part may be used separately -under those permissions, but the entire Program remains governed by -this License without regard to the additional permissions. - - When you convey a copy of a covered work, you may at your option -remove any additional permissions from that copy, or from any part of -it. (Additional permissions may be written to require their own -removal in certain cases when you modify the work.) You may place -additional permissions on material, added by you to a covered work, -for which you have or can give appropriate copyright permission. - - Notwithstanding any other provision of this License, for material you -add to a covered work, you may (if authorized by the copyright holders of -that material) supplement the terms of this License with terms: - - a) Disclaiming warranty or limiting liability differently from the - terms of sections 15 and 16 of this License; or - - b) Requiring preservation of specified reasonable legal notices or - author attributions in that material or in the Appropriate Legal - Notices displayed by works containing it; or - - c) Prohibiting misrepresentation of the origin of that material, or - requiring that modified versions of such material be marked in - reasonable ways as different from the original version; or - - d) Limiting the use for publicity purposes of names of licensors or - authors of the material; or - - e) Declining to grant rights under trademark law for use of some - trade names, trademarks, or service marks; or - - f) Requiring indemnification of licensors and authors of that - material by anyone who conveys the material (or modified versions of - it) with contractual assumptions of liability to the recipient, for - any liability that these contractual assumptions directly impose on - those licensors and authors. - - All other non-permissive additional terms are considered "further -restrictions" within the meaning of section 10. If the Program as you -received it, or any part of it, contains a notice stating that it is -governed by this License along with a term that is a further -restriction, you may remove that term. If a license document contains -a further restriction but permits relicensing or conveying under this -License, you may add to a covered work material governed by the terms -of that license document, provided that the further restriction does -not survive such relicensing or conveying. - - If you add terms to a covered work in accord with this section, you -must place, in the relevant source files, a statement of the -additional terms that apply to those files, or a notice indicating -where to find the applicable terms. - - Additional terms, permissive or non-permissive, may be stated in the -form of a separately written license, or stated as exceptions; -the above requirements apply either way. - - 8. Termination. - - You may not propagate or modify a covered work except as expressly -provided under this License. Any attempt otherwise to propagate or -modify it is void, and will automatically terminate your rights under -this License (including any patent licenses granted under the third -paragraph of section 11). - - However, if you cease all violation of this License, then your -license from a particular copyright holder is reinstated (a) -provisionally, unless and until the copyright holder explicitly and -finally terminates your license, and (b) permanently, if the copyright -holder fails to notify you of the violation by some reasonable means -prior to 60 days after the cessation. - - Moreover, your license from a particular copyright holder is -reinstated permanently if the copyright holder notifies you of the -violation by some reasonable means, this is the first time you have -received notice of violation of this License (for any work) from that -copyright holder, and you cure the violation prior to 30 days after -your receipt of the notice. - - Termination of your rights under this section does not terminate the -licenses of parties who have received copies or rights from you under -this License. If your rights have been terminated and not permanently -reinstated, you do not qualify to receive new licenses for the same -material under section 10. - - 9. Acceptance Not Required for Having Copies. - - You are not required to accept this License in order to receive or -run a copy of the Program. Ancillary propagation of a covered work -occurring solely as a consequence of using peer-to-peer transmission -to receive a copy likewise does not require acceptance. However, -nothing other than this License grants you permission to propagate or -modify any covered work. These actions infringe copyright if you do -not accept this License. Therefore, by modifying or propagating a -covered work, you indicate your acceptance of this License to do so. - - 10. Automatic Licensing of Downstream Recipients. - - Each time you convey a covered work, the recipient automatically -receives a license from the original licensors, to run, modify and -propagate that work, subject to this License. You are not responsible -for enforcing compliance by third parties with this License. - - An "entity transaction" is a transaction transferring control of an -organization, or substantially all assets of one, or subdividing an -organization, or merging organizations. If propagation of a covered -work results from an entity transaction, each party to that -transaction who receives a copy of the work also receives whatever -licenses to the work the party's predecessor in interest had or could -give under the previous paragraph, plus a right to possession of the -Corresponding Source of the work from the predecessor in interest, if -the predecessor has it or can get it with reasonable efforts. - - You may not impose any further restrictions on the exercise of the -rights granted or affirmed under this License. For example, you may -not impose a license fee, royalty, or other charge for exercise of -rights granted under this License, and you may not initiate litigation -(including a cross-claim or counterclaim in a lawsuit) alleging that -any patent claim is infringed by making, using, selling, offering for -sale, or importing the Program or any portion of it. - - 11. Patents. - - A "contributor" is a copyright holder who authorizes use under this -License of the Program or a work on which the Program is based. The -work thus licensed is called the contributor's "contributor version". - - A contributor's "essential patent claims" are all patent claims -owned or controlled by the contributor, whether already acquired or -hereafter acquired, that would be infringed by some manner, permitted -by this License, of making, using, or selling its contributor version, -but do not include claims that would be infringed only as a -consequence of further modification of the contributor version. For -purposes of this definition, "control" includes the right to grant -patent sublicenses in a manner consistent with the requirements of -this License. - - Each contributor grants you a non-exclusive, worldwide, royalty-free -patent license under the contributor's essential patent claims, to -make, use, sell, offer for sale, import and otherwise run, modify and -propagate the contents of its contributor version. - - In the following three paragraphs, a "patent license" is any express -agreement or commitment, however denominated, not to enforce a patent -(such as an express permission to practice a patent or covenant not to -sue for patent infringement). To "grant" such a patent license to a -party means to make such an agreement or commitment not to enforce a -patent against the party. - - If you convey a covered work, knowingly relying on a patent license, -and the Corresponding Source of the work is not available for anyone -to copy, free of charge and under the terms of this License, through a -publicly available network server or other readily accessible means, -then you must either (1) cause the Corresponding Source to be so -available, or (2) arrange to deprive yourself of the benefit of the -patent license for this particular work, or (3) arrange, in a manner -consistent with the requirements of this License, to extend the patent -license to downstream recipients. "Knowingly relying" means you have -actual knowledge that, but for the patent license, your conveying the -covered work in a country, or your recipient's use of the covered work -in a country, would infringe one or more identifiable patents in that -country that you have reason to believe are valid. - - If, pursuant to or in connection with a single transaction or -arrangement, you convey, or propagate by procuring conveyance of, a -covered work, and grant a patent license to some of the parties -receiving the covered work authorizing them to use, propagate, modify -or convey a specific copy of the covered work, then the patent license -you grant is automatically extended to all recipients of the covered -work and works based on it. - - A patent license is "discriminatory" if it does not include within -the scope of its coverage, prohibits the exercise of, or is -conditioned on the non-exercise of one or more of the rights that are -specifically granted under this License. You may not convey a covered -work if you are a party to an arrangement with a third party that is -in the business of distributing software, under which you make payment -to the third party based on the extent of your activity of conveying -the work, and under which the third party grants, to any of the -parties who would receive the covered work from you, a discriminatory -patent license (a) in connection with copies of the covered work -conveyed by you (or copies made from those copies), or (b) primarily -for and in connection with specific products or compilations that -contain the covered work, unless you entered into that arrangement, -or that patent license was granted, prior to 28 March 2007. - - Nothing in this License shall be construed as excluding or limiting -any implied license or other defenses to infringement that may -otherwise be available to you under applicable patent law. - - 12. No Surrender of Others' Freedom. - - If conditions are imposed on you (whether by court order, agreement or -otherwise) that contradict the conditions of this License, they do not -excuse you from the conditions of this License. If you cannot convey a -covered work so as to satisfy simultaneously your obligations under this -License and any other pertinent obligations, then as a consequence you may -not convey it at all. For example, if you agree to terms that obligate you -to collect a royalty for further conveying from those to whom you convey -the Program, the only way you could satisfy both those terms and this -License would be to refrain entirely from conveying the Program. - - 13. Use with the GNU Affero General Public License. - - Notwithstanding any other provision of this License, you have -permission to link or combine any covered work with a work licensed -under version 3 of the GNU Affero General Public License into a single -combined work, and to convey the resulting work. The terms of this -License will continue to apply to the part which is the covered work, -but the special requirements of the GNU Affero General Public License, -section 13, concerning interaction through a network will apply to the -combination as such. - - 14. Revised Versions of this License. - - The Free Software Foundation may publish revised and/or new versions of -the GNU General Public License from time to time. Such new versions will -be similar in spirit to the present version, but may differ in detail to -address new problems or concerns. - - Each version is given a distinguishing version number. If the -Program specifies that a certain numbered version of the GNU General -Public License "or any later version" applies to it, you have the -option of following the terms and conditions either of that numbered -version or of any later version published by the Free Software -Foundation. If the Program does not specify a version number of the -GNU General Public License, you may choose any version ever published -by the Free Software Foundation. - - If the Program specifies that a proxy can decide which future -versions of the GNU General Public License can be used, that proxy's -public statement of acceptance of a version permanently authorizes you -to choose that version for the Program. - - Later license versions may give you additional or different -permissions. However, no additional obligations are imposed on any -author or copyright holder as a result of your choosing to follow a -later version. - - 15. Disclaimer of Warranty. - - THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY -APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT -HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY -OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, -THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR -PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM -IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF -ALL NECESSARY SERVICING, REPAIR OR CORRECTION. - - 16. Limitation of Liability. - - IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING -WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS -THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY -GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE -USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF -DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD -PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), -EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF -SUCH DAMAGES. - - 17. Interpretation of Sections 15 and 16. - - If the disclaimer of warranty and limitation of liability provided -above cannot be given local legal effect according to their terms, -reviewing courts shall apply local law that most closely approximates -an absolute waiver of all civil liability in connection with the -Program, unless a warranty or assumption of liability accompanies a -copy of the Program in return for a fee. - - END OF TERMS AND CONDITIONS - - How to Apply These Terms to Your New Programs - - If you develop a new program, and you want it to be of the greatest -possible use to the public, the best way to achieve this is to make it -free software which everyone can redistribute and change under these terms. - - To do so, attach the following notices to the program. It is safest -to attach them to the start of each source file to most effectively -state the exclusion of warranty; and each file should have at least -the "copyright" line and a pointer to where the full notice is found. - - <one line to give the program's name and a brief idea of what it does.> - Copyright (C) <year> <name of author> - - This program is free software: you can redistribute it and/or modify - it under the terms of the GNU General Public License as published by - the Free Software Foundation, either version 3 of the License, or - (at your option) any later version. - - This program is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. - - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. - -Also add information on how to contact you by electronic and paper mail. - - If the program does terminal interaction, make it output a short -notice like this when it starts in an interactive mode: - - <program> Copyright (C) <year> <name of author> - This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. - This is free software, and you are welcome to redistribute it - under certain conditions; type `show c' for details. - -The hypothetical commands `show w' and `show c' should show the appropriate -parts of the General Public License. Of course, your program's commands -might be different; for a GUI interface, you would use an "about box". - - You should also get your employer (if you work as a programmer) or school, -if any, to sign a "copyright disclaimer" for the program, if necessary. -For more information on this, and how to apply and follow the GNU GPL, see -<http://www.gnu.org/licenses/>. - - The GNU General Public License does not permit incorporating your program -into proprietary programs. If your program is a subroutine library, you -may consider it more useful to permit linking proprietary applications with -the library. If this is what you want to do, use the GNU Lesser General -Public License instead of this License. But first, please read -<http://www.gnu.org/philosophy/why-not-lgpl.html>. diff --git a/src/main/jni/ffmpeg/COPYING.LGPLv2.1 b/src/main/jni/ffmpeg/COPYING.LGPLv2.1 deleted file mode 100644 index 58af0d378..000000000 --- a/src/main/jni/ffmpeg/COPYING.LGPLv2.1 +++ /dev/null @@ -1,502 +0,0 @@ - GNU LESSER GENERAL PUBLIC LICENSE - Version 2.1, February 1999 - - Copyright (C) 1991, 1999 Free Software Foundation, Inc. - 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - Everyone is permitted to copy and distribute verbatim copies - of this license document, but changing it is not allowed. - -[This is the first released version of the Lesser GPL. It also counts - as the successor of the GNU Library Public License, version 2, hence - the version number 2.1.] - - Preamble - - The licenses for most software are designed to take away your -freedom to share and change it. By contrast, the GNU General Public -Licenses are intended to guarantee your freedom to share and change -free software--to make sure the software is free for all its users. - - This license, the Lesser General Public License, applies to some -specially designated software packages--typically libraries--of the -Free Software Foundation and other authors who decide to use it. You -can use it too, but we suggest you first think carefully about whether -this license or the ordinary General Public License is the better -strategy to use in any particular case, based on the explanations below. - - When we speak of free software, we are referring to freedom of use, -not price. Our General Public Licenses are designed to make sure that -you have the freedom to distribute copies of free software (and charge -for this service if you wish); that you receive source code or can get -it if you want it; that you can change the software and use pieces of -it in new free programs; and that you are informed that you can do -these things. - - To protect your rights, we need to make restrictions that forbid -distributors to deny you these rights or to ask you to surrender these -rights. These restrictions translate to certain responsibilities for -you if you distribute copies of the library or if you modify it. - - For example, if you distribute copies of the library, whether gratis -or for a fee, you must give the recipients all the rights that we gave -you. You must make sure that they, too, receive or can get the source -code. If you link other code with the library, you must provide -complete object files to the recipients, so that they can relink them -with the library after making changes to the library and recompiling -it. And you must show them these terms so they know their rights. - - We protect your rights with a two-step method: (1) we copyright the -library, and (2) we offer you this license, which gives you legal -permission to copy, distribute and/or modify the library. - - To protect each distributor, we want to make it very clear that -there is no warranty for the free library. Also, if the library is -modified by someone else and passed on, the recipients should know -that what they have is not the original version, so that the original -author's reputation will not be affected by problems that might be -introduced by others. - - Finally, software patents pose a constant threat to the existence of -any free program. We wish to make sure that a company cannot -effectively restrict the users of a free program by obtaining a -restrictive license from a patent holder. Therefore, we insist that -any patent license obtained for a version of the library must be -consistent with the full freedom of use specified in this license. - - Most GNU software, including some libraries, is covered by the -ordinary GNU General Public License. This license, the GNU Lesser -General Public License, applies to certain designated libraries, and -is quite different from the ordinary General Public License. We use -this license for certain libraries in order to permit linking those -libraries into non-free programs. - - When a program is linked with a library, whether statically or using -a shared library, the combination of the two is legally speaking a -combined work, a derivative of the original library. The ordinary -General Public License therefore permits such linking only if the -entire combination fits its criteria of freedom. The Lesser General -Public License permits more lax criteria for linking other code with -the library. - - We call this license the "Lesser" General Public License because it -does Less to protect the user's freedom than the ordinary General -Public License. It also provides other free software developers Less -of an advantage over competing non-free programs. These disadvantages -are the reason we use the ordinary General Public License for many -libraries. However, the Lesser license provides advantages in certain -special circumstances. - - For example, on rare occasions, there may be a special need to -encourage the widest possible use of a certain library, so that it becomes -a de-facto standard. To achieve this, non-free programs must be -allowed to use the library. A more frequent case is that a free -library does the same job as widely used non-free libraries. In this -case, there is little to gain by limiting the free library to free -software only, so we use the Lesser General Public License. - - In other cases, permission to use a particular library in non-free -programs enables a greater number of people to use a large body of -free software. For example, permission to use the GNU C Library in -non-free programs enables many more people to use the whole GNU -operating system, as well as its variant, the GNU/Linux operating -system. - - Although the Lesser General Public License is Less protective of the -users' freedom, it does ensure that the user of a program that is -linked with the Library has the freedom and the wherewithal to run -that program using a modified version of the Library. - - The precise terms and conditions for copying, distribution and -modification follow. Pay close attention to the difference between a -"work based on the library" and a "work that uses the library". The -former contains code derived from the library, whereas the latter must -be combined with the library in order to run. - - GNU LESSER GENERAL PUBLIC LICENSE - TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION - - 0. This License Agreement applies to any software library or other -program which contains a notice placed by the copyright holder or -other authorized party saying it may be distributed under the terms of -this Lesser General Public License (also called "this License"). -Each licensee is addressed as "you". - - A "library" means a collection of software functions and/or data -prepared so as to be conveniently linked with application programs -(which use some of those functions and data) to form executables. - - The "Library", below, refers to any such software library or work -which has been distributed under these terms. A "work based on the -Library" means either the Library or any derivative work under -copyright law: that is to say, a work containing the Library or a -portion of it, either verbatim or with modifications and/or translated -straightforwardly into another language. (Hereinafter, translation is -included without limitation in the term "modification".) - - "Source code" for a work means the preferred form of the work for -making modifications to it. For a library, complete source code means -all the source code for all modules it contains, plus any associated -interface definition files, plus the scripts used to control compilation -and installation of the library. - - Activities other than copying, distribution and modification are not -covered by this License; they are outside its scope. The act of -running a program using the Library is not restricted, and output from -such a program is covered only if its contents constitute a work based -on the Library (independent of the use of the Library in a tool for -writing it). Whether that is true depends on what the Library does -and what the program that uses the Library does. - - 1. You may copy and distribute verbatim copies of the Library's -complete source code as you receive it, in any medium, provided that -you conspicuously and appropriately publish on each copy an -appropriate copyright notice and disclaimer of warranty; keep intact -all the notices that refer to this License and to the absence of any -warranty; and distribute a copy of this License along with the -Library. - - You may charge a fee for the physical act of transferring a copy, -and you may at your option offer warranty protection in exchange for a -fee. - - 2. You may modify your copy or copies of the Library or any portion -of it, thus forming a work based on the Library, and copy and -distribute such modifications or work under the terms of Section 1 -above, provided that you also meet all of these conditions: - - a) The modified work must itself be a software library. - - b) You must cause the files modified to carry prominent notices - stating that you changed the files and the date of any change. - - c) You must cause the whole of the work to be licensed at no - charge to all third parties under the terms of this License. - - d) If a facility in the modified Library refers to a function or a - table of data to be supplied by an application program that uses - the facility, other than as an argument passed when the facility - is invoked, then you must make a good faith effort to ensure that, - in the event an application does not supply such function or - table, the facility still operates, and performs whatever part of - its purpose remains meaningful. - - (For example, a function in a library to compute square roots has - a purpose that is entirely well-defined independent of the - application. Therefore, Subsection 2d requires that any - application-supplied function or table used by this function must - be optional: if the application does not supply it, the square - root function must still compute square roots.) - -These requirements apply to the modified work as a whole. If -identifiable sections of that work are not derived from the Library, -and can be reasonably considered independent and separate works in -themselves, then this License, and its terms, do not apply to those -sections when you distribute them as separate works. But when you -distribute the same sections as part of a whole which is a work based -on the Library, the distribution of the whole must be on the terms of -this License, whose permissions for other licensees extend to the -entire whole, and thus to each and every part regardless of who wrote -it. - -Thus, it is not the intent of this section to claim rights or contest -your rights to work written entirely by you; rather, the intent is to -exercise the right to control the distribution of derivative or -collective works based on the Library. - -In addition, mere aggregation of another work not based on the Library -with the Library (or with a work based on the Library) on a volume of -a storage or distribution medium does not bring the other work under -the scope of this License. - - 3. You may opt to apply the terms of the ordinary GNU General Public -License instead of this License to a given copy of the Library. To do -this, you must alter all the notices that refer to this License, so -that they refer to the ordinary GNU General Public License, version 2, -instead of to this License. (If a newer version than version 2 of the -ordinary GNU General Public License has appeared, then you can specify -that version instead if you wish.) Do not make any other change in -these notices. - - Once this change is made in a given copy, it is irreversible for -that copy, so the ordinary GNU General Public License applies to all -subsequent copies and derivative works made from that copy. - - This option is useful when you wish to copy part of the code of -the Library into a program that is not a library. - - 4. You may copy and distribute the Library (or a portion or -derivative of it, under Section 2) in object code or executable form -under the terms of Sections 1 and 2 above provided that you accompany -it with the complete corresponding machine-readable source code, which -must be distributed under the terms of Sections 1 and 2 above on a -medium customarily used for software interchange. - - If distribution of object code is made by offering access to copy -from a designated place, then offering equivalent access to copy the -source code from the same place satisfies the requirement to -distribute the source code, even though third parties are not -compelled to copy the source along with the object code. - - 5. A program that contains no derivative of any portion of the -Library, but is designed to work with the Library by being compiled or -linked with it, is called a "work that uses the Library". Such a -work, in isolation, is not a derivative work of the Library, and -therefore falls outside the scope of this License. - - However, linking a "work that uses the Library" with the Library -creates an executable that is a derivative of the Library (because it -contains portions of the Library), rather than a "work that uses the -library". The executable is therefore covered by this License. -Section 6 states terms for distribution of such executables. - - When a "work that uses the Library" uses material from a header file -that is part of the Library, the object code for the work may be a -derivative work of the Library even though the source code is not. -Whether this is true is especially significant if the work can be -linked without the Library, or if the work is itself a library. The -threshold for this to be true is not precisely defined by law. - - If such an object file uses only numerical parameters, data -structure layouts and accessors, and small macros and small inline -functions (ten lines or less in length), then the use of the object -file is unrestricted, regardless of whether it is legally a derivative -work. (Executables containing this object code plus portions of the -Library will still fall under Section 6.) - - Otherwise, if the work is a derivative of the Library, you may -distribute the object code for the work under the terms of Section 6. -Any executables containing that work also fall under Section 6, -whether or not they are linked directly with the Library itself. - - 6. As an exception to the Sections above, you may also combine or -link a "work that uses the Library" with the Library to produce a -work containing portions of the Library, and distribute that work -under terms of your choice, provided that the terms permit -modification of the work for the customer's own use and reverse -engineering for debugging such modifications. - - You must give prominent notice with each copy of the work that the -Library is used in it and that the Library and its use are covered by -this License. You must supply a copy of this License. If the work -during execution displays copyright notices, you must include the -copyright notice for the Library among them, as well as a reference -directing the user to the copy of this License. Also, you must do one -of these things: - - a) Accompany the work with the complete corresponding - machine-readable source code for the Library including whatever - changes were used in the work (which must be distributed under - Sections 1 and 2 above); and, if the work is an executable linked - with the Library, with the complete machine-readable "work that - uses the Library", as object code and/or source code, so that the - user can modify the Library and then relink to produce a modified - executable containing the modified Library. (It is understood - that the user who changes the contents of definitions files in the - Library will not necessarily be able to recompile the application - to use the modified definitions.) - - b) Use a suitable shared library mechanism for linking with the - Library. A suitable mechanism is one that (1) uses at run time a - copy of the library already present on the user's computer system, - rather than copying library functions into the executable, and (2) - will operate properly with a modified version of the library, if - the user installs one, as long as the modified version is - interface-compatible with the version that the work was made with. - - c) Accompany the work with a written offer, valid for at - least three years, to give the same user the materials - specified in Subsection 6a, above, for a charge no more - than the cost of performing this distribution. - - d) If distribution of the work is made by offering access to copy - from a designated place, offer equivalent access to copy the above - specified materials from the same place. - - e) Verify that the user has already received a copy of these - materials or that you have already sent this user a copy. - - For an executable, the required form of the "work that uses the -Library" must include any data and utility programs needed for -reproducing the executable from it. However, as a special exception, -the materials to be distributed need not include anything that is -normally distributed (in either source or binary form) with the major -components (compiler, kernel, and so on) of the operating system on -which the executable runs, unless that component itself accompanies -the executable. - - It may happen that this requirement contradicts the license -restrictions of other proprietary libraries that do not normally -accompany the operating system. Such a contradiction means you cannot -use both them and the Library together in an executable that you -distribute. - - 7. You may place library facilities that are a work based on the -Library side-by-side in a single library together with other library -facilities not covered by this License, and distribute such a combined -library, provided that the separate distribution of the work based on -the Library and of the other library facilities is otherwise -permitted, and provided that you do these two things: - - a) Accompany the combined library with a copy of the same work - based on the Library, uncombined with any other library - facilities. This must be distributed under the terms of the - Sections above. - - b) Give prominent notice with the combined library of the fact - that part of it is a work based on the Library, and explaining - where to find the accompanying uncombined form of the same work. - - 8. You may not copy, modify, sublicense, link with, or distribute -the Library except as expressly provided under this License. Any -attempt otherwise to copy, modify, sublicense, link with, or -distribute the Library is void, and will automatically terminate your -rights under this License. However, parties who have received copies, -or rights, from you under this License will not have their licenses -terminated so long as such parties remain in full compliance. - - 9. You are not required to accept this License, since you have not -signed it. However, nothing else grants you permission to modify or -distribute the Library or its derivative works. These actions are -prohibited by law if you do not accept this License. Therefore, by -modifying or distributing the Library (or any work based on the -Library), you indicate your acceptance of this License to do so, and -all its terms and conditions for copying, distributing or modifying -the Library or works based on it. - - 10. Each time you redistribute the Library (or any work based on the -Library), the recipient automatically receives a license from the -original licensor to copy, distribute, link with or modify the Library -subject to these terms and conditions. You may not impose any further -restrictions on the recipients' exercise of the rights granted herein. -You are not responsible for enforcing compliance by third parties with -this License. - - 11. If, as a consequence of a court judgment or allegation of patent -infringement or for any other reason (not limited to patent issues), -conditions are imposed on you (whether by court order, agreement or -otherwise) that contradict the conditions of this License, they do not -excuse you from the conditions of this License. If you cannot -distribute so as to satisfy simultaneously your obligations under this -License and any other pertinent obligations, then as a consequence you -may not distribute the Library at all. For example, if a patent -license would not permit royalty-free redistribution of the Library by -all those who receive copies directly or indirectly through you, then -the only way you could satisfy both it and this License would be to -refrain entirely from distribution of the Library. - -If any portion of this section is held invalid or unenforceable under any -particular circumstance, the balance of the section is intended to apply, -and the section as a whole is intended to apply in other circumstances. - -It is not the purpose of this section to induce you to infringe any -patents or other property right claims or to contest validity of any -such claims; this section has the sole purpose of protecting the -integrity of the free software distribution system which is -implemented by public license practices. Many people have made -generous contributions to the wide range of software distributed -through that system in reliance on consistent application of that -system; it is up to the author/donor to decide if he or she is willing -to distribute software through any other system and a licensee cannot -impose that choice. - -This section is intended to make thoroughly clear what is believed to -be a consequence of the rest of this License. - - 12. If the distribution and/or use of the Library is restricted in -certain countries either by patents or by copyrighted interfaces, the -original copyright holder who places the Library under this License may add -an explicit geographical distribution limitation excluding those countries, -so that distribution is permitted only in or among countries not thus -excluded. In such case, this License incorporates the limitation as if -written in the body of this License. - - 13. The Free Software Foundation may publish revised and/or new -versions of the Lesser General Public License from time to time. -Such new versions will be similar in spirit to the present version, -but may differ in detail to address new problems or concerns. - -Each version is given a distinguishing version number. If the Library -specifies a version number of this License which applies to it and -"any later version", you have the option of following the terms and -conditions either of that version or of any later version published by -the Free Software Foundation. If the Library does not specify a -license version number, you may choose any version ever published by -the Free Software Foundation. - - 14. If you wish to incorporate parts of the Library into other free -programs whose distribution conditions are incompatible with these, -write to the author to ask for permission. For software which is -copyrighted by the Free Software Foundation, write to the Free -Software Foundation; we sometimes make exceptions for this. Our -decision will be guided by the two goals of preserving the free status -of all derivatives of our free software and of promoting the sharing -and reuse of software generally. - - NO WARRANTY - - 15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO -WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE LAW. -EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR -OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY OF ANY -KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR -PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE -LIBRARY IS WITH YOU. SHOULD THE LIBRARY PROVE DEFECTIVE, YOU ASSUME -THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. - - 16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN -WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY -AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE LIABLE TO YOU -FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR -CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE -LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING -RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A -FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF -SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH -DAMAGES. - - END OF TERMS AND CONDITIONS - - How to Apply These Terms to Your New Libraries - - If you develop a new library, and you want it to be of the greatest -possible use to the public, we recommend making it free software that -everyone can redistribute and change. You can do so by permitting -redistribution under these terms (or, alternatively, under the terms of the -ordinary General Public License). - - To apply these terms, attach the following notices to the library. It is -safest to attach them to the start of each source file to most effectively -convey the exclusion of warranty; and each file should have at least the -"copyright" line and a pointer to where the full notice is found. - - <one line to give the library's name and a brief idea of what it does.> - Copyright (C) <year> <name of author> - - This library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - This library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with this library; if not, write to the Free Software - Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - -Also add information on how to contact you by electronic and paper mail. - -You should also get your employer (if you work as a programmer) or your -school, if any, to sign a "copyright disclaimer" for the library, if -necessary. Here is a sample; alter the names: - - Yoyodyne, Inc., hereby disclaims all copyright interest in the - library `Frob' (a library for tweaking knobs) written by James Random Hacker. - - <signature of Ty Coon>, 1 April 1990 - Ty Coon, President of Vice - -That's all there is to it! diff --git a/src/main/jni/ffmpeg/COPYING.LGPLv3 b/src/main/jni/ffmpeg/COPYING.LGPLv3 deleted file mode 100644 index 65c5ca88a..000000000 --- a/src/main/jni/ffmpeg/COPYING.LGPLv3 +++ /dev/null @@ -1,165 +0,0 @@ - GNU LESSER GENERAL PUBLIC LICENSE - Version 3, 29 June 2007 - - Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/> - Everyone is permitted to copy and distribute verbatim copies - of this license document, but changing it is not allowed. - - - This version of the GNU Lesser General Public License incorporates -the terms and conditions of version 3 of the GNU General Public -License, supplemented by the additional permissions listed below. - - 0. Additional Definitions. - - As used herein, "this License" refers to version 3 of the GNU Lesser -General Public License, and the "GNU GPL" refers to version 3 of the GNU -General Public License. - - "The Library" refers to a covered work governed by this License, -other than an Application or a Combined Work as defined below. - - An "Application" is any work that makes use of an interface provided -by the Library, but which is not otherwise based on the Library. -Defining a subclass of a class defined by the Library is deemed a mode -of using an interface provided by the Library. - - A "Combined Work" is a work produced by combining or linking an -Application with the Library. The particular version of the Library -with which the Combined Work was made is also called the "Linked -Version". - - The "Minimal Corresponding Source" for a Combined Work means the -Corresponding Source for the Combined Work, excluding any source code -for portions of the Combined Work that, considered in isolation, are -based on the Application, and not on the Linked Version. - - The "Corresponding Application Code" for a Combined Work means the -object code and/or source code for the Application, including any data -and utility programs needed for reproducing the Combined Work from the -Application, but excluding the System Libraries of the Combined Work. - - 1. Exception to Section 3 of the GNU GPL. - - You may convey a covered work under sections 3 and 4 of this License -without being bound by section 3 of the GNU GPL. - - 2. Conveying Modified Versions. - - If you modify a copy of the Library, and, in your modifications, a -facility refers to a function or data to be supplied by an Application -that uses the facility (other than as an argument passed when the -facility is invoked), then you may convey a copy of the modified -version: - - a) under this License, provided that you make a good faith effort to - ensure that, in the event an Application does not supply the - function or data, the facility still operates, and performs - whatever part of its purpose remains meaningful, or - - b) under the GNU GPL, with none of the additional permissions of - this License applicable to that copy. - - 3. Object Code Incorporating Material from Library Header Files. - - The object code form of an Application may incorporate material from -a header file that is part of the Library. You may convey such object -code under terms of your choice, provided that, if the incorporated -material is not limited to numerical parameters, data structure -layouts and accessors, or small macros, inline functions and templates -(ten or fewer lines in length), you do both of the following: - - a) Give prominent notice with each copy of the object code that the - Library is used in it and that the Library and its use are - covered by this License. - - b) Accompany the object code with a copy of the GNU GPL and this license - document. - - 4. Combined Works. - - You may convey a Combined Work under terms of your choice that, -taken together, effectively do not restrict modification of the -portions of the Library contained in the Combined Work and reverse -engineering for debugging such modifications, if you also do each of -the following: - - a) Give prominent notice with each copy of the Combined Work that - the Library is used in it and that the Library and its use are - covered by this License. - - b) Accompany the Combined Work with a copy of the GNU GPL and this license - document. - - c) For a Combined Work that displays copyright notices during - execution, include the copyright notice for the Library among - these notices, as well as a reference directing the user to the - copies of the GNU GPL and this license document. - - d) Do one of the following: - - 0) Convey the Minimal Corresponding Source under the terms of this - License, and the Corresponding Application Code in a form - suitable for, and under terms that permit, the user to - recombine or relink the Application with a modified version of - the Linked Version to produce a modified Combined Work, in the - manner specified by section 6 of the GNU GPL for conveying - Corresponding Source. - - 1) Use a suitable shared library mechanism for linking with the - Library. A suitable mechanism is one that (a) uses at run time - a copy of the Library already present on the user's computer - system, and (b) will operate properly with a modified version - of the Library that is interface-compatible with the Linked - Version. - - e) Provide Installation Information, but only if you would otherwise - be required to provide such information under section 6 of the - GNU GPL, and only to the extent that such information is - necessary to install and execute a modified version of the - Combined Work produced by recombining or relinking the - Application with a modified version of the Linked Version. (If - you use option 4d0, the Installation Information must accompany - the Minimal Corresponding Source and Corresponding Application - Code. If you use option 4d1, you must provide the Installation - Information in the manner specified by section 6 of the GNU GPL - for conveying Corresponding Source.) - - 5. Combined Libraries. - - You may place library facilities that are a work based on the -Library side by side in a single library together with other library -facilities that are not Applications and are not covered by this -License, and convey such a combined library under terms of your -choice, if you do both of the following: - - a) Accompany the combined library with a copy of the same work based - on the Library, uncombined with any other library facilities, - conveyed under the terms of this License. - - b) Give prominent notice with the combined library that part of it - is a work based on the Library, and explaining where to find the - accompanying uncombined form of the same work. - - 6. Revised Versions of the GNU Lesser General Public License. - - The Free Software Foundation may publish revised and/or new versions -of the GNU Lesser General Public License from time to time. Such new -versions will be similar in spirit to the present version, but may -differ in detail to address new problems or concerns. - - Each version is given a distinguishing version number. If the -Library as you received it specifies that a certain numbered version -of the GNU Lesser General Public License "or any later version" -applies to it, you have the option of following the terms and -conditions either of that published version or of any later version -published by the Free Software Foundation. If the Library as you -received it does not specify a version number of the GNU Lesser -General Public License, you may choose any version of the GNU Lesser -General Public License ever published by the Free Software Foundation. - - If the Library as you received it specifies that a proxy can decide -whether future versions of the GNU Lesser General Public License shall -apply, that proxy's public statement of acceptance of any version is -permanent authorization for you to choose that version for the -Library. diff --git a/src/main/jni/ffmpeg/CREDITS b/src/main/jni/ffmpeg/CREDITS deleted file mode 100644 index e29f0b853..000000000 --- a/src/main/jni/ffmpeg/CREDITS +++ /dev/null @@ -1,6 +0,0 @@ -See the Git history of the project (git://source.ffmpeg.org/ffmpeg) to -get the names of people who have contributed to FFmpeg. - -To check the log, you can type the command "git log" in the FFmpeg -source directory, or browse the online repository at -http://source.ffmpeg.org. diff --git a/src/main/jni/ffmpeg/LICENSE.md b/src/main/jni/ffmpeg/LICENSE.md deleted file mode 100644 index 4c4a84534..000000000 --- a/src/main/jni/ffmpeg/LICENSE.md +++ /dev/null @@ -1,114 +0,0 @@ -#FFmpeg: - -Most files in FFmpeg are under the GNU Lesser General Public License version 2.1 -or later (LGPL v2.1+). Read the file `COPYING.LGPLv2.1` for details. Some other -files have MIT/X11/BSD-style licenses. In combination the LGPL v2.1+ applies to -FFmpeg. - -Some optional parts of FFmpeg are licensed under the GNU General Public License -version 2 or later (GPL v2+). See the file `COPYING.GPLv2` for details. None of -these parts are used by default, you have to explicitly pass `--enable-gpl` to -configure to activate them. In this case, FFmpeg's license changes to GPL v2+. - -Specifically, the GPL parts of FFmpeg are: - -- libpostproc -- optional x86 optimizations in the files - - `libavcodec/x86/flac_dsp_gpl.asm` - - `libavcodec/x86/idct_mmx.c` - - `libavfilter/x86/vf_removegrain.asm` -- libutvideo encoding/decoding wrappers in - `libavcodec/libutvideo*.cpp` -- the X11 grabber in `libavdevice/x11grab.c` -- the swresample test app in - `libswresample/swresample-test.c` -- the `texi2pod.pl` tool -- the following filters in libavfilter: - - `f_ebur128.c` - - `vf_blackframe.c` - - `vf_boxblur.c` - - `vf_colormatrix.c` - - `vf_cover_rect.c` - - `vf_cropdetect.c` - - `vf_delogo.c` - - `vf_eq.c` - - `vf_find_rect.c` - - `vf_fspp.c` - - `vf_geq.c` - - `vf_histeq.c` - - `vf_hqdn3d.c` - - `vf_interlace.c` - - `vf_kerndeint.c` - - `vf_mcdeint.c` - - `vf_mpdecimate.c` - - `vf_owdenoise.c` - - `vf_perspective.c` - - `vf_phase.c` - - `vf_pp.c` - - `vf_pp7.c` - - `vf_pullup.c` - - `vf_sab.c` - - `vf_smartblur.c` - - `vf_repeatfields.c` - - `vf_spp.c` - - `vf_stereo3d.c` - - `vf_super2xsai.c` - - `vf_tinterlace.c` - - `vf_uspp.c` - - `vsrc_mptestsrc.c` - -Should you, for whatever reason, prefer to use version 3 of the (L)GPL, then -the configure parameter `--enable-version3` will activate this licensing option -for you. Read the file `COPYING.LGPLv3` or, if you have enabled GPL parts, -`COPYING.GPLv3` to learn the exact legal terms that apply in this case. - -There are a handful of files under other licensing terms, namely: - -* The files `libavcodec/jfdctfst.c`, `libavcodec/jfdctint_template.c` and - `libavcodec/jrevdct.c` are taken from libjpeg, see the top of the files for - licensing details. Specifically note that you must credit the IJG in the - documentation accompanying your program if you only distribute executables. - You must also indicate any changes including additions and deletions to - those three files in the documentation. -* `tests/reference.pnm` is under the expat license. - - -external libraries -================== - -FFmpeg can be combined with a number of external libraries, which sometimes -affect the licensing of binaries resulting from the combination. - -compatible libraries --------------------- - -The following libraries are under GPL: -- frei0r -- libcdio -- librubberband -- libutvideo -- libvidstab -- libx264 -- libx265 -- libxavs -- libxvid - -When combining them with FFmpeg, FFmpeg needs to be licensed as GPL as well by -passing `--enable-gpl` to configure. - -The OpenCORE and VisualOn libraries are under the Apache License 2.0. That -license is incompatible with the LGPL v2.1 and the GPL v2, but not with -version 3 of those licenses. So to combine these libraries with FFmpeg, the -license version needs to be upgraded by passing `--enable-version3` to configure. - -incompatible libraries ----------------------- - -The Fraunhofer AAC library, FAAC and aacplus are under licenses which -are incompatible with the GPLv2 and v3. We do not know for certain if their -licenses are compatible with the LGPL. -If you wish to enable these libraries, pass `--enable-nonfree` to configure. -But note that if you enable any of these libraries the resulting binary will -be under a complex license mix that is more restrictive than the LGPL and that -may result in additional obligations. It is possible that these -restrictions cause the resulting binary to be unredistributeable. diff --git a/src/main/jni/ffmpeg/README.md b/src/main/jni/ffmpeg/README.md deleted file mode 100644 index 241919193..000000000 --- a/src/main/jni/ffmpeg/README.md +++ /dev/null @@ -1,49 +0,0 @@ -FFmpeg README -============= - -FFmpeg is a collection of libraries and tools to process multimedia content -such as audio, video, subtitles and related metadata. - -## Libraries - -* `libavcodec` provides implementation of a wider range of codecs. -* `libavformat` implements streaming protocols, container formats and basic I/O access. -* `libavutil` includes hashers, decompressors and miscellaneous utility functions. -* `libavfilter` provides a mean to alter decoded Audio and Video through chain of filters. -* `libavdevice` provides an abstraction to access capture and playback devices. -* `libswresample` implements audio mixing and resampling routines. -* `libswscale` implements color conversion and scaling routines. - -## Tools - -* [ffmpeg](https://ffmpeg.org/ffmpeg.html) is a command line toolbox to - manipulate, convert and stream multimedia content. -* [ffplay](https://ffmpeg.org/ffplay.html) is a minimalistic multimedia player. -* [ffprobe](https://ffmpeg.org/ffprobe.html) is a simple analysis tool to inspect - multimedia content. -* [ffserver](https://ffmpeg.org/ffserver.html) is a multimedia streaming server - for live broadcasts. -* Additional small tools such as `aviocat`, `ismindex` and `qt-faststart`. - -## Documentation - -The offline documentation is available in the **doc/** directory. - -The online documentation is available in the main [website](https://ffmpeg.org) -and in the [wiki](https://trac.ffmpeg.org). - -### Examples - -Coding examples are available in the **doc/examples** directory. - -## License - -FFmpeg codebase is mainly LGPL-licensed with optional components licensed under -GPL. Please refer to the LICENSE file for detailed information. - -## Contributing - -Patches should be submitted to the ffmpeg-devel mailing list using -`git format-patch` or `git send-email`. Github pull requests should be -avoided because they are not part of our review process. Few developers -follow pull requests so they will likely be ignored. diff --git a/src/main/jni/ffmpeg/armv5te/libavcodec.a b/src/main/jni/ffmpeg/armv5te/libavcodec.a Binary files differdeleted file mode 100644 index 35b5dbdb1..000000000 --- a/src/main/jni/ffmpeg/armv5te/libavcodec.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/armv5te/libavformat.a b/src/main/jni/ffmpeg/armv5te/libavformat.a Binary files differdeleted file mode 100644 index a93225ee0..000000000 --- a/src/main/jni/ffmpeg/armv5te/libavformat.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/armv5te/libavutil.a b/src/main/jni/ffmpeg/armv5te/libavutil.a Binary files differdeleted file mode 100644 index 4131c6c55..000000000 --- a/src/main/jni/ffmpeg/armv5te/libavutil.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/armv7-a/libavcodec.a b/src/main/jni/ffmpeg/armv7-a/libavcodec.a Binary files differdeleted file mode 100644 index 68bc1ddd9..000000000 --- a/src/main/jni/ffmpeg/armv7-a/libavcodec.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/armv7-a/libavformat.a b/src/main/jni/ffmpeg/armv7-a/libavformat.a Binary files differdeleted file mode 100644 index 080afa5ef..000000000 --- a/src/main/jni/ffmpeg/armv7-a/libavformat.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/armv7-a/libavutil.a b/src/main/jni/ffmpeg/armv7-a/libavutil.a Binary files differdeleted file mode 100644 index 7136db32d..000000000 --- a/src/main/jni/ffmpeg/armv7-a/libavutil.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/build_ffmpeg_android.sh b/src/main/jni/ffmpeg/build_ffmpeg_android.sh deleted file mode 100644 index 7703afa84..000000000 --- a/src/main/jni/ffmpeg/build_ffmpeg_android.sh +++ /dev/null @@ -1,106 +0,0 @@ -#!/bin/bash -#apply fix http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/203198 - -function build_one { - -echo "Cleaning..." -make clean - -echo "Configuring..." - -./configure \ ---cc=$CC \ ---nm=$NM \ ---enable-stripping \ ---arch=$ARCH \ ---cpu=$CPU \ ---target-os=linux \ ---enable-cross-compile \ ---yasmexe=$NDK/prebuilt/darwin-x86_64/bin/yasm \ ---prefix=$PREFIX \ ---enable-pic \ ---disable-shared \ ---enable-static \ ---cross-prefix=$CROSS_PREFIX \ ---sysroot=$PLATFORM \ ---extra-cflags="-Os -DANDROID $OPTIMIZE_CFLAGS -fPIE -pie --static" \ ---extra-ldflags="-Wl,-rpath-link=$PLATFORM/usr/lib -L$PLATFORM/usr/lib -nostdlib -lc -lm -ldl" \ ---extra-libs="-lgcc" \ -\ ---enable-version3 \ ---enable-gpl \ -\ ---disable-doc \ ---disable-avx \ -\ ---disable-everything \ ---disable-network \ ---disable-zlib \ ---disable-avfilter \ ---disable-avdevice \ ---disable-postproc \ ---disable-debug \ ---disable-programs \ ---disable-network \ -\ ---enable-pthreads \ ---enable-protocol=file \ ---enable-decoder=h264 \ ---enable-decoder=gif \ ---enable-demuxer=mov \ ---enable-demuxer=gif \ ---enable-hwaccels \ ---enable-runtime-cpudetect \ ---enable-asm \ -$ADDITIONAL_CONFIGURE_FLAG - -#echo "continue?" -#read -make -j8 install - -} - -NDK=/Applications/sdk/ndk-bundle - -#arm platform -PLATFORM=$NDK/platforms/android-9/arch-arm -PREBUILT=$NDK/toolchains/arm-linux-androideabi-4.8/prebuilt/darwin-x86_64 -LD=$PREBUILT/bin/arm-linux-androideabi-ld -AR=$PREBUILT/bin/arm-linux-androideabi-ar -NM=$PREBUILT/bin/arm-linux-androideabi-nm -GCCLIB=$PREBUILT/lib/gcc/arm-linux-androideabi/4.8/libgcc.a -ARCH=arm -CC=$PREBUILT/bin/arm-linux-androideabi-gcc -CROSS_PREFIX=$PREBUILT/bin/arm-linux-androideabi- - -#arm v5 -CPU=armv5te -OPTIMIZE_CFLAGS="-marm -march=$CPU" -PREFIX=./android/$CPU -ADDITIONAL_CONFIGURE_FLAG="--disable-armv6 --disable-armv6t2 --disable-vfp --disable-neon" -build_one - -#arm v7n -CPU=armv7-a -OPTIMIZE_CFLAGS="-marm -march=$CPU" -PREFIX=./android/$CPU -ADDITIONAL_CONFIGURE_FLAG=--enable-neon -build_one - -#x86 platform -PLATFORM=$NDK/platforms/android-9/arch-x86 -PREBUILT=$NDK/toolchains/x86-4.8/prebuilt/darwin-x86_64 -LD=$PREBUILT/bin/i686-linux-android-ld -AR=$PREBUILT/bin/i686-linux-android-ar -NM=$PREBUILT/bin/i686-linux-android-nm -GCCLIB=$PREBUILT/lib/gcc/i686-linux-android/4.8/libgcc.a -ARCH=x86 -CC=$PREBUILT/bin/i686-linux-android-gcc -CROSS_PREFIX=$PREBUILT/bin/i686-linux-android- - -CPU=i686 -OPTIMIZE_CFLAGS="-march=$CPU" -PREFIX=./android/$CPU -ADDITIONAL_CONFIGURE_FLAG="--disable-mmx --disable-yasm" -build_one - diff --git a/src/main/jni/ffmpeg/i686/libavcodec.a b/src/main/jni/ffmpeg/i686/libavcodec.a Binary files differdeleted file mode 100644 index 4fbf65ca4..000000000 --- a/src/main/jni/ffmpeg/i686/libavcodec.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/i686/libavformat.a b/src/main/jni/ffmpeg/i686/libavformat.a Binary files differdeleted file mode 100644 index 30e6bfad4..000000000 --- a/src/main/jni/ffmpeg/i686/libavformat.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/i686/libavutil.a b/src/main/jni/ffmpeg/i686/libavutil.a Binary files differdeleted file mode 100644 index 9c0e0f869..000000000 --- a/src/main/jni/ffmpeg/i686/libavutil.a +++ /dev/null diff --git a/src/main/jni/ffmpeg/include/libavcodec/avcodec.h b/src/main/jni/ffmpeg/include/libavcodec/avcodec.h deleted file mode 100644 index f365775f0..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/avcodec.h +++ /dev/null @@ -1,5418 +0,0 @@ -/* - * copyright (c) 2001 Fabrice Bellard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_AVCODEC_H -#define AVCODEC_AVCODEC_H - -/** - * @file - * @ingroup libavc - * Libavcodec external API header - */ - -#include <errno.h> -#include "libavutil/samplefmt.h" -#include "libavutil/attributes.h" -#include "libavutil/avutil.h" -#include "libavutil/buffer.h" -#include "libavutil/cpu.h" -#include "libavutil/channel_layout.h" -#include "libavutil/dict.h" -#include "libavutil/frame.h" -#include "libavutil/log.h" -#include "libavutil/pixfmt.h" -#include "libavutil/rational.h" - -#include "version.h" - -/** - * @defgroup libavc Encoding/Decoding Library - * @{ - * - * @defgroup lavc_decoding Decoding - * @{ - * @} - * - * @defgroup lavc_encoding Encoding - * @{ - * @} - * - * @defgroup lavc_codec Codecs - * @{ - * @defgroup lavc_codec_native Native Codecs - * @{ - * @} - * @defgroup lavc_codec_wrappers External library wrappers - * @{ - * @} - * @defgroup lavc_codec_hwaccel Hardware Accelerators bridge - * @{ - * @} - * @} - * @defgroup lavc_internal Internal - * @{ - * @} - * @} - * - */ - -/** - * @defgroup lavc_core Core functions/structures. - * @ingroup libavc - * - * Basic definitions, functions for querying libavcodec capabilities, - * allocating core structures, etc. - * @{ - */ - - -/** - * Identify the syntax and semantics of the bitstream. - * The principle is roughly: - * Two decoders with the same ID can decode the same streams. - * Two encoders with the same ID can encode compatible streams. - * There may be slight deviations from the principle due to implementation - * details. - * - * If you add a codec ID to this list, add it so that - * 1. no value of a existing codec ID changes (that would break ABI), - * 2. it is as close as possible to similar codecs - * - * After adding new codec IDs, do not forget to add an entry to the codec - * descriptor list and bump libavcodec minor version. - */ -enum AVCodecID { - AV_CODEC_ID_NONE, - - /* video codecs */ - AV_CODEC_ID_MPEG1VIDEO, - AV_CODEC_ID_MPEG2VIDEO, ///< preferred ID for MPEG-1/2 video decoding -#if FF_API_XVMC - AV_CODEC_ID_MPEG2VIDEO_XVMC, -#endif /* FF_API_XVMC */ - AV_CODEC_ID_H261, - AV_CODEC_ID_H263, - AV_CODEC_ID_RV10, - AV_CODEC_ID_RV20, - AV_CODEC_ID_MJPEG, - AV_CODEC_ID_MJPEGB, - AV_CODEC_ID_LJPEG, - AV_CODEC_ID_SP5X, - AV_CODEC_ID_JPEGLS, - AV_CODEC_ID_MPEG4, - AV_CODEC_ID_RAWVIDEO, - AV_CODEC_ID_MSMPEG4V1, - AV_CODEC_ID_MSMPEG4V2, - AV_CODEC_ID_MSMPEG4V3, - AV_CODEC_ID_WMV1, - AV_CODEC_ID_WMV2, - AV_CODEC_ID_H263P, - AV_CODEC_ID_H263I, - AV_CODEC_ID_FLV1, - AV_CODEC_ID_SVQ1, - AV_CODEC_ID_SVQ3, - AV_CODEC_ID_DVVIDEO, - AV_CODEC_ID_HUFFYUV, - AV_CODEC_ID_CYUV, - AV_CODEC_ID_H264, - AV_CODEC_ID_INDEO3, - AV_CODEC_ID_VP3, - AV_CODEC_ID_THEORA, - AV_CODEC_ID_ASV1, - AV_CODEC_ID_ASV2, - AV_CODEC_ID_FFV1, - AV_CODEC_ID_4XM, - AV_CODEC_ID_VCR1, - AV_CODEC_ID_CLJR, - AV_CODEC_ID_MDEC, - AV_CODEC_ID_ROQ, - AV_CODEC_ID_INTERPLAY_VIDEO, - AV_CODEC_ID_XAN_WC3, - AV_CODEC_ID_XAN_WC4, - AV_CODEC_ID_RPZA, - AV_CODEC_ID_CINEPAK, - AV_CODEC_ID_WS_VQA, - AV_CODEC_ID_MSRLE, - AV_CODEC_ID_MSVIDEO1, - AV_CODEC_ID_IDCIN, - AV_CODEC_ID_8BPS, - AV_CODEC_ID_SMC, - AV_CODEC_ID_FLIC, - AV_CODEC_ID_TRUEMOTION1, - AV_CODEC_ID_VMDVIDEO, - AV_CODEC_ID_MSZH, - AV_CODEC_ID_ZLIB, - AV_CODEC_ID_QTRLE, - AV_CODEC_ID_TSCC, - AV_CODEC_ID_ULTI, - AV_CODEC_ID_QDRAW, - AV_CODEC_ID_VIXL, - AV_CODEC_ID_QPEG, - AV_CODEC_ID_PNG, - AV_CODEC_ID_PPM, - AV_CODEC_ID_PBM, - AV_CODEC_ID_PGM, - AV_CODEC_ID_PGMYUV, - AV_CODEC_ID_PAM, - AV_CODEC_ID_FFVHUFF, - AV_CODEC_ID_RV30, - AV_CODEC_ID_RV40, - AV_CODEC_ID_VC1, - AV_CODEC_ID_WMV3, - AV_CODEC_ID_LOCO, - AV_CODEC_ID_WNV1, - AV_CODEC_ID_AASC, - AV_CODEC_ID_INDEO2, - AV_CODEC_ID_FRAPS, - AV_CODEC_ID_TRUEMOTION2, - AV_CODEC_ID_BMP, - AV_CODEC_ID_CSCD, - AV_CODEC_ID_MMVIDEO, - AV_CODEC_ID_ZMBV, - AV_CODEC_ID_AVS, - AV_CODEC_ID_SMACKVIDEO, - AV_CODEC_ID_NUV, - AV_CODEC_ID_KMVC, - AV_CODEC_ID_FLASHSV, - AV_CODEC_ID_CAVS, - AV_CODEC_ID_JPEG2000, - AV_CODEC_ID_VMNC, - AV_CODEC_ID_VP5, - AV_CODEC_ID_VP6, - AV_CODEC_ID_VP6F, - AV_CODEC_ID_TARGA, - AV_CODEC_ID_DSICINVIDEO, - AV_CODEC_ID_TIERTEXSEQVIDEO, - AV_CODEC_ID_TIFF, - AV_CODEC_ID_GIF, - AV_CODEC_ID_DXA, - AV_CODEC_ID_DNXHD, - AV_CODEC_ID_THP, - AV_CODEC_ID_SGI, - AV_CODEC_ID_C93, - AV_CODEC_ID_BETHSOFTVID, - AV_CODEC_ID_PTX, - AV_CODEC_ID_TXD, - AV_CODEC_ID_VP6A, - AV_CODEC_ID_AMV, - AV_CODEC_ID_VB, - AV_CODEC_ID_PCX, - AV_CODEC_ID_SUNRAST, - AV_CODEC_ID_INDEO4, - AV_CODEC_ID_INDEO5, - AV_CODEC_ID_MIMIC, - AV_CODEC_ID_RL2, - AV_CODEC_ID_ESCAPE124, - AV_CODEC_ID_DIRAC, - AV_CODEC_ID_BFI, - AV_CODEC_ID_CMV, - AV_CODEC_ID_MOTIONPIXELS, - AV_CODEC_ID_TGV, - AV_CODEC_ID_TGQ, - AV_CODEC_ID_TQI, - AV_CODEC_ID_AURA, - AV_CODEC_ID_AURA2, - AV_CODEC_ID_V210X, - AV_CODEC_ID_TMV, - AV_CODEC_ID_V210, - AV_CODEC_ID_DPX, - AV_CODEC_ID_MAD, - AV_CODEC_ID_FRWU, - AV_CODEC_ID_FLASHSV2, - AV_CODEC_ID_CDGRAPHICS, - AV_CODEC_ID_R210, - AV_CODEC_ID_ANM, - AV_CODEC_ID_BINKVIDEO, - AV_CODEC_ID_IFF_ILBM, -#define AV_CODEC_ID_IFF_BYTERUN1 AV_CODEC_ID_IFF_ILBM - AV_CODEC_ID_KGV1, - AV_CODEC_ID_YOP, - AV_CODEC_ID_VP8, - AV_CODEC_ID_PICTOR, - AV_CODEC_ID_ANSI, - AV_CODEC_ID_A64_MULTI, - AV_CODEC_ID_A64_MULTI5, - AV_CODEC_ID_R10K, - AV_CODEC_ID_MXPEG, - AV_CODEC_ID_LAGARITH, - AV_CODEC_ID_PRORES, - AV_CODEC_ID_JV, - AV_CODEC_ID_DFA, - AV_CODEC_ID_WMV3IMAGE, - AV_CODEC_ID_VC1IMAGE, - AV_CODEC_ID_UTVIDEO, - AV_CODEC_ID_BMV_VIDEO, - AV_CODEC_ID_VBLE, - AV_CODEC_ID_DXTORY, - AV_CODEC_ID_V410, - AV_CODEC_ID_XWD, - AV_CODEC_ID_CDXL, - AV_CODEC_ID_XBM, - AV_CODEC_ID_ZEROCODEC, - AV_CODEC_ID_MSS1, - AV_CODEC_ID_MSA1, - AV_CODEC_ID_TSCC2, - AV_CODEC_ID_MTS2, - AV_CODEC_ID_CLLC, - AV_CODEC_ID_MSS2, - AV_CODEC_ID_VP9, - AV_CODEC_ID_AIC, - AV_CODEC_ID_ESCAPE130, - AV_CODEC_ID_G2M, - AV_CODEC_ID_WEBP, - AV_CODEC_ID_HNM4_VIDEO, - AV_CODEC_ID_HEVC, -#define AV_CODEC_ID_H265 AV_CODEC_ID_HEVC - AV_CODEC_ID_FIC, - AV_CODEC_ID_ALIAS_PIX, - AV_CODEC_ID_BRENDER_PIX, - AV_CODEC_ID_PAF_VIDEO, - AV_CODEC_ID_EXR, - AV_CODEC_ID_VP7, - AV_CODEC_ID_SANM, - AV_CODEC_ID_SGIRLE, - AV_CODEC_ID_MVC1, - AV_CODEC_ID_MVC2, - AV_CODEC_ID_HQX, - AV_CODEC_ID_TDSC, - AV_CODEC_ID_HQ_HQA, - AV_CODEC_ID_HAP, - AV_CODEC_ID_DDS, - AV_CODEC_ID_DXV, - AV_CODEC_ID_SCREENPRESSO, - AV_CODEC_ID_RSCC, - - AV_CODEC_ID_Y41P = 0x8000, - AV_CODEC_ID_AVRP, - AV_CODEC_ID_012V, - AV_CODEC_ID_AVUI, - AV_CODEC_ID_AYUV, - AV_CODEC_ID_TARGA_Y216, - AV_CODEC_ID_V308, - AV_CODEC_ID_V408, - AV_CODEC_ID_YUV4, - AV_CODEC_ID_AVRN, - AV_CODEC_ID_CPIA, - AV_CODEC_ID_XFACE, - AV_CODEC_ID_SNOW, - AV_CODEC_ID_SMVJPEG, - AV_CODEC_ID_APNG, - AV_CODEC_ID_DAALA, - - /* various PCM "codecs" */ - AV_CODEC_ID_FIRST_AUDIO = 0x10000, ///< A dummy id pointing at the start of audio codecs - AV_CODEC_ID_PCM_S16LE = 0x10000, - AV_CODEC_ID_PCM_S16BE, - AV_CODEC_ID_PCM_U16LE, - AV_CODEC_ID_PCM_U16BE, - AV_CODEC_ID_PCM_S8, - AV_CODEC_ID_PCM_U8, - AV_CODEC_ID_PCM_MULAW, - AV_CODEC_ID_PCM_ALAW, - AV_CODEC_ID_PCM_S32LE, - AV_CODEC_ID_PCM_S32BE, - AV_CODEC_ID_PCM_U32LE, - AV_CODEC_ID_PCM_U32BE, - AV_CODEC_ID_PCM_S24LE, - AV_CODEC_ID_PCM_S24BE, - AV_CODEC_ID_PCM_U24LE, - AV_CODEC_ID_PCM_U24BE, - AV_CODEC_ID_PCM_S24DAUD, - AV_CODEC_ID_PCM_ZORK, - AV_CODEC_ID_PCM_S16LE_PLANAR, - AV_CODEC_ID_PCM_DVD, - AV_CODEC_ID_PCM_F32BE, - AV_CODEC_ID_PCM_F32LE, - AV_CODEC_ID_PCM_F64BE, - AV_CODEC_ID_PCM_F64LE, - AV_CODEC_ID_PCM_BLURAY, - AV_CODEC_ID_PCM_LXF, - AV_CODEC_ID_S302M, - AV_CODEC_ID_PCM_S8_PLANAR, - AV_CODEC_ID_PCM_S24LE_PLANAR, - AV_CODEC_ID_PCM_S32LE_PLANAR, - AV_CODEC_ID_PCM_S16BE_PLANAR, - /* new PCM "codecs" should be added right below this line starting with - * an explicit value of for example 0x10800 - */ - - /* various ADPCM codecs */ - AV_CODEC_ID_ADPCM_IMA_QT = 0x11000, - AV_CODEC_ID_ADPCM_IMA_WAV, - AV_CODEC_ID_ADPCM_IMA_DK3, - AV_CODEC_ID_ADPCM_IMA_DK4, - AV_CODEC_ID_ADPCM_IMA_WS, - AV_CODEC_ID_ADPCM_IMA_SMJPEG, - AV_CODEC_ID_ADPCM_MS, - AV_CODEC_ID_ADPCM_4XM, - AV_CODEC_ID_ADPCM_XA, - AV_CODEC_ID_ADPCM_ADX, - AV_CODEC_ID_ADPCM_EA, - AV_CODEC_ID_ADPCM_G726, - AV_CODEC_ID_ADPCM_CT, - AV_CODEC_ID_ADPCM_SWF, - AV_CODEC_ID_ADPCM_YAMAHA, - AV_CODEC_ID_ADPCM_SBPRO_4, - AV_CODEC_ID_ADPCM_SBPRO_3, - AV_CODEC_ID_ADPCM_SBPRO_2, - AV_CODEC_ID_ADPCM_THP, - AV_CODEC_ID_ADPCM_IMA_AMV, - AV_CODEC_ID_ADPCM_EA_R1, - AV_CODEC_ID_ADPCM_EA_R3, - AV_CODEC_ID_ADPCM_EA_R2, - AV_CODEC_ID_ADPCM_IMA_EA_SEAD, - AV_CODEC_ID_ADPCM_IMA_EA_EACS, - AV_CODEC_ID_ADPCM_EA_XAS, - AV_CODEC_ID_ADPCM_EA_MAXIS_XA, - AV_CODEC_ID_ADPCM_IMA_ISS, - AV_CODEC_ID_ADPCM_G722, - AV_CODEC_ID_ADPCM_IMA_APC, - AV_CODEC_ID_ADPCM_VIMA, -#if FF_API_VIMA_DECODER - AV_CODEC_ID_VIMA = AV_CODEC_ID_ADPCM_VIMA, -#endif - - AV_CODEC_ID_ADPCM_AFC = 0x11800, - AV_CODEC_ID_ADPCM_IMA_OKI, - AV_CODEC_ID_ADPCM_DTK, - AV_CODEC_ID_ADPCM_IMA_RAD, - AV_CODEC_ID_ADPCM_G726LE, - AV_CODEC_ID_ADPCM_THP_LE, - AV_CODEC_ID_ADPCM_PSX, - AV_CODEC_ID_ADPCM_AICA, - - /* AMR */ - AV_CODEC_ID_AMR_NB = 0x12000, - AV_CODEC_ID_AMR_WB, - - /* RealAudio codecs*/ - AV_CODEC_ID_RA_144 = 0x13000, - AV_CODEC_ID_RA_288, - - /* various DPCM codecs */ - AV_CODEC_ID_ROQ_DPCM = 0x14000, - AV_CODEC_ID_INTERPLAY_DPCM, - AV_CODEC_ID_XAN_DPCM, - AV_CODEC_ID_SOL_DPCM, - - AV_CODEC_ID_SDX2_DPCM = 0x14800, - - /* audio codecs */ - AV_CODEC_ID_MP2 = 0x15000, - AV_CODEC_ID_MP3, ///< preferred ID for decoding MPEG audio layer 1, 2 or 3 - AV_CODEC_ID_AAC, - AV_CODEC_ID_AC3, - AV_CODEC_ID_DTS, - AV_CODEC_ID_VORBIS, - AV_CODEC_ID_DVAUDIO, - AV_CODEC_ID_WMAV1, - AV_CODEC_ID_WMAV2, - AV_CODEC_ID_MACE3, - AV_CODEC_ID_MACE6, - AV_CODEC_ID_VMDAUDIO, - AV_CODEC_ID_FLAC, - AV_CODEC_ID_MP3ADU, - AV_CODEC_ID_MP3ON4, - AV_CODEC_ID_SHORTEN, - AV_CODEC_ID_ALAC, - AV_CODEC_ID_WESTWOOD_SND1, - AV_CODEC_ID_GSM, ///< as in Berlin toast format - AV_CODEC_ID_QDM2, - AV_CODEC_ID_COOK, - AV_CODEC_ID_TRUESPEECH, - AV_CODEC_ID_TTA, - AV_CODEC_ID_SMACKAUDIO, - AV_CODEC_ID_QCELP, - AV_CODEC_ID_WAVPACK, - AV_CODEC_ID_DSICINAUDIO, - AV_CODEC_ID_IMC, - AV_CODEC_ID_MUSEPACK7, - AV_CODEC_ID_MLP, - AV_CODEC_ID_GSM_MS, /* as found in WAV */ - AV_CODEC_ID_ATRAC3, -#if FF_API_VOXWARE - AV_CODEC_ID_VOXWARE, -#endif - AV_CODEC_ID_APE, - AV_CODEC_ID_NELLYMOSER, - AV_CODEC_ID_MUSEPACK8, - AV_CODEC_ID_SPEEX, - AV_CODEC_ID_WMAVOICE, - AV_CODEC_ID_WMAPRO, - AV_CODEC_ID_WMALOSSLESS, - AV_CODEC_ID_ATRAC3P, - AV_CODEC_ID_EAC3, - AV_CODEC_ID_SIPR, - AV_CODEC_ID_MP1, - AV_CODEC_ID_TWINVQ, - AV_CODEC_ID_TRUEHD, - AV_CODEC_ID_MP4ALS, - AV_CODEC_ID_ATRAC1, - AV_CODEC_ID_BINKAUDIO_RDFT, - AV_CODEC_ID_BINKAUDIO_DCT, - AV_CODEC_ID_AAC_LATM, - AV_CODEC_ID_QDMC, - AV_CODEC_ID_CELT, - AV_CODEC_ID_G723_1, - AV_CODEC_ID_G729, - AV_CODEC_ID_8SVX_EXP, - AV_CODEC_ID_8SVX_FIB, - AV_CODEC_ID_BMV_AUDIO, - AV_CODEC_ID_RALF, - AV_CODEC_ID_IAC, - AV_CODEC_ID_ILBC, - AV_CODEC_ID_OPUS, - AV_CODEC_ID_COMFORT_NOISE, - AV_CODEC_ID_TAK, - AV_CODEC_ID_METASOUND, - AV_CODEC_ID_PAF_AUDIO, - AV_CODEC_ID_ON2AVC, - AV_CODEC_ID_DSS_SP, - - AV_CODEC_ID_FFWAVESYNTH = 0x15800, - AV_CODEC_ID_SONIC, - AV_CODEC_ID_SONIC_LS, - AV_CODEC_ID_EVRC, - AV_CODEC_ID_SMV, - AV_CODEC_ID_DSD_LSBF, - AV_CODEC_ID_DSD_MSBF, - AV_CODEC_ID_DSD_LSBF_PLANAR, - AV_CODEC_ID_DSD_MSBF_PLANAR, - AV_CODEC_ID_4GV, - AV_CODEC_ID_INTERPLAY_ACM, - AV_CODEC_ID_XMA1, - AV_CODEC_ID_XMA2, - - /* subtitle codecs */ - AV_CODEC_ID_FIRST_SUBTITLE = 0x17000, ///< A dummy ID pointing at the start of subtitle codecs. - AV_CODEC_ID_DVD_SUBTITLE = 0x17000, - AV_CODEC_ID_DVB_SUBTITLE, - AV_CODEC_ID_TEXT, ///< raw UTF-8 text - AV_CODEC_ID_XSUB, - AV_CODEC_ID_SSA, - AV_CODEC_ID_MOV_TEXT, - AV_CODEC_ID_HDMV_PGS_SUBTITLE, - AV_CODEC_ID_DVB_TELETEXT, - AV_CODEC_ID_SRT, - - AV_CODEC_ID_MICRODVD = 0x17800, - AV_CODEC_ID_EIA_608, - AV_CODEC_ID_JACOSUB, - AV_CODEC_ID_SAMI, - AV_CODEC_ID_REALTEXT, - AV_CODEC_ID_STL, - AV_CODEC_ID_SUBVIEWER1, - AV_CODEC_ID_SUBVIEWER, - AV_CODEC_ID_SUBRIP, - AV_CODEC_ID_WEBVTT, - AV_CODEC_ID_MPL2, - AV_CODEC_ID_VPLAYER, - AV_CODEC_ID_PJS, - AV_CODEC_ID_ASS, - AV_CODEC_ID_HDMV_TEXT_SUBTITLE, - - /* other specific kind of codecs (generally used for attachments) */ - AV_CODEC_ID_FIRST_UNKNOWN = 0x18000, ///< A dummy ID pointing at the start of various fake codecs. - AV_CODEC_ID_TTF = 0x18000, - - AV_CODEC_ID_BINTEXT = 0x18800, - AV_CODEC_ID_XBIN, - AV_CODEC_ID_IDF, - AV_CODEC_ID_OTF, - AV_CODEC_ID_SMPTE_KLV, - AV_CODEC_ID_DVD_NAV, - AV_CODEC_ID_TIMED_ID3, - AV_CODEC_ID_BIN_DATA, - - - AV_CODEC_ID_PROBE = 0x19000, ///< codec_id is not known (like AV_CODEC_ID_NONE) but lavf should attempt to identify it - - AV_CODEC_ID_MPEG2TS = 0x20000, /**< _FAKE_ codec to indicate a raw MPEG-2 TS - * stream (only used by libavformat) */ - AV_CODEC_ID_MPEG4SYSTEMS = 0x20001, /**< _FAKE_ codec to indicate a MPEG-4 Systems - * stream (only used by libavformat) */ - AV_CODEC_ID_FFMETADATA = 0x21000, ///< Dummy codec for streams containing only metadata information. - AV_CODEC_ID_WRAPPED_AVFRAME = 0x21001, ///< Passthrough codec, AVFrames wrapped in AVPacket -}; - -/** - * This struct describes the properties of a single codec described by an - * AVCodecID. - * @see avcodec_descriptor_get() - */ -typedef struct AVCodecDescriptor { - enum AVCodecID id; - enum AVMediaType type; - /** - * Name of the codec described by this descriptor. It is non-empty and - * unique for each codec descriptor. It should contain alphanumeric - * characters and '_' only. - */ - const char *name; - /** - * A more descriptive name for this codec. May be NULL. - */ - const char *long_name; - /** - * Codec properties, a combination of AV_CODEC_PROP_* flags. - */ - int props; - /** - * MIME type(s) associated with the codec. - * May be NULL; if not, a NULL-terminated array of MIME types. - * The first item is always non-NULL and is the preferred MIME type. - */ - const char *const *mime_types; - /** - * If non-NULL, an array of profiles recognized for this codec. - * Terminated with FF_PROFILE_UNKNOWN. - */ - const struct AVProfile *profiles; -} AVCodecDescriptor; - -/** - * Codec uses only intra compression. - * Video codecs only. - */ -#define AV_CODEC_PROP_INTRA_ONLY (1 << 0) -/** - * Codec supports lossy compression. Audio and video codecs only. - * @note a codec may support both lossy and lossless - * compression modes - */ -#define AV_CODEC_PROP_LOSSY (1 << 1) -/** - * Codec supports lossless compression. Audio and video codecs only. - */ -#define AV_CODEC_PROP_LOSSLESS (1 << 2) -/** - * Codec supports frame reordering. That is, the coded order (the order in which - * the encoded packets are output by the encoders / stored / input to the - * decoders) may be different from the presentation order of the corresponding - * frames. - * - * For codecs that do not have this property set, PTS and DTS should always be - * equal. - */ -#define AV_CODEC_PROP_REORDER (1 << 3) -/** - * Subtitle codec is bitmap based - * Decoded AVSubtitle data can be read from the AVSubtitleRect->pict field. - */ -#define AV_CODEC_PROP_BITMAP_SUB (1 << 16) -/** - * Subtitle codec is text based. - * Decoded AVSubtitle data can be read from the AVSubtitleRect->ass field. - */ -#define AV_CODEC_PROP_TEXT_SUB (1 << 17) - -/** - * @ingroup lavc_decoding - * Required number of additionally allocated bytes at the end of the input bitstream for decoding. - * This is mainly needed because some optimized bitstream readers read - * 32 or 64 bit at once and could read over the end.<br> - * Note: If the first 23 bits of the additional bytes are not 0, then damaged - * MPEG bitstreams could cause overread and segfault. - */ -#define AV_INPUT_BUFFER_PADDING_SIZE 32 - -/** - * @ingroup lavc_encoding - * minimum encoding buffer size - * Used to avoid some checks during header writing. - */ -#define AV_INPUT_BUFFER_MIN_SIZE 16384 - -#if FF_API_WITHOUT_PREFIX -/** - * @deprecated use AV_INPUT_BUFFER_PADDING_SIZE instead - */ -#define FF_INPUT_BUFFER_PADDING_SIZE 32 - -/** - * @deprecated use AV_INPUT_BUFFER_MIN_SIZE instead - */ -#define FF_MIN_BUFFER_SIZE 16384 -#endif /* FF_API_WITHOUT_PREFIX */ - -/** - * @ingroup lavc_encoding - * motion estimation type. - * @deprecated use codec private option instead - */ -#if FF_API_MOTION_EST -enum Motion_Est_ID { - ME_ZERO = 1, ///< no search, that is use 0,0 vector whenever one is needed - ME_FULL, - ME_LOG, - ME_PHODS, - ME_EPZS, ///< enhanced predictive zonal search - ME_X1, ///< reserved for experiments - ME_HEX, ///< hexagon based search - ME_UMH, ///< uneven multi-hexagon search - ME_TESA, ///< transformed exhaustive search algorithm - ME_ITER=50, ///< iterative search -}; -#endif - -/** - * @ingroup lavc_decoding - */ -enum AVDiscard{ - /* We leave some space between them for extensions (drop some - * keyframes for intra-only or drop just some bidir frames). */ - AVDISCARD_NONE =-16, ///< discard nothing - AVDISCARD_DEFAULT = 0, ///< discard useless packets like 0 size packets in avi - AVDISCARD_NONREF = 8, ///< discard all non reference - AVDISCARD_BIDIR = 16, ///< discard all bidirectional frames - AVDISCARD_NONINTRA= 24, ///< discard all non intra frames - AVDISCARD_NONKEY = 32, ///< discard all frames except keyframes - AVDISCARD_ALL = 48, ///< discard all -}; - -enum AVAudioServiceType { - AV_AUDIO_SERVICE_TYPE_MAIN = 0, - AV_AUDIO_SERVICE_TYPE_EFFECTS = 1, - AV_AUDIO_SERVICE_TYPE_VISUALLY_IMPAIRED = 2, - AV_AUDIO_SERVICE_TYPE_HEARING_IMPAIRED = 3, - AV_AUDIO_SERVICE_TYPE_DIALOGUE = 4, - AV_AUDIO_SERVICE_TYPE_COMMENTARY = 5, - AV_AUDIO_SERVICE_TYPE_EMERGENCY = 6, - AV_AUDIO_SERVICE_TYPE_VOICE_OVER = 7, - AV_AUDIO_SERVICE_TYPE_KARAOKE = 8, - AV_AUDIO_SERVICE_TYPE_NB , ///< Not part of ABI -}; - -/** - * @ingroup lavc_encoding - */ -typedef struct RcOverride{ - int start_frame; - int end_frame; - int qscale; // If this is 0 then quality_factor will be used instead. - float quality_factor; -} RcOverride; - -#if FF_API_MAX_BFRAMES -/** - * @deprecated there is no libavcodec-wide limit on the number of B-frames - */ -#define FF_MAX_B_FRAMES 16 -#endif - -/* encoding support - These flags can be passed in AVCodecContext.flags before initialization. - Note: Not everything is supported yet. -*/ - -/** - * Allow decoders to produce frames with data planes that are not aligned - * to CPU requirements (e.g. due to cropping). - */ -#define AV_CODEC_FLAG_UNALIGNED (1 << 0) -/** - * Use fixed qscale. - */ -#define AV_CODEC_FLAG_QSCALE (1 << 1) -/** - * 4 MV per MB allowed / advanced prediction for H.263. - */ -#define AV_CODEC_FLAG_4MV (1 << 2) -/** - * Output even those frames that might be corrupted. - */ -#define AV_CODEC_FLAG_OUTPUT_CORRUPT (1 << 3) -/** - * Use qpel MC. - */ -#define AV_CODEC_FLAG_QPEL (1 << 4) -/** - * Use internal 2pass ratecontrol in first pass mode. - */ -#define AV_CODEC_FLAG_PASS1 (1 << 9) -/** - * Use internal 2pass ratecontrol in second pass mode. - */ -#define AV_CODEC_FLAG_PASS2 (1 << 10) -/** - * loop filter. - */ -#define AV_CODEC_FLAG_LOOP_FILTER (1 << 11) -/** - * Only decode/encode grayscale. - */ -#define AV_CODEC_FLAG_GRAY (1 << 13) -/** - * error[?] variables will be set during encoding. - */ -#define AV_CODEC_FLAG_PSNR (1 << 15) -/** - * Input bitstream might be truncated at a random location - * instead of only at frame boundaries. - */ -#define AV_CODEC_FLAG_TRUNCATED (1 << 16) -/** - * Use interlaced DCT. - */ -#define AV_CODEC_FLAG_INTERLACED_DCT (1 << 18) -/** - * Force low delay. - */ -#define AV_CODEC_FLAG_LOW_DELAY (1 << 19) -/** - * Place global headers in extradata instead of every keyframe. - */ -#define AV_CODEC_FLAG_GLOBAL_HEADER (1 << 22) -/** - * Use only bitexact stuff (except (I)DCT). - */ -#define AV_CODEC_FLAG_BITEXACT (1 << 23) -/* Fx : Flag for h263+ extra options */ -/** - * H.263 advanced intra coding / MPEG-4 AC prediction - */ -#define AV_CODEC_FLAG_AC_PRED (1 << 24) -/** - * interlaced motion estimation - */ -#define AV_CODEC_FLAG_INTERLACED_ME (1 << 29) -#define AV_CODEC_FLAG_CLOSED_GOP (1U << 31) - -/** - * Allow non spec compliant speedup tricks. - */ -#define AV_CODEC_FLAG2_FAST (1 << 0) -/** - * Skip bitstream encoding. - */ -#define AV_CODEC_FLAG2_NO_OUTPUT (1 << 2) -/** - * Place global headers at every keyframe instead of in extradata. - */ -#define AV_CODEC_FLAG2_LOCAL_HEADER (1 << 3) - -/** - * timecode is in drop frame format. DEPRECATED!!!! - */ -#define AV_CODEC_FLAG2_DROP_FRAME_TIMECODE (1 << 13) - -/** - * Input bitstream might be truncated at a packet boundaries - * instead of only at frame boundaries. - */ -#define AV_CODEC_FLAG2_CHUNKS (1 << 15) -/** - * Discard cropping information from SPS. - */ -#define AV_CODEC_FLAG2_IGNORE_CROP (1 << 16) - -/** - * Show all frames before the first keyframe - */ -#define AV_CODEC_FLAG2_SHOW_ALL (1 << 22) -/** - * Export motion vectors through frame side data - */ -#define AV_CODEC_FLAG2_EXPORT_MVS (1 << 28) -/** - * Do not skip samples and export skip information as frame side data - */ -#define AV_CODEC_FLAG2_SKIP_MANUAL (1 << 29) - -/* Unsupported options : - * Syntax Arithmetic coding (SAC) - * Reference Picture Selection - * Independent Segment Decoding */ -/* /Fx */ -/* codec capabilities */ - -/** - * Decoder can use draw_horiz_band callback. - */ -#define AV_CODEC_CAP_DRAW_HORIZ_BAND (1 << 0) -/** - * Codec uses get_buffer() for allocating buffers and supports custom allocators. - * If not set, it might not use get_buffer() at all or use operations that - * assume the buffer was allocated by avcodec_default_get_buffer. - */ -#define AV_CODEC_CAP_DR1 (1 << 1) -#define AV_CODEC_CAP_TRUNCATED (1 << 3) -/** - * Encoder or decoder requires flushing with NULL input at the end in order to - * give the complete and correct output. - * - * NOTE: If this flag is not set, the codec is guaranteed to never be fed with - * with NULL data. The user can still send NULL data to the public encode - * or decode function, but libavcodec will not pass it along to the codec - * unless this flag is set. - * - * Decoders: - * The decoder has a non-zero delay and needs to be fed with avpkt->data=NULL, - * avpkt->size=0 at the end to get the delayed data until the decoder no longer - * returns frames. - * - * Encoders: - * The encoder needs to be fed with NULL data at the end of encoding until the - * encoder no longer returns data. - * - * NOTE: For encoders implementing the AVCodec.encode2() function, setting this - * flag also means that the encoder must set the pts and duration for - * each output packet. If this flag is not set, the pts and duration will - * be determined by libavcodec from the input frame. - */ -#define AV_CODEC_CAP_DELAY (1 << 5) -/** - * Codec can be fed a final frame with a smaller size. - * This can be used to prevent truncation of the last audio samples. - */ -#define AV_CODEC_CAP_SMALL_LAST_FRAME (1 << 6) - -#if FF_API_CAP_VDPAU -/** - * Codec can export data for HW decoding (VDPAU). - */ -#define AV_CODEC_CAP_HWACCEL_VDPAU (1 << 7) -#endif - -/** - * Codec can output multiple frames per AVPacket - * Normally demuxers return one frame at a time, demuxers which do not do - * are connected to a parser to split what they return into proper frames. - * This flag is reserved to the very rare category of codecs which have a - * bitstream that cannot be split into frames without timeconsuming - * operations like full decoding. Demuxers carring such bitstreams thus - * may return multiple frames in a packet. This has many disadvantages like - * prohibiting stream copy in many cases thus it should only be considered - * as a last resort. - */ -#define AV_CODEC_CAP_SUBFRAMES (1 << 8) -/** - * Codec is experimental and is thus avoided in favor of non experimental - * encoders - */ -#define AV_CODEC_CAP_EXPERIMENTAL (1 << 9) -/** - * Codec should fill in channel configuration and samplerate instead of container - */ -#define AV_CODEC_CAP_CHANNEL_CONF (1 << 10) -/** - * Codec supports frame-level multithreading. - */ -#define AV_CODEC_CAP_FRAME_THREADS (1 << 12) -/** - * Codec supports slice-based (or partition-based) multithreading. - */ -#define AV_CODEC_CAP_SLICE_THREADS (1 << 13) -/** - * Codec supports changed parameters at any point. - */ -#define AV_CODEC_CAP_PARAM_CHANGE (1 << 14) -/** - * Codec supports avctx->thread_count == 0 (auto). - */ -#define AV_CODEC_CAP_AUTO_THREADS (1 << 15) -/** - * Audio encoder supports receiving a different number of samples in each call. - */ -#define AV_CODEC_CAP_VARIABLE_FRAME_SIZE (1 << 16) -/** - * Codec is intra only. - */ -#define AV_CODEC_CAP_INTRA_ONLY 0x40000000 -/** - * Codec is lossless. - */ -#define AV_CODEC_CAP_LOSSLESS 0x80000000 - - -#if FF_API_WITHOUT_PREFIX -/** - * Allow decoders to produce frames with data planes that are not aligned - * to CPU requirements (e.g. due to cropping). - */ -#define CODEC_FLAG_UNALIGNED AV_CODEC_FLAG_UNALIGNED -#define CODEC_FLAG_QSCALE AV_CODEC_FLAG_QSCALE -#define CODEC_FLAG_4MV AV_CODEC_FLAG_4MV -#define CODEC_FLAG_OUTPUT_CORRUPT AV_CODEC_FLAG_OUTPUT_CORRUPT -#define CODEC_FLAG_QPEL AV_CODEC_FLAG_QPEL -#if FF_API_GMC -/** - * @deprecated use the "gmc" private option of the libxvid encoder - */ -#define CODEC_FLAG_GMC 0x0020 ///< Use GMC. -#endif -#if FF_API_MV0 -/** - * @deprecated use the flag "mv0" in the "mpv_flags" private option of the - * mpegvideo encoders - */ -#define CODEC_FLAG_MV0 0x0040 -#endif -#if FF_API_INPUT_PRESERVED -/** - * @deprecated passing reference-counted frames to the encoders replaces this - * flag - */ -#define CODEC_FLAG_INPUT_PRESERVED 0x0100 -#endif -#define CODEC_FLAG_PASS1 AV_CODEC_FLAG_PASS1 -#define CODEC_FLAG_PASS2 AV_CODEC_FLAG_PASS2 -#define CODEC_FLAG_GRAY AV_CODEC_FLAG_GRAY -#if FF_API_EMU_EDGE -/** - * @deprecated edges are not used/required anymore. I.e. this flag is now always - * set. - */ -#define CODEC_FLAG_EMU_EDGE 0x4000 -#endif -#define CODEC_FLAG_PSNR AV_CODEC_FLAG_PSNR -#define CODEC_FLAG_TRUNCATED AV_CODEC_FLAG_TRUNCATED - -#if FF_API_NORMALIZE_AQP -/** - * @deprecated use the flag "naq" in the "mpv_flags" private option of the - * mpegvideo encoders - */ -#define CODEC_FLAG_NORMALIZE_AQP 0x00020000 -#endif -#define CODEC_FLAG_INTERLACED_DCT AV_CODEC_FLAG_INTERLACED_DCT -#define CODEC_FLAG_LOW_DELAY AV_CODEC_FLAG_LOW_DELAY -#define CODEC_FLAG_GLOBAL_HEADER AV_CODEC_FLAG_GLOBAL_HEADER -#define CODEC_FLAG_BITEXACT AV_CODEC_FLAG_BITEXACT -#define CODEC_FLAG_AC_PRED AV_CODEC_FLAG_AC_PRED -#define CODEC_FLAG_LOOP_FILTER AV_CODEC_FLAG_LOOP_FILTER -#define CODEC_FLAG_INTERLACED_ME AV_CODEC_FLAG_INTERLACED_ME -#define CODEC_FLAG_CLOSED_GOP AV_CODEC_FLAG_CLOSED_GOP -#define CODEC_FLAG2_FAST AV_CODEC_FLAG2_FAST -#define CODEC_FLAG2_NO_OUTPUT AV_CODEC_FLAG2_NO_OUTPUT -#define CODEC_FLAG2_LOCAL_HEADER AV_CODEC_FLAG2_LOCAL_HEADER -#define CODEC_FLAG2_DROP_FRAME_TIMECODE AV_CODEC_FLAG2_DROP_FRAME_TIMECODE -#define CODEC_FLAG2_IGNORE_CROP AV_CODEC_FLAG2_IGNORE_CROP - -#define CODEC_FLAG2_CHUNKS AV_CODEC_FLAG2_CHUNKS -#define CODEC_FLAG2_SHOW_ALL AV_CODEC_FLAG2_SHOW_ALL -#define CODEC_FLAG2_EXPORT_MVS AV_CODEC_FLAG2_EXPORT_MVS -#define CODEC_FLAG2_SKIP_MANUAL AV_CODEC_FLAG2_SKIP_MANUAL - -/* Unsupported options : - * Syntax Arithmetic coding (SAC) - * Reference Picture Selection - * Independent Segment Decoding */ -/* /Fx */ -/* codec capabilities */ - -#define CODEC_CAP_DRAW_HORIZ_BAND AV_CODEC_CAP_DRAW_HORIZ_BAND ///< Decoder can use draw_horiz_band callback. -/** - * Codec uses get_buffer() for allocating buffers and supports custom allocators. - * If not set, it might not use get_buffer() at all or use operations that - * assume the buffer was allocated by avcodec_default_get_buffer. - */ -#define CODEC_CAP_DR1 AV_CODEC_CAP_DR1 -#define CODEC_CAP_TRUNCATED AV_CODEC_CAP_TRUNCATED -#if FF_API_XVMC -/* Codec can export data for HW decoding. This flag indicates that - * the codec would call get_format() with list that might contain HW accelerated - * pixel formats (XvMC, VDPAU, VAAPI, etc). The application can pick any of them - * including raw image format. - * The application can use the passed context to determine bitstream version, - * chroma format, resolution etc. - */ -#define CODEC_CAP_HWACCEL 0x0010 -#endif /* FF_API_XVMC */ -/** - * Encoder or decoder requires flushing with NULL input at the end in order to - * give the complete and correct output. - * - * NOTE: If this flag is not set, the codec is guaranteed to never be fed with - * with NULL data. The user can still send NULL data to the public encode - * or decode function, but libavcodec will not pass it along to the codec - * unless this flag is set. - * - * Decoders: - * The decoder has a non-zero delay and needs to be fed with avpkt->data=NULL, - * avpkt->size=0 at the end to get the delayed data until the decoder no longer - * returns frames. - * - * Encoders: - * The encoder needs to be fed with NULL data at the end of encoding until the - * encoder no longer returns data. - * - * NOTE: For encoders implementing the AVCodec.encode2() function, setting this - * flag also means that the encoder must set the pts and duration for - * each output packet. If this flag is not set, the pts and duration will - * be determined by libavcodec from the input frame. - */ -#define CODEC_CAP_DELAY AV_CODEC_CAP_DELAY -/** - * Codec can be fed a final frame with a smaller size. - * This can be used to prevent truncation of the last audio samples. - */ -#define CODEC_CAP_SMALL_LAST_FRAME AV_CODEC_CAP_SMALL_LAST_FRAME -#if FF_API_CAP_VDPAU -/** - * Codec can export data for HW decoding (VDPAU). - */ -#define CODEC_CAP_HWACCEL_VDPAU AV_CODEC_CAP_HWACCEL_VDPAU -#endif -/** - * Codec can output multiple frames per AVPacket - * Normally demuxers return one frame at a time, demuxers which do not do - * are connected to a parser to split what they return into proper frames. - * This flag is reserved to the very rare category of codecs which have a - * bitstream that cannot be split into frames without timeconsuming - * operations like full decoding. Demuxers carring such bitstreams thus - * may return multiple frames in a packet. This has many disadvantages like - * prohibiting stream copy in many cases thus it should only be considered - * as a last resort. - */ -#define CODEC_CAP_SUBFRAMES AV_CODEC_CAP_SUBFRAMES -/** - * Codec is experimental and is thus avoided in favor of non experimental - * encoders - */ -#define CODEC_CAP_EXPERIMENTAL AV_CODEC_CAP_EXPERIMENTAL -/** - * Codec should fill in channel configuration and samplerate instead of container - */ -#define CODEC_CAP_CHANNEL_CONF AV_CODEC_CAP_CHANNEL_CONF -#if FF_API_NEG_LINESIZES -/** - * @deprecated no codecs use this capability - */ -#define CODEC_CAP_NEG_LINESIZES 0x0800 -#endif -/** - * Codec supports frame-level multithreading. - */ -#define CODEC_CAP_FRAME_THREADS AV_CODEC_CAP_FRAME_THREADS -/** - * Codec supports slice-based (or partition-based) multithreading. - */ -#define CODEC_CAP_SLICE_THREADS AV_CODEC_CAP_SLICE_THREADS -/** - * Codec supports changed parameters at any point. - */ -#define CODEC_CAP_PARAM_CHANGE AV_CODEC_CAP_PARAM_CHANGE -/** - * Codec supports avctx->thread_count == 0 (auto). - */ -#define CODEC_CAP_AUTO_THREADS AV_CODEC_CAP_AUTO_THREADS -/** - * Audio encoder supports receiving a different number of samples in each call. - */ -#define CODEC_CAP_VARIABLE_FRAME_SIZE AV_CODEC_CAP_VARIABLE_FRAME_SIZE -/** - * Codec is intra only. - */ -#define CODEC_CAP_INTRA_ONLY AV_CODEC_CAP_INTRA_ONLY -/** - * Codec is lossless. - */ -#define CODEC_CAP_LOSSLESS AV_CODEC_CAP_LOSSLESS - -/** - * HWAccel is experimental and is thus avoided in favor of non experimental - * codecs - */ -#define HWACCEL_CODEC_CAP_EXPERIMENTAL 0x0200 -#endif /* FF_API_WITHOUT_PREFIX */ - -#if FF_API_MB_TYPE -//The following defines may change, don't expect compatibility if you use them. -#define MB_TYPE_INTRA4x4 0x0001 -#define MB_TYPE_INTRA16x16 0x0002 //FIXME H.264-specific -#define MB_TYPE_INTRA_PCM 0x0004 //FIXME H.264-specific -#define MB_TYPE_16x16 0x0008 -#define MB_TYPE_16x8 0x0010 -#define MB_TYPE_8x16 0x0020 -#define MB_TYPE_8x8 0x0040 -#define MB_TYPE_INTERLACED 0x0080 -#define MB_TYPE_DIRECT2 0x0100 //FIXME -#define MB_TYPE_ACPRED 0x0200 -#define MB_TYPE_GMC 0x0400 -#define MB_TYPE_SKIP 0x0800 -#define MB_TYPE_P0L0 0x1000 -#define MB_TYPE_P1L0 0x2000 -#define MB_TYPE_P0L1 0x4000 -#define MB_TYPE_P1L1 0x8000 -#define MB_TYPE_L0 (MB_TYPE_P0L0 | MB_TYPE_P1L0) -#define MB_TYPE_L1 (MB_TYPE_P0L1 | MB_TYPE_P1L1) -#define MB_TYPE_L0L1 (MB_TYPE_L0 | MB_TYPE_L1) -#define MB_TYPE_QUANT 0x00010000 -#define MB_TYPE_CBP 0x00020000 -//Note bits 24-31 are reserved for codec specific use (h264 ref0, mpeg1 0mv, ...) -#endif - -/** - * Pan Scan area. - * This specifies the area which should be displayed. - * Note there may be multiple such areas for one frame. - */ -typedef struct AVPanScan{ - /** - * id - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - int id; - - /** - * width and height in 1/16 pel - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - int width; - int height; - - /** - * position of the top left corner in 1/16 pel for up to 3 fields/frames - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - int16_t position[3][2]; -}AVPanScan; - -/** - * This structure describes the bitrate properties of an encoded bitstream. It - * roughly corresponds to a subset the VBV parameters for MPEG-2 or HRD - * parameters for H.264/HEVC. - */ -typedef struct AVCPBProperties { - /** - * Maximum bitrate of the stream, in bits per second. - * Zero if unknown or unspecified. - */ - int max_bitrate; - /** - * Minimum bitrate of the stream, in bits per second. - * Zero if unknown or unspecified. - */ - int min_bitrate; - /** - * Average bitrate of the stream, in bits per second. - * Zero if unknown or unspecified. - */ - int avg_bitrate; - - /** - * The size of the buffer to which the ratecontrol is applied, in bits. - * Zero if unknown or unspecified. - */ - int buffer_size; - - /** - * The delay between the time the packet this structure is associated with - * is received and the time when it should be decoded, in periods of a 27MHz - * clock. - * - * UINT64_MAX when unknown or unspecified. - */ - uint64_t vbv_delay; -} AVCPBProperties; - -#if FF_API_QSCALE_TYPE -#define FF_QSCALE_TYPE_MPEG1 0 -#define FF_QSCALE_TYPE_MPEG2 1 -#define FF_QSCALE_TYPE_H264 2 -#define FF_QSCALE_TYPE_VP56 3 -#endif - -/** - * The decoder will keep a reference to the frame and may reuse it later. - */ -#define AV_GET_BUFFER_FLAG_REF (1 << 0) - -/** - * @defgroup lavc_packet AVPacket - * - * Types and functions for working with AVPacket. - * @{ - */ -enum AVPacketSideDataType { - AV_PKT_DATA_PALETTE, - AV_PKT_DATA_NEW_EXTRADATA, - - /** - * An AV_PKT_DATA_PARAM_CHANGE side data packet is laid out as follows: - * @code - * u32le param_flags - * if (param_flags & AV_SIDE_DATA_PARAM_CHANGE_CHANNEL_COUNT) - * s32le channel_count - * if (param_flags & AV_SIDE_DATA_PARAM_CHANGE_CHANNEL_LAYOUT) - * u64le channel_layout - * if (param_flags & AV_SIDE_DATA_PARAM_CHANGE_SAMPLE_RATE) - * s32le sample_rate - * if (param_flags & AV_SIDE_DATA_PARAM_CHANGE_DIMENSIONS) - * s32le width - * s32le height - * @endcode - */ - AV_PKT_DATA_PARAM_CHANGE, - - /** - * An AV_PKT_DATA_H263_MB_INFO side data packet contains a number of - * structures with info about macroblocks relevant to splitting the - * packet into smaller packets on macroblock edges (e.g. as for RFC 2190). - * That is, it does not necessarily contain info about all macroblocks, - * as long as the distance between macroblocks in the info is smaller - * than the target payload size. - * Each MB info structure is 12 bytes, and is laid out as follows: - * @code - * u32le bit offset from the start of the packet - * u8 current quantizer at the start of the macroblock - * u8 GOB number - * u16le macroblock address within the GOB - * u8 horizontal MV predictor - * u8 vertical MV predictor - * u8 horizontal MV predictor for block number 3 - * u8 vertical MV predictor for block number 3 - * @endcode - */ - AV_PKT_DATA_H263_MB_INFO, - - /** - * This side data should be associated with an audio stream and contains - * ReplayGain information in form of the AVReplayGain struct. - */ - AV_PKT_DATA_REPLAYGAIN, - - /** - * This side data contains a 3x3 transformation matrix describing an affine - * transformation that needs to be applied to the decoded video frames for - * correct presentation. - * - * See libavutil/display.h for a detailed description of the data. - */ - AV_PKT_DATA_DISPLAYMATRIX, - - /** - * This side data should be associated with a video stream and contains - * Stereoscopic 3D information in form of the AVStereo3D struct. - */ - AV_PKT_DATA_STEREO3D, - - /** - * This side data should be associated with an audio stream and corresponds - * to enum AVAudioServiceType. - */ - AV_PKT_DATA_AUDIO_SERVICE_TYPE, - - /** - * This side data contains quality related information from the encoder. - * @code - * u32le quality factor of the compressed frame. Allowed range is between 1 (good) and FF_LAMBDA_MAX (bad). - * u8 picture type - * u8 error count - * u16 reserved - * u64le[error count] sum of squared differences between encoder in and output - * @endcode - */ - AV_PKT_DATA_QUALITY_STATS, - - /** - * This side data contains an integer value representing the stream index - * of a "fallback" track. A fallback track indicates an alternate - * track to use when the current track can not be decoded for some reason. - * e.g. no decoder available for codec. - */ - AV_PKT_DATA_FALLBACK_TRACK, - - /** - * This side data corresponds to the AVCPBProperties struct. - */ - AV_PKT_DATA_CPB_PROPERTIES, - - /** - * Recommmends skipping the specified number of samples - * @code - * u32le number of samples to skip from start of this packet - * u32le number of samples to skip from end of this packet - * u8 reason for start skip - * u8 reason for end skip (0=padding silence, 1=convergence) - * @endcode - */ - AV_PKT_DATA_SKIP_SAMPLES=70, - - /** - * An AV_PKT_DATA_JP_DUALMONO side data packet indicates that - * the packet may contain "dual mono" audio specific to Japanese DTV - * and if it is true, recommends only the selected channel to be used. - * @code - * u8 selected channels (0=mail/left, 1=sub/right, 2=both) - * @endcode - */ - AV_PKT_DATA_JP_DUALMONO, - - /** - * A list of zero terminated key/value strings. There is no end marker for - * the list, so it is required to rely on the side data size to stop. - */ - AV_PKT_DATA_STRINGS_METADATA, - - /** - * Subtitle event position - * @code - * u32le x1 - * u32le y1 - * u32le x2 - * u32le y2 - * @endcode - */ - AV_PKT_DATA_SUBTITLE_POSITION, - - /** - * Data found in BlockAdditional element of matroska container. There is - * no end marker for the data, so it is required to rely on the side data - * size to recognize the end. 8 byte id (as found in BlockAddId) followed - * by data. - */ - AV_PKT_DATA_MATROSKA_BLOCKADDITIONAL, - - /** - * The optional first identifier line of a WebVTT cue. - */ - AV_PKT_DATA_WEBVTT_IDENTIFIER, - - /** - * The optional settings (rendering instructions) that immediately - * follow the timestamp specifier of a WebVTT cue. - */ - AV_PKT_DATA_WEBVTT_SETTINGS, - - /** - * A list of zero terminated key/value strings. There is no end marker for - * the list, so it is required to rely on the side data size to stop. This - * side data includes updated metadata which appeared in the stream. - */ - AV_PKT_DATA_METADATA_UPDATE, -}; - -#define AV_PKT_DATA_QUALITY_FACTOR AV_PKT_DATA_QUALITY_STATS //DEPRECATED - -typedef struct AVPacketSideData { - uint8_t *data; - int size; - enum AVPacketSideDataType type; -} AVPacketSideData; - -/** - * This structure stores compressed data. It is typically exported by demuxers - * and then passed as input to decoders, or received as output from encoders and - * then passed to muxers. - * - * For video, it should typically contain one compressed frame. For audio it may - * contain several compressed frames. Encoders are allowed to output empty - * packets, with no compressed data, containing only side data - * (e.g. to update some stream parameters at the end of encoding). - * - * AVPacket is one of the few structs in FFmpeg, whose size is a part of public - * ABI. Thus it may be allocated on stack and no new fields can be added to it - * without libavcodec and libavformat major bump. - * - * The semantics of data ownership depends on the buf field. - * If it is set, the packet data is dynamically allocated and is - * valid indefinitely until a call to av_packet_unref() reduces the - * reference count to 0. - * - * If the buf field is not set av_packet_ref() would make a copy instead - * of increasing the reference count. - * - * The side data is always allocated with av_malloc(), copied by - * av_packet_ref() and freed by av_packet_unref(). - * - * @see av_packet_ref - * @see av_packet_unref - */ -typedef struct AVPacket { - /** - * A reference to the reference-counted buffer where the packet data is - * stored. - * May be NULL, then the packet data is not reference-counted. - */ - AVBufferRef *buf; - /** - * Presentation timestamp in AVStream->time_base units; the time at which - * the decompressed packet will be presented to the user. - * Can be AV_NOPTS_VALUE if it is not stored in the file. - * pts MUST be larger or equal to dts as presentation cannot happen before - * decompression, unless one wants to view hex dumps. Some formats misuse - * the terms dts and pts/cts to mean something different. Such timestamps - * must be converted to true pts/dts before they are stored in AVPacket. - */ - int64_t pts; - /** - * Decompression timestamp in AVStream->time_base units; the time at which - * the packet is decompressed. - * Can be AV_NOPTS_VALUE if it is not stored in the file. - */ - int64_t dts; - uint8_t *data; - int size; - int stream_index; - /** - * A combination of AV_PKT_FLAG values - */ - int flags; - /** - * Additional packet data that can be provided by the container. - * Packet can contain several types of side information. - */ - AVPacketSideData *side_data; - int side_data_elems; - - /** - * Duration of this packet in AVStream->time_base units, 0 if unknown. - * Equals next_pts - this_pts in presentation order. - */ - int64_t duration; - - int64_t pos; ///< byte position in stream, -1 if unknown - -#if FF_API_CONVERGENCE_DURATION - /** - * @deprecated Same as the duration field, but as int64_t. This was required - * for Matroska subtitles, whose duration values could overflow when the - * duration field was still an int. - */ - attribute_deprecated - int64_t convergence_duration; -#endif -} AVPacket; -#define AV_PKT_FLAG_KEY 0x0001 ///< The packet contains a keyframe -#define AV_PKT_FLAG_CORRUPT 0x0002 ///< The packet content is corrupted - -enum AVSideDataParamChangeFlags { - AV_SIDE_DATA_PARAM_CHANGE_CHANNEL_COUNT = 0x0001, - AV_SIDE_DATA_PARAM_CHANGE_CHANNEL_LAYOUT = 0x0002, - AV_SIDE_DATA_PARAM_CHANGE_SAMPLE_RATE = 0x0004, - AV_SIDE_DATA_PARAM_CHANGE_DIMENSIONS = 0x0008, -}; -/** - * @} - */ - -struct AVCodecInternal; - -enum AVFieldOrder { - AV_FIELD_UNKNOWN, - AV_FIELD_PROGRESSIVE, - AV_FIELD_TT, //< Top coded_first, top displayed first - AV_FIELD_BB, //< Bottom coded first, bottom displayed first - AV_FIELD_TB, //< Top coded first, bottom displayed first - AV_FIELD_BT, //< Bottom coded first, top displayed first -}; - -/** - * main external API structure. - * New fields can be added to the end with minor version bumps. - * Removal, reordering and changes to existing fields require a major - * version bump. - * Please use AVOptions (av_opt* / av_set/get*()) to access these fields from user - * applications. - * sizeof(AVCodecContext) must not be used outside libav*. - */ -typedef struct AVCodecContext { - /** - * information on struct for av_log - * - set by avcodec_alloc_context3 - */ - const AVClass *av_class; - int log_level_offset; - - enum AVMediaType codec_type; /* see AVMEDIA_TYPE_xxx */ - const struct AVCodec *codec; -#if FF_API_CODEC_NAME - /** - * @deprecated this field is not used for anything in libavcodec - */ - attribute_deprecated - char codec_name[32]; -#endif - enum AVCodecID codec_id; /* see AV_CODEC_ID_xxx */ - - /** - * fourcc (LSB first, so "ABCD" -> ('D'<<24) + ('C'<<16) + ('B'<<8) + 'A'). - * This is used to work around some encoder bugs. - * A demuxer should set this to what is stored in the field used to identify the codec. - * If there are multiple such fields in a container then the demuxer should choose the one - * which maximizes the information about the used codec. - * If the codec tag field in a container is larger than 32 bits then the demuxer should - * remap the longer ID to 32 bits with a table or other structure. Alternatively a new - * extra_codec_tag + size could be added but for this a clear advantage must be demonstrated - * first. - * - encoding: Set by user, if not then the default based on codec_id will be used. - * - decoding: Set by user, will be converted to uppercase by libavcodec during init. - */ - unsigned int codec_tag; - -#if FF_API_STREAM_CODEC_TAG - /** - * @deprecated this field is unused - */ - attribute_deprecated - unsigned int stream_codec_tag; -#endif - - void *priv_data; - - /** - * Private context used for internal data. - * - * Unlike priv_data, this is not codec-specific. It is used in general - * libavcodec functions. - */ - struct AVCodecInternal *internal; - - /** - * Private data of the user, can be used to carry app specific stuff. - * - encoding: Set by user. - * - decoding: Set by user. - */ - void *opaque; - - /** - * the average bitrate - * - encoding: Set by user; unused for constant quantizer encoding. - * - decoding: Set by user, may be overwritten by libavcodec - * if this info is available in the stream - */ - int64_t bit_rate; - - /** - * number of bits the bitstream is allowed to diverge from the reference. - * the reference can be CBR (for CBR pass1) or VBR (for pass2) - * - encoding: Set by user; unused for constant quantizer encoding. - * - decoding: unused - */ - int bit_rate_tolerance; - - /** - * Global quality for codecs which cannot change it per frame. - * This should be proportional to MPEG-1/2/4 qscale. - * - encoding: Set by user. - * - decoding: unused - */ - int global_quality; - - /** - * - encoding: Set by user. - * - decoding: unused - */ - int compression_level; -#define FF_COMPRESSION_DEFAULT -1 - - /** - * AV_CODEC_FLAG_*. - * - encoding: Set by user. - * - decoding: Set by user. - */ - int flags; - - /** - * AV_CODEC_FLAG2_* - * - encoding: Set by user. - * - decoding: Set by user. - */ - int flags2; - - /** - * some codecs need / can use extradata like Huffman tables. - * mjpeg: Huffman tables - * rv10: additional flags - * mpeg4: global headers (they can be in the bitstream or here) - * The allocated memory should be AV_INPUT_BUFFER_PADDING_SIZE bytes larger - * than extradata_size to avoid problems if it is read with the bitstream reader. - * The bytewise contents of extradata must not depend on the architecture or CPU endianness. - * - encoding: Set/allocated/freed by libavcodec. - * - decoding: Set/allocated/freed by user. - */ - uint8_t *extradata; - int extradata_size; - - /** - * This is the fundamental unit of time (in seconds) in terms - * of which frame timestamps are represented. For fixed-fps content, - * timebase should be 1/framerate and timestamp increments should be - * identically 1. - * This often, but not always is the inverse of the frame rate or field rate - * for video. - * - encoding: MUST be set by user. - * - decoding: the use of this field for decoding is deprecated. - * Use framerate instead. - */ - AVRational time_base; - - /** - * For some codecs, the time base is closer to the field rate than the frame rate. - * Most notably, H.264 and MPEG-2 specify time_base as half of frame duration - * if no telecine is used ... - * - * Set to time_base ticks per frame. Default 1, e.g., H.264/MPEG-2 set it to 2. - */ - int ticks_per_frame; - - /** - * Codec delay. - * - * Encoding: Number of frames delay there will be from the encoder input to - * the decoder output. (we assume the decoder matches the spec) - * Decoding: Number of frames delay in addition to what a standard decoder - * as specified in the spec would produce. - * - * Video: - * Number of frames the decoded output will be delayed relative to the - * encoded input. - * - * Audio: - * For encoding, this field is unused (see initial_padding). - * - * For decoding, this is the number of samples the decoder needs to - * output before the decoder's output is valid. When seeking, you should - * start decoding this many samples prior to your desired seek point. - * - * - encoding: Set by libavcodec. - * - decoding: Set by libavcodec. - */ - int delay; - - - /* video only */ - /** - * picture width / height. - * - * @note Those fields may not match the values of the last - * AVFrame outputted by avcodec_decode_video2 due frame - * reordering. - * - * - encoding: MUST be set by user. - * - decoding: May be set by the user before opening the decoder if known e.g. - * from the container. Some decoders will require the dimensions - * to be set by the caller. During decoding, the decoder may - * overwrite those values as required while parsing the data. - */ - int width, height; - - /** - * Bitstream width / height, may be different from width/height e.g. when - * the decoded frame is cropped before being output or lowres is enabled. - * - * @note Those field may not match the value of the last - * AVFrame outputted by avcodec_decode_video2 due frame - * reordering. - * - * - encoding: unused - * - decoding: May be set by the user before opening the decoder if known - * e.g. from the container. During decoding, the decoder may - * overwrite those values as required while parsing the data. - */ - int coded_width, coded_height; - -#if FF_API_ASPECT_EXTENDED -#define FF_ASPECT_EXTENDED 15 -#endif - - /** - * the number of pictures in a group of pictures, or 0 for intra_only - * - encoding: Set by user. - * - decoding: unused - */ - int gop_size; - - /** - * Pixel format, see AV_PIX_FMT_xxx. - * May be set by the demuxer if known from headers. - * May be overridden by the decoder if it knows better. - * - * @note This field may not match the value of the last - * AVFrame outputted by avcodec_decode_video2 due frame - * reordering. - * - * - encoding: Set by user. - * - decoding: Set by user if known, overridden by libavcodec while - * parsing the data. - */ - enum AVPixelFormat pix_fmt; - -#if FF_API_MOTION_EST - /** - * This option does nothing - * @deprecated use codec private options instead - */ - attribute_deprecated int me_method; -#endif - - /** - * If non NULL, 'draw_horiz_band' is called by the libavcodec - * decoder to draw a horizontal band. It improves cache usage. Not - * all codecs can do that. You must check the codec capabilities - * beforehand. - * When multithreading is used, it may be called from multiple threads - * at the same time; threads might draw different parts of the same AVFrame, - * or multiple AVFrames, and there is no guarantee that slices will be drawn - * in order. - * The function is also used by hardware acceleration APIs. - * It is called at least once during frame decoding to pass - * the data needed for hardware render. - * In that mode instead of pixel data, AVFrame points to - * a structure specific to the acceleration API. The application - * reads the structure and can change some fields to indicate progress - * or mark state. - * - encoding: unused - * - decoding: Set by user. - * @param height the height of the slice - * @param y the y position of the slice - * @param type 1->top field, 2->bottom field, 3->frame - * @param offset offset into the AVFrame.data from which the slice should be read - */ - void (*draw_horiz_band)(struct AVCodecContext *s, - const AVFrame *src, int offset[AV_NUM_DATA_POINTERS], - int y, int type, int height); - - /** - * callback to negotiate the pixelFormat - * @param fmt is the list of formats which are supported by the codec, - * it is terminated by -1 as 0 is a valid format, the formats are ordered by quality. - * The first is always the native one. - * @note The callback may be called again immediately if initialization for - * the selected (hardware-accelerated) pixel format failed. - * @warning Behavior is undefined if the callback returns a value not - * in the fmt list of formats. - * @return the chosen format - * - encoding: unused - * - decoding: Set by user, if not set the native format will be chosen. - */ - enum AVPixelFormat (*get_format)(struct AVCodecContext *s, const enum AVPixelFormat * fmt); - - /** - * maximum number of B-frames between non-B-frames - * Note: The output will be delayed by max_b_frames+1 relative to the input. - * - encoding: Set by user. - * - decoding: unused - */ - int max_b_frames; - - /** - * qscale factor between IP and B-frames - * If > 0 then the last P-frame quantizer will be used (q= lastp_q*factor+offset). - * If < 0 then normal ratecontrol will be done (q= -normal_q*factor+offset). - * - encoding: Set by user. - * - decoding: unused - */ - float b_quant_factor; - -#if FF_API_RC_STRATEGY - /** @deprecated use codec private option instead */ - attribute_deprecated int rc_strategy; -#define FF_RC_STRATEGY_XVID 1 -#endif - - int b_frame_strategy; - - /** - * qscale offset between IP and B-frames - * - encoding: Set by user. - * - decoding: unused - */ - float b_quant_offset; - - /** - * Size of the frame reordering buffer in the decoder. - * For MPEG-2 it is 1 IPB or 0 low delay IP. - * - encoding: Set by libavcodec. - * - decoding: Set by libavcodec. - */ - int has_b_frames; - - /** - * 0-> h263 quant 1-> mpeg quant - * - encoding: Set by user. - * - decoding: unused - */ - int mpeg_quant; - - /** - * qscale factor between P and I-frames - * If > 0 then the last p frame quantizer will be used (q= lastp_q*factor+offset). - * If < 0 then normal ratecontrol will be done (q= -normal_q*factor+offset). - * - encoding: Set by user. - * - decoding: unused - */ - float i_quant_factor; - - /** - * qscale offset between P and I-frames - * - encoding: Set by user. - * - decoding: unused - */ - float i_quant_offset; - - /** - * luminance masking (0-> disabled) - * - encoding: Set by user. - * - decoding: unused - */ - float lumi_masking; - - /** - * temporary complexity masking (0-> disabled) - * - encoding: Set by user. - * - decoding: unused - */ - float temporal_cplx_masking; - - /** - * spatial complexity masking (0-> disabled) - * - encoding: Set by user. - * - decoding: unused - */ - float spatial_cplx_masking; - - /** - * p block masking (0-> disabled) - * - encoding: Set by user. - * - decoding: unused - */ - float p_masking; - - /** - * darkness masking (0-> disabled) - * - encoding: Set by user. - * - decoding: unused - */ - float dark_masking; - - /** - * slice count - * - encoding: Set by libavcodec. - * - decoding: Set by user (or 0). - */ - int slice_count; - /** - * prediction method (needed for huffyuv) - * - encoding: Set by user. - * - decoding: unused - */ - int prediction_method; -#define FF_PRED_LEFT 0 -#define FF_PRED_PLANE 1 -#define FF_PRED_MEDIAN 2 - - /** - * slice offsets in the frame in bytes - * - encoding: Set/allocated by libavcodec. - * - decoding: Set/allocated by user (or NULL). - */ - int *slice_offset; - - /** - * sample aspect ratio (0 if unknown) - * That is the width of a pixel divided by the height of the pixel. - * Numerator and denominator must be relatively prime and smaller than 256 for some video standards. - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - AVRational sample_aspect_ratio; - - /** - * motion estimation comparison function - * - encoding: Set by user. - * - decoding: unused - */ - int me_cmp; - /** - * subpixel motion estimation comparison function - * - encoding: Set by user. - * - decoding: unused - */ - int me_sub_cmp; - /** - * macroblock comparison function (not supported yet) - * - encoding: Set by user. - * - decoding: unused - */ - int mb_cmp; - /** - * interlaced DCT comparison function - * - encoding: Set by user. - * - decoding: unused - */ - int ildct_cmp; -#define FF_CMP_SAD 0 -#define FF_CMP_SSE 1 -#define FF_CMP_SATD 2 -#define FF_CMP_DCT 3 -#define FF_CMP_PSNR 4 -#define FF_CMP_BIT 5 -#define FF_CMP_RD 6 -#define FF_CMP_ZERO 7 -#define FF_CMP_VSAD 8 -#define FF_CMP_VSSE 9 -#define FF_CMP_NSSE 10 -#define FF_CMP_W53 11 -#define FF_CMP_W97 12 -#define FF_CMP_DCTMAX 13 -#define FF_CMP_DCT264 14 -#define FF_CMP_CHROMA 256 - - /** - * ME diamond size & shape - * - encoding: Set by user. - * - decoding: unused - */ - int dia_size; - - /** - * amount of previous MV predictors (2a+1 x 2a+1 square) - * - encoding: Set by user. - * - decoding: unused - */ - int last_predictor_count; - - /** - * prepass for motion estimation - * - encoding: Set by user. - * - decoding: unused - */ - int pre_me; - - /** - * motion estimation prepass comparison function - * - encoding: Set by user. - * - decoding: unused - */ - int me_pre_cmp; - - /** - * ME prepass diamond size & shape - * - encoding: Set by user. - * - decoding: unused - */ - int pre_dia_size; - - /** - * subpel ME quality - * - encoding: Set by user. - * - decoding: unused - */ - int me_subpel_quality; - -#if FF_API_AFD - /** - * DTG active format information (additional aspect ratio - * information only used in DVB MPEG-2 transport streams) - * 0 if not set. - * - * - encoding: unused - * - decoding: Set by decoder. - * @deprecated Deprecated in favor of AVSideData - */ - attribute_deprecated int dtg_active_format; -#define FF_DTG_AFD_SAME 8 -#define FF_DTG_AFD_4_3 9 -#define FF_DTG_AFD_16_9 10 -#define FF_DTG_AFD_14_9 11 -#define FF_DTG_AFD_4_3_SP_14_9 13 -#define FF_DTG_AFD_16_9_SP_14_9 14 -#define FF_DTG_AFD_SP_4_3 15 -#endif /* FF_API_AFD */ - - /** - * maximum motion estimation search range in subpel units - * If 0 then no limit. - * - * - encoding: Set by user. - * - decoding: unused - */ - int me_range; - -#if FF_API_QUANT_BIAS - /** - * @deprecated use encoder private option instead - */ - attribute_deprecated int intra_quant_bias; -#define FF_DEFAULT_QUANT_BIAS 999999 - - /** - * @deprecated use encoder private option instead - */ - attribute_deprecated int inter_quant_bias; -#endif - - /** - * slice flags - * - encoding: unused - * - decoding: Set by user. - */ - int slice_flags; -#define SLICE_FLAG_CODED_ORDER 0x0001 ///< draw_horiz_band() is called in coded order instead of display -#define SLICE_FLAG_ALLOW_FIELD 0x0002 ///< allow draw_horiz_band() with field slices (MPEG2 field pics) -#define SLICE_FLAG_ALLOW_PLANE 0x0004 ///< allow draw_horiz_band() with 1 component at a time (SVQ1) - -#if FF_API_XVMC - /** - * XVideo Motion Acceleration - * - encoding: forbidden - * - decoding: set by decoder - * @deprecated XvMC doesn't need it anymore. - */ - attribute_deprecated int xvmc_acceleration; -#endif /* FF_API_XVMC */ - - /** - * macroblock decision mode - * - encoding: Set by user. - * - decoding: unused - */ - int mb_decision; -#define FF_MB_DECISION_SIMPLE 0 ///< uses mb_cmp -#define FF_MB_DECISION_BITS 1 ///< chooses the one which needs the fewest bits -#define FF_MB_DECISION_RD 2 ///< rate distortion - - /** - * custom intra quantization matrix - * - encoding: Set by user, can be NULL. - * - decoding: Set by libavcodec. - */ - uint16_t *intra_matrix; - - /** - * custom inter quantization matrix - * - encoding: Set by user, can be NULL. - * - decoding: Set by libavcodec. - */ - uint16_t *inter_matrix; - - /** - * scene change detection threshold - * 0 is default, larger means fewer detected scene changes. - * - encoding: Set by user. - * - decoding: unused - */ - int scenechange_threshold; - - /** - * noise reduction strength - * - encoding: Set by user. - * - decoding: unused - */ - int noise_reduction; - -#if FF_API_MPV_OPT - /** - * @deprecated this field is unused - */ - attribute_deprecated - int me_threshold; - - /** - * @deprecated this field is unused - */ - attribute_deprecated - int mb_threshold; -#endif - - /** - * precision of the intra DC coefficient - 8 - * - encoding: Set by user. - * - decoding: Set by libavcodec - */ - int intra_dc_precision; - - /** - * Number of macroblock rows at the top which are skipped. - * - encoding: unused - * - decoding: Set by user. - */ - int skip_top; - - /** - * Number of macroblock rows at the bottom which are skipped. - * - encoding: unused - * - decoding: Set by user. - */ - int skip_bottom; - -#if FF_API_MPV_OPT - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - float border_masking; -#endif - - /** - * minimum MB lagrange multipler - * - encoding: Set by user. - * - decoding: unused - */ - int mb_lmin; - - /** - * maximum MB lagrange multipler - * - encoding: Set by user. - * - decoding: unused - */ - int mb_lmax; - - /** - * - * - encoding: Set by user. - * - decoding: unused - */ - int me_penalty_compensation; - - /** - * - * - encoding: Set by user. - * - decoding: unused - */ - int bidir_refine; - - /** - * - * - encoding: Set by user. - * - decoding: unused - */ - int brd_scale; - - /** - * minimum GOP size - * - encoding: Set by user. - * - decoding: unused - */ - int keyint_min; - - /** - * number of reference frames - * - encoding: Set by user. - * - decoding: Set by lavc. - */ - int refs; - - /** - * chroma qp offset from luma - * - encoding: Set by user. - * - decoding: unused - */ - int chromaoffset; - -#if FF_API_UNUSED_MEMBERS - /** - * Multiplied by qscale for each frame and added to scene_change_score. - * - encoding: Set by user. - * - decoding: unused - */ - attribute_deprecated int scenechange_factor; -#endif - - /** - * - * Note: Value depends upon the compare function used for fullpel ME. - * - encoding: Set by user. - * - decoding: unused - */ - int mv0_threshold; - - /** - * Adjust sensitivity of b_frame_strategy 1. - * - encoding: Set by user. - * - decoding: unused - */ - int b_sensitivity; - - /** - * Chromaticity coordinates of the source primaries. - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVColorPrimaries color_primaries; - - /** - * Color Transfer Characteristic. - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVColorTransferCharacteristic color_trc; - - /** - * YUV colorspace type. - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVColorSpace colorspace; - - /** - * MPEG vs JPEG YUV range. - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVColorRange color_range; - - /** - * This defines the location of chroma samples. - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVChromaLocation chroma_sample_location; - - /** - * Number of slices. - * Indicates number of picture subdivisions. Used for parallelized - * decoding. - * - encoding: Set by user - * - decoding: unused - */ - int slices; - - /** Field order - * - encoding: set by libavcodec - * - decoding: Set by user. - */ - enum AVFieldOrder field_order; - - /* audio only */ - int sample_rate; ///< samples per second - int channels; ///< number of audio channels - - /** - * audio sample format - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - enum AVSampleFormat sample_fmt; ///< sample format - - /* The following data should not be initialized. */ - /** - * Number of samples per channel in an audio frame. - * - * - encoding: set by libavcodec in avcodec_open2(). Each submitted frame - * except the last must contain exactly frame_size samples per channel. - * May be 0 when the codec has AV_CODEC_CAP_VARIABLE_FRAME_SIZE set, then the - * frame size is not restricted. - * - decoding: may be set by some decoders to indicate constant frame size - */ - int frame_size; - - /** - * Frame counter, set by libavcodec. - * - * - decoding: total number of frames returned from the decoder so far. - * - encoding: total number of frames passed to the encoder so far. - * - * @note the counter is not incremented if encoding/decoding resulted in - * an error. - */ - int frame_number; - - /** - * number of bytes per packet if constant and known or 0 - * Used by some WAV based audio codecs. - */ - int block_align; - - /** - * Audio cutoff bandwidth (0 means "automatic") - * - encoding: Set by user. - * - decoding: unused - */ - int cutoff; - - /** - * Audio channel layout. - * - encoding: set by user. - * - decoding: set by user, may be overwritten by libavcodec. - */ - uint64_t channel_layout; - - /** - * Request decoder to use this channel layout if it can (0 for default) - * - encoding: unused - * - decoding: Set by user. - */ - uint64_t request_channel_layout; - - /** - * Type of service that the audio stream conveys. - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - enum AVAudioServiceType audio_service_type; - - /** - * desired sample format - * - encoding: Not used. - * - decoding: Set by user. - * Decoder will decode to this format if it can. - */ - enum AVSampleFormat request_sample_fmt; - - /** - * This callback is called at the beginning of each frame to get data - * buffer(s) for it. There may be one contiguous buffer for all the data or - * there may be a buffer per each data plane or anything in between. What - * this means is, you may set however many entries in buf[] you feel necessary. - * Each buffer must be reference-counted using the AVBuffer API (see description - * of buf[] below). - * - * The following fields will be set in the frame before this callback is - * called: - * - format - * - width, height (video only) - * - sample_rate, channel_layout, nb_samples (audio only) - * Their values may differ from the corresponding values in - * AVCodecContext. This callback must use the frame values, not the codec - * context values, to calculate the required buffer size. - * - * This callback must fill the following fields in the frame: - * - data[] - * - linesize[] - * - extended_data: - * * if the data is planar audio with more than 8 channels, then this - * callback must allocate and fill extended_data to contain all pointers - * to all data planes. data[] must hold as many pointers as it can. - * extended_data must be allocated with av_malloc() and will be freed in - * av_frame_unref(). - * * otherwise exended_data must point to data - * - buf[] must contain one or more pointers to AVBufferRef structures. Each of - * the frame's data and extended_data pointers must be contained in these. That - * is, one AVBufferRef for each allocated chunk of memory, not necessarily one - * AVBufferRef per data[] entry. See: av_buffer_create(), av_buffer_alloc(), - * and av_buffer_ref(). - * - extended_buf and nb_extended_buf must be allocated with av_malloc() by - * this callback and filled with the extra buffers if there are more - * buffers than buf[] can hold. extended_buf will be freed in - * av_frame_unref(). - * - * If AV_CODEC_CAP_DR1 is not set then get_buffer2() must call - * avcodec_default_get_buffer2() instead of providing buffers allocated by - * some other means. - * - * Each data plane must be aligned to the maximum required by the target - * CPU. - * - * @see avcodec_default_get_buffer2() - * - * Video: - * - * If AV_GET_BUFFER_FLAG_REF is set in flags then the frame may be reused - * (read and/or written to if it is writable) later by libavcodec. - * - * avcodec_align_dimensions2() should be used to find the required width and - * height, as they normally need to be rounded up to the next multiple of 16. - * - * Some decoders do not support linesizes changing between frames. - * - * If frame multithreading is used and thread_safe_callbacks is set, - * this callback may be called from a different thread, but not from more - * than one at once. Does not need to be reentrant. - * - * @see avcodec_align_dimensions2() - * - * Audio: - * - * Decoders request a buffer of a particular size by setting - * AVFrame.nb_samples prior to calling get_buffer2(). The decoder may, - * however, utilize only part of the buffer by setting AVFrame.nb_samples - * to a smaller value in the output frame. - * - * As a convenience, av_samples_get_buffer_size() and - * av_samples_fill_arrays() in libavutil may be used by custom get_buffer2() - * functions to find the required data size and to fill data pointers and - * linesize. In AVFrame.linesize, only linesize[0] may be set for audio - * since all planes must be the same size. - * - * @see av_samples_get_buffer_size(), av_samples_fill_arrays() - * - * - encoding: unused - * - decoding: Set by libavcodec, user can override. - */ - int (*get_buffer2)(struct AVCodecContext *s, AVFrame *frame, int flags); - - /** - * If non-zero, the decoded audio and video frames returned from - * avcodec_decode_video2() and avcodec_decode_audio4() are reference-counted - * and are valid indefinitely. The caller must free them with - * av_frame_unref() when they are not needed anymore. - * Otherwise, the decoded frames must not be freed by the caller and are - * only valid until the next decode call. - * - * - encoding: unused - * - decoding: set by the caller before avcodec_open2(). - */ - int refcounted_frames; - - /* - encoding parameters */ - float qcompress; ///< amount of qscale change between easy & hard scenes (0.0-1.0) - float qblur; ///< amount of qscale smoothing over time (0.0-1.0) - - /** - * minimum quantizer - * - encoding: Set by user. - * - decoding: unused - */ - int qmin; - - /** - * maximum quantizer - * - encoding: Set by user. - * - decoding: unused - */ - int qmax; - - /** - * maximum quantizer difference between frames - * - encoding: Set by user. - * - decoding: unused - */ - int max_qdiff; - -#if FF_API_MPV_OPT - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - float rc_qsquish; - - attribute_deprecated - float rc_qmod_amp; - attribute_deprecated - int rc_qmod_freq; -#endif - - /** - * decoder bitstream buffer size - * - encoding: Set by user. - * - decoding: unused - */ - int rc_buffer_size; - - /** - * ratecontrol override, see RcOverride - * - encoding: Allocated/set/freed by user. - * - decoding: unused - */ - int rc_override_count; - RcOverride *rc_override; - -#if FF_API_MPV_OPT - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - const char *rc_eq; -#endif - - /** - * maximum bitrate - * - encoding: Set by user. - * - decoding: Set by user, may be overwritten by libavcodec. - */ - int64_t rc_max_rate; - - /** - * minimum bitrate - * - encoding: Set by user. - * - decoding: unused - */ - int64_t rc_min_rate; - -#if FF_API_MPV_OPT - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - float rc_buffer_aggressivity; - - attribute_deprecated - float rc_initial_cplx; -#endif - - /** - * Ratecontrol attempt to use, at maximum, <value> of what can be used without an underflow. - * - encoding: Set by user. - * - decoding: unused. - */ - float rc_max_available_vbv_use; - - /** - * Ratecontrol attempt to use, at least, <value> times the amount needed to prevent a vbv overflow. - * - encoding: Set by user. - * - decoding: unused. - */ - float rc_min_vbv_overflow_use; - - /** - * Number of bits which should be loaded into the rc buffer before decoding starts. - * - encoding: Set by user. - * - decoding: unused - */ - int rc_initial_buffer_occupancy; - -#if FF_API_CODER_TYPE -#define FF_CODER_TYPE_VLC 0 -#define FF_CODER_TYPE_AC 1 -#define FF_CODER_TYPE_RAW 2 -#define FF_CODER_TYPE_RLE 3 -#if FF_API_UNUSED_MEMBERS -#define FF_CODER_TYPE_DEFLATE 4 -#endif /* FF_API_UNUSED_MEMBERS */ - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - int coder_type; -#endif /* FF_API_CODER_TYPE */ - - /** - * context model - * - encoding: Set by user. - * - decoding: unused - */ - int context_model; - -#if FF_API_MPV_OPT - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - int lmin; - - /** - * @deprecated use encoder private options instead - */ - attribute_deprecated - int lmax; -#endif - - /** - * frame skip threshold - * - encoding: Set by user. - * - decoding: unused - */ - int frame_skip_threshold; - - /** - * frame skip factor - * - encoding: Set by user. - * - decoding: unused - */ - int frame_skip_factor; - - /** - * frame skip exponent - * - encoding: Set by user. - * - decoding: unused - */ - int frame_skip_exp; - - /** - * frame skip comparison function - * - encoding: Set by user. - * - decoding: unused - */ - int frame_skip_cmp; - - /** - * trellis RD quantization - * - encoding: Set by user. - * - decoding: unused - */ - int trellis; - - /** - * - encoding: Set by user. - * - decoding: unused - */ - int min_prediction_order; - - /** - * - encoding: Set by user. - * - decoding: unused - */ - int max_prediction_order; - - /** - * GOP timecode frame start number - * - encoding: Set by user, in non drop frame format - * - decoding: Set by libavcodec (timecode in the 25 bits format, -1 if unset) - */ - int64_t timecode_frame_start; - -#if FF_API_RTP_CALLBACK - /** - * @deprecated unused - */ - /* The RTP callback: This function is called */ - /* every time the encoder has a packet to send. */ - /* It depends on the encoder if the data starts */ - /* with a Start Code (it should). H.263 does. */ - /* mb_nb contains the number of macroblocks */ - /* encoded in the RTP payload. */ - attribute_deprecated - void (*rtp_callback)(struct AVCodecContext *avctx, void *data, int size, int mb_nb); -#endif - - int rtp_payload_size; /* The size of the RTP payload: the coder will */ - /* do its best to deliver a chunk with size */ - /* below rtp_payload_size, the chunk will start */ - /* with a start code on some codecs like H.263. */ - /* This doesn't take account of any particular */ - /* headers inside the transmitted RTP payload. */ - -#if FF_API_STAT_BITS - /* statistics, used for 2-pass encoding */ - attribute_deprecated - int mv_bits; - attribute_deprecated - int header_bits; - attribute_deprecated - int i_tex_bits; - attribute_deprecated - int p_tex_bits; - attribute_deprecated - int i_count; - attribute_deprecated - int p_count; - attribute_deprecated - int skip_count; - attribute_deprecated - int misc_bits; - - /** @deprecated this field is unused */ - attribute_deprecated - int frame_bits; -#endif - - /** - * pass1 encoding statistics output buffer - * - encoding: Set by libavcodec. - * - decoding: unused - */ - char *stats_out; - - /** - * pass2 encoding statistics input buffer - * Concatenated stuff from stats_out of pass1 should be placed here. - * - encoding: Allocated/set/freed by user. - * - decoding: unused - */ - char *stats_in; - - /** - * Work around bugs in encoders which sometimes cannot be detected automatically. - * - encoding: Set by user - * - decoding: Set by user - */ - int workaround_bugs; -#define FF_BUG_AUTODETECT 1 ///< autodetection -#if FF_API_OLD_MSMPEG4 -#define FF_BUG_OLD_MSMPEG4 2 -#endif -#define FF_BUG_XVID_ILACE 4 -#define FF_BUG_UMP4 8 -#define FF_BUG_NO_PADDING 16 -#define FF_BUG_AMV 32 -#if FF_API_AC_VLC -#define FF_BUG_AC_VLC 0 ///< Will be removed, libavcodec can now handle these non-compliant files by default. -#endif -#define FF_BUG_QPEL_CHROMA 64 -#define FF_BUG_STD_QPEL 128 -#define FF_BUG_QPEL_CHROMA2 256 -#define FF_BUG_DIRECT_BLOCKSIZE 512 -#define FF_BUG_EDGE 1024 -#define FF_BUG_HPEL_CHROMA 2048 -#define FF_BUG_DC_CLIP 4096 -#define FF_BUG_MS 8192 ///< Work around various bugs in Microsoft's broken decoders. -#define FF_BUG_TRUNCATED 16384 - - /** - * strictly follow the standard (MPEG4, ...). - * - encoding: Set by user. - * - decoding: Set by user. - * Setting this to STRICT or higher means the encoder and decoder will - * generally do stupid things, whereas setting it to unofficial or lower - * will mean the encoder might produce output that is not supported by all - * spec-compliant decoders. Decoders don't differentiate between normal, - * unofficial and experimental (that is, they always try to decode things - * when they can) unless they are explicitly asked to behave stupidly - * (=strictly conform to the specs) - */ - int strict_std_compliance; -#define FF_COMPLIANCE_VERY_STRICT 2 ///< Strictly conform to an older more strict version of the spec or reference software. -#define FF_COMPLIANCE_STRICT 1 ///< Strictly conform to all the things in the spec no matter what consequences. -#define FF_COMPLIANCE_NORMAL 0 -#define FF_COMPLIANCE_UNOFFICIAL -1 ///< Allow unofficial extensions -#define FF_COMPLIANCE_EXPERIMENTAL -2 ///< Allow nonstandardized experimental things. - - /** - * error concealment flags - * - encoding: unused - * - decoding: Set by user. - */ - int error_concealment; -#define FF_EC_GUESS_MVS 1 -#define FF_EC_DEBLOCK 2 -#define FF_EC_FAVOR_INTER 256 - - /** - * debug - * - encoding: Set by user. - * - decoding: Set by user. - */ - int debug; -#define FF_DEBUG_PICT_INFO 1 -#define FF_DEBUG_RC 2 -#define FF_DEBUG_BITSTREAM 4 -#define FF_DEBUG_MB_TYPE 8 -#define FF_DEBUG_QP 16 -#if FF_API_DEBUG_MV -/** - * @deprecated this option does nothing - */ -#define FF_DEBUG_MV 32 -#endif -#define FF_DEBUG_DCT_COEFF 0x00000040 -#define FF_DEBUG_SKIP 0x00000080 -#define FF_DEBUG_STARTCODE 0x00000100 -#if FF_API_UNUSED_MEMBERS -#define FF_DEBUG_PTS 0x00000200 -#endif /* FF_API_UNUSED_MEMBERS */ -#define FF_DEBUG_ER 0x00000400 -#define FF_DEBUG_MMCO 0x00000800 -#define FF_DEBUG_BUGS 0x00001000 -#if FF_API_DEBUG_MV -#define FF_DEBUG_VIS_QP 0x00002000 ///< only access through AVOptions from outside libavcodec -#define FF_DEBUG_VIS_MB_TYPE 0x00004000 ///< only access through AVOptions from outside libavcodec -#endif -#define FF_DEBUG_BUFFERS 0x00008000 -#define FF_DEBUG_THREADS 0x00010000 -#define FF_DEBUG_GREEN_MD 0x00800000 -#define FF_DEBUG_NOMC 0x01000000 - -#if FF_API_DEBUG_MV - /** - * debug - * Code outside libavcodec should access this field using AVOptions - * - encoding: Set by user. - * - decoding: Set by user. - */ - int debug_mv; -#define FF_DEBUG_VIS_MV_P_FOR 0x00000001 //visualize forward predicted MVs of P frames -#define FF_DEBUG_VIS_MV_B_FOR 0x00000002 //visualize forward predicted MVs of B frames -#define FF_DEBUG_VIS_MV_B_BACK 0x00000004 //visualize backward predicted MVs of B frames -#endif - - /** - * Error recognition; may misdetect some more or less valid parts as errors. - * - encoding: unused - * - decoding: Set by user. - */ - int err_recognition; - -/** - * Verify checksums embedded in the bitstream (could be of either encoded or - * decoded data, depending on the codec) and print an error message on mismatch. - * If AV_EF_EXPLODE is also set, a mismatching checksum will result in the - * decoder returning an error. - */ -#define AV_EF_CRCCHECK (1<<0) -#define AV_EF_BITSTREAM (1<<1) ///< detect bitstream specification deviations -#define AV_EF_BUFFER (1<<2) ///< detect improper bitstream length -#define AV_EF_EXPLODE (1<<3) ///< abort decoding on minor error detection - -#define AV_EF_IGNORE_ERR (1<<15) ///< ignore errors and continue -#define AV_EF_CAREFUL (1<<16) ///< consider things that violate the spec, are fast to calculate and have not been seen in the wild as errors -#define AV_EF_COMPLIANT (1<<17) ///< consider all spec non compliances as errors -#define AV_EF_AGGRESSIVE (1<<18) ///< consider things that a sane encoder should not do as an error - - - /** - * opaque 64bit number (generally a PTS) that will be reordered and - * output in AVFrame.reordered_opaque - * - encoding: unused - * - decoding: Set by user. - */ - int64_t reordered_opaque; - - /** - * Hardware accelerator in use - * - encoding: unused. - * - decoding: Set by libavcodec - */ - struct AVHWAccel *hwaccel; - - /** - * Hardware accelerator context. - * For some hardware accelerators, a global context needs to be - * provided by the user. In that case, this holds display-dependent - * data FFmpeg cannot instantiate itself. Please refer to the - * FFmpeg HW accelerator documentation to know how to fill this - * is. e.g. for VA API, this is a struct vaapi_context. - * - encoding: unused - * - decoding: Set by user - */ - void *hwaccel_context; - - /** - * error - * - encoding: Set by libavcodec if flags & AV_CODEC_FLAG_PSNR. - * - decoding: unused - */ - uint64_t error[AV_NUM_DATA_POINTERS]; - - /** - * DCT algorithm, see FF_DCT_* below - * - encoding: Set by user. - * - decoding: unused - */ - int dct_algo; -#define FF_DCT_AUTO 0 -#define FF_DCT_FASTINT 1 -#define FF_DCT_INT 2 -#define FF_DCT_MMX 3 -#define FF_DCT_ALTIVEC 5 -#define FF_DCT_FAAN 6 - - /** - * IDCT algorithm, see FF_IDCT_* below. - * - encoding: Set by user. - * - decoding: Set by user. - */ - int idct_algo; -#define FF_IDCT_AUTO 0 -#define FF_IDCT_INT 1 -#define FF_IDCT_SIMPLE 2 -#define FF_IDCT_SIMPLEMMX 3 -#define FF_IDCT_ARM 7 -#define FF_IDCT_ALTIVEC 8 -#if FF_API_ARCH_SH4 -#define FF_IDCT_SH4 9 -#endif -#define FF_IDCT_SIMPLEARM 10 -#if FF_API_UNUSED_MEMBERS -#define FF_IDCT_IPP 13 -#endif /* FF_API_UNUSED_MEMBERS */ -#define FF_IDCT_XVID 14 -#if FF_API_IDCT_XVIDMMX -#define FF_IDCT_XVIDMMX 14 -#endif /* FF_API_IDCT_XVIDMMX */ -#define FF_IDCT_SIMPLEARMV5TE 16 -#define FF_IDCT_SIMPLEARMV6 17 -#if FF_API_ARCH_SPARC -#define FF_IDCT_SIMPLEVIS 18 -#endif -#define FF_IDCT_FAAN 20 -#define FF_IDCT_SIMPLENEON 22 -#if FF_API_ARCH_ALPHA -#define FF_IDCT_SIMPLEALPHA 23 -#endif -#define FF_IDCT_SIMPLEAUTO 128 - - /** - * bits per sample/pixel from the demuxer (needed for huffyuv). - * - encoding: Set by libavcodec. - * - decoding: Set by user. - */ - int bits_per_coded_sample; - - /** - * Bits per sample/pixel of internal libavcodec pixel/sample format. - * - encoding: set by user. - * - decoding: set by libavcodec. - */ - int bits_per_raw_sample; - -#if FF_API_LOWRES - /** - * low resolution decoding, 1-> 1/2 size, 2->1/4 size - * - encoding: unused - * - decoding: Set by user. - * Code outside libavcodec should access this field using: - * av_codec_{get,set}_lowres(avctx) - */ - int lowres; -#endif - -#if FF_API_CODED_FRAME - /** - * the picture in the bitstream - * - encoding: Set by libavcodec. - * - decoding: unused - * - * @deprecated use the quality factor packet side data instead - */ - attribute_deprecated AVFrame *coded_frame; -#endif - - /** - * thread count - * is used to decide how many independent tasks should be passed to execute() - * - encoding: Set by user. - * - decoding: Set by user. - */ - int thread_count; - - /** - * Which multithreading methods to use. - * Use of FF_THREAD_FRAME will increase decoding delay by one frame per thread, - * so clients which cannot provide future frames should not use it. - * - * - encoding: Set by user, otherwise the default is used. - * - decoding: Set by user, otherwise the default is used. - */ - int thread_type; -#define FF_THREAD_FRAME 1 ///< Decode more than one frame at once -#define FF_THREAD_SLICE 2 ///< Decode more than one part of a single frame at once - - /** - * Which multithreading methods are in use by the codec. - * - encoding: Set by libavcodec. - * - decoding: Set by libavcodec. - */ - int active_thread_type; - - /** - * Set by the client if its custom get_buffer() callback can be called - * synchronously from another thread, which allows faster multithreaded decoding. - * draw_horiz_band() will be called from other threads regardless of this setting. - * Ignored if the default get_buffer() is used. - * - encoding: Set by user. - * - decoding: Set by user. - */ - int thread_safe_callbacks; - - /** - * The codec may call this to execute several independent things. - * It will return only after finishing all tasks. - * The user may replace this with some multithreaded implementation, - * the default implementation will execute the parts serially. - * @param count the number of things to execute - * - encoding: Set by libavcodec, user can override. - * - decoding: Set by libavcodec, user can override. - */ - int (*execute)(struct AVCodecContext *c, int (*func)(struct AVCodecContext *c2, void *arg), void *arg2, int *ret, int count, int size); - - /** - * The codec may call this to execute several independent things. - * It will return only after finishing all tasks. - * The user may replace this with some multithreaded implementation, - * the default implementation will execute the parts serially. - * Also see avcodec_thread_init and e.g. the --enable-pthread configure option. - * @param c context passed also to func - * @param count the number of things to execute - * @param arg2 argument passed unchanged to func - * @param ret return values of executed functions, must have space for "count" values. May be NULL. - * @param func function that will be called count times, with jobnr from 0 to count-1. - * threadnr will be in the range 0 to c->thread_count-1 < MAX_THREADS and so that no - * two instances of func executing at the same time will have the same threadnr. - * @return always 0 currently, but code should handle a future improvement where when any call to func - * returns < 0 no further calls to func may be done and < 0 is returned. - * - encoding: Set by libavcodec, user can override. - * - decoding: Set by libavcodec, user can override. - */ - int (*execute2)(struct AVCodecContext *c, int (*func)(struct AVCodecContext *c2, void *arg, int jobnr, int threadnr), void *arg2, int *ret, int count); - - /** - * noise vs. sse weight for the nsse comparison function - * - encoding: Set by user. - * - decoding: unused - */ - int nsse_weight; - - /** - * profile - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - int profile; -#define FF_PROFILE_UNKNOWN -99 -#define FF_PROFILE_RESERVED -100 - -#define FF_PROFILE_AAC_MAIN 0 -#define FF_PROFILE_AAC_LOW 1 -#define FF_PROFILE_AAC_SSR 2 -#define FF_PROFILE_AAC_LTP 3 -#define FF_PROFILE_AAC_HE 4 -#define FF_PROFILE_AAC_HE_V2 28 -#define FF_PROFILE_AAC_LD 22 -#define FF_PROFILE_AAC_ELD 38 -#define FF_PROFILE_MPEG2_AAC_LOW 128 -#define FF_PROFILE_MPEG2_AAC_HE 131 - -#define FF_PROFILE_DTS 20 -#define FF_PROFILE_DTS_ES 30 -#define FF_PROFILE_DTS_96_24 40 -#define FF_PROFILE_DTS_HD_HRA 50 -#define FF_PROFILE_DTS_HD_MA 60 -#define FF_PROFILE_DTS_EXPRESS 70 - -#define FF_PROFILE_MPEG2_422 0 -#define FF_PROFILE_MPEG2_HIGH 1 -#define FF_PROFILE_MPEG2_SS 2 -#define FF_PROFILE_MPEG2_SNR_SCALABLE 3 -#define FF_PROFILE_MPEG2_MAIN 4 -#define FF_PROFILE_MPEG2_SIMPLE 5 - -#define FF_PROFILE_H264_CONSTRAINED (1<<9) // 8+1; constraint_set1_flag -#define FF_PROFILE_H264_INTRA (1<<11) // 8+3; constraint_set3_flag - -#define FF_PROFILE_H264_BASELINE 66 -#define FF_PROFILE_H264_CONSTRAINED_BASELINE (66|FF_PROFILE_H264_CONSTRAINED) -#define FF_PROFILE_H264_MAIN 77 -#define FF_PROFILE_H264_EXTENDED 88 -#define FF_PROFILE_H264_HIGH 100 -#define FF_PROFILE_H264_HIGH_10 110 -#define FF_PROFILE_H264_HIGH_10_INTRA (110|FF_PROFILE_H264_INTRA) -#define FF_PROFILE_H264_HIGH_422 122 -#define FF_PROFILE_H264_HIGH_422_INTRA (122|FF_PROFILE_H264_INTRA) -#define FF_PROFILE_H264_HIGH_444 144 -#define FF_PROFILE_H264_HIGH_444_PREDICTIVE 244 -#define FF_PROFILE_H264_HIGH_444_INTRA (244|FF_PROFILE_H264_INTRA) -#define FF_PROFILE_H264_CAVLC_444 44 - -#define FF_PROFILE_VC1_SIMPLE 0 -#define FF_PROFILE_VC1_MAIN 1 -#define FF_PROFILE_VC1_COMPLEX 2 -#define FF_PROFILE_VC1_ADVANCED 3 - -#define FF_PROFILE_MPEG4_SIMPLE 0 -#define FF_PROFILE_MPEG4_SIMPLE_SCALABLE 1 -#define FF_PROFILE_MPEG4_CORE 2 -#define FF_PROFILE_MPEG4_MAIN 3 -#define FF_PROFILE_MPEG4_N_BIT 4 -#define FF_PROFILE_MPEG4_SCALABLE_TEXTURE 5 -#define FF_PROFILE_MPEG4_SIMPLE_FACE_ANIMATION 6 -#define FF_PROFILE_MPEG4_BASIC_ANIMATED_TEXTURE 7 -#define FF_PROFILE_MPEG4_HYBRID 8 -#define FF_PROFILE_MPEG4_ADVANCED_REAL_TIME 9 -#define FF_PROFILE_MPEG4_CORE_SCALABLE 10 -#define FF_PROFILE_MPEG4_ADVANCED_CODING 11 -#define FF_PROFILE_MPEG4_ADVANCED_CORE 12 -#define FF_PROFILE_MPEG4_ADVANCED_SCALABLE_TEXTURE 13 -#define FF_PROFILE_MPEG4_SIMPLE_STUDIO 14 -#define FF_PROFILE_MPEG4_ADVANCED_SIMPLE 15 - -#define FF_PROFILE_JPEG2000_CSTREAM_RESTRICTION_0 0 -#define FF_PROFILE_JPEG2000_CSTREAM_RESTRICTION_1 1 -#define FF_PROFILE_JPEG2000_CSTREAM_NO_RESTRICTION 2 -#define FF_PROFILE_JPEG2000_DCINEMA_2K 3 -#define FF_PROFILE_JPEG2000_DCINEMA_4K 4 - -#define FF_PROFILE_VP9_0 0 -#define FF_PROFILE_VP9_1 1 -#define FF_PROFILE_VP9_2 2 -#define FF_PROFILE_VP9_3 3 - -#define FF_PROFILE_HEVC_MAIN 1 -#define FF_PROFILE_HEVC_MAIN_10 2 -#define FF_PROFILE_HEVC_MAIN_STILL_PICTURE 3 -#define FF_PROFILE_HEVC_REXT 4 - - /** - * level - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - int level; -#define FF_LEVEL_UNKNOWN -99 - - /** - * Skip loop filtering for selected frames. - * - encoding: unused - * - decoding: Set by user. - */ - enum AVDiscard skip_loop_filter; - - /** - * Skip IDCT/dequantization for selected frames. - * - encoding: unused - * - decoding: Set by user. - */ - enum AVDiscard skip_idct; - - /** - * Skip decoding for selected frames. - * - encoding: unused - * - decoding: Set by user. - */ - enum AVDiscard skip_frame; - - /** - * Header containing style information for text subtitles. - * For SUBTITLE_ASS subtitle type, it should contain the whole ASS - * [Script Info] and [V4+ Styles] section, plus the [Events] line and - * the Format line following. It shouldn't include any Dialogue line. - * - encoding: Set/allocated/freed by user (before avcodec_open2()) - * - decoding: Set/allocated/freed by libavcodec (by avcodec_open2()) - */ - uint8_t *subtitle_header; - int subtitle_header_size; - -#if FF_API_ERROR_RATE - /** - * @deprecated use the 'error_rate' private AVOption of the mpegvideo - * encoders - */ - attribute_deprecated - int error_rate; -#endif - -#if FF_API_VBV_DELAY - /** - * VBV delay coded in the last frame (in periods of a 27 MHz clock). - * Used for compliant TS muxing. - * - encoding: Set by libavcodec. - * - decoding: unused. - * @deprecated this value is now exported as a part of - * AV_PKT_DATA_CPB_PROPERTIES packet side data - */ - attribute_deprecated - uint64_t vbv_delay; -#endif - -#if FF_API_SIDEDATA_ONLY_PKT - /** - * Encoding only and set by default. Allow encoders to output packets - * that do not contain any encoded data, only side data. - * - * Some encoders need to output such packets, e.g. to update some stream - * parameters at the end of encoding. - * - * @deprecated this field disables the default behaviour and - * it is kept only for compatibility. - */ - attribute_deprecated - int side_data_only_packets; -#endif - - /** - * Audio only. The number of "priming" samples (padding) inserted by the - * encoder at the beginning of the audio. I.e. this number of leading - * decoded samples must be discarded by the caller to get the original audio - * without leading padding. - * - * - decoding: unused - * - encoding: Set by libavcodec. The timestamps on the output packets are - * adjusted by the encoder so that they always refer to the - * first sample of the data actually contained in the packet, - * including any added padding. E.g. if the timebase is - * 1/samplerate and the timestamp of the first input sample is - * 0, the timestamp of the first output packet will be - * -initial_padding. - */ - int initial_padding; - - /** - * - decoding: For codecs that store a framerate value in the compressed - * bitstream, the decoder may export it here. { 0, 1} when - * unknown. - * - encoding: unused - */ - AVRational framerate; - - /** - * Nominal unaccelerated pixel format, see AV_PIX_FMT_xxx. - * - encoding: unused. - * - decoding: Set by libavcodec before calling get_format() - */ - enum AVPixelFormat sw_pix_fmt; - - /** - * Timebase in which pkt_dts/pts and AVPacket.dts/pts are. - * Code outside libavcodec should access this field using: - * av_codec_{get,set}_pkt_timebase(avctx) - * - encoding unused. - * - decoding set by user. - */ - AVRational pkt_timebase; - - /** - * AVCodecDescriptor - * Code outside libavcodec should access this field using: - * av_codec_{get,set}_codec_descriptor(avctx) - * - encoding: unused. - * - decoding: set by libavcodec. - */ - const AVCodecDescriptor *codec_descriptor; - -#if !FF_API_LOWRES - /** - * low resolution decoding, 1-> 1/2 size, 2->1/4 size - * - encoding: unused - * - decoding: Set by user. - * Code outside libavcodec should access this field using: - * av_codec_{get,set}_lowres(avctx) - */ - int lowres; -#endif - - /** - * Current statistics for PTS correction. - * - decoding: maintained and used by libavcodec, not intended to be used by user apps - * - encoding: unused - */ - int64_t pts_correction_num_faulty_pts; /// Number of incorrect PTS values so far - int64_t pts_correction_num_faulty_dts; /// Number of incorrect DTS values so far - int64_t pts_correction_last_pts; /// PTS of the last frame - int64_t pts_correction_last_dts; /// DTS of the last frame - - /** - * Character encoding of the input subtitles file. - * - decoding: set by user - * - encoding: unused - */ - char *sub_charenc; - - /** - * Subtitles character encoding mode. Formats or codecs might be adjusting - * this setting (if they are doing the conversion themselves for instance). - * - decoding: set by libavcodec - * - encoding: unused - */ - int sub_charenc_mode; -#define FF_SUB_CHARENC_MODE_DO_NOTHING -1 ///< do nothing (demuxer outputs a stream supposed to be already in UTF-8, or the codec is bitmap for instance) -#define FF_SUB_CHARENC_MODE_AUTOMATIC 0 ///< libavcodec will select the mode itself -#define FF_SUB_CHARENC_MODE_PRE_DECODER 1 ///< the AVPacket data needs to be recoded to UTF-8 before being fed to the decoder, requires iconv - - /** - * Skip processing alpha if supported by codec. - * Note that if the format uses pre-multiplied alpha (common with VP6, - * and recommended due to better video quality/compression) - * the image will look as if alpha-blended onto a black background. - * However for formats that do not use pre-multiplied alpha - * there might be serious artefacts (though e.g. libswscale currently - * assumes pre-multiplied alpha anyway). - * Code outside libavcodec should access this field using AVOptions - * - * - decoding: set by user - * - encoding: unused - */ - int skip_alpha; - - /** - * Number of samples to skip after a discontinuity - * - decoding: unused - * - encoding: set by libavcodec - */ - int seek_preroll; - -#if !FF_API_DEBUG_MV - /** - * debug motion vectors - * Code outside libavcodec should access this field using AVOptions - * - encoding: Set by user. - * - decoding: Set by user. - */ - int debug_mv; -#define FF_DEBUG_VIS_MV_P_FOR 0x00000001 //visualize forward predicted MVs of P frames -#define FF_DEBUG_VIS_MV_B_FOR 0x00000002 //visualize forward predicted MVs of B frames -#define FF_DEBUG_VIS_MV_B_BACK 0x00000004 //visualize backward predicted MVs of B frames -#endif - - /** - * custom intra quantization matrix - * Code outside libavcodec should access this field using av_codec_g/set_chroma_intra_matrix() - * - encoding: Set by user, can be NULL. - * - decoding: unused. - */ - uint16_t *chroma_intra_matrix; - - /** - * dump format separator. - * can be ", " or "\n " or anything else - * Code outside libavcodec should access this field using AVOptions - * (NO direct access). - * - encoding: Set by user. - * - decoding: Set by user. - */ - uint8_t *dump_separator; - - /** - * ',' separated list of allowed decoders. - * If NULL then all are allowed - * - encoding: unused - * - decoding: set by user through AVOPtions (NO direct access) - */ - char *codec_whitelist; - - /* - * Properties of the stream that gets decoded - * To be accessed through av_codec_get_properties() (NO direct access) - * - encoding: unused - * - decoding: set by libavcodec - */ - unsigned properties; -#define FF_CODEC_PROPERTY_LOSSLESS 0x00000001 -#define FF_CODEC_PROPERTY_CLOSED_CAPTIONS 0x00000002 - - /** - * Additional data associated with the entire coded stream. - * - * - decoding: unused - * - encoding: may be set by libavcodec after avcodec_open2(). - */ - AVPacketSideData *coded_side_data; - int nb_coded_side_data; - -} AVCodecContext; - -AVRational av_codec_get_pkt_timebase (const AVCodecContext *avctx); -void av_codec_set_pkt_timebase (AVCodecContext *avctx, AVRational val); - -const AVCodecDescriptor *av_codec_get_codec_descriptor(const AVCodecContext *avctx); -void av_codec_set_codec_descriptor(AVCodecContext *avctx, const AVCodecDescriptor *desc); - -unsigned av_codec_get_codec_properties(const AVCodecContext *avctx); - -int av_codec_get_lowres(const AVCodecContext *avctx); -void av_codec_set_lowres(AVCodecContext *avctx, int val); - -int av_codec_get_seek_preroll(const AVCodecContext *avctx); -void av_codec_set_seek_preroll(AVCodecContext *avctx, int val); - -uint16_t *av_codec_get_chroma_intra_matrix(const AVCodecContext *avctx); -void av_codec_set_chroma_intra_matrix(AVCodecContext *avctx, uint16_t *val); - -/** - * AVProfile. - */ -typedef struct AVProfile { - int profile; - const char *name; ///< short name for the profile -} AVProfile; - -typedef struct AVCodecDefault AVCodecDefault; - -struct AVSubtitle; - -/** - * AVCodec. - */ -typedef struct AVCodec { - /** - * Name of the codec implementation. - * The name is globally unique among encoders and among decoders (but an - * encoder and a decoder can share the same name). - * This is the primary way to find a codec from the user perspective. - */ - const char *name; - /** - * Descriptive name for the codec, meant to be more human readable than name. - * You should use the NULL_IF_CONFIG_SMALL() macro to define it. - */ - const char *long_name; - enum AVMediaType type; - enum AVCodecID id; - /** - * Codec capabilities. - * see AV_CODEC_CAP_* - */ - int capabilities; - const AVRational *supported_framerates; ///< array of supported framerates, or NULL if any, array is terminated by {0,0} - const enum AVPixelFormat *pix_fmts; ///< array of supported pixel formats, or NULL if unknown, array is terminated by -1 - const int *supported_samplerates; ///< array of supported audio samplerates, or NULL if unknown, array is terminated by 0 - const enum AVSampleFormat *sample_fmts; ///< array of supported sample formats, or NULL if unknown, array is terminated by -1 - const uint64_t *channel_layouts; ///< array of support channel layouts, or NULL if unknown. array is terminated by 0 - uint8_t max_lowres; ///< maximum value for lowres supported by the decoder, no direct access, use av_codec_get_max_lowres() - const AVClass *priv_class; ///< AVClass for the private context - const AVProfile *profiles; ///< array of recognized profiles, or NULL if unknown, array is terminated by {FF_PROFILE_UNKNOWN} - - /***************************************************************** - * No fields below this line are part of the public API. They - * may not be used outside of libavcodec and can be changed and - * removed at will. - * New public fields should be added right above. - ***************************************************************** - */ - int priv_data_size; - struct AVCodec *next; - /** - * @name Frame-level threading support functions - * @{ - */ - /** - * If defined, called on thread contexts when they are created. - * If the codec allocates writable tables in init(), re-allocate them here. - * priv_data will be set to a copy of the original. - */ - int (*init_thread_copy)(AVCodecContext *); - /** - * Copy necessary context variables from a previous thread context to the current one. - * If not defined, the next thread will start automatically; otherwise, the codec - * must call ff_thread_finish_setup(). - * - * dst and src will (rarely) point to the same context, in which case memcpy should be skipped. - */ - int (*update_thread_context)(AVCodecContext *dst, const AVCodecContext *src); - /** @} */ - - /** - * Private codec-specific defaults. - */ - const AVCodecDefault *defaults; - - /** - * Initialize codec static data, called from avcodec_register(). - */ - void (*init_static_data)(struct AVCodec *codec); - - int (*init)(AVCodecContext *); - int (*encode_sub)(AVCodecContext *, uint8_t *buf, int buf_size, - const struct AVSubtitle *sub); - /** - * Encode data to an AVPacket. - * - * @param avctx codec context - * @param avpkt output AVPacket (may contain a user-provided buffer) - * @param[in] frame AVFrame containing the raw data to be encoded - * @param[out] got_packet_ptr encoder sets to 0 or 1 to indicate that a - * non-empty packet was returned in avpkt. - * @return 0 on success, negative error code on failure - */ - int (*encode2)(AVCodecContext *avctx, AVPacket *avpkt, const AVFrame *frame, - int *got_packet_ptr); - int (*decode)(AVCodecContext *, void *outdata, int *outdata_size, AVPacket *avpkt); - int (*close)(AVCodecContext *); - /** - * Flush buffers. - * Will be called when seeking - */ - void (*flush)(AVCodecContext *); - /** - * Internal codec capabilities. - * See FF_CODEC_CAP_* in internal.h - */ - int caps_internal; -} AVCodec; - -int av_codec_get_max_lowres(const AVCodec *codec); - -struct MpegEncContext; - -/** - * @defgroup lavc_hwaccel AVHWAccel - * @{ - */ -typedef struct AVHWAccel { - /** - * Name of the hardware accelerated codec. - * The name is globally unique among encoders and among decoders (but an - * encoder and a decoder can share the same name). - */ - const char *name; - - /** - * Type of codec implemented by the hardware accelerator. - * - * See AVMEDIA_TYPE_xxx - */ - enum AVMediaType type; - - /** - * Codec implemented by the hardware accelerator. - * - * See AV_CODEC_ID_xxx - */ - enum AVCodecID id; - - /** - * Supported pixel format. - * - * Only hardware accelerated formats are supported here. - */ - enum AVPixelFormat pix_fmt; - - /** - * Hardware accelerated codec capabilities. - * see HWACCEL_CODEC_CAP_* - */ - int capabilities; - - /***************************************************************** - * No fields below this line are part of the public API. They - * may not be used outside of libavcodec and can be changed and - * removed at will. - * New public fields should be added right above. - ***************************************************************** - */ - struct AVHWAccel *next; - - /** - * Allocate a custom buffer - */ - int (*alloc_frame)(AVCodecContext *avctx, AVFrame *frame); - - /** - * Called at the beginning of each frame or field picture. - * - * Meaningful frame information (codec specific) is guaranteed to - * be parsed at this point. This function is mandatory. - * - * Note that buf can be NULL along with buf_size set to 0. - * Otherwise, this means the whole frame is available at this point. - * - * @param avctx the codec context - * @param buf the frame data buffer base - * @param buf_size the size of the frame in bytes - * @return zero if successful, a negative value otherwise - */ - int (*start_frame)(AVCodecContext *avctx, const uint8_t *buf, uint32_t buf_size); - - /** - * Callback for each slice. - * - * Meaningful slice information (codec specific) is guaranteed to - * be parsed at this point. This function is mandatory. - * The only exception is XvMC, that works on MB level. - * - * @param avctx the codec context - * @param buf the slice data buffer base - * @param buf_size the size of the slice in bytes - * @return zero if successful, a negative value otherwise - */ - int (*decode_slice)(AVCodecContext *avctx, const uint8_t *buf, uint32_t buf_size); - - /** - * Called at the end of each frame or field picture. - * - * The whole picture is parsed at this point and can now be sent - * to the hardware accelerator. This function is mandatory. - * - * @param avctx the codec context - * @return zero if successful, a negative value otherwise - */ - int (*end_frame)(AVCodecContext *avctx); - - /** - * Size of per-frame hardware accelerator private data. - * - * Private data is allocated with av_mallocz() before - * AVCodecContext.get_buffer() and deallocated after - * AVCodecContext.release_buffer(). - */ - int frame_priv_data_size; - - /** - * Called for every Macroblock in a slice. - * - * XvMC uses it to replace the ff_mpv_decode_mb(). - * Instead of decoding to raw picture, MB parameters are - * stored in an array provided by the video driver. - * - * @param s the mpeg context - */ - void (*decode_mb)(struct MpegEncContext *s); - - /** - * Initialize the hwaccel private data. - * - * This will be called from ff_get_format(), after hwaccel and - * hwaccel_context are set and the hwaccel private data in AVCodecInternal - * is allocated. - */ - int (*init)(AVCodecContext *avctx); - - /** - * Uninitialize the hwaccel private data. - * - * This will be called from get_format() or avcodec_close(), after hwaccel - * and hwaccel_context are already uninitialized. - */ - int (*uninit)(AVCodecContext *avctx); - - /** - * Size of the private data to allocate in - * AVCodecInternal.hwaccel_priv_data. - */ - int priv_data_size; -} AVHWAccel; - -/** - * Hardware acceleration should be used for decoding even if the codec level - * used is unknown or higher than the maximum supported level reported by the - * hardware driver. - * - * It's generally a good idea to pass this flag unless you have a specific - * reason not to, as hardware tends to under-report supported levels. - */ -#define AV_HWACCEL_FLAG_IGNORE_LEVEL (1 << 0) - -/** - * Hardware acceleration can output YUV pixel formats with a different chroma - * sampling than 4:2:0 and/or other than 8 bits per component. - */ -#define AV_HWACCEL_FLAG_ALLOW_HIGH_DEPTH (1 << 1) - -/** - * @} - */ - -#if FF_API_AVPICTURE -/** - * @defgroup lavc_picture AVPicture - * - * Functions for working with AVPicture - * @{ - */ - -/** - * Picture data structure. - * - * Up to four components can be stored into it, the last component is - * alpha. - * @deprecated use AVFrame or imgutils functions instead - */ -typedef struct AVPicture { - attribute_deprecated - uint8_t *data[AV_NUM_DATA_POINTERS]; ///< pointers to the image data planes - attribute_deprecated - int linesize[AV_NUM_DATA_POINTERS]; ///< number of bytes per line -} AVPicture; - -/** - * @} - */ -#endif - -enum AVSubtitleType { - SUBTITLE_NONE, - - SUBTITLE_BITMAP, ///< A bitmap, pict will be set - - /** - * Plain text, the text field must be set by the decoder and is - * authoritative. ass and pict fields may contain approximations. - */ - SUBTITLE_TEXT, - - /** - * Formatted text, the ass field must be set by the decoder and is - * authoritative. pict and text fields may contain approximations. - */ - SUBTITLE_ASS, -}; - -#define AV_SUBTITLE_FLAG_FORCED 0x00000001 - -typedef struct AVSubtitleRect { - int x; ///< top left corner of pict, undefined when pict is not set - int y; ///< top left corner of pict, undefined when pict is not set - int w; ///< width of pict, undefined when pict is not set - int h; ///< height of pict, undefined when pict is not set - int nb_colors; ///< number of colors in pict, undefined when pict is not set - -#if FF_API_AVPICTURE - /** - * @deprecated unused - */ - attribute_deprecated - AVPicture pict; -#endif - /** - * data+linesize for the bitmap of this subtitle. - * Can be set for text/ass as well once they are rendered. - */ - uint8_t *data[4]; - int linesize[4]; - - enum AVSubtitleType type; - - char *text; ///< 0 terminated plain UTF-8 text - - /** - * 0 terminated ASS/SSA compatible event line. - * The presentation of this is unaffected by the other values in this - * struct. - */ - char *ass; - - int flags; -} AVSubtitleRect; - -typedef struct AVSubtitle { - uint16_t format; /* 0 = graphics */ - uint32_t start_display_time; /* relative to packet pts, in ms */ - uint32_t end_display_time; /* relative to packet pts, in ms */ - unsigned num_rects; - AVSubtitleRect **rects; - int64_t pts; ///< Same as packet pts, in AV_TIME_BASE -} AVSubtitle; - -/** - * If c is NULL, returns the first registered codec, - * if c is non-NULL, returns the next registered codec after c, - * or NULL if c is the last one. - */ -AVCodec *av_codec_next(const AVCodec *c); - -/** - * Return the LIBAVCODEC_VERSION_INT constant. - */ -unsigned avcodec_version(void); - -/** - * Return the libavcodec build-time configuration. - */ -const char *avcodec_configuration(void); - -/** - * Return the libavcodec license. - */ -const char *avcodec_license(void); - -/** - * Register the codec codec and initialize libavcodec. - * - * @warning either this function or avcodec_register_all() must be called - * before any other libavcodec functions. - * - * @see avcodec_register_all() - */ -void avcodec_register(AVCodec *codec); - -/** - * Register all the codecs, parsers and bitstream filters which were enabled at - * configuration time. If you do not call this function you can select exactly - * which formats you want to support, by using the individual registration - * functions. - * - * @see avcodec_register - * @see av_register_codec_parser - * @see av_register_bitstream_filter - */ -void avcodec_register_all(void); - -/** - * Allocate an AVCodecContext and set its fields to default values. The - * resulting struct should be freed with avcodec_free_context(). - * - * @param codec if non-NULL, allocate private data and initialize defaults - * for the given codec. It is illegal to then call avcodec_open2() - * with a different codec. - * If NULL, then the codec-specific defaults won't be initialized, - * which may result in suboptimal default settings (this is - * important mainly for encoders, e.g. libx264). - * - * @return An AVCodecContext filled with default values or NULL on failure. - * @see avcodec_get_context_defaults - */ -AVCodecContext *avcodec_alloc_context3(const AVCodec *codec); - -/** - * Free the codec context and everything associated with it and write NULL to - * the provided pointer. - */ -void avcodec_free_context(AVCodecContext **avctx); - -/** - * Set the fields of the given AVCodecContext to default values corresponding - * to the given codec (defaults may be codec-dependent). - * - * Do not call this function if a non-NULL codec has been passed - * to avcodec_alloc_context3() that allocated this AVCodecContext. - * If codec is non-NULL, it is illegal to call avcodec_open2() with a - * different codec on this AVCodecContext. - */ -int avcodec_get_context_defaults3(AVCodecContext *s, const AVCodec *codec); - -/** - * Get the AVClass for AVCodecContext. It can be used in combination with - * AV_OPT_SEARCH_FAKE_OBJ for examining options. - * - * @see av_opt_find(). - */ -const AVClass *avcodec_get_class(void); - -/** - * Get the AVClass for AVFrame. It can be used in combination with - * AV_OPT_SEARCH_FAKE_OBJ for examining options. - * - * @see av_opt_find(). - */ -const AVClass *avcodec_get_frame_class(void); - -/** - * Get the AVClass for AVSubtitleRect. It can be used in combination with - * AV_OPT_SEARCH_FAKE_OBJ for examining options. - * - * @see av_opt_find(). - */ -const AVClass *avcodec_get_subtitle_rect_class(void); - -/** - * Copy the settings of the source AVCodecContext into the destination - * AVCodecContext. The resulting destination codec context will be - * unopened, i.e. you are required to call avcodec_open2() before you - * can use this AVCodecContext to decode/encode video/audio data. - * - * @param dest target codec context, should be initialized with - * avcodec_alloc_context3(NULL), but otherwise uninitialized - * @param src source codec context - * @return AVERROR() on error (e.g. memory allocation error), 0 on success - */ -int avcodec_copy_context(AVCodecContext *dest, const AVCodecContext *src); - -/** - * Initialize the AVCodecContext to use the given AVCodec. Prior to using this - * function the context has to be allocated with avcodec_alloc_context3(). - * - * The functions avcodec_find_decoder_by_name(), avcodec_find_encoder_by_name(), - * avcodec_find_decoder() and avcodec_find_encoder() provide an easy way for - * retrieving a codec. - * - * @warning This function is not thread safe! - * - * @note Always call this function before using decoding routines (such as - * @ref avcodec_decode_video2()). - * - * @code - * avcodec_register_all(); - * av_dict_set(&opts, "b", "2.5M", 0); - * codec = avcodec_find_decoder(AV_CODEC_ID_H264); - * if (!codec) - * exit(1); - * - * context = avcodec_alloc_context3(codec); - * - * if (avcodec_open2(context, codec, opts) < 0) - * exit(1); - * @endcode - * - * @param avctx The context to initialize. - * @param codec The codec to open this context for. If a non-NULL codec has been - * previously passed to avcodec_alloc_context3() or - * avcodec_get_context_defaults3() for this context, then this - * parameter MUST be either NULL or equal to the previously passed - * codec. - * @param options A dictionary filled with AVCodecContext and codec-private options. - * On return this object will be filled with options that were not found. - * - * @return zero on success, a negative value on error - * @see avcodec_alloc_context3(), avcodec_find_decoder(), avcodec_find_encoder(), - * av_dict_set(), av_opt_find(). - */ -int avcodec_open2(AVCodecContext *avctx, const AVCodec *codec, AVDictionary **options); - -/** - * Close a given AVCodecContext and free all the data associated with it - * (but not the AVCodecContext itself). - * - * Calling this function on an AVCodecContext that hasn't been opened will free - * the codec-specific data allocated in avcodec_alloc_context3() / - * avcodec_get_context_defaults3() with a non-NULL codec. Subsequent calls will - * do nothing. - */ -int avcodec_close(AVCodecContext *avctx); - -/** - * Free all allocated data in the given subtitle struct. - * - * @param sub AVSubtitle to free. - */ -void avsubtitle_free(AVSubtitle *sub); - -/** - * @} - */ - -/** - * @addtogroup lavc_packet - * @{ - */ - -/** - * Allocate an AVPacket and set its fields to default values. The resulting - * struct must be freed using av_packet_free(). - * - * @return An AVPacket filled with default values or NULL on failure. - * - * @note this only allocates the AVPacket itself, not the data buffers. Those - * must be allocated through other means such as av_new_packet. - * - * @see av_new_packet - */ -AVPacket *av_packet_alloc(void); - -/** - * Create a new packet that references the same data as src. - * - * This is a shortcut for av_packet_alloc()+av_packet_ref(). - * - * @return newly created AVPacket on success, NULL on error. - * - * @see av_packet_alloc - * @see av_packet_ref - */ -AVPacket *av_packet_clone(AVPacket *src); - -/** - * Free the packet, if the packet is reference counted, it will be - * unreferenced first. - * - * @param packet packet to be freed. The pointer will be set to NULL. - * @note passing NULL is a no-op. - */ -void av_packet_free(AVPacket **pkt); - -/** - * Initialize optional fields of a packet with default values. - * - * Note, this does not touch the data and size members, which have to be - * initialized separately. - * - * @param pkt packet - */ -void av_init_packet(AVPacket *pkt); - -/** - * Allocate the payload of a packet and initialize its fields with - * default values. - * - * @param pkt packet - * @param size wanted payload size - * @return 0 if OK, AVERROR_xxx otherwise - */ -int av_new_packet(AVPacket *pkt, int size); - -/** - * Reduce packet size, correctly zeroing padding - * - * @param pkt packet - * @param size new size - */ -void av_shrink_packet(AVPacket *pkt, int size); - -/** - * Increase packet size, correctly zeroing padding - * - * @param pkt packet - * @param grow_by number of bytes by which to increase the size of the packet - */ -int av_grow_packet(AVPacket *pkt, int grow_by); - -/** - * Initialize a reference-counted packet from av_malloc()ed data. - * - * @param pkt packet to be initialized. This function will set the data, size, - * buf and destruct fields, all others are left untouched. - * @param data Data allocated by av_malloc() to be used as packet data. If this - * function returns successfully, the data is owned by the underlying AVBuffer. - * The caller may not access the data through other means. - * @param size size of data in bytes, without the padding. I.e. the full buffer - * size is assumed to be size + AV_INPUT_BUFFER_PADDING_SIZE. - * - * @return 0 on success, a negative AVERROR on error - */ -int av_packet_from_data(AVPacket *pkt, uint8_t *data, int size); - -#if FF_API_AVPACKET_OLD_API -/** - * @warning This is a hack - the packet memory allocation stuff is broken. The - * packet is allocated if it was not really allocated. - * - * @deprecated Use av_packet_ref - */ -attribute_deprecated -int av_dup_packet(AVPacket *pkt); -/** - * Copy packet, including contents - * - * @return 0 on success, negative AVERROR on fail - */ -int av_copy_packet(AVPacket *dst, const AVPacket *src); - -/** - * Copy packet side data - * - * @return 0 on success, negative AVERROR on fail - */ -int av_copy_packet_side_data(AVPacket *dst, const AVPacket *src); - -/** - * Free a packet. - * - * @deprecated Use av_packet_unref - * - * @param pkt packet to free - */ -attribute_deprecated -void av_free_packet(AVPacket *pkt); -#endif -/** - * Allocate new information of a packet. - * - * @param pkt packet - * @param type side information type - * @param size side information size - * @return pointer to fresh allocated data or NULL otherwise - */ -uint8_t* av_packet_new_side_data(AVPacket *pkt, enum AVPacketSideDataType type, - int size); - -/** - * Wrap an existing array as a packet side data. - * - * @param pkt packet - * @param type side information type - * @param data the side data array. It must be allocated with the av_malloc() - * family of functions. The ownership of the data is transferred to - * pkt. - * @param size side information size - * @return a non-negative number on success, a negative AVERROR code on - * failure. On failure, the packet is unchanged and the data remains - * owned by the caller. - */ -int av_packet_add_side_data(AVPacket *pkt, enum AVPacketSideDataType type, - uint8_t *data, size_t size); - -/** - * Shrink the already allocated side data buffer - * - * @param pkt packet - * @param type side information type - * @param size new side information size - * @return 0 on success, < 0 on failure - */ -int av_packet_shrink_side_data(AVPacket *pkt, enum AVPacketSideDataType type, - int size); - -/** - * Get side information from packet. - * - * @param pkt packet - * @param type desired side information type - * @param size pointer for side information size to store (optional) - * @return pointer to data if present or NULL otherwise - */ -uint8_t* av_packet_get_side_data(AVPacket *pkt, enum AVPacketSideDataType type, - int *size); - -int av_packet_merge_side_data(AVPacket *pkt); - -int av_packet_split_side_data(AVPacket *pkt); - -const char *av_packet_side_data_name(enum AVPacketSideDataType type); - -/** - * Pack a dictionary for use in side_data. - * - * @param dict The dictionary to pack. - * @param size pointer to store the size of the returned data - * @return pointer to data if successful, NULL otherwise - */ -uint8_t *av_packet_pack_dictionary(AVDictionary *dict, int *size); -/** - * Unpack a dictionary from side_data. - * - * @param data data from side_data - * @param size size of the data - * @param dict the metadata storage dictionary - * @return 0 on success, < 0 on failure - */ -int av_packet_unpack_dictionary(const uint8_t *data, int size, AVDictionary **dict); - - -/** - * Convenience function to free all the side data stored. - * All the other fields stay untouched. - * - * @param pkt packet - */ -void av_packet_free_side_data(AVPacket *pkt); - -/** - * Setup a new reference to the data described by a given packet - * - * If src is reference-counted, setup dst as a new reference to the - * buffer in src. Otherwise allocate a new buffer in dst and copy the - * data from src into it. - * - * All the other fields are copied from src. - * - * @see av_packet_unref - * - * @param dst Destination packet - * @param src Source packet - * - * @return 0 on success, a negative AVERROR on error. - */ -int av_packet_ref(AVPacket *dst, const AVPacket *src); - -/** - * Wipe the packet. - * - * Unreference the buffer referenced by the packet and reset the - * remaining packet fields to their default values. - * - * @param pkt The packet to be unreferenced. - */ -void av_packet_unref(AVPacket *pkt); - -/** - * Move every field in src to dst and reset src. - * - * @see av_packet_unref - * - * @param src Source packet, will be reset - * @param dst Destination packet - */ -void av_packet_move_ref(AVPacket *dst, AVPacket *src); - -/** - * Copy only "properties" fields from src to dst. - * - * Properties for the purpose of this function are all the fields - * beside those related to the packet data (buf, data, size) - * - * @param dst Destination packet - * @param src Source packet - * - * @return 0 on success AVERROR on failure. - * - */ -int av_packet_copy_props(AVPacket *dst, const AVPacket *src); - -/** - * Convert valid timing fields (timestamps / durations) in a packet from one - * timebase to another. Timestamps with unknown values (AV_NOPTS_VALUE) will be - * ignored. - * - * @param pkt packet on which the conversion will be performed - * @param tb_src source timebase, in which the timing fields in pkt are - * expressed - * @param tb_dst destination timebase, to which the timing fields will be - * converted - */ -void av_packet_rescale_ts(AVPacket *pkt, AVRational tb_src, AVRational tb_dst); - -/** - * @} - */ - -/** - * @addtogroup lavc_decoding - * @{ - */ - -/** - * Find a registered decoder with a matching codec ID. - * - * @param id AVCodecID of the requested decoder - * @return A decoder if one was found, NULL otherwise. - */ -AVCodec *avcodec_find_decoder(enum AVCodecID id); - -/** - * Find a registered decoder with the specified name. - * - * @param name name of the requested decoder - * @return A decoder if one was found, NULL otherwise. - */ -AVCodec *avcodec_find_decoder_by_name(const char *name); - -/** - * The default callback for AVCodecContext.get_buffer2(). It is made public so - * it can be called by custom get_buffer2() implementations for decoders without - * AV_CODEC_CAP_DR1 set. - */ -int avcodec_default_get_buffer2(AVCodecContext *s, AVFrame *frame, int flags); - -#if FF_API_EMU_EDGE -/** - * Return the amount of padding in pixels which the get_buffer callback must - * provide around the edge of the image for codecs which do not have the - * CODEC_FLAG_EMU_EDGE flag. - * - * @return Required padding in pixels. - * - * @deprecated CODEC_FLAG_EMU_EDGE is deprecated, so this function is no longer - * needed - */ -attribute_deprecated -unsigned avcodec_get_edge_width(void); -#endif - -/** - * Modify width and height values so that they will result in a memory - * buffer that is acceptable for the codec if you do not use any horizontal - * padding. - * - * May only be used if a codec with AV_CODEC_CAP_DR1 has been opened. - */ -void avcodec_align_dimensions(AVCodecContext *s, int *width, int *height); - -/** - * Modify width and height values so that they will result in a memory - * buffer that is acceptable for the codec if you also ensure that all - * line sizes are a multiple of the respective linesize_align[i]. - * - * May only be used if a codec with AV_CODEC_CAP_DR1 has been opened. - */ -void avcodec_align_dimensions2(AVCodecContext *s, int *width, int *height, - int linesize_align[AV_NUM_DATA_POINTERS]); - -/** - * Converts AVChromaLocation to swscale x/y chroma position. - * - * The positions represent the chroma (0,0) position in a coordinates system - * with luma (0,0) representing the origin and luma(1,1) representing 256,256 - * - * @param xpos horizontal chroma sample position - * @param ypos vertical chroma sample position - */ -int avcodec_enum_to_chroma_pos(int *xpos, int *ypos, enum AVChromaLocation pos); - -/** - * Converts swscale x/y chroma position to AVChromaLocation. - * - * The positions represent the chroma (0,0) position in a coordinates system - * with luma (0,0) representing the origin and luma(1,1) representing 256,256 - * - * @param xpos horizontal chroma sample position - * @param ypos vertical chroma sample position - */ -enum AVChromaLocation avcodec_chroma_pos_to_enum(int xpos, int ypos); - -/** - * Decode the audio frame of size avpkt->size from avpkt->data into frame. - * - * Some decoders may support multiple frames in a single AVPacket. Such - * decoders would then just decode the first frame and the return value would be - * less than the packet size. In this case, avcodec_decode_audio4 has to be - * called again with an AVPacket containing the remaining data in order to - * decode the second frame, etc... Even if no frames are returned, the packet - * needs to be fed to the decoder with remaining data until it is completely - * consumed or an error occurs. - * - * Some decoders (those marked with AV_CODEC_CAP_DELAY) have a delay between input - * and output. This means that for some packets they will not immediately - * produce decoded output and need to be flushed at the end of decoding to get - * all the decoded data. Flushing is done by calling this function with packets - * with avpkt->data set to NULL and avpkt->size set to 0 until it stops - * returning samples. It is safe to flush even those decoders that are not - * marked with AV_CODEC_CAP_DELAY, then no samples will be returned. - * - * @warning The input buffer, avpkt->data must be AV_INPUT_BUFFER_PADDING_SIZE - * larger than the actual read bytes because some optimized bitstream - * readers read 32 or 64 bits at once and could read over the end. - * - * @note The AVCodecContext MUST have been opened with @ref avcodec_open2() - * before packets may be fed to the decoder. - * - * @param avctx the codec context - * @param[out] frame The AVFrame in which to store decoded audio samples. - * The decoder will allocate a buffer for the decoded frame by - * calling the AVCodecContext.get_buffer2() callback. - * When AVCodecContext.refcounted_frames is set to 1, the frame is - * reference counted and the returned reference belongs to the - * caller. The caller must release the frame using av_frame_unref() - * when the frame is no longer needed. The caller may safely write - * to the frame if av_frame_is_writable() returns 1. - * When AVCodecContext.refcounted_frames is set to 0, the returned - * reference belongs to the decoder and is valid only until the - * next call to this function or until closing or flushing the - * decoder. The caller may not write to it. - * @param[out] got_frame_ptr Zero if no frame could be decoded, otherwise it is - * non-zero. Note that this field being set to zero - * does not mean that an error has occurred. For - * decoders with AV_CODEC_CAP_DELAY set, no given decode - * call is guaranteed to produce a frame. - * @param[in] avpkt The input AVPacket containing the input buffer. - * At least avpkt->data and avpkt->size should be set. Some - * decoders might also require additional fields to be set. - * @return A negative error code is returned if an error occurred during - * decoding, otherwise the number of bytes consumed from the input - * AVPacket is returned. - */ -int avcodec_decode_audio4(AVCodecContext *avctx, AVFrame *frame, - int *got_frame_ptr, const AVPacket *avpkt); - -/** - * Decode the video frame of size avpkt->size from avpkt->data into picture. - * Some decoders may support multiple frames in a single AVPacket, such - * decoders would then just decode the first frame. - * - * @warning The input buffer must be AV_INPUT_BUFFER_PADDING_SIZE larger than - * the actual read bytes because some optimized bitstream readers read 32 or 64 - * bits at once and could read over the end. - * - * @warning The end of the input buffer buf should be set to 0 to ensure that - * no overreading happens for damaged MPEG streams. - * - * @note Codecs which have the AV_CODEC_CAP_DELAY capability set have a delay - * between input and output, these need to be fed with avpkt->data=NULL, - * avpkt->size=0 at the end to return the remaining frames. - * - * @note The AVCodecContext MUST have been opened with @ref avcodec_open2() - * before packets may be fed to the decoder. - * - * @param avctx the codec context - * @param[out] picture The AVFrame in which the decoded video frame will be stored. - * Use av_frame_alloc() to get an AVFrame. The codec will - * allocate memory for the actual bitmap by calling the - * AVCodecContext.get_buffer2() callback. - * When AVCodecContext.refcounted_frames is set to 1, the frame is - * reference counted and the returned reference belongs to the - * caller. The caller must release the frame using av_frame_unref() - * when the frame is no longer needed. The caller may safely write - * to the frame if av_frame_is_writable() returns 1. - * When AVCodecContext.refcounted_frames is set to 0, the returned - * reference belongs to the decoder and is valid only until the - * next call to this function or until closing or flushing the - * decoder. The caller may not write to it. - * - * @param[in] avpkt The input AVPacket containing the input buffer. - * You can create such packet with av_init_packet() and by then setting - * data and size, some decoders might in addition need other fields like - * flags&AV_PKT_FLAG_KEY. All decoders are designed to use the least - * fields possible. - * @param[in,out] got_picture_ptr Zero if no frame could be decompressed, otherwise, it is nonzero. - * @return On error a negative value is returned, otherwise the number of bytes - * used or zero if no frame could be decompressed. - */ -int avcodec_decode_video2(AVCodecContext *avctx, AVFrame *picture, - int *got_picture_ptr, - const AVPacket *avpkt); - -/** - * Decode a subtitle message. - * Return a negative value on error, otherwise return the number of bytes used. - * If no subtitle could be decompressed, got_sub_ptr is zero. - * Otherwise, the subtitle is stored in *sub. - * Note that AV_CODEC_CAP_DR1 is not available for subtitle codecs. This is for - * simplicity, because the performance difference is expect to be negligible - * and reusing a get_buffer written for video codecs would probably perform badly - * due to a potentially very different allocation pattern. - * - * Some decoders (those marked with CODEC_CAP_DELAY) have a delay between input - * and output. This means that for some packets they will not immediately - * produce decoded output and need to be flushed at the end of decoding to get - * all the decoded data. Flushing is done by calling this function with packets - * with avpkt->data set to NULL and avpkt->size set to 0 until it stops - * returning subtitles. It is safe to flush even those decoders that are not - * marked with CODEC_CAP_DELAY, then no subtitles will be returned. - * - * @note The AVCodecContext MUST have been opened with @ref avcodec_open2() - * before packets may be fed to the decoder. - * - * @param avctx the codec context - * @param[out] sub The Preallocated AVSubtitle in which the decoded subtitle will be stored, - * must be freed with avsubtitle_free if *got_sub_ptr is set. - * @param[in,out] got_sub_ptr Zero if no subtitle could be decompressed, otherwise, it is nonzero. - * @param[in] avpkt The input AVPacket containing the input buffer. - */ -int avcodec_decode_subtitle2(AVCodecContext *avctx, AVSubtitle *sub, - int *got_sub_ptr, - AVPacket *avpkt); - -/** - * @defgroup lavc_parsing Frame parsing - * @{ - */ - -enum AVPictureStructure { - AV_PICTURE_STRUCTURE_UNKNOWN, //< unknown - AV_PICTURE_STRUCTURE_TOP_FIELD, //< coded as top field - AV_PICTURE_STRUCTURE_BOTTOM_FIELD, //< coded as bottom field - AV_PICTURE_STRUCTURE_FRAME, //< coded as frame -}; - -typedef struct AVCodecParserContext { - void *priv_data; - struct AVCodecParser *parser; - int64_t frame_offset; /* offset of the current frame */ - int64_t cur_offset; /* current offset - (incremented by each av_parser_parse()) */ - int64_t next_frame_offset; /* offset of the next frame */ - /* video info */ - int pict_type; /* XXX: Put it back in AVCodecContext. */ - /** - * This field is used for proper frame duration computation in lavf. - * It signals, how much longer the frame duration of the current frame - * is compared to normal frame duration. - * - * frame_duration = (1 + repeat_pict) * time_base - * - * It is used by codecs like H.264 to display telecined material. - */ - int repeat_pict; /* XXX: Put it back in AVCodecContext. */ - int64_t pts; /* pts of the current frame */ - int64_t dts; /* dts of the current frame */ - - /* private data */ - int64_t last_pts; - int64_t last_dts; - int fetch_timestamp; - -#define AV_PARSER_PTS_NB 4 - int cur_frame_start_index; - int64_t cur_frame_offset[AV_PARSER_PTS_NB]; - int64_t cur_frame_pts[AV_PARSER_PTS_NB]; - int64_t cur_frame_dts[AV_PARSER_PTS_NB]; - - int flags; -#define PARSER_FLAG_COMPLETE_FRAMES 0x0001 -#define PARSER_FLAG_ONCE 0x0002 -/// Set if the parser has a valid file offset -#define PARSER_FLAG_FETCHED_OFFSET 0x0004 -#define PARSER_FLAG_USE_CODEC_TS 0x1000 - - int64_t offset; ///< byte offset from starting packet start - int64_t cur_frame_end[AV_PARSER_PTS_NB]; - - /** - * Set by parser to 1 for key frames and 0 for non-key frames. - * It is initialized to -1, so if the parser doesn't set this flag, - * old-style fallback using AV_PICTURE_TYPE_I picture type as key frames - * will be used. - */ - int key_frame; - -#if FF_API_CONVERGENCE_DURATION - /** - * @deprecated unused - */ - attribute_deprecated - int64_t convergence_duration; -#endif - - // Timestamp generation support: - /** - * Synchronization point for start of timestamp generation. - * - * Set to >0 for sync point, 0 for no sync point and <0 for undefined - * (default). - * - * For example, this corresponds to presence of H.264 buffering period - * SEI message. - */ - int dts_sync_point; - - /** - * Offset of the current timestamp against last timestamp sync point in - * units of AVCodecContext.time_base. - * - * Set to INT_MIN when dts_sync_point unused. Otherwise, it must - * contain a valid timestamp offset. - * - * Note that the timestamp of sync point has usually a nonzero - * dts_ref_dts_delta, which refers to the previous sync point. Offset of - * the next frame after timestamp sync point will be usually 1. - * - * For example, this corresponds to H.264 cpb_removal_delay. - */ - int dts_ref_dts_delta; - - /** - * Presentation delay of current frame in units of AVCodecContext.time_base. - * - * Set to INT_MIN when dts_sync_point unused. Otherwise, it must - * contain valid non-negative timestamp delta (presentation time of a frame - * must not lie in the past). - * - * This delay represents the difference between decoding and presentation - * time of the frame. - * - * For example, this corresponds to H.264 dpb_output_delay. - */ - int pts_dts_delta; - - /** - * Position of the packet in file. - * - * Analogous to cur_frame_pts/dts - */ - int64_t cur_frame_pos[AV_PARSER_PTS_NB]; - - /** - * Byte position of currently parsed frame in stream. - */ - int64_t pos; - - /** - * Previous frame byte position. - */ - int64_t last_pos; - - /** - * Duration of the current frame. - * For audio, this is in units of 1 / AVCodecContext.sample_rate. - * For all other types, this is in units of AVCodecContext.time_base. - */ - int duration; - - enum AVFieldOrder field_order; - - /** - * Indicate whether a picture is coded as a frame, top field or bottom field. - * - * For example, H.264 field_pic_flag equal to 0 corresponds to - * AV_PICTURE_STRUCTURE_FRAME. An H.264 picture with field_pic_flag - * equal to 1 and bottom_field_flag equal to 0 corresponds to - * AV_PICTURE_STRUCTURE_TOP_FIELD. - */ - enum AVPictureStructure picture_structure; - - /** - * Picture number incremented in presentation or output order. - * This field may be reinitialized at the first picture of a new sequence. - * - * For example, this corresponds to H.264 PicOrderCnt. - */ - int output_picture_number; - - /** - * Dimensions of the decoded video intended for presentation. - */ - int width; - int height; - - /** - * Dimensions of the coded video. - */ - int coded_width; - int coded_height; - - /** - * The format of the coded data, corresponds to enum AVPixelFormat for video - * and for enum AVSampleFormat for audio. - * - * Note that a decoder can have considerable freedom in how exactly it - * decodes the data, so the format reported here might be different from the - * one returned by a decoder. - */ - int format; -} AVCodecParserContext; - -typedef struct AVCodecParser { - int codec_ids[5]; /* several codec IDs are permitted */ - int priv_data_size; - int (*parser_init)(AVCodecParserContext *s); - /* This callback never returns an error, a negative value means that - * the frame start was in a previous packet. */ - int (*parser_parse)(AVCodecParserContext *s, - AVCodecContext *avctx, - const uint8_t **poutbuf, int *poutbuf_size, - const uint8_t *buf, int buf_size); - void (*parser_close)(AVCodecParserContext *s); - int (*split)(AVCodecContext *avctx, const uint8_t *buf, int buf_size); - struct AVCodecParser *next; -} AVCodecParser; - -AVCodecParser *av_parser_next(const AVCodecParser *c); - -void av_register_codec_parser(AVCodecParser *parser); -AVCodecParserContext *av_parser_init(int codec_id); - -/** - * Parse a packet. - * - * @param s parser context. - * @param avctx codec context. - * @param poutbuf set to pointer to parsed buffer or NULL if not yet finished. - * @param poutbuf_size set to size of parsed buffer or zero if not yet finished. - * @param buf input buffer. - * @param buf_size input length, to signal EOF, this should be 0 (so that the last frame can be output). - * @param pts input presentation timestamp. - * @param dts input decoding timestamp. - * @param pos input byte position in stream. - * @return the number of bytes of the input bitstream used. - * - * Example: - * @code - * while(in_len){ - * len = av_parser_parse2(myparser, AVCodecContext, &data, &size, - * in_data, in_len, - * pts, dts, pos); - * in_data += len; - * in_len -= len; - * - * if(size) - * decode_frame(data, size); - * } - * @endcode - */ -int av_parser_parse2(AVCodecParserContext *s, - AVCodecContext *avctx, - uint8_t **poutbuf, int *poutbuf_size, - const uint8_t *buf, int buf_size, - int64_t pts, int64_t dts, - int64_t pos); - -/** - * @return 0 if the output buffer is a subset of the input, 1 if it is allocated and must be freed - * @deprecated use AVBitStreamFilter - */ -int av_parser_change(AVCodecParserContext *s, - AVCodecContext *avctx, - uint8_t **poutbuf, int *poutbuf_size, - const uint8_t *buf, int buf_size, int keyframe); -void av_parser_close(AVCodecParserContext *s); - -/** - * @} - * @} - */ - -/** - * @addtogroup lavc_encoding - * @{ - */ - -/** - * Find a registered encoder with a matching codec ID. - * - * @param id AVCodecID of the requested encoder - * @return An encoder if one was found, NULL otherwise. - */ -AVCodec *avcodec_find_encoder(enum AVCodecID id); - -/** - * Find a registered encoder with the specified name. - * - * @param name name of the requested encoder - * @return An encoder if one was found, NULL otherwise. - */ -AVCodec *avcodec_find_encoder_by_name(const char *name); - -/** - * Encode a frame of audio. - * - * Takes input samples from frame and writes the next output packet, if - * available, to avpkt. The output packet does not necessarily contain data for - * the most recent frame, as encoders can delay, split, and combine input frames - * internally as needed. - * - * @param avctx codec context - * @param avpkt output AVPacket. - * The user can supply an output buffer by setting - * avpkt->data and avpkt->size prior to calling the - * function, but if the size of the user-provided data is not - * large enough, encoding will fail. If avpkt->data and - * avpkt->size are set, avpkt->destruct must also be set. All - * other AVPacket fields will be reset by the encoder using - * av_init_packet(). If avpkt->data is NULL, the encoder will - * allocate it. The encoder will set avpkt->size to the size - * of the output packet. - * - * If this function fails or produces no output, avpkt will be - * freed using av_packet_unref(). - * @param[in] frame AVFrame containing the raw audio data to be encoded. - * May be NULL when flushing an encoder that has the - * AV_CODEC_CAP_DELAY capability set. - * If AV_CODEC_CAP_VARIABLE_FRAME_SIZE is set, then each frame - * can have any number of samples. - * If it is not set, frame->nb_samples must be equal to - * avctx->frame_size for all frames except the last. - * The final frame may be smaller than avctx->frame_size. - * @param[out] got_packet_ptr This field is set to 1 by libavcodec if the - * output packet is non-empty, and to 0 if it is - * empty. If the function returns an error, the - * packet can be assumed to be invalid, and the - * value of got_packet_ptr is undefined and should - * not be used. - * @return 0 on success, negative error code on failure - */ -int avcodec_encode_audio2(AVCodecContext *avctx, AVPacket *avpkt, - const AVFrame *frame, int *got_packet_ptr); - -/** - * Encode a frame of video. - * - * Takes input raw video data from frame and writes the next output packet, if - * available, to avpkt. The output packet does not necessarily contain data for - * the most recent frame, as encoders can delay and reorder input frames - * internally as needed. - * - * @param avctx codec context - * @param avpkt output AVPacket. - * The user can supply an output buffer by setting - * avpkt->data and avpkt->size prior to calling the - * function, but if the size of the user-provided data is not - * large enough, encoding will fail. All other AVPacket fields - * will be reset by the encoder using av_init_packet(). If - * avpkt->data is NULL, the encoder will allocate it. - * The encoder will set avpkt->size to the size of the - * output packet. The returned data (if any) belongs to the - * caller, he is responsible for freeing it. - * - * If this function fails or produces no output, avpkt will be - * freed using av_packet_unref(). - * @param[in] frame AVFrame containing the raw video data to be encoded. - * May be NULL when flushing an encoder that has the - * AV_CODEC_CAP_DELAY capability set. - * @param[out] got_packet_ptr This field is set to 1 by libavcodec if the - * output packet is non-empty, and to 0 if it is - * empty. If the function returns an error, the - * packet can be assumed to be invalid, and the - * value of got_packet_ptr is undefined and should - * not be used. - * @return 0 on success, negative error code on failure - */ -int avcodec_encode_video2(AVCodecContext *avctx, AVPacket *avpkt, - const AVFrame *frame, int *got_packet_ptr); - -int avcodec_encode_subtitle(AVCodecContext *avctx, uint8_t *buf, int buf_size, - const AVSubtitle *sub); - - -/** - * @} - */ - -#if FF_API_AVCODEC_RESAMPLE -/** - * @defgroup lavc_resample Audio resampling - * @ingroup libavc - * @deprecated use libswresample instead - * - * @{ - */ -struct ReSampleContext; -struct AVResampleContext; - -typedef struct ReSampleContext ReSampleContext; - -/** - * Initialize audio resampling context. - * - * @param output_channels number of output channels - * @param input_channels number of input channels - * @param output_rate output sample rate - * @param input_rate input sample rate - * @param sample_fmt_out requested output sample format - * @param sample_fmt_in input sample format - * @param filter_length length of each FIR filter in the filterbank relative to the cutoff frequency - * @param log2_phase_count log2 of the number of entries in the polyphase filterbank - * @param linear if 1 then the used FIR filter will be linearly interpolated - between the 2 closest, if 0 the closest will be used - * @param cutoff cutoff frequency, 1.0 corresponds to half the output sampling rate - * @return allocated ReSampleContext, NULL if error occurred - */ -attribute_deprecated -ReSampleContext *av_audio_resample_init(int output_channels, int input_channels, - int output_rate, int input_rate, - enum AVSampleFormat sample_fmt_out, - enum AVSampleFormat sample_fmt_in, - int filter_length, int log2_phase_count, - int linear, double cutoff); - -attribute_deprecated -int audio_resample(ReSampleContext *s, short *output, short *input, int nb_samples); - -/** - * Free resample context. - * - * @param s a non-NULL pointer to a resample context previously - * created with av_audio_resample_init() - */ -attribute_deprecated -void audio_resample_close(ReSampleContext *s); - - -/** - * Initialize an audio resampler. - * Note, if either rate is not an integer then simply scale both rates up so they are. - * @param filter_length length of each FIR filter in the filterbank relative to the cutoff freq - * @param log2_phase_count log2 of the number of entries in the polyphase filterbank - * @param linear If 1 then the used FIR filter will be linearly interpolated - between the 2 closest, if 0 the closest will be used - * @param cutoff cutoff frequency, 1.0 corresponds to half the output sampling rate - */ -attribute_deprecated -struct AVResampleContext *av_resample_init(int out_rate, int in_rate, int filter_length, int log2_phase_count, int linear, double cutoff); - -/** - * Resample an array of samples using a previously configured context. - * @param src an array of unconsumed samples - * @param consumed the number of samples of src which have been consumed are returned here - * @param src_size the number of unconsumed samples available - * @param dst_size the amount of space in samples available in dst - * @param update_ctx If this is 0 then the context will not be modified, that way several channels can be resampled with the same context. - * @return the number of samples written in dst or -1 if an error occurred - */ -attribute_deprecated -int av_resample(struct AVResampleContext *c, short *dst, short *src, int *consumed, int src_size, int dst_size, int update_ctx); - - -/** - * Compensate samplerate/timestamp drift. The compensation is done by changing - * the resampler parameters, so no audible clicks or similar distortions occur - * @param compensation_distance distance in output samples over which the compensation should be performed - * @param sample_delta number of output samples which should be output less - * - * example: av_resample_compensate(c, 10, 500) - * here instead of 510 samples only 500 samples would be output - * - * note, due to rounding the actual compensation might be slightly different, - * especially if the compensation_distance is large and the in_rate used during init is small - */ -attribute_deprecated -void av_resample_compensate(struct AVResampleContext *c, int sample_delta, int compensation_distance); -attribute_deprecated -void av_resample_close(struct AVResampleContext *c); - -/** - * @} - */ -#endif - -#if FF_API_AVPICTURE -/** - * @addtogroup lavc_picture - * @{ - */ - -/** - * @deprecated unused - */ -attribute_deprecated -int avpicture_alloc(AVPicture *picture, enum AVPixelFormat pix_fmt, int width, int height); - -/** - * @deprecated unused - */ -attribute_deprecated -void avpicture_free(AVPicture *picture); - -/** - * @deprecated use av_image_fill_arrays() instead. - */ -attribute_deprecated -int avpicture_fill(AVPicture *picture, const uint8_t *ptr, - enum AVPixelFormat pix_fmt, int width, int height); - -/** - * @deprecated use av_image_copy_to_buffer() instead. - */ -attribute_deprecated -int avpicture_layout(const AVPicture *src, enum AVPixelFormat pix_fmt, - int width, int height, - unsigned char *dest, int dest_size); - -/** - * @deprecated use av_image_get_buffer_size() instead. - */ -attribute_deprecated -int avpicture_get_size(enum AVPixelFormat pix_fmt, int width, int height); - -/** - * @deprecated av_image_copy() instead. - */ -attribute_deprecated -void av_picture_copy(AVPicture *dst, const AVPicture *src, - enum AVPixelFormat pix_fmt, int width, int height); - -/** - * @deprecated unused - */ -attribute_deprecated -int av_picture_crop(AVPicture *dst, const AVPicture *src, - enum AVPixelFormat pix_fmt, int top_band, int left_band); - -/** - * @deprecated unused - */ -attribute_deprecated -int av_picture_pad(AVPicture *dst, const AVPicture *src, int height, int width, enum AVPixelFormat pix_fmt, - int padtop, int padbottom, int padleft, int padright, int *color); - -/** - * @} - */ -#endif - -/** - * @defgroup lavc_misc Utility functions - * @ingroup libavc - * - * Miscellaneous utility functions related to both encoding and decoding - * (or neither). - * @{ - */ - -/** - * @defgroup lavc_misc_pixfmt Pixel formats - * - * Functions for working with pixel formats. - * @{ - */ - -/** - * Utility function to access log2_chroma_w log2_chroma_h from - * the pixel format AVPixFmtDescriptor. - * - * This function asserts that pix_fmt is valid. See av_pix_fmt_get_chroma_sub_sample - * for one that returns a failure code and continues in case of invalid - * pix_fmts. - * - * @param[in] pix_fmt the pixel format - * @param[out] h_shift store log2_chroma_w - * @param[out] v_shift store log2_chroma_h - * - * @see av_pix_fmt_get_chroma_sub_sample - */ - -void avcodec_get_chroma_sub_sample(enum AVPixelFormat pix_fmt, int *h_shift, int *v_shift); - -/** - * Return a value representing the fourCC code associated to the - * pixel format pix_fmt, or 0 if no associated fourCC code can be - * found. - */ -unsigned int avcodec_pix_fmt_to_codec_tag(enum AVPixelFormat pix_fmt); - -/** - * @deprecated see av_get_pix_fmt_loss() - */ -int avcodec_get_pix_fmt_loss(enum AVPixelFormat dst_pix_fmt, enum AVPixelFormat src_pix_fmt, - int has_alpha); - -/** - * Find the best pixel format to convert to given a certain source pixel - * format. When converting from one pixel format to another, information loss - * may occur. For example, when converting from RGB24 to GRAY, the color - * information will be lost. Similarly, other losses occur when converting from - * some formats to other formats. avcodec_find_best_pix_fmt_of_2() searches which of - * the given pixel formats should be used to suffer the least amount of loss. - * The pixel formats from which it chooses one, are determined by the - * pix_fmt_list parameter. - * - * - * @param[in] pix_fmt_list AV_PIX_FMT_NONE terminated array of pixel formats to choose from - * @param[in] src_pix_fmt source pixel format - * @param[in] has_alpha Whether the source pixel format alpha channel is used. - * @param[out] loss_ptr Combination of flags informing you what kind of losses will occur. - * @return The best pixel format to convert to or -1 if none was found. - */ -enum AVPixelFormat avcodec_find_best_pix_fmt_of_list(const enum AVPixelFormat *pix_fmt_list, - enum AVPixelFormat src_pix_fmt, - int has_alpha, int *loss_ptr); - -/** - * @deprecated see av_find_best_pix_fmt_of_2() - */ -enum AVPixelFormat avcodec_find_best_pix_fmt_of_2(enum AVPixelFormat dst_pix_fmt1, enum AVPixelFormat dst_pix_fmt2, - enum AVPixelFormat src_pix_fmt, int has_alpha, int *loss_ptr); - -attribute_deprecated -#if AV_HAVE_INCOMPATIBLE_LIBAV_ABI -enum AVPixelFormat avcodec_find_best_pix_fmt2(const enum AVPixelFormat *pix_fmt_list, - enum AVPixelFormat src_pix_fmt, - int has_alpha, int *loss_ptr); -#else -enum AVPixelFormat avcodec_find_best_pix_fmt2(enum AVPixelFormat dst_pix_fmt1, enum AVPixelFormat dst_pix_fmt2, - enum AVPixelFormat src_pix_fmt, int has_alpha, int *loss_ptr); -#endif - - -enum AVPixelFormat avcodec_default_get_format(struct AVCodecContext *s, const enum AVPixelFormat * fmt); - -/** - * @} - */ - -#if FF_API_SET_DIMENSIONS -/** - * @deprecated this function is not supposed to be used from outside of lavc - */ -attribute_deprecated -void avcodec_set_dimensions(AVCodecContext *s, int width, int height); -#endif - -/** - * Put a string representing the codec tag codec_tag in buf. - * - * @param buf buffer to place codec tag in - * @param buf_size size in bytes of buf - * @param codec_tag codec tag to assign - * @return the length of the string that would have been generated if - * enough space had been available, excluding the trailing null - */ -size_t av_get_codec_tag_string(char *buf, size_t buf_size, unsigned int codec_tag); - -void avcodec_string(char *buf, int buf_size, AVCodecContext *enc, int encode); - -/** - * Return a name for the specified profile, if available. - * - * @param codec the codec that is searched for the given profile - * @param profile the profile value for which a name is requested - * @return A name for the profile if found, NULL otherwise. - */ -const char *av_get_profile_name(const AVCodec *codec, int profile); - -/** - * Return a name for the specified profile, if available. - * - * @param codec_id the ID of the codec to which the requested profile belongs - * @param profile the profile value for which a name is requested - * @return A name for the profile if found, NULL otherwise. - * - * @note unlike av_get_profile_name(), which searches a list of profiles - * supported by a specific decoder or encoder implementation, this - * function searches the list of profiles from the AVCodecDescriptor - */ -const char *avcodec_profile_name(enum AVCodecID codec_id, int profile); - -int avcodec_default_execute(AVCodecContext *c, int (*func)(AVCodecContext *c2, void *arg2),void *arg, int *ret, int count, int size); -int avcodec_default_execute2(AVCodecContext *c, int (*func)(AVCodecContext *c2, void *arg2, int, int),void *arg, int *ret, int count); -//FIXME func typedef - -/** - * Fill AVFrame audio data and linesize pointers. - * - * The buffer buf must be a preallocated buffer with a size big enough - * to contain the specified samples amount. The filled AVFrame data - * pointers will point to this buffer. - * - * AVFrame extended_data channel pointers are allocated if necessary for - * planar audio. - * - * @param frame the AVFrame - * frame->nb_samples must be set prior to calling the - * function. This function fills in frame->data, - * frame->extended_data, frame->linesize[0]. - * @param nb_channels channel count - * @param sample_fmt sample format - * @param buf buffer to use for frame data - * @param buf_size size of buffer - * @param align plane size sample alignment (0 = default) - * @return >=0 on success, negative error code on failure - * @todo return the size in bytes required to store the samples in - * case of success, at the next libavutil bump - */ -int avcodec_fill_audio_frame(AVFrame *frame, int nb_channels, - enum AVSampleFormat sample_fmt, const uint8_t *buf, - int buf_size, int align); - -/** - * Reset the internal decoder state / flush internal buffers. Should be called - * e.g. when seeking or when switching to a different stream. - * - * @note when refcounted frames are not used (i.e. avctx->refcounted_frames is 0), - * this invalidates the frames previously returned from the decoder. When - * refcounted frames are used, the decoder just releases any references it might - * keep internally, but the caller's reference remains valid. - */ -void avcodec_flush_buffers(AVCodecContext *avctx); - -/** - * Return codec bits per sample. - * - * @param[in] codec_id the codec - * @return Number of bits per sample or zero if unknown for the given codec. - */ -int av_get_bits_per_sample(enum AVCodecID codec_id); - -/** - * Return the PCM codec associated with a sample format. - * @param be endianness, 0 for little, 1 for big, - * -1 (or anything else) for native - * @return AV_CODEC_ID_PCM_* or AV_CODEC_ID_NONE - */ -enum AVCodecID av_get_pcm_codec(enum AVSampleFormat fmt, int be); - -/** - * Return codec bits per sample. - * Only return non-zero if the bits per sample is exactly correct, not an - * approximation. - * - * @param[in] codec_id the codec - * @return Number of bits per sample or zero if unknown for the given codec. - */ -int av_get_exact_bits_per_sample(enum AVCodecID codec_id); - -/** - * Return audio frame duration. - * - * @param avctx codec context - * @param frame_bytes size of the frame, or 0 if unknown - * @return frame duration, in samples, if known. 0 if not able to - * determine. - */ -int av_get_audio_frame_duration(AVCodecContext *avctx, int frame_bytes); - - -typedef struct AVBitStreamFilterContext { - void *priv_data; - struct AVBitStreamFilter *filter; - AVCodecParserContext *parser; - struct AVBitStreamFilterContext *next; - /** - * Internal default arguments, used if NULL is passed to av_bitstream_filter_filter(). - * Not for access by library users. - */ - char *args; -} AVBitStreamFilterContext; - - -typedef struct AVBitStreamFilter { - const char *name; - int priv_data_size; - int (*filter)(AVBitStreamFilterContext *bsfc, - AVCodecContext *avctx, const char *args, - uint8_t **poutbuf, int *poutbuf_size, - const uint8_t *buf, int buf_size, int keyframe); - void (*close)(AVBitStreamFilterContext *bsfc); - struct AVBitStreamFilter *next; -} AVBitStreamFilter; - -/** - * Register a bitstream filter. - * - * The filter will be accessible to the application code through - * av_bitstream_filter_next() or can be directly initialized with - * av_bitstream_filter_init(). - * - * @see avcodec_register_all() - */ -void av_register_bitstream_filter(AVBitStreamFilter *bsf); - -/** - * Create and initialize a bitstream filter context given a bitstream - * filter name. - * - * The returned context must be freed with av_bitstream_filter_close(). - * - * @param name the name of the bitstream filter - * @return a bitstream filter context if a matching filter was found - * and successfully initialized, NULL otherwise - */ -AVBitStreamFilterContext *av_bitstream_filter_init(const char *name); - -/** - * Filter bitstream. - * - * This function filters the buffer buf with size buf_size, and places the - * filtered buffer in the buffer pointed to by poutbuf. - * - * The output buffer must be freed by the caller. - * - * @param bsfc bitstream filter context created by av_bitstream_filter_init() - * @param avctx AVCodecContext accessed by the filter, may be NULL. - * If specified, this must point to the encoder context of the - * output stream the packet is sent to. - * @param args arguments which specify the filter configuration, may be NULL - * @param poutbuf pointer which is updated to point to the filtered buffer - * @param poutbuf_size pointer which is updated to the filtered buffer size in bytes - * @param buf buffer containing the data to filter - * @param buf_size size in bytes of buf - * @param keyframe set to non-zero if the buffer to filter corresponds to a key-frame packet data - * @return >= 0 in case of success, or a negative error code in case of failure - * - * If the return value is positive, an output buffer is allocated and - * is available in *poutbuf, and is distinct from the input buffer. - * - * If the return value is 0, the output buffer is not allocated and - * should be considered identical to the input buffer, or in case - * *poutbuf was set it points to the input buffer (not necessarily to - * its starting address). - */ -int av_bitstream_filter_filter(AVBitStreamFilterContext *bsfc, - AVCodecContext *avctx, const char *args, - uint8_t **poutbuf, int *poutbuf_size, - const uint8_t *buf, int buf_size, int keyframe); - -/** - * Release bitstream filter context. - * - * @param bsf the bitstream filter context created with - * av_bitstream_filter_init(), can be NULL - */ -void av_bitstream_filter_close(AVBitStreamFilterContext *bsf); - -/** - * If f is NULL, return the first registered bitstream filter, - * if f is non-NULL, return the next registered bitstream filter - * after f, or NULL if f is the last one. - * - * This function can be used to iterate over all registered bitstream - * filters. - */ -AVBitStreamFilter *av_bitstream_filter_next(const AVBitStreamFilter *f); - -/* memory */ - -/** - * Same behaviour av_fast_malloc but the buffer has additional - * AV_INPUT_BUFFER_PADDING_SIZE at the end which will always be 0. - * - * In addition the whole buffer will initially and after resizes - * be 0-initialized so that no uninitialized data will ever appear. - */ -void av_fast_padded_malloc(void *ptr, unsigned int *size, size_t min_size); - -/** - * Same behaviour av_fast_padded_malloc except that buffer will always - * be 0-initialized after call. - */ -void av_fast_padded_mallocz(void *ptr, unsigned int *size, size_t min_size); - -/** - * Encode extradata length to a buffer. Used by xiph codecs. - * - * @param s buffer to write to; must be at least (v/255+1) bytes long - * @param v size of extradata in bytes - * @return number of bytes written to the buffer. - */ -unsigned int av_xiphlacing(unsigned char *s, unsigned int v); - -#if FF_API_MISSING_SAMPLE -/** - * Log a generic warning message about a missing feature. This function is - * intended to be used internally by FFmpeg (libavcodec, libavformat, etc.) - * only, and would normally not be used by applications. - * @param[in] avc a pointer to an arbitrary struct of which the first field is - * a pointer to an AVClass struct - * @param[in] feature string containing the name of the missing feature - * @param[in] want_sample indicates if samples are wanted which exhibit this feature. - * If want_sample is non-zero, additional verbage will be added to the log - * message which tells the user how to report samples to the development - * mailing list. - * @deprecated Use avpriv_report_missing_feature() instead. - */ -attribute_deprecated -void av_log_missing_feature(void *avc, const char *feature, int want_sample); - -/** - * Log a generic warning message asking for a sample. This function is - * intended to be used internally by FFmpeg (libavcodec, libavformat, etc.) - * only, and would normally not be used by applications. - * @param[in] avc a pointer to an arbitrary struct of which the first field is - * a pointer to an AVClass struct - * @param[in] msg string containing an optional message, or NULL if no message - * @deprecated Use avpriv_request_sample() instead. - */ -attribute_deprecated -void av_log_ask_for_sample(void *avc, const char *msg, ...) av_printf_format(2, 3); -#endif /* FF_API_MISSING_SAMPLE */ - -/** - * Register the hardware accelerator hwaccel. - */ -void av_register_hwaccel(AVHWAccel *hwaccel); - -/** - * If hwaccel is NULL, returns the first registered hardware accelerator, - * if hwaccel is non-NULL, returns the next registered hardware accelerator - * after hwaccel, or NULL if hwaccel is the last one. - */ -AVHWAccel *av_hwaccel_next(const AVHWAccel *hwaccel); - - -/** - * Lock operation used by lockmgr - */ -enum AVLockOp { - AV_LOCK_CREATE, ///< Create a mutex - AV_LOCK_OBTAIN, ///< Lock the mutex - AV_LOCK_RELEASE, ///< Unlock the mutex - AV_LOCK_DESTROY, ///< Free mutex resources -}; - -/** - * Register a user provided lock manager supporting the operations - * specified by AVLockOp. The "mutex" argument to the function points - * to a (void *) where the lockmgr should store/get a pointer to a user - * allocated mutex. It is NULL upon AV_LOCK_CREATE and equal to the - * value left by the last call for all other ops. If the lock manager is - * unable to perform the op then it should leave the mutex in the same - * state as when it was called and return a non-zero value. However, - * when called with AV_LOCK_DESTROY the mutex will always be assumed to - * have been successfully destroyed. If av_lockmgr_register succeeds - * it will return a non-negative value, if it fails it will return a - * negative value and destroy all mutex and unregister all callbacks. - * av_lockmgr_register is not thread-safe, it must be called from a - * single thread before any calls which make use of locking are used. - * - * @param cb User defined callback. av_lockmgr_register invokes calls - * to this callback and the previously registered callback. - * The callback will be used to create more than one mutex - * each of which must be backed by its own underlying locking - * mechanism (i.e. do not use a single static object to - * implement your lock manager). If cb is set to NULL the - * lockmgr will be unregistered. - */ -int av_lockmgr_register(int (*cb)(void **mutex, enum AVLockOp op)); - -/** - * Get the type of the given codec. - */ -enum AVMediaType avcodec_get_type(enum AVCodecID codec_id); - -/** - * Get the name of a codec. - * @return a static string identifying the codec; never NULL - */ -const char *avcodec_get_name(enum AVCodecID id); - -/** - * @return a positive value if s is open (i.e. avcodec_open2() was called on it - * with no corresponding avcodec_close()), 0 otherwise. - */ -int avcodec_is_open(AVCodecContext *s); - -/** - * @return a non-zero number if codec is an encoder, zero otherwise - */ -int av_codec_is_encoder(const AVCodec *codec); - -/** - * @return a non-zero number if codec is a decoder, zero otherwise - */ -int av_codec_is_decoder(const AVCodec *codec); - -/** - * @return descriptor for given codec ID or NULL if no descriptor exists. - */ -const AVCodecDescriptor *avcodec_descriptor_get(enum AVCodecID id); - -/** - * Iterate over all codec descriptors known to libavcodec. - * - * @param prev previous descriptor. NULL to get the first descriptor. - * - * @return next descriptor or NULL after the last descriptor - */ -const AVCodecDescriptor *avcodec_descriptor_next(const AVCodecDescriptor *prev); - -/** - * @return codec descriptor with the given name or NULL if no such descriptor - * exists. - */ -const AVCodecDescriptor *avcodec_descriptor_get_by_name(const char *name); - -/** - * Allocate a CPB properties structure and initialize its fields to default - * values. - * - * @param size if non-NULL, the size of the allocated struct will be written - * here. This is useful for embedding it in side data. - * - * @return the newly allocated struct or NULL on failure - */ -AVCPBProperties *av_cpb_properties_alloc(size_t *size); - -/** - * @} - */ - -#endif /* AVCODEC_AVCODEC_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/avdct.h b/src/main/jni/ffmpeg/include/libavcodec/avdct.h deleted file mode 100644 index 272422e44..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/avdct.h +++ /dev/null @@ -1,84 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_AVDCT_H -#define AVCODEC_AVDCT_H - -#include "libavutil/opt.h" - -/** - * AVDCT context. - * @note function pointers can be NULL if the specific features have been - * disabled at build time. - */ -typedef struct AVDCT { - const AVClass *av_class; - - void (*idct)(int16_t *block /* align 16 */); - - /** - * IDCT input permutation. - * Several optimized IDCTs need a permutated input (relative to the - * normal order of the reference IDCT). - * This permutation must be performed before the idct_put/add. - * Note, normally this can be merged with the zigzag/alternate scan<br> - * An example to avoid confusion: - * - (->decode coeffs -> zigzag reorder -> dequant -> reference IDCT -> ...) - * - (x -> reference DCT -> reference IDCT -> x) - * - (x -> reference DCT -> simple_mmx_perm = idct_permutation - * -> simple_idct_mmx -> x) - * - (-> decode coeffs -> zigzag reorder -> simple_mmx_perm -> dequant - * -> simple_idct_mmx -> ...) - */ - uint8_t idct_permutation[64]; - - void (*fdct)(int16_t *block /* align 16 */); - - - /** - * DCT algorithm. - * must use AVOptions to set this field. - */ - int dct_algo; - - /** - * IDCT algorithm. - * must use AVOptions to set this field. - */ - int idct_algo; - - void (*get_pixels)(int16_t *block /* align 16 */, - const uint8_t *pixels /* align 8 */, - ptrdiff_t line_size); - - int bits_per_sample; -} AVDCT; - -/** - * Allocates a AVDCT context. - * This needs to be initialized with avcodec_dct_init() after optionally - * configuring it with AVOptions. - * - * To free it use av_free() - */ -AVDCT *avcodec_dct_alloc(void); -int avcodec_dct_init(AVDCT *); - -const AVClass *avcodec_dct_get_class(void); - -#endif /* AVCODEC_AVDCT_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/avfft.h b/src/main/jni/ffmpeg/include/libavcodec/avfft.h deleted file mode 100644 index 0c0f9b8d8..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/avfft.h +++ /dev/null @@ -1,118 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_AVFFT_H -#define AVCODEC_AVFFT_H - -/** - * @file - * @ingroup lavc_fft - * FFT functions - */ - -/** - * @defgroup lavc_fft FFT functions - * @ingroup lavc_misc - * - * @{ - */ - -typedef float FFTSample; - -typedef struct FFTComplex { - FFTSample re, im; -} FFTComplex; - -typedef struct FFTContext FFTContext; - -/** - * Set up a complex FFT. - * @param nbits log2 of the length of the input array - * @param inverse if 0 perform the forward transform, if 1 perform the inverse - */ -FFTContext *av_fft_init(int nbits, int inverse); - -/** - * Do the permutation needed BEFORE calling ff_fft_calc(). - */ -void av_fft_permute(FFTContext *s, FFTComplex *z); - -/** - * Do a complex FFT with the parameters defined in av_fft_init(). The - * input data must be permuted before. No 1.0/sqrt(n) normalization is done. - */ -void av_fft_calc(FFTContext *s, FFTComplex *z); - -void av_fft_end(FFTContext *s); - -FFTContext *av_mdct_init(int nbits, int inverse, double scale); -void av_imdct_calc(FFTContext *s, FFTSample *output, const FFTSample *input); -void av_imdct_half(FFTContext *s, FFTSample *output, const FFTSample *input); -void av_mdct_calc(FFTContext *s, FFTSample *output, const FFTSample *input); -void av_mdct_end(FFTContext *s); - -/* Real Discrete Fourier Transform */ - -enum RDFTransformType { - DFT_R2C, - IDFT_C2R, - IDFT_R2C, - DFT_C2R, -}; - -typedef struct RDFTContext RDFTContext; - -/** - * Set up a real FFT. - * @param nbits log2 of the length of the input array - * @param trans the type of transform - */ -RDFTContext *av_rdft_init(int nbits, enum RDFTransformType trans); -void av_rdft_calc(RDFTContext *s, FFTSample *data); -void av_rdft_end(RDFTContext *s); - -/* Discrete Cosine Transform */ - -typedef struct DCTContext DCTContext; - -enum DCTTransformType { - DCT_II = 0, - DCT_III, - DCT_I, - DST_I, -}; - -/** - * Set up DCT. - * - * @param nbits size of the input array: - * (1 << nbits) for DCT-II, DCT-III and DST-I - * (1 << nbits) + 1 for DCT-I - * @param type the type of transform - * - * @note the first element of the input of DST-I is ignored - */ -DCTContext *av_dct_init(int nbits, enum DCTTransformType type); -void av_dct_calc(DCTContext *s, FFTSample *data); -void av_dct_end (DCTContext *s); - -/** - * @} - */ - -#endif /* AVCODEC_AVFFT_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/d3d11va.h b/src/main/jni/ffmpeg/include/libavcodec/d3d11va.h deleted file mode 100644 index 6816b6c1e..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/d3d11va.h +++ /dev/null @@ -1,112 +0,0 @@ -/* - * Direct3D11 HW acceleration - * - * copyright (c) 2009 Laurent Aimar - * copyright (c) 2015 Steve Lhomme - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_D3D11VA_H -#define AVCODEC_D3D11VA_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_d3d11va - * Public libavcodec D3D11VA header. - */ - -#if !defined(_WIN32_WINNT) || _WIN32_WINNT < 0x0602 -#undef _WIN32_WINNT -#define _WIN32_WINNT 0x0602 -#endif - -#include <stdint.h> -#include <d3d11.h> - -/** - * @defgroup lavc_codec_hwaccel_d3d11va Direct3D11 - * @ingroup lavc_codec_hwaccel - * - * @{ - */ - -#define FF_DXVA2_WORKAROUND_SCALING_LIST_ZIGZAG 1 ///< Work around for Direct3D11 and old UVD/UVD+ ATI video cards -#define FF_DXVA2_WORKAROUND_INTEL_CLEARVIDEO 2 ///< Work around for Direct3D11 and old Intel GPUs with ClearVideo interface - -/** - * This structure is used to provides the necessary configurations and data - * to the Direct3D11 FFmpeg HWAccel implementation. - * - * The application must make it available as AVCodecContext.hwaccel_context. - * - * Use av_d3d11va_alloc_context() exclusively to allocate an AVD3D11VAContext. - */ -typedef struct AVD3D11VAContext { - /** - * D3D11 decoder object - */ - ID3D11VideoDecoder *decoder; - - /** - * D3D11 VideoContext - */ - ID3D11VideoContext *video_context; - - /** - * D3D11 configuration used to create the decoder - */ - D3D11_VIDEO_DECODER_CONFIG *cfg; - - /** - * The number of surface in the surface array - */ - unsigned surface_count; - - /** - * The array of Direct3D surfaces used to create the decoder - */ - ID3D11VideoDecoderOutputView **surface; - - /** - * A bit field configuring the workarounds needed for using the decoder - */ - uint64_t workaround; - - /** - * Private to the FFmpeg AVHWAccel implementation - */ - unsigned report_id; - - /** - * Mutex to access video_context - */ - HANDLE context_mutex; -} AVD3D11VAContext; - -/** - * Allocate an AVD3D11VAContext. - * - * @return Newly-allocated AVD3D11VAContext or NULL on failure. - */ -AVD3D11VAContext *av_d3d11va_alloc_context(void); - -/** - * @} - */ - -#endif /* AVCODEC_D3D11VA_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/dirac.h b/src/main/jni/ffmpeg/include/libavcodec/dirac.h deleted file mode 100644 index 8d7953d9b..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/dirac.h +++ /dev/null @@ -1,113 +0,0 @@ -/* - * Copyright (C) 2007 Marco Gerards <marco@gnu.org> - * Copyright (C) 2009 David Conrad - * Copyright (C) 2011 Jordi Ortiz - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_DIRAC_H -#define AVCODEC_DIRAC_H - -/** - * @file - * Interface to Dirac Decoder/Encoder - * @author Marco Gerards <marco@gnu.org> - * @author David Conrad - * @author Jordi Ortiz - */ - -#include "avcodec.h" - -/** - * Parse code values: - * - * Dirac Specification -> - * 9.6.1 Table 9.1 - * - * VC-2 Specification -> - * 10.4.1 Table 10.1 - */ - -enum DiracParseCodes { - DIRAC_PCODE_SEQ_HEADER = 0x00, - DIRAC_PCODE_END_SEQ = 0x10, - DIRAC_PCODE_AUX = 0x20, - DIRAC_PCODE_PAD = 0x30, - DIRAC_PCODE_PICTURE_CODED = 0x08, - DIRAC_PCODE_PICTURE_RAW = 0x48, - DIRAC_PCODE_PICTURE_LOW_DEL = 0xC8, - DIRAC_PCODE_PICTURE_HQ = 0xE8, - DIRAC_PCODE_MAGIC = 0x42424344, -}; - -typedef struct DiracVersionInfo { - int major; - int minor; -} DiracVersionInfo; - -typedef struct AVDiracSeqHeader { - unsigned width; - unsigned height; - uint8_t chroma_format; ///< 0: 444 1: 422 2: 420 - - uint8_t interlaced; - uint8_t top_field_first; - - uint8_t frame_rate_index; ///< index into dirac_frame_rate[] - uint8_t aspect_ratio_index; ///< index into dirac_aspect_ratio[] - - uint16_t clean_width; - uint16_t clean_height; - uint16_t clean_left_offset; - uint16_t clean_right_offset; - - uint8_t pixel_range_index; ///< index into dirac_pixel_range_presets[] - uint8_t color_spec_index; ///< index into dirac_color_spec_presets[] - - int profile; - int level; - - AVRational framerate; - AVRational sample_aspect_ratio; - - enum AVPixelFormat pix_fmt; - enum AVColorRange color_range; - enum AVColorPrimaries color_primaries; - enum AVColorTransferCharacteristic color_trc; - enum AVColorSpace colorspace; - - DiracVersionInfo version; - int bit_depth; -} AVDiracSeqHeader; - -/** - * Parse a Dirac sequence header. - * - * @param dsh this function will allocate and fill an AVDiracSeqHeader struct - * and write it into this pointer. The caller must free it with - * av_free(). - * @param buf the data buffer - * @param buf_size the size of the data buffer in bytes - * @param log_ctx if non-NULL, this function will log errors here - * @return 0 on success, a negative AVERROR code on failure - */ -int av_dirac_parse_sequence_header(AVDiracSeqHeader **dsh, - const uint8_t *buf, size_t buf_size, - void *log_ctx); - -#endif /* AVCODEC_DIRAC_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/dv_profile.h b/src/main/jni/ffmpeg/include/libavcodec/dv_profile.h deleted file mode 100644 index 9380a66f0..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/dv_profile.h +++ /dev/null @@ -1,83 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_DV_PROFILE_H -#define AVCODEC_DV_PROFILE_H - -#include <stdint.h> - -#include "libavutil/pixfmt.h" -#include "libavutil/rational.h" -#include "avcodec.h" - -/* minimum number of bytes to read from a DV stream in order to - * determine the profile */ -#define DV_PROFILE_BYTES (6 * 80) /* 6 DIF blocks */ - - -/* - * AVDVProfile is used to express the differences between various - * DV flavors. For now it's primarily used for differentiating - * 525/60 and 625/50, but the plans are to use it for various - * DV specs as well (e.g. SMPTE314M vs. IEC 61834). - */ -typedef struct AVDVProfile { - int dsf; /* value of the dsf in the DV header */ - int video_stype; /* stype for VAUX source pack */ - int frame_size; /* total size of one frame in bytes */ - int difseg_size; /* number of DIF segments per DIF channel */ - int n_difchan; /* number of DIF channels per frame */ - AVRational time_base; /* 1/framerate */ - int ltc_divisor; /* FPS from the LTS standpoint */ - int height; /* picture height in pixels */ - int width; /* picture width in pixels */ - AVRational sar[2]; /* sample aspect ratios for 4:3 and 16:9 */ - enum AVPixelFormat pix_fmt; /* picture pixel format */ - int bpm; /* blocks per macroblock */ - const uint8_t *block_sizes; /* AC block sizes, in bits */ - int audio_stride; /* size of audio_shuffle table */ - int audio_min_samples[3]; /* min amount of audio samples */ - /* for 48kHz, 44.1kHz and 32kHz */ - int audio_samples_dist[5]; /* how many samples are supposed to be */ - /* in each frame in a 5 frames window */ - const uint8_t (*audio_shuffle)[9]; /* PCM shuffling table */ -} AVDVProfile; - -/** - * Get a DV profile for the provided compressed frame. - * - * @param sys the profile used for the previous frame, may be NULL - * @param frame the compressed data buffer - * @param buf_size size of the buffer in bytes - * @return the DV profile for the supplied data or NULL on failure - */ -const AVDVProfile *av_dv_frame_profile(const AVDVProfile *sys, - const uint8_t *frame, unsigned buf_size); - -/** - * Get a DV profile for the provided stream parameters. - */ -const AVDVProfile *av_dv_codec_profile(int width, int height, enum AVPixelFormat pix_fmt); - -/** - * Get a DV profile for the provided stream parameters. - * The frame rate is used as a best-effort parameter. - */ -const AVDVProfile *av_dv_codec_profile2(int width, int height, enum AVPixelFormat pix_fmt, AVRational frame_rate); - -#endif /* AVCODEC_DV_PROFILE_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/dxva2.h b/src/main/jni/ffmpeg/include/libavcodec/dxva2.h deleted file mode 100644 index 9e3ab86a8..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/dxva2.h +++ /dev/null @@ -1,93 +0,0 @@ -/* - * DXVA2 HW acceleration - * - * copyright (c) 2009 Laurent Aimar - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_DXVA_H -#define AVCODEC_DXVA_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_dxva2 - * Public libavcodec DXVA2 header. - */ - -#if !defined(_WIN32_WINNT) || _WIN32_WINNT < 0x0602 -#undef _WIN32_WINNT -#define _WIN32_WINNT 0x0602 -#endif - -#include <stdint.h> -#include <d3d9.h> -#include <dxva2api.h> - -/** - * @defgroup lavc_codec_hwaccel_dxva2 DXVA2 - * @ingroup lavc_codec_hwaccel - * - * @{ - */ - -#define FF_DXVA2_WORKAROUND_SCALING_LIST_ZIGZAG 1 ///< Work around for DXVA2 and old UVD/UVD+ ATI video cards -#define FF_DXVA2_WORKAROUND_INTEL_CLEARVIDEO 2 ///< Work around for DXVA2 and old Intel GPUs with ClearVideo interface - -/** - * This structure is used to provides the necessary configurations and data - * to the DXVA2 FFmpeg HWAccel implementation. - * - * The application must make it available as AVCodecContext.hwaccel_context. - */ -struct dxva_context { - /** - * DXVA2 decoder object - */ - IDirectXVideoDecoder *decoder; - - /** - * DXVA2 configuration used to create the decoder - */ - const DXVA2_ConfigPictureDecode *cfg; - - /** - * The number of surface in the surface array - */ - unsigned surface_count; - - /** - * The array of Direct3D surfaces used to create the decoder - */ - LPDIRECT3DSURFACE9 *surface; - - /** - * A bit field configuring the workarounds needed for using the decoder - */ - uint64_t workaround; - - /** - * Private to the FFmpeg AVHWAccel implementation - */ - unsigned report_id; -}; - -/** - * @} - */ - -#endif /* AVCODEC_DXVA_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/qsv.h b/src/main/jni/ffmpeg/include/libavcodec/qsv.h deleted file mode 100644 index b77158ec2..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/qsv.h +++ /dev/null @@ -1,107 +0,0 @@ -/* - * Intel MediaSDK QSV public API - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_QSV_H -#define AVCODEC_QSV_H - -#include <mfx/mfxvideo.h> - -#include "libavutil/buffer.h" - -/** - * This struct is used for communicating QSV parameters between libavcodec and - * the caller. It is managed by the caller and must be assigned to - * AVCodecContext.hwaccel_context. - * - decoding: hwaccel_context must be set on return from the get_format() - * callback - * - encoding: hwaccel_context must be set before avcodec_open2() - */ -typedef struct AVQSVContext { - /** - * If non-NULL, the session to use for encoding or decoding. - * Otherwise, libavcodec will try to create an internal session. - */ - mfxSession session; - - /** - * The IO pattern to use. - */ - int iopattern; - - /** - * Extra buffers to pass to encoder or decoder initialization. - */ - mfxExtBuffer **ext_buffers; - int nb_ext_buffers; - - /** - * Encoding only. If this field is set to non-zero by the caller, libavcodec - * will create an mfxExtOpaqueSurfaceAlloc extended buffer and pass it to - * the encoder initialization. This only makes sense if iopattern is also - * set to MFX_IOPATTERN_IN_OPAQUE_MEMORY. - * - * The number of allocated opaque surfaces will be the sum of the number - * required by the encoder and the user-provided value nb_opaque_surfaces. - * The array of the opaque surfaces will be exported to the caller through - * the opaque_surfaces field. - */ - int opaque_alloc; - - /** - * Encoding only, and only if opaque_alloc is set to non-zero. Before - * calling avcodec_open2(), the caller should set this field to the number - * of extra opaque surfaces to allocate beyond what is required by the - * encoder. - * - * On return from avcodec_open2(), this field will be set by libavcodec to - * the total number of allocated opaque surfaces. - */ - int nb_opaque_surfaces; - - /** - * Encoding only, and only if opaque_alloc is set to non-zero. On return - * from avcodec_open2(), this field will be used by libavcodec to export the - * array of the allocated opaque surfaces to the caller, so they can be - * passed to other parts of the pipeline. - * - * The buffer reference exported here is owned and managed by libavcodec, - * the callers should make their own reference with av_buffer_ref() and free - * it with av_buffer_unref() when it is no longer needed. - * - * The buffer data is an nb_opaque_surfaces-sized array of mfxFrameSurface1. - */ - AVBufferRef *opaque_surfaces; - - /** - * Encoding only, and only if opaque_alloc is set to non-zero. On return - * from avcodec_open2(), this field will be set to the surface type used in - * the opaque allocation request. - */ - int opaque_alloc_type; -} AVQSVContext; - -/** - * Allocate a new context. - * - * It must be freed by the caller with av_free(). - */ -AVQSVContext *av_qsv_alloc_context(void); - -#endif /* AVCODEC_QSV_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/vaapi.h b/src/main/jni/ffmpeg/include/libavcodec/vaapi.h deleted file mode 100644 index 7a29f6f88..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/vaapi.h +++ /dev/null @@ -1,189 +0,0 @@ -/* - * Video Acceleration API (shared data between FFmpeg and the video player) - * HW decode acceleration for MPEG-2, MPEG-4, H.264 and VC-1 - * - * Copyright (C) 2008-2009 Splitted-Desktop Systems - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_VAAPI_H -#define AVCODEC_VAAPI_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_vaapi - * Public libavcodec VA API header. - */ - -#include <stdint.h> -#include "libavutil/attributes.h" -#include "version.h" - -/** - * @defgroup lavc_codec_hwaccel_vaapi VA API Decoding - * @ingroup lavc_codec_hwaccel - * @{ - */ - -/** - * This structure is used to share data between the FFmpeg library and - * the client video application. - * This shall be zero-allocated and available as - * AVCodecContext.hwaccel_context. All user members can be set once - * during initialization or through each AVCodecContext.get_buffer() - * function call. In any case, they must be valid prior to calling - * decoding functions. - */ -struct vaapi_context { - /** - * Window system dependent data - * - * - encoding: unused - * - decoding: Set by user - */ - void *display; - - /** - * Configuration ID - * - * - encoding: unused - * - decoding: Set by user - */ - uint32_t config_id; - - /** - * Context ID (video decode pipeline) - * - * - encoding: unused - * - decoding: Set by user - */ - uint32_t context_id; - -#if FF_API_VAAPI_CONTEXT - /** - * VAPictureParameterBuffer ID - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - uint32_t pic_param_buf_id; - - /** - * VAIQMatrixBuffer ID - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - uint32_t iq_matrix_buf_id; - - /** - * VABitPlaneBuffer ID (for VC-1 decoding) - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - uint32_t bitplane_buf_id; - - /** - * Slice parameter/data buffer IDs - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - uint32_t *slice_buf_ids; - - /** - * Number of effective slice buffer IDs to send to the HW - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - unsigned int n_slice_buf_ids; - - /** - * Size of pre-allocated slice_buf_ids - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - unsigned int slice_buf_ids_alloc; - - /** - * Pointer to VASliceParameterBuffers - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - void *slice_params; - - /** - * Size of a VASliceParameterBuffer element - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - unsigned int slice_param_size; - - /** - * Size of pre-allocated slice_params - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - unsigned int slice_params_alloc; - - /** - * Number of slices currently filled in - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - unsigned int slice_count; - - /** - * Pointer to slice data buffer base - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - const uint8_t *slice_data; - - /** - * Current size of slice data - * - * - encoding: unused - * - decoding: Set by libavcodec - */ - attribute_deprecated - uint32_t slice_data_size; -#endif -}; - -/* @} */ - -#endif /* AVCODEC_VAAPI_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/vda.h b/src/main/jni/ffmpeg/include/libavcodec/vda.h deleted file mode 100644 index bde14e31d..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/vda.h +++ /dev/null @@ -1,230 +0,0 @@ -/* - * VDA HW acceleration - * - * copyright (c) 2011 Sebastien Zwickert - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_VDA_H -#define AVCODEC_VDA_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_vda - * Public libavcodec VDA header. - */ - -#include "libavcodec/avcodec.h" - -#include <stdint.h> - -// emmintrin.h is unable to compile with -std=c99 -Werror=missing-prototypes -// http://openradar.appspot.com/8026390 -#undef __GNUC_STDC_INLINE__ - -#define Picture QuickdrawPicture -#include <VideoDecodeAcceleration/VDADecoder.h> -#undef Picture - -#include "libavcodec/version.h" - -// extra flags not defined in VDADecoder.h -enum { - kVDADecodeInfo_Asynchronous = 1UL << 0, - kVDADecodeInfo_FrameDropped = 1UL << 1 -}; - -/** - * @defgroup lavc_codec_hwaccel_vda VDA - * @ingroup lavc_codec_hwaccel - * - * @{ - */ - -/** - * This structure is used to provide the necessary configurations and data - * to the VDA FFmpeg HWAccel implementation. - * - * The application must make it available as AVCodecContext.hwaccel_context. - */ -struct vda_context { - /** - * VDA decoder object. - * - * - encoding: unused - * - decoding: Set/Unset by libavcodec. - */ - VDADecoder decoder; - - /** - * The Core Video pixel buffer that contains the current image data. - * - * encoding: unused - * decoding: Set by libavcodec. Unset by user. - */ - CVPixelBufferRef cv_buffer; - - /** - * Use the hardware decoder in synchronous mode. - * - * encoding: unused - * decoding: Set by user. - */ - int use_sync_decoding; - - /** - * The frame width. - * - * - encoding: unused - * - decoding: Set/Unset by user. - */ - int width; - - /** - * The frame height. - * - * - encoding: unused - * - decoding: Set/Unset by user. - */ - int height; - - /** - * The frame format. - * - * - encoding: unused - * - decoding: Set/Unset by user. - */ - int format; - - /** - * The pixel format for output image buffers. - * - * - encoding: unused - * - decoding: Set/Unset by user. - */ - OSType cv_pix_fmt_type; - - /** - * unused - */ - uint8_t *priv_bitstream; - - /** - * unused - */ - int priv_bitstream_size; - - /** - * unused - */ - int priv_allocated_size; - - /** - * Use av_buffer to manage buffer. - * When the flag is set, the CVPixelBuffers returned by the decoder will - * be released automatically, so you have to retain them if necessary. - * Not setting this flag may cause memory leak. - * - * encoding: unused - * decoding: Set by user. - */ - int use_ref_buffer; -}; - -/** Create the video decoder. */ -int ff_vda_create_decoder(struct vda_context *vda_ctx, - uint8_t *extradata, - int extradata_size); - -/** Destroy the video decoder. */ -int ff_vda_destroy_decoder(struct vda_context *vda_ctx); - -/** - * This struct holds all the information that needs to be passed - * between the caller and libavcodec for initializing VDA decoding. - * Its size is not a part of the public ABI, it must be allocated with - * av_vda_alloc_context() and freed with av_free(). - */ -typedef struct AVVDAContext { - /** - * VDA decoder object. Created and freed by the caller. - */ - VDADecoder decoder; - - /** - * The output callback that must be passed to VDADecoderCreate. - * Set by av_vda_alloc_context(). - */ - VDADecoderOutputCallback output_callback; - - /** - * CVPixelBuffer Format Type that VDA will use for decoded frames; set by - * the caller. - */ - OSType cv_pix_fmt_type; -} AVVDAContext; - -/** - * Allocate and initialize a VDA context. - * - * This function should be called from the get_format() callback when the caller - * selects the AV_PIX_FMT_VDA format. The caller must then create the decoder - * object (using the output callback provided by libavcodec) that will be used - * for VDA-accelerated decoding. - * - * When decoding with VDA is finished, the caller must destroy the decoder - * object and free the VDA context using av_free(). - * - * @return the newly allocated context or NULL on failure - */ -AVVDAContext *av_vda_alloc_context(void); - -/** - * This is a convenience function that creates and sets up the VDA context using - * an internal implementation. - * - * @param avctx the corresponding codec context - * - * @return >= 0 on success, a negative AVERROR code on failure - */ -int av_vda_default_init(AVCodecContext *avctx); - -/** - * This is a convenience function that creates and sets up the VDA context using - * an internal implementation. - * - * @param avctx the corresponding codec context - * @param vdactx the VDA context to use - * - * @return >= 0 on success, a negative AVERROR code on failure - */ -int av_vda_default_init2(AVCodecContext *avctx, AVVDAContext *vdactx); - -/** - * This function must be called to free the VDA context initialized with - * av_vda_default_init(). - * - * @param avctx the corresponding codec context - */ -void av_vda_default_free(AVCodecContext *avctx); - -/** - * @} - */ - -#endif /* AVCODEC_VDA_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/vdpau.h b/src/main/jni/ffmpeg/include/libavcodec/vdpau.h deleted file mode 100644 index e85e4d9e9..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/vdpau.h +++ /dev/null @@ -1,253 +0,0 @@ -/* - * The Video Decode and Presentation API for UNIX (VDPAU) is used for - * hardware-accelerated decoding of MPEG-1/2, H.264 and VC-1. - * - * Copyright (C) 2008 NVIDIA - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_VDPAU_H -#define AVCODEC_VDPAU_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_vdpau - * Public libavcodec VDPAU header. - */ - - -/** - * @defgroup lavc_codec_hwaccel_vdpau VDPAU Decoder and Renderer - * @ingroup lavc_codec_hwaccel - * - * VDPAU hardware acceleration has two modules - * - VDPAU decoding - * - VDPAU presentation - * - * The VDPAU decoding module parses all headers using FFmpeg - * parsing mechanisms and uses VDPAU for the actual decoding. - * - * As per the current implementation, the actual decoding - * and rendering (API calls) are done as part of the VDPAU - * presentation (vo_vdpau.c) module. - * - * @{ - */ - -#include <vdpau/vdpau.h> -#include <vdpau/vdpau_x11.h> -#include "libavutil/avconfig.h" -#include "libavutil/attributes.h" - -#include "avcodec.h" -#include "version.h" - -#if FF_API_BUFS_VDPAU -union AVVDPAUPictureInfo { - VdpPictureInfoH264 h264; - VdpPictureInfoMPEG1Or2 mpeg; - VdpPictureInfoVC1 vc1; - VdpPictureInfoMPEG4Part2 mpeg4; -}; -#endif - -struct AVCodecContext; -struct AVFrame; - -typedef int (*AVVDPAU_Render2)(struct AVCodecContext *, struct AVFrame *, - const VdpPictureInfo *, uint32_t, - const VdpBitstreamBuffer *); - -/** - * This structure is used to share data between the libavcodec library and - * the client video application. - * The user shall allocate the structure via the av_alloc_vdpau_hwaccel - * function and make it available as - * AVCodecContext.hwaccel_context. Members can be set by the user once - * during initialization or through each AVCodecContext.get_buffer() - * function call. In any case, they must be valid prior to calling - * decoding functions. - * - * The size of this structure is not a part of the public ABI and must not - * be used outside of libavcodec. Use av_vdpau_alloc_context() to allocate an - * AVVDPAUContext. - */ -typedef struct AVVDPAUContext { - /** - * VDPAU decoder handle - * - * Set by user. - */ - VdpDecoder decoder; - - /** - * VDPAU decoder render callback - * - * Set by the user. - */ - VdpDecoderRender *render; - -#if FF_API_BUFS_VDPAU - /** - * VDPAU picture information - * - * Set by libavcodec. - */ - attribute_deprecated - union AVVDPAUPictureInfo info; - - /** - * Allocated size of the bitstream_buffers table. - * - * Set by libavcodec. - */ - attribute_deprecated - int bitstream_buffers_allocated; - - /** - * Useful bitstream buffers in the bitstream buffers table. - * - * Set by libavcodec. - */ - attribute_deprecated - int bitstream_buffers_used; - - /** - * Table of bitstream buffers. - * The user is responsible for freeing this buffer using av_freep(). - * - * Set by libavcodec. - */ - attribute_deprecated - VdpBitstreamBuffer *bitstream_buffers; -#endif - AVVDPAU_Render2 render2; -} AVVDPAUContext; - -/** - * @brief allocation function for AVVDPAUContext - * - * Allows extending the struct without breaking API/ABI - */ -AVVDPAUContext *av_alloc_vdpaucontext(void); - -AVVDPAU_Render2 av_vdpau_hwaccel_get_render2(const AVVDPAUContext *); -void av_vdpau_hwaccel_set_render2(AVVDPAUContext *, AVVDPAU_Render2); - -/** - * Associate a VDPAU device with a codec context for hardware acceleration. - * This function is meant to be called from the get_format() codec callback, - * or earlier. It can also be called after avcodec_flush_buffers() to change - * the underlying VDPAU device mid-stream (e.g. to recover from non-transparent - * display preemption). - * - * @note get_format() must return AV_PIX_FMT_VDPAU if this function completes - * successfully. - * - * @param avctx decoding context whose get_format() callback is invoked - * @param device VDPAU device handle to use for hardware acceleration - * @param get_proc_address VDPAU device driver - * @param flags zero of more OR'd AV_HWACCEL_FLAG_* flags - * - * @return 0 on success, an AVERROR code on failure. - */ -int av_vdpau_bind_context(AVCodecContext *avctx, VdpDevice device, - VdpGetProcAddress *get_proc_address, unsigned flags); - -/** - * Gets the parameters to create an adequate VDPAU video surface for the codec - * context using VDPAU hardware decoding acceleration. - * - * @note Behavior is undefined if the context was not successfully bound to a - * VDPAU device using av_vdpau_bind_context(). - * - * @param avctx the codec context being used for decoding the stream - * @param type storage space for the VDPAU video surface chroma type - * (or NULL to ignore) - * @param width storage space for the VDPAU video surface pixel width - * (or NULL to ignore) - * @param height storage space for the VDPAU video surface pixel height - * (or NULL to ignore) - * - * @return 0 on success, a negative AVERROR code on failure. - */ -int av_vdpau_get_surface_parameters(AVCodecContext *avctx, VdpChromaType *type, - uint32_t *width, uint32_t *height); - -/** - * Allocate an AVVDPAUContext. - * - * @return Newly-allocated AVVDPAUContext or NULL on failure. - */ -AVVDPAUContext *av_vdpau_alloc_context(void); - -#if FF_API_VDPAU_PROFILE -/** - * Get a decoder profile that should be used for initializing a VDPAU decoder. - * Should be called from the AVCodecContext.get_format() callback. - * - * @deprecated Use av_vdpau_bind_context() instead. - * - * @param avctx the codec context being used for decoding the stream - * @param profile a pointer into which the result will be written on success. - * The contents of profile are undefined if this function returns - * an error. - * - * @return 0 on success (non-negative), a negative AVERROR on failure. - */ -attribute_deprecated -int av_vdpau_get_profile(AVCodecContext *avctx, VdpDecoderProfile *profile); -#endif - -#if FF_API_CAP_VDPAU -/** @brief The videoSurface is used for rendering. */ -#define FF_VDPAU_STATE_USED_FOR_RENDER 1 - -/** - * @brief The videoSurface is needed for reference/prediction. - * The codec manipulates this. - */ -#define FF_VDPAU_STATE_USED_FOR_REFERENCE 2 - -/** - * @brief This structure is used as a callback between the FFmpeg - * decoder (vd_) and presentation (vo_) module. - * This is used for defining a video frame containing surface, - * picture parameter, bitstream information etc which are passed - * between the FFmpeg decoder and its clients. - */ -struct vdpau_render_state { - VdpVideoSurface surface; ///< Used as rendered surface, never changed. - - int state; ///< Holds FF_VDPAU_STATE_* values. - - /** picture parameter information for all supported codecs */ - union AVVDPAUPictureInfo info; - - /** Describe size/location of the compressed video data. - Set to 0 when freeing bitstream_buffers. */ - int bitstream_buffers_allocated; - int bitstream_buffers_used; - /** The user is responsible for freeing this buffer using av_freep(). */ - VdpBitstreamBuffer *bitstream_buffers; -}; -#endif - -/* @}*/ - -#endif /* AVCODEC_VDPAU_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/version.h b/src/main/jni/ffmpeg/include/libavcodec/version.h deleted file mode 100644 index 9846275e0..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/version.h +++ /dev/null @@ -1,210 +0,0 @@ -/* - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_VERSION_H -#define AVCODEC_VERSION_H - -/** - * @file - * @ingroup libavc - * Libavcodec version macros. - */ - -#include "libavutil/version.h" - -#define LIBAVCODEC_VERSION_MAJOR 57 -#define LIBAVCODEC_VERSION_MINOR 21 -#define LIBAVCODEC_VERSION_MICRO 100 - -#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \ - LIBAVCODEC_VERSION_MINOR, \ - LIBAVCODEC_VERSION_MICRO) -#define LIBAVCODEC_VERSION AV_VERSION(LIBAVCODEC_VERSION_MAJOR, \ - LIBAVCODEC_VERSION_MINOR, \ - LIBAVCODEC_VERSION_MICRO) -#define LIBAVCODEC_BUILD LIBAVCODEC_VERSION_INT - -#define LIBAVCODEC_IDENT "Lavc" AV_STRINGIFY(LIBAVCODEC_VERSION) - -/** - * FF_API_* defines may be placed below to indicate public API that will be - * dropped at a future version bump. The defines themselves are not part of - * the public API and may change, break or disappear at any time. - * - * @note, when bumping the major version it is recommended to manually - * disable each FF_API_* in its own commit instead of disabling them all - * at once through the bump. This improves the git bisect-ability of the change. - */ - -#ifndef FF_API_VIMA_DECODER -#define FF_API_VIMA_DECODER (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_AUDIO_CONVERT -#define FF_API_AUDIO_CONVERT (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_AVCODEC_RESAMPLE -#define FF_API_AVCODEC_RESAMPLE FF_API_AUDIO_CONVERT -#endif -#ifndef FF_API_MISSING_SAMPLE -#define FF_API_MISSING_SAMPLE (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_LOWRES -#define FF_API_LOWRES (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_CAP_VDPAU -#define FF_API_CAP_VDPAU (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_BUFS_VDPAU -#define FF_API_BUFS_VDPAU (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_VOXWARE -#define FF_API_VOXWARE (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_SET_DIMENSIONS -#define FF_API_SET_DIMENSIONS (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_DEBUG_MV -#define FF_API_DEBUG_MV (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_AC_VLC -#define FF_API_AC_VLC (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_OLD_MSMPEG4 -#define FF_API_OLD_MSMPEG4 (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_ASPECT_EXTENDED -#define FF_API_ASPECT_EXTENDED (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_ARCH_ALPHA -#define FF_API_ARCH_ALPHA (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_XVMC -#define FF_API_XVMC (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_ERROR_RATE -#define FF_API_ERROR_RATE (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_QSCALE_TYPE -#define FF_API_QSCALE_TYPE (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_MB_TYPE -#define FF_API_MB_TYPE (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_MAX_BFRAMES -#define FF_API_MAX_BFRAMES (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_NEG_LINESIZES -#define FF_API_NEG_LINESIZES (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_EMU_EDGE -#define FF_API_EMU_EDGE (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_ARCH_SH4 -#define FF_API_ARCH_SH4 (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_ARCH_SPARC -#define FF_API_ARCH_SPARC (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_UNUSED_MEMBERS -#define FF_API_UNUSED_MEMBERS (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_IDCT_XVIDMMX -#define FF_API_IDCT_XVIDMMX (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_INPUT_PRESERVED -#define FF_API_INPUT_PRESERVED (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_NORMALIZE_AQP -#define FF_API_NORMALIZE_AQP (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_GMC -#define FF_API_GMC (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_MV0 -#define FF_API_MV0 (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_CODEC_NAME -#define FF_API_CODEC_NAME (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_AFD -#define FF_API_AFD (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_VISMV -/* XXX: don't forget to drop the -vismv documentation */ -#define FF_API_VISMV (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_AUDIOENC_DELAY -#define FF_API_AUDIOENC_DELAY (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_VAAPI_CONTEXT -#define FF_API_VAAPI_CONTEXT (LIBAVCODEC_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_AVCTX_TIMEBASE -#define FF_API_AVCTX_TIMEBASE (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_MPV_OPT -#define FF_API_MPV_OPT (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_STREAM_CODEC_TAG -#define FF_API_STREAM_CODEC_TAG (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_QUANT_BIAS -#define FF_API_QUANT_BIAS (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_RC_STRATEGY -#define FF_API_RC_STRATEGY (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_CODED_FRAME -#define FF_API_CODED_FRAME (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_MOTION_EST -#define FF_API_MOTION_EST (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_WITHOUT_PREFIX -#define FF_API_WITHOUT_PREFIX (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_SIDEDATA_ONLY_PKT -#define FF_API_SIDEDATA_ONLY_PKT (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_VDPAU_PROFILE -#define FF_API_VDPAU_PROFILE (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_CONVERGENCE_DURATION -#define FF_API_CONVERGENCE_DURATION (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_AVPICTURE -#define FF_API_AVPICTURE (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_AVPACKET_OLD_API -#define FF_API_AVPACKET_OLD_API (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_RTP_CALLBACK -#define FF_API_RTP_CALLBACK (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_VBV_DELAY -#define FF_API_VBV_DELAY (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_CODER_TYPE -#define FF_API_CODER_TYPE (LIBAVCODEC_VERSION_MAJOR < 59) -#endif -#ifndef FF_API_STAT_BITS -#define FF_API_STAT_BITS (LIBAVCODEC_VERSION_MAJOR < 59) -#endif - -#endif /* AVCODEC_VERSION_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/videotoolbox.h b/src/main/jni/ffmpeg/include/libavcodec/videotoolbox.h deleted file mode 100644 index a48638e2b..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/videotoolbox.h +++ /dev/null @@ -1,126 +0,0 @@ -/* - * Videotoolbox hardware acceleration - * - * copyright (c) 2012 Sebastien Zwickert - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_VIDEOTOOLBOX_H -#define AVCODEC_VIDEOTOOLBOX_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_videotoolbox - * Public libavcodec Videotoolbox header. - */ - -#include <stdint.h> - -#define Picture QuickdrawPicture -#include <VideoToolbox/VideoToolbox.h> -#undef Picture - -#include "libavcodec/avcodec.h" - -/** - * This struct holds all the information that needs to be passed - * between the caller and libavcodec for initializing Videotoolbox decoding. - * Its size is not a part of the public ABI, it must be allocated with - * av_videotoolbox_alloc_context() and freed with av_free(). - */ -typedef struct AVVideotoolboxContext { - /** - * Videotoolbox decompression session object. - * Created and freed the caller. - */ - VTDecompressionSessionRef session; - - /** - * The output callback that must be passed to the session. - * Set by av_videottoolbox_default_init() - */ - VTDecompressionOutputCallback output_callback; - - /** - * CVPixelBuffer Format Type that Videotoolbox will use for decoded frames. - * set by the caller. - */ - OSType cv_pix_fmt_type; - - /** - * CoreMedia Format Description that Videotoolbox will use to create the decompression session. - * Set by the caller. - */ - CMVideoFormatDescriptionRef cm_fmt_desc; - - /** - * CoreMedia codec type that Videotoolbox will use to create the decompression session. - * Set by the caller. - */ - int cm_codec_type; -} AVVideotoolboxContext; - -/** - * Allocate and initialize a Videotoolbox context. - * - * This function should be called from the get_format() callback when the caller - * selects the AV_PIX_FMT_VIDETOOLBOX format. The caller must then create - * the decoder object (using the output callback provided by libavcodec) that - * will be used for Videotoolbox-accelerated decoding. - * - * When decoding with Videotoolbox is finished, the caller must destroy the decoder - * object and free the Videotoolbox context using av_free(). - * - * @return the newly allocated context or NULL on failure - */ -AVVideotoolboxContext *av_videotoolbox_alloc_context(void); - -/** - * This is a convenience function that creates and sets up the Videotoolbox context using - * an internal implementation. - * - * @param avctx the corresponding codec context - * - * @return >= 0 on success, a negative AVERROR code on failure - */ -int av_videotoolbox_default_init(AVCodecContext *avctx); - -/** - * This is a convenience function that creates and sets up the Videotoolbox context using - * an internal implementation. - * - * @param avctx the corresponding codec context - * @param vtctx the Videotoolbox context to use - * - * @return >= 0 on success, a negative AVERROR code on failure - */ -int av_videotoolbox_default_init2(AVCodecContext *avctx, AVVideotoolboxContext *vtctx); - -/** - * This function must be called to free the Videotoolbox context initialized with - * av_videotoolbox_default_init(). - * - * @param avctx the corresponding codec context - */ -void av_videotoolbox_default_free(AVCodecContext *avctx); - -/** - * @} - */ - -#endif /* AVCODEC_VIDEOTOOLBOX_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/vorbis_parser.h b/src/main/jni/ffmpeg/include/libavcodec/vorbis_parser.h deleted file mode 100644 index 06e48bd3b..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/vorbis_parser.h +++ /dev/null @@ -1,78 +0,0 @@ -/* - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * A public API for Vorbis parsing - * - * Determines the duration for each packet. - */ - -#ifndef AVCODEC_VORBIS_PARSE_H -#define AVCODEC_VORBIS_PARSE_H - -#include <stdint.h> - -typedef struct AVVorbisParseContext AVVorbisParseContext; - -/** - * Allocate and initialize the Vorbis parser using headers in the extradata. - * - * @param avctx codec context - * @param s Vorbis parser context - */ -AVVorbisParseContext *av_vorbis_parse_init(const uint8_t *extradata, - int extradata_size); - -/** - * Free the parser and everything associated with it. - */ -void av_vorbis_parse_free(AVVorbisParseContext **s); - -#define VORBIS_FLAG_HEADER 0x00000001 -#define VORBIS_FLAG_COMMENT 0x00000002 -#define VORBIS_FLAG_SETUP 0x00000004 - -/** - * Get the duration for a Vorbis packet. - * - * If @p flags is @c NULL, - * special frames are considered invalid. - * - * @param s Vorbis parser context - * @param buf buffer containing a Vorbis frame - * @param buf_size size of the buffer - * @param flags flags for special frames - */ -int av_vorbis_parse_frame_flags(AVVorbisParseContext *s, const uint8_t *buf, - int buf_size, int *flags); - -/** - * Get the duration for a Vorbis packet. - * - * @param s Vorbis parser context - * @param buf buffer containing a Vorbis frame - * @param buf_size size of the buffer - */ -int av_vorbis_parse_frame(AVVorbisParseContext *s, const uint8_t *buf, - int buf_size); - -void av_vorbis_parse_reset(AVVorbisParseContext *s); - -#endif /* AVCODEC_VORBIS_PARSE_H */ diff --git a/src/main/jni/ffmpeg/include/libavcodec/xvmc.h b/src/main/jni/ffmpeg/include/libavcodec/xvmc.h deleted file mode 100644 index 465ee78d6..000000000 --- a/src/main/jni/ffmpeg/include/libavcodec/xvmc.h +++ /dev/null @@ -1,170 +0,0 @@ -/* - * Copyright (C) 2003 Ivan Kalvachev - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVCODEC_XVMC_H -#define AVCODEC_XVMC_H - -/** - * @file - * @ingroup lavc_codec_hwaccel_xvmc - * Public libavcodec XvMC header. - */ - -#include <X11/extensions/XvMC.h> - -#include "libavutil/attributes.h" -#include "version.h" -#include "avcodec.h" - -/** - * @defgroup lavc_codec_hwaccel_xvmc XvMC - * @ingroup lavc_codec_hwaccel - * - * @{ - */ - -#define AV_XVMC_ID 0x1DC711C0 /**< special value to ensure that regular pixel routines haven't corrupted the struct - the number is 1337 speak for the letters IDCT MCo (motion compensation) */ - -struct attribute_deprecated xvmc_pix_fmt { - /** The field contains the special constant value AV_XVMC_ID. - It is used as a test that the application correctly uses the API, - and that there is no corruption caused by pixel routines. - - application - set during initialization - - libavcodec - unchanged - */ - int xvmc_id; - - /** Pointer to the block array allocated by XvMCCreateBlocks(). - The array has to be freed by XvMCDestroyBlocks(). - Each group of 64 values represents one data block of differential - pixel information (in MoCo mode) or coefficients for IDCT. - - application - set the pointer during initialization - - libavcodec - fills coefficients/pixel data into the array - */ - short* data_blocks; - - /** Pointer to the macroblock description array allocated by - XvMCCreateMacroBlocks() and freed by XvMCDestroyMacroBlocks(). - - application - set the pointer during initialization - - libavcodec - fills description data into the array - */ - XvMCMacroBlock* mv_blocks; - - /** Number of macroblock descriptions that can be stored in the mv_blocks - array. - - application - set during initialization - - libavcodec - unchanged - */ - int allocated_mv_blocks; - - /** Number of blocks that can be stored at once in the data_blocks array. - - application - set during initialization - - libavcodec - unchanged - */ - int allocated_data_blocks; - - /** Indicate that the hardware would interpret data_blocks as IDCT - coefficients and perform IDCT on them. - - application - set during initialization - - libavcodec - unchanged - */ - int idct; - - /** In MoCo mode it indicates that intra macroblocks are assumed to be in - unsigned format; same as the XVMC_INTRA_UNSIGNED flag. - - application - set during initialization - - libavcodec - unchanged - */ - int unsigned_intra; - - /** Pointer to the surface allocated by XvMCCreateSurface(). - It has to be freed by XvMCDestroySurface() on application exit. - It identifies the frame and its state on the video hardware. - - application - set during initialization - - libavcodec - unchanged - */ - XvMCSurface* p_surface; - -/** Set by the decoder before calling ff_draw_horiz_band(), - needed by the XvMCRenderSurface function. */ -//@{ - /** Pointer to the surface used as past reference - - application - unchanged - - libavcodec - set - */ - XvMCSurface* p_past_surface; - - /** Pointer to the surface used as future reference - - application - unchanged - - libavcodec - set - */ - XvMCSurface* p_future_surface; - - /** top/bottom field or frame - - application - unchanged - - libavcodec - set - */ - unsigned int picture_structure; - - /** XVMC_SECOND_FIELD - 1st or 2nd field in the sequence - - application - unchanged - - libavcodec - set - */ - unsigned int flags; -//}@ - - /** Number of macroblock descriptions in the mv_blocks array - that have already been passed to the hardware. - - application - zeroes it on get_buffer(). - A successful ff_draw_horiz_band() may increment it - with filled_mb_block_num or zero both. - - libavcodec - unchanged - */ - int start_mv_blocks_num; - - /** Number of new macroblock descriptions in the mv_blocks array (after - start_mv_blocks_num) that are filled by libavcodec and have to be - passed to the hardware. - - application - zeroes it on get_buffer() or after successful - ff_draw_horiz_band(). - - libavcodec - increment with one of each stored MB - */ - int filled_mv_blocks_num; - - /** Number of the next free data block; one data block consists of - 64 short values in the data_blocks array. - All blocks before this one have already been claimed by placing their - position into the corresponding block description structure field, - that are part of the mv_blocks array. - - application - zeroes it on get_buffer(). - A successful ff_draw_horiz_band() may zero it together - with start_mb_blocks_num. - - libavcodec - each decoded macroblock increases it by the number - of coded blocks it contains. - */ - int next_free_data_block_num; -}; - -/** - * @} - */ - -#endif /* AVCODEC_XVMC_H */ diff --git a/src/main/jni/ffmpeg/include/libavformat/avformat.h b/src/main/jni/ffmpeg/include/libavformat/avformat.h deleted file mode 100644 index 95a645b35..000000000 --- a/src/main/jni/ffmpeg/include/libavformat/avformat.h +++ /dev/null @@ -1,2806 +0,0 @@ -/* - * copyright (c) 2001 Fabrice Bellard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVFORMAT_AVFORMAT_H -#define AVFORMAT_AVFORMAT_H - -/** - * @file - * @ingroup libavf - * Main libavformat public API header - */ - -/** - * @defgroup libavf I/O and Muxing/Demuxing Library - * @{ - * - * Libavformat (lavf) is a library for dealing with various media container - * formats. Its main two purposes are demuxing - i.e. splitting a media file - * into component streams, and the reverse process of muxing - writing supplied - * data in a specified container format. It also has an @ref lavf_io - * "I/O module" which supports a number of protocols for accessing the data (e.g. - * file, tcp, http and others). Before using lavf, you need to call - * av_register_all() to register all compiled muxers, demuxers and protocols. - * Unless you are absolutely sure you won't use libavformat's network - * capabilities, you should also call avformat_network_init(). - * - * A supported input format is described by an AVInputFormat struct, conversely - * an output format is described by AVOutputFormat. You can iterate over all - * registered input/output formats using the av_iformat_next() / - * av_oformat_next() functions. The protocols layer is not part of the public - * API, so you can only get the names of supported protocols with the - * avio_enum_protocols() function. - * - * Main lavf structure used for both muxing and demuxing is AVFormatContext, - * which exports all information about the file being read or written. As with - * most Libavformat structures, its size is not part of public ABI, so it cannot be - * allocated on stack or directly with av_malloc(). To create an - * AVFormatContext, use avformat_alloc_context() (some functions, like - * avformat_open_input() might do that for you). - * - * Most importantly an AVFormatContext contains: - * @li the @ref AVFormatContext.iformat "input" or @ref AVFormatContext.oformat - * "output" format. It is either autodetected or set by user for input; - * always set by user for output. - * @li an @ref AVFormatContext.streams "array" of AVStreams, which describe all - * elementary streams stored in the file. AVStreams are typically referred to - * using their index in this array. - * @li an @ref AVFormatContext.pb "I/O context". It is either opened by lavf or - * set by user for input, always set by user for output (unless you are dealing - * with an AVFMT_NOFILE format). - * - * @section lavf_options Passing options to (de)muxers - * It is possible to configure lavf muxers and demuxers using the @ref avoptions - * mechanism. Generic (format-independent) libavformat options are provided by - * AVFormatContext, they can be examined from a user program by calling - * av_opt_next() / av_opt_find() on an allocated AVFormatContext (or its AVClass - * from avformat_get_class()). Private (format-specific) options are provided by - * AVFormatContext.priv_data if and only if AVInputFormat.priv_class / - * AVOutputFormat.priv_class of the corresponding format struct is non-NULL. - * Further options may be provided by the @ref AVFormatContext.pb "I/O context", - * if its AVClass is non-NULL, and the protocols layer. See the discussion on - * nesting in @ref avoptions documentation to learn how to access those. - * - * @defgroup lavf_decoding Demuxing - * @{ - * Demuxers read a media file and split it into chunks of data (@em packets). A - * @ref AVPacket "packet" contains one or more encoded frames which belongs to a - * single elementary stream. In the lavf API this process is represented by the - * avformat_open_input() function for opening a file, av_read_frame() for - * reading a single packet and finally avformat_close_input(), which does the - * cleanup. - * - * @section lavf_decoding_open Opening a media file - * The minimum information required to open a file is its URL or filename, which - * is passed to avformat_open_input(), as in the following code: - * @code - * const char *url = "in.mp3"; - * AVFormatContext *s = NULL; - * int ret = avformat_open_input(&s, url, NULL, NULL); - * if (ret < 0) - * abort(); - * @endcode - * The above code attempts to allocate an AVFormatContext, open the - * specified file (autodetecting the format) and read the header, exporting the - * information stored there into s. Some formats do not have a header or do not - * store enough information there, so it is recommended that you call the - * avformat_find_stream_info() function which tries to read and decode a few - * frames to find missing information. - * - * In some cases you might want to preallocate an AVFormatContext yourself with - * avformat_alloc_context() and do some tweaking on it before passing it to - * avformat_open_input(). One such case is when you want to use custom functions - * for reading input data instead of lavf internal I/O layer. - * To do that, create your own AVIOContext with avio_alloc_context(), passing - * your reading callbacks to it. Then set the @em pb field of your - * AVFormatContext to newly created AVIOContext. - * - * Since the format of the opened file is in general not known until after - * avformat_open_input() has returned, it is not possible to set demuxer private - * options on a preallocated context. Instead, the options should be passed to - * avformat_open_input() wrapped in an AVDictionary: - * @code - * AVDictionary *options = NULL; - * av_dict_set(&options, "video_size", "640x480", 0); - * av_dict_set(&options, "pixel_format", "rgb24", 0); - * - * if (avformat_open_input(&s, url, NULL, &options) < 0) - * abort(); - * av_dict_free(&options); - * @endcode - * This code passes the private options 'video_size' and 'pixel_format' to the - * demuxer. They would be necessary for e.g. the rawvideo demuxer, since it - * cannot know how to interpret raw video data otherwise. If the format turns - * out to be something different than raw video, those options will not be - * recognized by the demuxer and therefore will not be applied. Such unrecognized - * options are then returned in the options dictionary (recognized options are - * consumed). The calling program can handle such unrecognized options as it - * wishes, e.g. - * @code - * AVDictionaryEntry *e; - * if (e = av_dict_get(options, "", NULL, AV_DICT_IGNORE_SUFFIX)) { - * fprintf(stderr, "Option %s not recognized by the demuxer.\n", e->key); - * abort(); - * } - * @endcode - * - * After you have finished reading the file, you must close it with - * avformat_close_input(). It will free everything associated with the file. - * - * @section lavf_decoding_read Reading from an opened file - * Reading data from an opened AVFormatContext is done by repeatedly calling - * av_read_frame() on it. Each call, if successful, will return an AVPacket - * containing encoded data for one AVStream, identified by - * AVPacket.stream_index. This packet may be passed straight into the libavcodec - * decoding functions avcodec_decode_video2(), avcodec_decode_audio4() or - * avcodec_decode_subtitle2() if the caller wishes to decode the data. - * - * AVPacket.pts, AVPacket.dts and AVPacket.duration timing information will be - * set if known. They may also be unset (i.e. AV_NOPTS_VALUE for - * pts/dts, 0 for duration) if the stream does not provide them. The timing - * information will be in AVStream.time_base units, i.e. it has to be - * multiplied by the timebase to convert them to seconds. - * - * If AVPacket.buf is set on the returned packet, then the packet is - * allocated dynamically and the user may keep it indefinitely. - * Otherwise, if AVPacket.buf is NULL, the packet data is backed by a - * static storage somewhere inside the demuxer and the packet is only valid - * until the next av_read_frame() call or closing the file. If the caller - * requires a longer lifetime, av_dup_packet() will make an av_malloc()ed copy - * of it. - * In both cases, the packet must be freed with av_packet_unref() when it is no - * longer needed. - * - * @section lavf_decoding_seek Seeking - * @} - * - * @defgroup lavf_encoding Muxing - * @{ - * Muxers take encoded data in the form of @ref AVPacket "AVPackets" and write - * it into files or other output bytestreams in the specified container format. - * - * The main API functions for muxing are avformat_write_header() for writing the - * file header, av_write_frame() / av_interleaved_write_frame() for writing the - * packets and av_write_trailer() for finalizing the file. - * - * At the beginning of the muxing process, the caller must first call - * avformat_alloc_context() to create a muxing context. The caller then sets up - * the muxer by filling the various fields in this context: - * - * - The @ref AVFormatContext.oformat "oformat" field must be set to select the - * muxer that will be used. - * - Unless the format is of the AVFMT_NOFILE type, the @ref AVFormatContext.pb - * "pb" field must be set to an opened IO context, either returned from - * avio_open2() or a custom one. - * - Unless the format is of the AVFMT_NOSTREAMS type, at least one stream must - * be created with the avformat_new_stream() function. The caller should fill - * the @ref AVStream.codec "stream codec context" information, such as the - * codec @ref AVCodecContext.codec_type "type", @ref AVCodecContext.codec_id - * "id" and other parameters (e.g. width / height, the pixel or sample format, - * etc.) as known. The @ref AVStream.time_base "stream timebase" should - * be set to the timebase that the caller desires to use for this stream (note - * that the timebase actually used by the muxer can be different, as will be - * described later). - * - It is advised to manually initialize only the relevant fields in - * AVCodecContext, rather than using @ref avcodec_copy_context() during - * remuxing: there is no guarantee that the codec context values remain valid - * for both input and output format contexts. - * - The caller may fill in additional information, such as @ref - * AVFormatContext.metadata "global" or @ref AVStream.metadata "per-stream" - * metadata, @ref AVFormatContext.chapters "chapters", @ref - * AVFormatContext.programs "programs", etc. as described in the - * AVFormatContext documentation. Whether such information will actually be - * stored in the output depends on what the container format and the muxer - * support. - * - * When the muxing context is fully set up, the caller must call - * avformat_write_header() to initialize the muxer internals and write the file - * header. Whether anything actually is written to the IO context at this step - * depends on the muxer, but this function must always be called. Any muxer - * private options must be passed in the options parameter to this function. - * - * The data is then sent to the muxer by repeatedly calling av_write_frame() or - * av_interleaved_write_frame() (consult those functions' documentation for - * discussion on the difference between them; only one of them may be used with - * a single muxing context, they should not be mixed). Do note that the timing - * information on the packets sent to the muxer must be in the corresponding - * AVStream's timebase. That timebase is set by the muxer (in the - * avformat_write_header() step) and may be different from the timebase - * requested by the caller. - * - * Once all the data has been written, the caller must call av_write_trailer() - * to flush any buffered packets and finalize the output file, then close the IO - * context (if any) and finally free the muxing context with - * avformat_free_context(). - * @} - * - * @defgroup lavf_io I/O Read/Write - * @{ - * @section lavf_io_dirlist Directory listing - * The directory listing API makes it possible to list files on remote servers. - * - * Some of possible use cases: - * - an "open file" dialog to choose files from a remote location, - * - a recursive media finder providing a player with an ability to play all - * files from a given directory. - * - * @subsection lavf_io_dirlist_open Opening a directory - * At first, a directory needs to be opened by calling avio_open_dir() - * supplied with a URL and, optionally, ::AVDictionary containing - * protocol-specific parameters. The function returns zero or positive - * integer and allocates AVIODirContext on success. - * - * @code - * AVIODirContext *ctx = NULL; - * if (avio_open_dir(&ctx, "smb://example.com/some_dir", NULL) < 0) { - * fprintf(stderr, "Cannot open directory.\n"); - * abort(); - * } - * @endcode - * - * This code tries to open a sample directory using smb protocol without - * any additional parameters. - * - * @subsection lavf_io_dirlist_read Reading entries - * Each directory's entry (i.e. file, another directory, anything else - * within ::AVIODirEntryType) is represented by AVIODirEntry. - * Reading consecutive entries from an opened AVIODirContext is done by - * repeatedly calling avio_read_dir() on it. Each call returns zero or - * positive integer if successful. Reading can be stopped right after the - * NULL entry has been read -- it means there are no entries left to be - * read. The following code reads all entries from a directory associated - * with ctx and prints their names to standard output. - * @code - * AVIODirEntry *entry = NULL; - * for (;;) { - * if (avio_read_dir(ctx, &entry) < 0) { - * fprintf(stderr, "Cannot list directory.\n"); - * abort(); - * } - * if (!entry) - * break; - * printf("%s\n", entry->name); - * avio_free_directory_entry(&entry); - * } - * @endcode - * @} - * - * @defgroup lavf_codec Demuxers - * @{ - * @defgroup lavf_codec_native Native Demuxers - * @{ - * @} - * @defgroup lavf_codec_wrappers External library wrappers - * @{ - * @} - * @} - * @defgroup lavf_protos I/O Protocols - * @{ - * @} - * @defgroup lavf_internal Internal - * @{ - * @} - * @} - * - */ - -#include <time.h> -#include <stdio.h> /* FILE */ -#include "libavcodec/avcodec.h" -#include "libavutil/dict.h" -#include "libavutil/log.h" - -#include "avio.h" -#include "libavformat/version.h" - -struct AVFormatContext; - -struct AVDeviceInfoList; -struct AVDeviceCapabilitiesQuery; - -/** - * @defgroup metadata_api Public Metadata API - * @{ - * @ingroup libavf - * The metadata API allows libavformat to export metadata tags to a client - * application when demuxing. Conversely it allows a client application to - * set metadata when muxing. - * - * Metadata is exported or set as pairs of key/value strings in the 'metadata' - * fields of the AVFormatContext, AVStream, AVChapter and AVProgram structs - * using the @ref lavu_dict "AVDictionary" API. Like all strings in FFmpeg, - * metadata is assumed to be UTF-8 encoded Unicode. Note that metadata - * exported by demuxers isn't checked to be valid UTF-8 in most cases. - * - * Important concepts to keep in mind: - * - Keys are unique; there can never be 2 tags with the same key. This is - * also meant semantically, i.e., a demuxer should not knowingly produce - * several keys that are literally different but semantically identical. - * E.g., key=Author5, key=Author6. In this example, all authors must be - * placed in the same tag. - * - Metadata is flat, not hierarchical; there are no subtags. If you - * want to store, e.g., the email address of the child of producer Alice - * and actor Bob, that could have key=alice_and_bobs_childs_email_address. - * - Several modifiers can be applied to the tag name. This is done by - * appending a dash character ('-') and the modifier name in the order - * they appear in the list below -- e.g. foo-eng-sort, not foo-sort-eng. - * - language -- a tag whose value is localized for a particular language - * is appended with the ISO 639-2/B 3-letter language code. - * For example: Author-ger=Michael, Author-eng=Mike - * The original/default language is in the unqualified "Author" tag. - * A demuxer should set a default if it sets any translated tag. - * - sorting -- a modified version of a tag that should be used for - * sorting will have '-sort' appended. E.g. artist="The Beatles", - * artist-sort="Beatles, The". - * - Some protocols and demuxers support metadata updates. After a successful - * call to av_read_packet(), AVFormatContext.event_flags or AVStream.event_flags - * will be updated to indicate if metadata changed. In order to detect metadata - * changes on a stream, you need to loop through all streams in the AVFormatContext - * and check their individual event_flags. - * - * - Demuxers attempt to export metadata in a generic format, however tags - * with no generic equivalents are left as they are stored in the container. - * Follows a list of generic tag names: - * - @verbatim - album -- name of the set this work belongs to - album_artist -- main creator of the set/album, if different from artist. - e.g. "Various Artists" for compilation albums. - artist -- main creator of the work - comment -- any additional description of the file. - composer -- who composed the work, if different from artist. - copyright -- name of copyright holder. - creation_time-- date when the file was created, preferably in ISO 8601. - date -- date when the work was created, preferably in ISO 8601. - disc -- number of a subset, e.g. disc in a multi-disc collection. - encoder -- name/settings of the software/hardware that produced the file. - encoded_by -- person/group who created the file. - filename -- original name of the file. - genre -- <self-evident>. - language -- main language in which the work is performed, preferably - in ISO 639-2 format. Multiple languages can be specified by - separating them with commas. - performer -- artist who performed the work, if different from artist. - E.g for "Also sprach Zarathustra", artist would be "Richard - Strauss" and performer "London Philharmonic Orchestra". - publisher -- name of the label/publisher. - service_name -- name of the service in broadcasting (channel name). - service_provider -- name of the service provider in broadcasting. - title -- name of the work. - track -- number of this work in the set, can be in form current/total. - variant_bitrate -- the total bitrate of the bitrate variant that the current stream is part of - @endverbatim - * - * Look in the examples section for an application example how to use the Metadata API. - * - * @} - */ - -/* packet functions */ - - -/** - * Allocate and read the payload of a packet and initialize its - * fields with default values. - * - * @param s associated IO context - * @param pkt packet - * @param size desired payload size - * @return >0 (read size) if OK, AVERROR_xxx otherwise - */ -int av_get_packet(AVIOContext *s, AVPacket *pkt, int size); - - -/** - * Read data and append it to the current content of the AVPacket. - * If pkt->size is 0 this is identical to av_get_packet. - * Note that this uses av_grow_packet and thus involves a realloc - * which is inefficient. Thus this function should only be used - * when there is no reasonable way to know (an upper bound of) - * the final size. - * - * @param s associated IO context - * @param pkt packet - * @param size amount of data to read - * @return >0 (read size) if OK, AVERROR_xxx otherwise, previous data - * will not be lost even if an error occurs. - */ -int av_append_packet(AVIOContext *s, AVPacket *pkt, int size); - -#if FF_API_LAVF_FRAC -/*************************************************/ -/* fractional numbers for exact pts handling */ - -/** - * The exact value of the fractional number is: 'val + num / den'. - * num is assumed to be 0 <= num < den. - */ -typedef struct AVFrac { - int64_t val, num, den; -} AVFrac; -#endif - -/*************************************************/ -/* input/output formats */ - -struct AVCodecTag; - -/** - * This structure contains the data a format has to probe a file. - */ -typedef struct AVProbeData { - const char *filename; - unsigned char *buf; /**< Buffer must have AVPROBE_PADDING_SIZE of extra allocated bytes filled with zero. */ - int buf_size; /**< Size of buf except extra allocated bytes */ - const char *mime_type; /**< mime_type, when known. */ -} AVProbeData; - -#define AVPROBE_SCORE_RETRY (AVPROBE_SCORE_MAX/4) -#define AVPROBE_SCORE_STREAM_RETRY (AVPROBE_SCORE_MAX/4-1) - -#define AVPROBE_SCORE_EXTENSION 50 ///< score for file extension -#define AVPROBE_SCORE_MIME 75 ///< score for file mime type -#define AVPROBE_SCORE_MAX 100 ///< maximum score - -#define AVPROBE_PADDING_SIZE 32 ///< extra allocated bytes at the end of the probe buffer - -/// Demuxer will use avio_open, no opened file should be provided by the caller. -#define AVFMT_NOFILE 0x0001 -#define AVFMT_NEEDNUMBER 0x0002 /**< Needs '%d' in filename. */ -#define AVFMT_SHOW_IDS 0x0008 /**< Show format stream IDs numbers. */ -#if FF_API_LAVF_FMT_RAWPICTURE -#define AVFMT_RAWPICTURE 0x0020 /**< Format wants AVPicture structure for - raw picture data. @deprecated Not used anymore */ -#endif -#define AVFMT_GLOBALHEADER 0x0040 /**< Format wants global header. */ -#define AVFMT_NOTIMESTAMPS 0x0080 /**< Format does not need / have any timestamps. */ -#define AVFMT_GENERIC_INDEX 0x0100 /**< Use generic index building code. */ -#define AVFMT_TS_DISCONT 0x0200 /**< Format allows timestamp discontinuities. Note, muxers always require valid (monotone) timestamps */ -#define AVFMT_VARIABLE_FPS 0x0400 /**< Format allows variable fps. */ -#define AVFMT_NODIMENSIONS 0x0800 /**< Format does not need width/height */ -#define AVFMT_NOSTREAMS 0x1000 /**< Format does not require any streams */ -#define AVFMT_NOBINSEARCH 0x2000 /**< Format does not allow to fall back on binary search via read_timestamp */ -#define AVFMT_NOGENSEARCH 0x4000 /**< Format does not allow to fall back on generic search */ -#define AVFMT_NO_BYTE_SEEK 0x8000 /**< Format does not allow seeking by bytes */ -#define AVFMT_ALLOW_FLUSH 0x10000 /**< Format allows flushing. If not set, the muxer will not receive a NULL packet in the write_packet function. */ -#define AVFMT_TS_NONSTRICT 0x20000 /**< Format does not require strictly - increasing timestamps, but they must - still be monotonic */ -#define AVFMT_TS_NEGATIVE 0x40000 /**< Format allows muxing negative - timestamps. If not set the timestamp - will be shifted in av_write_frame and - av_interleaved_write_frame so they - start from 0. - The user or muxer can override this through - AVFormatContext.avoid_negative_ts - */ - -#define AVFMT_SEEK_TO_PTS 0x4000000 /**< Seeking is based on PTS */ - -/** - * @addtogroup lavf_encoding - * @{ - */ -typedef struct AVOutputFormat { - const char *name; - /** - * Descriptive name for the format, meant to be more human-readable - * than name. You should use the NULL_IF_CONFIG_SMALL() macro - * to define it. - */ - const char *long_name; - const char *mime_type; - const char *extensions; /**< comma-separated filename extensions */ - /* output support */ - enum AVCodecID audio_codec; /**< default audio codec */ - enum AVCodecID video_codec; /**< default video codec */ - enum AVCodecID subtitle_codec; /**< default subtitle codec */ - /** - * can use flags: AVFMT_NOFILE, AVFMT_NEEDNUMBER, - * AVFMT_GLOBALHEADER, AVFMT_NOTIMESTAMPS, AVFMT_VARIABLE_FPS, - * AVFMT_NODIMENSIONS, AVFMT_NOSTREAMS, AVFMT_ALLOW_FLUSH, - * AVFMT_TS_NONSTRICT - */ - int flags; - - /** - * List of supported codec_id-codec_tag pairs, ordered by "better - * choice first". The arrays are all terminated by AV_CODEC_ID_NONE. - */ - const struct AVCodecTag * const *codec_tag; - - - const AVClass *priv_class; ///< AVClass for the private context - - /***************************************************************** - * No fields below this line are part of the public API. They - * may not be used outside of libavformat and can be changed and - * removed at will. - * New public fields should be added right above. - ***************************************************************** - */ - struct AVOutputFormat *next; - /** - * size of private data so that it can be allocated in the wrapper - */ - int priv_data_size; - - int (*write_header)(struct AVFormatContext *); - /** - * Write a packet. If AVFMT_ALLOW_FLUSH is set in flags, - * pkt can be NULL in order to flush data buffered in the muxer. - * When flushing, return 0 if there still is more data to flush, - * or 1 if everything was flushed and there is no more buffered - * data. - */ - int (*write_packet)(struct AVFormatContext *, AVPacket *pkt); - int (*write_trailer)(struct AVFormatContext *); - /** - * Currently only used to set pixel format if not YUV420P. - */ - int (*interleave_packet)(struct AVFormatContext *, AVPacket *out, - AVPacket *in, int flush); - /** - * Test if the given codec can be stored in this container. - * - * @return 1 if the codec is supported, 0 if it is not. - * A negative number if unknown. - * MKTAG('A', 'P', 'I', 'C') if the codec is only supported as AV_DISPOSITION_ATTACHED_PIC - */ - int (*query_codec)(enum AVCodecID id, int std_compliance); - - void (*get_output_timestamp)(struct AVFormatContext *s, int stream, - int64_t *dts, int64_t *wall); - /** - * Allows sending messages from application to device. - */ - int (*control_message)(struct AVFormatContext *s, int type, - void *data, size_t data_size); - - /** - * Write an uncoded AVFrame. - * - * See av_write_uncoded_frame() for details. - * - * The library will free *frame afterwards, but the muxer can prevent it - * by setting the pointer to NULL. - */ - int (*write_uncoded_frame)(struct AVFormatContext *, int stream_index, - AVFrame **frame, unsigned flags); - /** - * Returns device list with it properties. - * @see avdevice_list_devices() for more details. - */ - int (*get_device_list)(struct AVFormatContext *s, struct AVDeviceInfoList *device_list); - /** - * Initialize device capabilities submodule. - * @see avdevice_capabilities_create() for more details. - */ - int (*create_device_capabilities)(struct AVFormatContext *s, struct AVDeviceCapabilitiesQuery *caps); - /** - * Free device capabilities submodule. - * @see avdevice_capabilities_free() for more details. - */ - int (*free_device_capabilities)(struct AVFormatContext *s, struct AVDeviceCapabilitiesQuery *caps); - enum AVCodecID data_codec; /**< default data codec */ - /** - * Initialize format. May allocate data here, and set any AVFormatContext or - * AVStream parameters that need to be set before packets are sent. - * This method must not write output. - * - * Any allocations made here must be freed in deinit(). - */ - int (*init)(struct AVFormatContext *); - /** - * Deinitialize format. If present, this is called whenever the muxer is being - * destroyed, regardless of whether or not the header has been written. - * - * If a trailer is being written, this is called after write_trailer(). - * - * This is called if init() fails as well. - */ - void (*deinit)(struct AVFormatContext *); - /** - * Set up any necessary bitstream filtering and extract any extra data needed - * for the global header. - * Return 0 if more packets from this stream must be checked; 1 if not. - */ - int (*check_bitstream)(struct AVFormatContext *, const AVPacket *pkt); -} AVOutputFormat; -/** - * @} - */ - -/** - * @addtogroup lavf_decoding - * @{ - */ -typedef struct AVInputFormat { - /** - * A comma separated list of short names for the format. New names - * may be appended with a minor bump. - */ - const char *name; - - /** - * Descriptive name for the format, meant to be more human-readable - * than name. You should use the NULL_IF_CONFIG_SMALL() macro - * to define it. - */ - const char *long_name; - - /** - * Can use flags: AVFMT_NOFILE, AVFMT_NEEDNUMBER, AVFMT_SHOW_IDS, - * AVFMT_GENERIC_INDEX, AVFMT_TS_DISCONT, AVFMT_NOBINSEARCH, - * AVFMT_NOGENSEARCH, AVFMT_NO_BYTE_SEEK, AVFMT_SEEK_TO_PTS. - */ - int flags; - - /** - * If extensions are defined, then no probe is done. You should - * usually not use extension format guessing because it is not - * reliable enough - */ - const char *extensions; - - const struct AVCodecTag * const *codec_tag; - - const AVClass *priv_class; ///< AVClass for the private context - - /** - * Comma-separated list of mime types. - * It is used check for matching mime types while probing. - * @see av_probe_input_format2 - */ - const char *mime_type; - - /***************************************************************** - * No fields below this line are part of the public API. They - * may not be used outside of libavformat and can be changed and - * removed at will. - * New public fields should be added right above. - ***************************************************************** - */ - struct AVInputFormat *next; - - /** - * Raw demuxers store their codec ID here. - */ - int raw_codec_id; - - /** - * Size of private data so that it can be allocated in the wrapper. - */ - int priv_data_size; - - /** - * Tell if a given file has a chance of being parsed as this format. - * The buffer provided is guaranteed to be AVPROBE_PADDING_SIZE bytes - * big so you do not have to check for that unless you need more. - */ - int (*read_probe)(AVProbeData *); - - /** - * Read the format header and initialize the AVFormatContext - * structure. Return 0 if OK. 'avformat_new_stream' should be - * called to create new streams. - */ - int (*read_header)(struct AVFormatContext *); - - /** - * Read one packet and put it in 'pkt'. pts and flags are also - * set. 'avformat_new_stream' can be called only if the flag - * AVFMTCTX_NOHEADER is used and only in the calling thread (not in a - * background thread). - * @return 0 on success, < 0 on error. - * When returning an error, pkt must not have been allocated - * or must be freed before returning - */ - int (*read_packet)(struct AVFormatContext *, AVPacket *pkt); - - /** - * Close the stream. The AVFormatContext and AVStreams are not - * freed by this function - */ - int (*read_close)(struct AVFormatContext *); - - /** - * Seek to a given timestamp relative to the frames in - * stream component stream_index. - * @param stream_index Must not be -1. - * @param flags Selects which direction should be preferred if no exact - * match is available. - * @return >= 0 on success (but not necessarily the new offset) - */ - int (*read_seek)(struct AVFormatContext *, - int stream_index, int64_t timestamp, int flags); - - /** - * Get the next timestamp in stream[stream_index].time_base units. - * @return the timestamp or AV_NOPTS_VALUE if an error occurred - */ - int64_t (*read_timestamp)(struct AVFormatContext *s, int stream_index, - int64_t *pos, int64_t pos_limit); - - /** - * Start/resume playing - only meaningful if using a network-based format - * (RTSP). - */ - int (*read_play)(struct AVFormatContext *); - - /** - * Pause playing - only meaningful if using a network-based format - * (RTSP). - */ - int (*read_pause)(struct AVFormatContext *); - - /** - * Seek to timestamp ts. - * Seeking will be done so that the point from which all active streams - * can be presented successfully will be closest to ts and within min/max_ts. - * Active streams are all streams that have AVStream.discard < AVDISCARD_ALL. - */ - int (*read_seek2)(struct AVFormatContext *s, int stream_index, int64_t min_ts, int64_t ts, int64_t max_ts, int flags); - - /** - * Returns device list with it properties. - * @see avdevice_list_devices() for more details. - */ - int (*get_device_list)(struct AVFormatContext *s, struct AVDeviceInfoList *device_list); - - /** - * Initialize device capabilities submodule. - * @see avdevice_capabilities_create() for more details. - */ - int (*create_device_capabilities)(struct AVFormatContext *s, struct AVDeviceCapabilitiesQuery *caps); - - /** - * Free device capabilities submodule. - * @see avdevice_capabilities_free() for more details. - */ - int (*free_device_capabilities)(struct AVFormatContext *s, struct AVDeviceCapabilitiesQuery *caps); -} AVInputFormat; -/** - * @} - */ - -enum AVStreamParseType { - AVSTREAM_PARSE_NONE, - AVSTREAM_PARSE_FULL, /**< full parsing and repack */ - AVSTREAM_PARSE_HEADERS, /**< Only parse headers, do not repack. */ - AVSTREAM_PARSE_TIMESTAMPS, /**< full parsing and interpolation of timestamps for frames not starting on a packet boundary */ - AVSTREAM_PARSE_FULL_ONCE, /**< full parsing and repack of the first frame only, only implemented for H.264 currently */ - AVSTREAM_PARSE_FULL_RAW=MKTAG(0,'R','A','W'), /**< full parsing and repack with timestamp and position generation by parser for raw - this assumes that each packet in the file contains no demuxer level headers and - just codec level data, otherwise position generation would fail */ -}; - -typedef struct AVIndexEntry { - int64_t pos; - int64_t timestamp; /**< - * Timestamp in AVStream.time_base units, preferably the time from which on correctly decoded frames are available - * when seeking to this entry. That means preferable PTS on keyframe based formats. - * But demuxers can choose to store a different timestamp, if it is more convenient for the implementation or nothing better - * is known - */ -#define AVINDEX_KEYFRAME 0x0001 - int flags:2; - int size:30; //Yeah, trying to keep the size of this small to reduce memory requirements (it is 24 vs. 32 bytes due to possible 8-byte alignment). - int min_distance; /**< Minimum distance between this and the previous keyframe, used to avoid unneeded searching. */ -} AVIndexEntry; - -#define AV_DISPOSITION_DEFAULT 0x0001 -#define AV_DISPOSITION_DUB 0x0002 -#define AV_DISPOSITION_ORIGINAL 0x0004 -#define AV_DISPOSITION_COMMENT 0x0008 -#define AV_DISPOSITION_LYRICS 0x0010 -#define AV_DISPOSITION_KARAOKE 0x0020 - -/** - * Track should be used during playback by default. - * Useful for subtitle track that should be displayed - * even when user did not explicitly ask for subtitles. - */ -#define AV_DISPOSITION_FORCED 0x0040 -#define AV_DISPOSITION_HEARING_IMPAIRED 0x0080 /**< stream for hearing impaired audiences */ -#define AV_DISPOSITION_VISUAL_IMPAIRED 0x0100 /**< stream for visual impaired audiences */ -#define AV_DISPOSITION_CLEAN_EFFECTS 0x0200 /**< stream without voice */ -/** - * The stream is stored in the file as an attached picture/"cover art" (e.g. - * APIC frame in ID3v2). The single packet associated with it will be returned - * among the first few packets read from the file unless seeking takes place. - * It can also be accessed at any time in AVStream.attached_pic. - */ -#define AV_DISPOSITION_ATTACHED_PIC 0x0400 - -typedef struct AVStreamInternal AVStreamInternal; - -/** - * To specify text track kind (different from subtitles default). - */ -#define AV_DISPOSITION_CAPTIONS 0x10000 -#define AV_DISPOSITION_DESCRIPTIONS 0x20000 -#define AV_DISPOSITION_METADATA 0x40000 - -/** - * Options for behavior on timestamp wrap detection. - */ -#define AV_PTS_WRAP_IGNORE 0 ///< ignore the wrap -#define AV_PTS_WRAP_ADD_OFFSET 1 ///< add the format specific offset on wrap detection -#define AV_PTS_WRAP_SUB_OFFSET -1 ///< subtract the format specific offset on wrap detection - -/** - * Stream structure. - * New fields can be added to the end with minor version bumps. - * Removal, reordering and changes to existing fields require a major - * version bump. - * sizeof(AVStream) must not be used outside libav*. - */ -typedef struct AVStream { - int index; /**< stream index in AVFormatContext */ - /** - * Format-specific stream ID. - * decoding: set by libavformat - * encoding: set by the user, replaced by libavformat if left unset - */ - int id; - /** - * Codec context associated with this stream. Allocated and freed by - * libavformat. - * - * - decoding: The demuxer exports codec information stored in the headers - * here. - * - encoding: The user sets codec information, the muxer writes it to the - * output. Mandatory fields as specified in AVCodecContext - * documentation must be set even if this AVCodecContext is - * not actually used for encoding. - */ - AVCodecContext *codec; - void *priv_data; - -#if FF_API_LAVF_FRAC - /** - * @deprecated this field is unused - */ - attribute_deprecated - struct AVFrac pts; -#endif - - /** - * This is the fundamental unit of time (in seconds) in terms - * of which frame timestamps are represented. - * - * decoding: set by libavformat - * encoding: May be set by the caller before avformat_write_header() to - * provide a hint to the muxer about the desired timebase. In - * avformat_write_header(), the muxer will overwrite this field - * with the timebase that will actually be used for the timestamps - * written into the file (which may or may not be related to the - * user-provided one, depending on the format). - */ - AVRational time_base; - - /** - * Decoding: pts of the first frame of the stream in presentation order, in stream time base. - * Only set this if you are absolutely 100% sure that the value you set - * it to really is the pts of the first frame. - * This may be undefined (AV_NOPTS_VALUE). - * @note The ASF header does NOT contain a correct start_time the ASF - * demuxer must NOT set this. - */ - int64_t start_time; - - /** - * Decoding: duration of the stream, in stream time base. - * If a source file does not specify a duration, but does specify - * a bitrate, this value will be estimated from bitrate and file size. - */ - int64_t duration; - - int64_t nb_frames; ///< number of frames in this stream if known or 0 - - int disposition; /**< AV_DISPOSITION_* bit field */ - - enum AVDiscard discard; ///< Selects which packets can be discarded at will and do not need to be demuxed. - - /** - * sample aspect ratio (0 if unknown) - * - encoding: Set by user. - * - decoding: Set by libavformat. - */ - AVRational sample_aspect_ratio; - - AVDictionary *metadata; - - /** - * Average framerate - * - * - demuxing: May be set by libavformat when creating the stream or in - * avformat_find_stream_info(). - * - muxing: May be set by the caller before avformat_write_header(). - */ - AVRational avg_frame_rate; - - /** - * For streams with AV_DISPOSITION_ATTACHED_PIC disposition, this packet - * will contain the attached picture. - * - * decoding: set by libavformat, must not be modified by the caller. - * encoding: unused - */ - AVPacket attached_pic; - - /** - * An array of side data that applies to the whole stream (i.e. the - * container does not allow it to change between packets). - * - * There may be no overlap between the side data in this array and side data - * in the packets. I.e. a given side data is either exported by the muxer - * (demuxing) / set by the caller (muxing) in this array, then it never - * appears in the packets, or the side data is exported / sent through - * the packets (always in the first packet where the value becomes known or - * changes), then it does not appear in this array. - * - * - demuxing: Set by libavformat when the stream is created. - * - muxing: May be set by the caller before avformat_write_header(). - * - * Freed by libavformat in avformat_free_context(). - * - * @see av_format_inject_global_side_data() - */ - AVPacketSideData *side_data; - /** - * The number of elements in the AVStream.side_data array. - */ - int nb_side_data; - - /** - * Flags for the user to detect events happening on the stream. Flags must - * be cleared by the user once the event has been handled. - * A combination of AVSTREAM_EVENT_FLAG_*. - */ - int event_flags; -#define AVSTREAM_EVENT_FLAG_METADATA_UPDATED 0x0001 ///< The call resulted in updated metadata. - - /***************************************************************** - * All fields below this line are not part of the public API. They - * may not be used outside of libavformat and can be changed and - * removed at will. - * New public fields should be added right above. - ***************************************************************** - */ - - /** - * Stream information used internally by av_find_stream_info() - */ -#define MAX_STD_TIMEBASES (30*12+7+6) - struct { - int64_t last_dts; - int64_t duration_gcd; - int duration_count; - int64_t rfps_duration_sum; - double (*duration_error)[2][MAX_STD_TIMEBASES]; - int64_t codec_info_duration; - int64_t codec_info_duration_fields; - - /** - * 0 -> decoder has not been searched for yet. - * >0 -> decoder found - * <0 -> decoder with codec_id == -found_decoder has not been found - */ - int found_decoder; - - int64_t last_duration; - - /** - * Those are used for average framerate estimation. - */ - int64_t fps_first_dts; - int fps_first_dts_idx; - int64_t fps_last_dts; - int fps_last_dts_idx; - - } *info; - - int pts_wrap_bits; /**< number of bits in pts (used for wrapping control) */ - - // Timestamp generation support: - /** - * Timestamp corresponding to the last dts sync point. - * - * Initialized when AVCodecParserContext.dts_sync_point >= 0 and - * a DTS is received from the underlying container. Otherwise set to - * AV_NOPTS_VALUE by default. - */ - int64_t first_dts; - int64_t cur_dts; - int64_t last_IP_pts; - int last_IP_duration; - - /** - * Number of packets to buffer for codec probing - */ - int probe_packets; - - /** - * Number of frames that have been demuxed during av_find_stream_info() - */ - int codec_info_nb_frames; - - /* av_read_frame() support */ - enum AVStreamParseType need_parsing; - struct AVCodecParserContext *parser; - - /** - * last packet in packet_buffer for this stream when muxing. - */ - struct AVPacketList *last_in_packet_buffer; - AVProbeData probe_data; -#define MAX_REORDER_DELAY 16 - int64_t pts_buffer[MAX_REORDER_DELAY+1]; - - AVIndexEntry *index_entries; /**< Only used if the format does not - support seeking natively. */ - int nb_index_entries; - unsigned int index_entries_allocated_size; - - /** - * Real base framerate of the stream. - * This is the lowest framerate with which all timestamps can be - * represented accurately (it is the least common multiple of all - * framerates in the stream). Note, this value is just a guess! - * For example, if the time base is 1/90000 and all frames have either - * approximately 3600 or 1800 timer ticks, then r_frame_rate will be 50/1. - * - * Code outside avformat should access this field using: - * av_stream_get/set_r_frame_rate(stream) - */ - AVRational r_frame_rate; - - /** - * Stream Identifier - * This is the MPEG-TS stream identifier +1 - * 0 means unknown - */ - int stream_identifier; - - int64_t interleaver_chunk_size; - int64_t interleaver_chunk_duration; - - /** - * stream probing state - * -1 -> probing finished - * 0 -> no probing requested - * rest -> perform probing with request_probe being the minimum score to accept. - * NOT PART OF PUBLIC API - */ - int request_probe; - /** - * Indicates that everything up to the next keyframe - * should be discarded. - */ - int skip_to_keyframe; - - /** - * Number of samples to skip at the start of the frame decoded from the next packet. - */ - int skip_samples; - - /** - * If not 0, the number of samples that should be skipped from the start of - * the stream (the samples are removed from packets with pts==0, which also - * assumes negative timestamps do not happen). - * Intended for use with formats such as mp3 with ad-hoc gapless audio - * support. - */ - int64_t start_skip_samples; - - /** - * If not 0, the first audio sample that should be discarded from the stream. - * This is broken by design (needs global sample count), but can't be - * avoided for broken by design formats such as mp3 with ad-hoc gapless - * audio support. - */ - int64_t first_discard_sample; - - /** - * The sample after last sample that is intended to be discarded after - * first_discard_sample. Works on frame boundaries only. Used to prevent - * early EOF if the gapless info is broken (considered concatenated mp3s). - */ - int64_t last_discard_sample; - - /** - * Number of internally decoded frames, used internally in libavformat, do not access - * its lifetime differs from info which is why it is not in that structure. - */ - int nb_decoded_frames; - - /** - * Timestamp offset added to timestamps before muxing - * NOT PART OF PUBLIC API - */ - int64_t mux_ts_offset; - - /** - * Internal data to check for wrapping of the time stamp - */ - int64_t pts_wrap_reference; - - /** - * Options for behavior, when a wrap is detected. - * - * Defined by AV_PTS_WRAP_ values. - * - * If correction is enabled, there are two possibilities: - * If the first time stamp is near the wrap point, the wrap offset - * will be subtracted, which will create negative time stamps. - * Otherwise the offset will be added. - */ - int pts_wrap_behavior; - - /** - * Internal data to prevent doing update_initial_durations() twice - */ - int update_initial_durations_done; - - /** - * Internal data to generate dts from pts - */ - int64_t pts_reorder_error[MAX_REORDER_DELAY+1]; - uint8_t pts_reorder_error_count[MAX_REORDER_DELAY+1]; - - /** - * Internal data to analyze DTS and detect faulty mpeg streams - */ - int64_t last_dts_for_order_check; - uint8_t dts_ordered; - uint8_t dts_misordered; - - /** - * Internal data to inject global side data - */ - int inject_global_side_data; - - /** - * String containing paris of key and values describing recommended encoder configuration. - * Paris are separated by ','. - * Keys are separated from values by '='. - */ - char *recommended_encoder_configuration; - - /** - * display aspect ratio (0 if unknown) - * - encoding: unused - * - decoding: Set by libavformat to calculate sample_aspect_ratio internally - */ - AVRational display_aspect_ratio; - - struct FFFrac *priv_pts; - - /** - * An opaque field for libavformat internal usage. - * Must not be accessed in any way by callers. - */ - AVStreamInternal *internal; -} AVStream; - -AVRational av_stream_get_r_frame_rate(const AVStream *s); -void av_stream_set_r_frame_rate(AVStream *s, AVRational r); -struct AVCodecParserContext *av_stream_get_parser(const AVStream *s); -char* av_stream_get_recommended_encoder_configuration(const AVStream *s); -void av_stream_set_recommended_encoder_configuration(AVStream *s, char *configuration); - -/** - * Returns the pts of the last muxed packet + its duration - * - * the retuned value is undefined when used with a demuxer. - */ -int64_t av_stream_get_end_pts(const AVStream *st); - -#define AV_PROGRAM_RUNNING 1 - -/** - * New fields can be added to the end with minor version bumps. - * Removal, reordering and changes to existing fields require a major - * version bump. - * sizeof(AVProgram) must not be used outside libav*. - */ -typedef struct AVProgram { - int id; - int flags; - enum AVDiscard discard; ///< selects which program to discard and which to feed to the caller - unsigned int *stream_index; - unsigned int nb_stream_indexes; - AVDictionary *metadata; - - int program_num; - int pmt_pid; - int pcr_pid; - - /***************************************************************** - * All fields below this line are not part of the public API. They - * may not be used outside of libavformat and can be changed and - * removed at will. - * New public fields should be added right above. - ***************************************************************** - */ - int64_t start_time; - int64_t end_time; - - int64_t pts_wrap_reference; ///< reference dts for wrap detection - int pts_wrap_behavior; ///< behavior on wrap detection -} AVProgram; - -#define AVFMTCTX_NOHEADER 0x0001 /**< signal that no header is present - (streams are added dynamically) */ - -typedef struct AVChapter { - int id; ///< unique ID to identify the chapter - AVRational time_base; ///< time base in which the start/end timestamps are specified - int64_t start, end; ///< chapter start/end time in time_base units - AVDictionary *metadata; -} AVChapter; - - -/** - * Callback used by devices to communicate with application. - */ -typedef int (*av_format_control_message)(struct AVFormatContext *s, int type, - void *data, size_t data_size); - -typedef int (*AVOpenCallback)(struct AVFormatContext *s, AVIOContext **pb, const char *url, int flags, - const AVIOInterruptCB *int_cb, AVDictionary **options); - -/** - * The duration of a video can be estimated through various ways, and this enum can be used - * to know how the duration was estimated. - */ -enum AVDurationEstimationMethod { - AVFMT_DURATION_FROM_PTS, ///< Duration accurately estimated from PTSes - AVFMT_DURATION_FROM_STREAM, ///< Duration estimated from a stream with a known duration - AVFMT_DURATION_FROM_BITRATE ///< Duration estimated from bitrate (less accurate) -}; - -typedef struct AVFormatInternal AVFormatInternal; - -/** - * Format I/O context. - * New fields can be added to the end with minor version bumps. - * Removal, reordering and changes to existing fields require a major - * version bump. - * sizeof(AVFormatContext) must not be used outside libav*, use - * avformat_alloc_context() to create an AVFormatContext. - */ -typedef struct AVFormatContext { - /** - * A class for logging and @ref avoptions. Set by avformat_alloc_context(). - * Exports (de)muxer private options if they exist. - */ - const AVClass *av_class; - - /** - * The input container format. - * - * Demuxing only, set by avformat_open_input(). - */ - struct AVInputFormat *iformat; - - /** - * The output container format. - * - * Muxing only, must be set by the caller before avformat_write_header(). - */ - struct AVOutputFormat *oformat; - - /** - * Format private data. This is an AVOptions-enabled struct - * if and only if iformat/oformat.priv_class is not NULL. - * - * - muxing: set by avformat_write_header() - * - demuxing: set by avformat_open_input() - */ - void *priv_data; - - /** - * I/O context. - * - * - demuxing: either set by the user before avformat_open_input() (then - * the user must close it manually) or set by avformat_open_input(). - * - muxing: set by the user before avformat_write_header(). The caller must - * take care of closing / freeing the IO context. - * - * Do NOT set this field if AVFMT_NOFILE flag is set in - * iformat/oformat.flags. In such a case, the (de)muxer will handle - * I/O in some other way and this field will be NULL. - */ - AVIOContext *pb; - - /* stream info */ - /** - * Flags signalling stream properties. A combination of AVFMTCTX_*. - * Set by libavformat. - */ - int ctx_flags; - - /** - * Number of elements in AVFormatContext.streams. - * - * Set by avformat_new_stream(), must not be modified by any other code. - */ - unsigned int nb_streams; - /** - * A list of all streams in the file. New streams are created with - * avformat_new_stream(). - * - * - demuxing: streams are created by libavformat in avformat_open_input(). - * If AVFMTCTX_NOHEADER is set in ctx_flags, then new streams may also - * appear in av_read_frame(). - * - muxing: streams are created by the user before avformat_write_header(). - * - * Freed by libavformat in avformat_free_context(). - */ - AVStream **streams; - - /** - * input or output filename - * - * - demuxing: set by avformat_open_input() - * - muxing: may be set by the caller before avformat_write_header() - */ - char filename[1024]; - - /** - * Position of the first frame of the component, in - * AV_TIME_BASE fractional seconds. NEVER set this value directly: - * It is deduced from the AVStream values. - * - * Demuxing only, set by libavformat. - */ - int64_t start_time; - - /** - * Duration of the stream, in AV_TIME_BASE fractional - * seconds. Only set this value if you know none of the individual stream - * durations and also do not set any of them. This is deduced from the - * AVStream values if not set. - * - * Demuxing only, set by libavformat. - */ - int64_t duration; - - /** - * Total stream bitrate in bit/s, 0 if not - * available. Never set it directly if the file_size and the - * duration are known as FFmpeg can compute it automatically. - */ - int64_t bit_rate; - - unsigned int packet_size; - int max_delay; - - /** - * Flags modifying the (de)muxer behaviour. A combination of AVFMT_FLAG_*. - * Set by the user before avformat_open_input() / avformat_write_header(). - */ - int flags; -#define AVFMT_FLAG_GENPTS 0x0001 ///< Generate missing pts even if it requires parsing future frames. -#define AVFMT_FLAG_IGNIDX 0x0002 ///< Ignore index. -#define AVFMT_FLAG_NONBLOCK 0x0004 ///< Do not block when reading packets from input. -#define AVFMT_FLAG_IGNDTS 0x0008 ///< Ignore DTS on frames that contain both DTS & PTS -#define AVFMT_FLAG_NOFILLIN 0x0010 ///< Do not infer any values from other values, just return what is stored in the container -#define AVFMT_FLAG_NOPARSE 0x0020 ///< Do not use AVParsers, you also must set AVFMT_FLAG_NOFILLIN as the fillin code works on frames and no parsing -> no frames. Also seeking to frames can not work if parsing to find frame boundaries has been disabled -#define AVFMT_FLAG_NOBUFFER 0x0040 ///< Do not buffer frames when possible -#define AVFMT_FLAG_CUSTOM_IO 0x0080 ///< The caller has supplied a custom AVIOContext, don't avio_close() it. -#define AVFMT_FLAG_DISCARD_CORRUPT 0x0100 ///< Discard frames marked corrupted -#define AVFMT_FLAG_FLUSH_PACKETS 0x0200 ///< Flush the AVIOContext every packet. -/** - * When muxing, try to avoid writing any random/volatile data to the output. - * This includes any random IDs, real-time timestamps/dates, muxer version, etc. - * - * This flag is mainly intended for testing. - */ -#define AVFMT_FLAG_BITEXACT 0x0400 -#define AVFMT_FLAG_MP4A_LATM 0x8000 ///< Enable RTP MP4A-LATM payload -#define AVFMT_FLAG_SORT_DTS 0x10000 ///< try to interleave outputted packets by dts (using this flag can slow demuxing down) -#define AVFMT_FLAG_PRIV_OPT 0x20000 ///< Enable use of private options by delaying codec open (this could be made default once all code is converted) -#define AVFMT_FLAG_KEEP_SIDE_DATA 0x40000 ///< Don't merge side data but keep it separate. -#define AVFMT_FLAG_FAST_SEEK 0x80000 ///< Enable fast, but inaccurate seeks for some formats - - /** - * Maximum size of the data read from input for determining - * the input container format. - * Demuxing only, set by the caller before avformat_open_input(). - */ - int64_t probesize; - - /** - * Maximum duration (in AV_TIME_BASE units) of the data read - * from input in avformat_find_stream_info(). - * Demuxing only, set by the caller before avformat_find_stream_info(). - * Can be set to 0 to let avformat choose using a heuristic. - */ - int64_t max_analyze_duration; - - const uint8_t *key; - int keylen; - - unsigned int nb_programs; - AVProgram **programs; - - /** - * Forced video codec_id. - * Demuxing: Set by user. - */ - enum AVCodecID video_codec_id; - - /** - * Forced audio codec_id. - * Demuxing: Set by user. - */ - enum AVCodecID audio_codec_id; - - /** - * Forced subtitle codec_id. - * Demuxing: Set by user. - */ - enum AVCodecID subtitle_codec_id; - - /** - * Maximum amount of memory in bytes to use for the index of each stream. - * If the index exceeds this size, entries will be discarded as - * needed to maintain a smaller size. This can lead to slower or less - * accurate seeking (depends on demuxer). - * Demuxers for which a full in-memory index is mandatory will ignore - * this. - * - muxing: unused - * - demuxing: set by user - */ - unsigned int max_index_size; - - /** - * Maximum amount of memory in bytes to use for buffering frames - * obtained from realtime capture devices. - */ - unsigned int max_picture_buffer; - - /** - * Number of chapters in AVChapter array. - * When muxing, chapters are normally written in the file header, - * so nb_chapters should normally be initialized before write_header - * is called. Some muxers (e.g. mov and mkv) can also write chapters - * in the trailer. To write chapters in the trailer, nb_chapters - * must be zero when write_header is called and non-zero when - * write_trailer is called. - * - muxing: set by user - * - demuxing: set by libavformat - */ - unsigned int nb_chapters; - AVChapter **chapters; - - /** - * Metadata that applies to the whole file. - * - * - demuxing: set by libavformat in avformat_open_input() - * - muxing: may be set by the caller before avformat_write_header() - * - * Freed by libavformat in avformat_free_context(). - */ - AVDictionary *metadata; - - /** - * Start time of the stream in real world time, in microseconds - * since the Unix epoch (00:00 1st January 1970). That is, pts=0 in the - * stream was captured at this real world time. - * - muxing: Set by the caller before avformat_write_header(). If set to - * either 0 or AV_NOPTS_VALUE, then the current wall-time will - * be used. - * - demuxing: Set by libavformat. AV_NOPTS_VALUE if unknown. Note that - * the value may become known after some number of frames - * have been received. - */ - int64_t start_time_realtime; - - /** - * The number of frames used for determining the framerate in - * avformat_find_stream_info(). - * Demuxing only, set by the caller before avformat_find_stream_info(). - */ - int fps_probe_size; - - /** - * Error recognition; higher values will detect more errors but may - * misdetect some more or less valid parts as errors. - * Demuxing only, set by the caller before avformat_open_input(). - */ - int error_recognition; - - /** - * Custom interrupt callbacks for the I/O layer. - * - * demuxing: set by the user before avformat_open_input(). - * muxing: set by the user before avformat_write_header() - * (mainly useful for AVFMT_NOFILE formats). The callback - * should also be passed to avio_open2() if it's used to - * open the file. - */ - AVIOInterruptCB interrupt_callback; - - /** - * Flags to enable debugging. - */ - int debug; -#define FF_FDEBUG_TS 0x0001 - - /** - * Maximum buffering duration for interleaving. - * - * To ensure all the streams are interleaved correctly, - * av_interleaved_write_frame() will wait until it has at least one packet - * for each stream before actually writing any packets to the output file. - * When some streams are "sparse" (i.e. there are large gaps between - * successive packets), this can result in excessive buffering. - * - * This field specifies the maximum difference between the timestamps of the - * first and the last packet in the muxing queue, above which libavformat - * will output a packet regardless of whether it has queued a packet for all - * the streams. - * - * Muxing only, set by the caller before avformat_write_header(). - */ - int64_t max_interleave_delta; - - /** - * Allow non-standard and experimental extension - * @see AVCodecContext.strict_std_compliance - */ - int strict_std_compliance; - - /** - * Flags for the user to detect events happening on the file. Flags must - * be cleared by the user once the event has been handled. - * A combination of AVFMT_EVENT_FLAG_*. - */ - int event_flags; -#define AVFMT_EVENT_FLAG_METADATA_UPDATED 0x0001 ///< The call resulted in updated metadata. - - /** - * Maximum number of packets to read while waiting for the first timestamp. - * Decoding only. - */ - int max_ts_probe; - - /** - * Avoid negative timestamps during muxing. - * Any value of the AVFMT_AVOID_NEG_TS_* constants. - * Note, this only works when using av_interleaved_write_frame. (interleave_packet_per_dts is in use) - * - muxing: Set by user - * - demuxing: unused - */ - int avoid_negative_ts; -#define AVFMT_AVOID_NEG_TS_AUTO -1 ///< Enabled when required by target format -#define AVFMT_AVOID_NEG_TS_MAKE_NON_NEGATIVE 1 ///< Shift timestamps so they are non negative -#define AVFMT_AVOID_NEG_TS_MAKE_ZERO 2 ///< Shift timestamps so that they start at 0 - - /** - * Transport stream id. - * This will be moved into demuxer private options. Thus no API/ABI compatibility - */ - int ts_id; - - /** - * Audio preload in microseconds. - * Note, not all formats support this and unpredictable things may happen if it is used when not supported. - * - encoding: Set by user via AVOptions (NO direct access) - * - decoding: unused - */ - int audio_preload; - - /** - * Max chunk time in microseconds. - * Note, not all formats support this and unpredictable things may happen if it is used when not supported. - * - encoding: Set by user via AVOptions (NO direct access) - * - decoding: unused - */ - int max_chunk_duration; - - /** - * Max chunk size in bytes - * Note, not all formats support this and unpredictable things may happen if it is used when not supported. - * - encoding: Set by user via AVOptions (NO direct access) - * - decoding: unused - */ - int max_chunk_size; - - /** - * forces the use of wallclock timestamps as pts/dts of packets - * This has undefined results in the presence of B frames. - * - encoding: unused - * - decoding: Set by user via AVOptions (NO direct access) - */ - int use_wallclock_as_timestamps; - - /** - * avio flags, used to force AVIO_FLAG_DIRECT. - * - encoding: unused - * - decoding: Set by user via AVOptions (NO direct access) - */ - int avio_flags; - - /** - * The duration field can be estimated through various ways, and this field can be used - * to know how the duration was estimated. - * - encoding: unused - * - decoding: Read by user via AVOptions (NO direct access) - */ - enum AVDurationEstimationMethod duration_estimation_method; - - /** - * Skip initial bytes when opening stream - * - encoding: unused - * - decoding: Set by user via AVOptions (NO direct access) - */ - int64_t skip_initial_bytes; - - /** - * Correct single timestamp overflows - * - encoding: unused - * - decoding: Set by user via AVOptions (NO direct access) - */ - unsigned int correct_ts_overflow; - - /** - * Force seeking to any (also non key) frames. - * - encoding: unused - * - decoding: Set by user via AVOptions (NO direct access) - */ - int seek2any; - - /** - * Flush the I/O context after each packet. - * - encoding: Set by user via AVOptions (NO direct access) - * - decoding: unused - */ - int flush_packets; - - /** - * format probing score. - * The maximal score is AVPROBE_SCORE_MAX, its set when the demuxer probes - * the format. - * - encoding: unused - * - decoding: set by avformat, read by user via av_format_get_probe_score() (NO direct access) - */ - int probe_score; - - /** - * number of bytes to read maximally to identify format. - * - encoding: unused - * - decoding: set by user through AVOPtions (NO direct access) - */ - int format_probesize; - - /** - * ',' separated list of allowed decoders. - * If NULL then all are allowed - * - encoding: unused - * - decoding: set by user through AVOptions (NO direct access) - */ - char *codec_whitelist; - - /** - * ',' separated list of allowed demuxers. - * If NULL then all are allowed - * - encoding: unused - * - decoding: set by user through AVOptions (NO direct access) - */ - char *format_whitelist; - - /** - * An opaque field for libavformat internal usage. - * Must not be accessed in any way by callers. - */ - AVFormatInternal *internal; - - /** - * IO repositioned flag. - * This is set by avformat when the underlaying IO context read pointer - * is repositioned, for example when doing byte based seeking. - * Demuxers can use the flag to detect such changes. - */ - int io_repositioned; - - /** - * Forced video codec. - * This allows forcing a specific decoder, even when there are multiple with - * the same codec_id. - * Demuxing: Set by user via av_format_set_video_codec (NO direct access). - */ - AVCodec *video_codec; - - /** - * Forced audio codec. - * This allows forcing a specific decoder, even when there are multiple with - * the same codec_id. - * Demuxing: Set by user via av_format_set_audio_codec (NO direct access). - */ - AVCodec *audio_codec; - - /** - * Forced subtitle codec. - * This allows forcing a specific decoder, even when there are multiple with - * the same codec_id. - * Demuxing: Set by user via av_format_set_subtitle_codec (NO direct access). - */ - AVCodec *subtitle_codec; - - /** - * Forced data codec. - * This allows forcing a specific decoder, even when there are multiple with - * the same codec_id. - * Demuxing: Set by user via av_format_set_data_codec (NO direct access). - */ - AVCodec *data_codec; - - /** - * Number of bytes to be written as padding in a metadata header. - * Demuxing: Unused. - * Muxing: Set by user via av_format_set_metadata_header_padding. - */ - int metadata_header_padding; - - /** - * User data. - * This is a place for some private data of the user. - * Mostly usable with control_message_cb or any future callbacks in device's context. - */ - void *opaque; - - /** - * Callback used by devices to communicate with application. - */ - av_format_control_message control_message_cb; - - /** - * Output timestamp offset, in microseconds. - * Muxing: set by user via AVOptions (NO direct access) - */ - int64_t output_ts_offset; - - /** - * dump format separator. - * can be ", " or "\n " or anything else - * Code outside libavformat should access this field using AVOptions - * (NO direct access). - * - muxing: Set by user. - * - demuxing: Set by user. - */ - uint8_t *dump_separator; - - /** - * Forced Data codec_id. - * Demuxing: Set by user. - */ - enum AVCodecID data_codec_id; - - /** - * Called to open further IO contexts when needed for demuxing. - * - * This can be set by the user application to perform security checks on - * the URLs before opening them. - * The function should behave like avio_open2(), AVFormatContext is provided - * as contextual information and to reach AVFormatContext.opaque. - * - * If NULL then some simple checks are used together with avio_open2(). - * - * Must not be accessed directly from outside avformat. - * @See av_format_set_open_cb() - * - * Demuxing: Set by user. - */ - int (*open_cb)(struct AVFormatContext *s, AVIOContext **p, const char *url, int flags, const AVIOInterruptCB *int_cb, AVDictionary **options); -} AVFormatContext; - -int av_format_get_probe_score(const AVFormatContext *s); -AVCodec * av_format_get_video_codec(const AVFormatContext *s); -void av_format_set_video_codec(AVFormatContext *s, AVCodec *c); -AVCodec * av_format_get_audio_codec(const AVFormatContext *s); -void av_format_set_audio_codec(AVFormatContext *s, AVCodec *c); -AVCodec * av_format_get_subtitle_codec(const AVFormatContext *s); -void av_format_set_subtitle_codec(AVFormatContext *s, AVCodec *c); -AVCodec * av_format_get_data_codec(const AVFormatContext *s); -void av_format_set_data_codec(AVFormatContext *s, AVCodec *c); -int av_format_get_metadata_header_padding(const AVFormatContext *s); -void av_format_set_metadata_header_padding(AVFormatContext *s, int c); -void * av_format_get_opaque(const AVFormatContext *s); -void av_format_set_opaque(AVFormatContext *s, void *opaque); -av_format_control_message av_format_get_control_message_cb(const AVFormatContext *s); -void av_format_set_control_message_cb(AVFormatContext *s, av_format_control_message callback); -AVOpenCallback av_format_get_open_cb(const AVFormatContext *s); -void av_format_set_open_cb(AVFormatContext *s, AVOpenCallback callback); - -/** - * This function will cause global side data to be injected in the next packet - * of each stream as well as after any subsequent seek. - */ -void av_format_inject_global_side_data(AVFormatContext *s); - -/** - * Returns the method used to set ctx->duration. - * - * @return AVFMT_DURATION_FROM_PTS, AVFMT_DURATION_FROM_STREAM, or AVFMT_DURATION_FROM_BITRATE. - */ -enum AVDurationEstimationMethod av_fmt_ctx_get_duration_estimation_method(const AVFormatContext* ctx); - -typedef struct AVPacketList { - AVPacket pkt; - struct AVPacketList *next; -} AVPacketList; - - -/** - * @defgroup lavf_core Core functions - * @ingroup libavf - * - * Functions for querying libavformat capabilities, allocating core structures, - * etc. - * @{ - */ - -/** - * Return the LIBAVFORMAT_VERSION_INT constant. - */ -unsigned avformat_version(void); - -/** - * Return the libavformat build-time configuration. - */ -const char *avformat_configuration(void); - -/** - * Return the libavformat license. - */ -const char *avformat_license(void); - -/** - * Initialize libavformat and register all the muxers, demuxers and - * protocols. If you do not call this function, then you can select - * exactly which formats you want to support. - * - * @see av_register_input_format() - * @see av_register_output_format() - */ -void av_register_all(void); - -void av_register_input_format(AVInputFormat *format); -void av_register_output_format(AVOutputFormat *format); - -/** - * Do global initialization of network components. This is optional, - * but recommended, since it avoids the overhead of implicitly - * doing the setup for each session. - * - * Calling this function will become mandatory if using network - * protocols at some major version bump. - */ -int avformat_network_init(void); - -/** - * Undo the initialization done by avformat_network_init. - */ -int avformat_network_deinit(void); - -/** - * If f is NULL, returns the first registered input format, - * if f is non-NULL, returns the next registered input format after f - * or NULL if f is the last one. - */ -AVInputFormat *av_iformat_next(const AVInputFormat *f); - -/** - * If f is NULL, returns the first registered output format, - * if f is non-NULL, returns the next registered output format after f - * or NULL if f is the last one. - */ -AVOutputFormat *av_oformat_next(const AVOutputFormat *f); - -/** - * Allocate an AVFormatContext. - * avformat_free_context() can be used to free the context and everything - * allocated by the framework within it. - */ -AVFormatContext *avformat_alloc_context(void); - -/** - * Free an AVFormatContext and all its streams. - * @param s context to free - */ -void avformat_free_context(AVFormatContext *s); - -/** - * Get the AVClass for AVFormatContext. It can be used in combination with - * AV_OPT_SEARCH_FAKE_OBJ for examining options. - * - * @see av_opt_find(). - */ -const AVClass *avformat_get_class(void); - -/** - * Add a new stream to a media file. - * - * When demuxing, it is called by the demuxer in read_header(). If the - * flag AVFMTCTX_NOHEADER is set in s.ctx_flags, then it may also - * be called in read_packet(). - * - * When muxing, should be called by the user before avformat_write_header(). - * - * User is required to call avcodec_close() and avformat_free_context() to - * clean up the allocation by avformat_new_stream(). - * - * @param s media file handle - * @param c If non-NULL, the AVCodecContext corresponding to the new stream - * will be initialized to use this codec. This is needed for e.g. codec-specific - * defaults to be set, so codec should be provided if it is known. - * - * @return newly created stream or NULL on error. - */ -AVStream *avformat_new_stream(AVFormatContext *s, const AVCodec *c); - -/** - * Allocate new information from stream. - * - * @param stream stream - * @param type desired side information type - * @param size side information size - * @return pointer to fresh allocated data or NULL otherwise - */ -uint8_t *av_stream_new_side_data(AVStream *stream, - enum AVPacketSideDataType type, int size); -/** - * Get side information from stream. - * - * @param stream stream - * @param type desired side information type - * @param size pointer for side information size to store (optional) - * @return pointer to data if present or NULL otherwise - */ -uint8_t *av_stream_get_side_data(AVStream *stream, - enum AVPacketSideDataType type, int *size); - -AVProgram *av_new_program(AVFormatContext *s, int id); - -/** - * @} - */ - - -/** - * Allocate an AVFormatContext for an output format. - * avformat_free_context() can be used to free the context and - * everything allocated by the framework within it. - * - * @param *ctx is set to the created format context, or to NULL in - * case of failure - * @param oformat format to use for allocating the context, if NULL - * format_name and filename are used instead - * @param format_name the name of output format to use for allocating the - * context, if NULL filename is used instead - * @param filename the name of the filename to use for allocating the - * context, may be NULL - * @return >= 0 in case of success, a negative AVERROR code in case of - * failure - */ -int avformat_alloc_output_context2(AVFormatContext **ctx, AVOutputFormat *oformat, - const char *format_name, const char *filename); - -/** - * @addtogroup lavf_decoding - * @{ - */ - -/** - * Find AVInputFormat based on the short name of the input format. - */ -AVInputFormat *av_find_input_format(const char *short_name); - -/** - * Guess the file format. - * - * @param pd data to be probed - * @param is_opened Whether the file is already opened; determines whether - * demuxers with or without AVFMT_NOFILE are probed. - */ -AVInputFormat *av_probe_input_format(AVProbeData *pd, int is_opened); - -/** - * Guess the file format. - * - * @param pd data to be probed - * @param is_opened Whether the file is already opened; determines whether - * demuxers with or without AVFMT_NOFILE are probed. - * @param score_max A probe score larger that this is required to accept a - * detection, the variable is set to the actual detection - * score afterwards. - * If the score is <= AVPROBE_SCORE_MAX / 4 it is recommended - * to retry with a larger probe buffer. - */ -AVInputFormat *av_probe_input_format2(AVProbeData *pd, int is_opened, int *score_max); - -/** - * Guess the file format. - * - * @param is_opened Whether the file is already opened; determines whether - * demuxers with or without AVFMT_NOFILE are probed. - * @param score_ret The score of the best detection. - */ -AVInputFormat *av_probe_input_format3(AVProbeData *pd, int is_opened, int *score_ret); - -/** - * Probe a bytestream to determine the input format. Each time a probe returns - * with a score that is too low, the probe buffer size is increased and another - * attempt is made. When the maximum probe size is reached, the input format - * with the highest score is returned. - * - * @param pb the bytestream to probe - * @param fmt the input format is put here - * @param filename the filename of the stream - * @param logctx the log context - * @param offset the offset within the bytestream to probe from - * @param max_probe_size the maximum probe buffer size (zero for default) - * @return the score in case of success, a negative value corresponding to an - * the maximal score is AVPROBE_SCORE_MAX - * AVERROR code otherwise - */ -int av_probe_input_buffer2(AVIOContext *pb, AVInputFormat **fmt, - const char *filename, void *logctx, - unsigned int offset, unsigned int max_probe_size); - -/** - * Like av_probe_input_buffer2() but returns 0 on success - */ -int av_probe_input_buffer(AVIOContext *pb, AVInputFormat **fmt, - const char *filename, void *logctx, - unsigned int offset, unsigned int max_probe_size); - -/** - * Open an input stream and read the header. The codecs are not opened. - * The stream must be closed with avformat_close_input(). - * - * @param ps Pointer to user-supplied AVFormatContext (allocated by avformat_alloc_context). - * May be a pointer to NULL, in which case an AVFormatContext is allocated by this - * function and written into ps. - * Note that a user-supplied AVFormatContext will be freed on failure. - * @param filename Name of the stream to open. - * @param fmt If non-NULL, this parameter forces a specific input format. - * Otherwise the format is autodetected. - * @param options A dictionary filled with AVFormatContext and demuxer-private options. - * On return this parameter will be destroyed and replaced with a dict containing - * options that were not found. May be NULL. - * - * @return 0 on success, a negative AVERROR on failure. - * - * @note If you want to use custom IO, preallocate the format context and set its pb field. - */ -int avformat_open_input(AVFormatContext **ps, const char *filename, AVInputFormat *fmt, AVDictionary **options); - -attribute_deprecated -int av_demuxer_open(AVFormatContext *ic); - -/** - * Read packets of a media file to get stream information. This - * is useful for file formats with no headers such as MPEG. This - * function also computes the real framerate in case of MPEG-2 repeat - * frame mode. - * The logical file position is not changed by this function; - * examined packets may be buffered for later processing. - * - * @param ic media file handle - * @param options If non-NULL, an ic.nb_streams long array of pointers to - * dictionaries, where i-th member contains options for - * codec corresponding to i-th stream. - * On return each dictionary will be filled with options that were not found. - * @return >=0 if OK, AVERROR_xxx on error - * - * @note this function isn't guaranteed to open all the codecs, so - * options being non-empty at return is a perfectly normal behavior. - * - * @todo Let the user decide somehow what information is needed so that - * we do not waste time getting stuff the user does not need. - */ -int avformat_find_stream_info(AVFormatContext *ic, AVDictionary **options); - -/** - * Find the programs which belong to a given stream. - * - * @param ic media file handle - * @param last the last found program, the search will start after this - * program, or from the beginning if it is NULL - * @param s stream index - * @return the next program which belongs to s, NULL if no program is found or - * the last program is not among the programs of ic. - */ -AVProgram *av_find_program_from_stream(AVFormatContext *ic, AVProgram *last, int s); - -void av_program_add_stream_index(AVFormatContext *ac, int progid, unsigned int idx); - -/** - * Find the "best" stream in the file. - * The best stream is determined according to various heuristics as the most - * likely to be what the user expects. - * If the decoder parameter is non-NULL, av_find_best_stream will find the - * default decoder for the stream's codec; streams for which no decoder can - * be found are ignored. - * - * @param ic media file handle - * @param type stream type: video, audio, subtitles, etc. - * @param wanted_stream_nb user-requested stream number, - * or -1 for automatic selection - * @param related_stream try to find a stream related (eg. in the same - * program) to this one, or -1 if none - * @param decoder_ret if non-NULL, returns the decoder for the - * selected stream - * @param flags flags; none are currently defined - * @return the non-negative stream number in case of success, - * AVERROR_STREAM_NOT_FOUND if no stream with the requested type - * could be found, - * AVERROR_DECODER_NOT_FOUND if streams were found but no decoder - * @note If av_find_best_stream returns successfully and decoder_ret is not - * NULL, then *decoder_ret is guaranteed to be set to a valid AVCodec. - */ -int av_find_best_stream(AVFormatContext *ic, - enum AVMediaType type, - int wanted_stream_nb, - int related_stream, - AVCodec **decoder_ret, - int flags); - -/** - * Return the next frame of a stream. - * This function returns what is stored in the file, and does not validate - * that what is there are valid frames for the decoder. It will split what is - * stored in the file into frames and return one for each call. It will not - * omit invalid data between valid frames so as to give the decoder the maximum - * information possible for decoding. - * - * If pkt->buf is NULL, then the packet is valid until the next - * av_read_frame() or until avformat_close_input(). Otherwise the packet - * is valid indefinitely. In both cases the packet must be freed with - * av_packet_unref when it is no longer needed. For video, the packet contains - * exactly one frame. For audio, it contains an integer number of frames if each - * frame has a known fixed size (e.g. PCM or ADPCM data). If the audio frames - * have a variable size (e.g. MPEG audio), then it contains one frame. - * - * pkt->pts, pkt->dts and pkt->duration are always set to correct - * values in AVStream.time_base units (and guessed if the format cannot - * provide them). pkt->pts can be AV_NOPTS_VALUE if the video format - * has B-frames, so it is better to rely on pkt->dts if you do not - * decompress the payload. - * - * @return 0 if OK, < 0 on error or end of file - */ -int av_read_frame(AVFormatContext *s, AVPacket *pkt); - -/** - * Seek to the keyframe at timestamp. - * 'timestamp' in 'stream_index'. - * - * @param s media file handle - * @param stream_index If stream_index is (-1), a default - * stream is selected, and timestamp is automatically converted - * from AV_TIME_BASE units to the stream specific time_base. - * @param timestamp Timestamp in AVStream.time_base units - * or, if no stream is specified, in AV_TIME_BASE units. - * @param flags flags which select direction and seeking mode - * @return >= 0 on success - */ -int av_seek_frame(AVFormatContext *s, int stream_index, int64_t timestamp, - int flags); - -/** - * Seek to timestamp ts. - * Seeking will be done so that the point from which all active streams - * can be presented successfully will be closest to ts and within min/max_ts. - * Active streams are all streams that have AVStream.discard < AVDISCARD_ALL. - * - * If flags contain AVSEEK_FLAG_BYTE, then all timestamps are in bytes and - * are the file position (this may not be supported by all demuxers). - * If flags contain AVSEEK_FLAG_FRAME, then all timestamps are in frames - * in the stream with stream_index (this may not be supported by all demuxers). - * Otherwise all timestamps are in units of the stream selected by stream_index - * or if stream_index is -1, in AV_TIME_BASE units. - * If flags contain AVSEEK_FLAG_ANY, then non-keyframes are treated as - * keyframes (this may not be supported by all demuxers). - * If flags contain AVSEEK_FLAG_BACKWARD, it is ignored. - * - * @param s media file handle - * @param stream_index index of the stream which is used as time base reference - * @param min_ts smallest acceptable timestamp - * @param ts target timestamp - * @param max_ts largest acceptable timestamp - * @param flags flags - * @return >=0 on success, error code otherwise - * - * @note This is part of the new seek API which is still under construction. - * Thus do not use this yet. It may change at any time, do not expect - * ABI compatibility yet! - */ -int avformat_seek_file(AVFormatContext *s, int stream_index, int64_t min_ts, int64_t ts, int64_t max_ts, int flags); - -/** - * Discard all internally buffered data. This can be useful when dealing with - * discontinuities in the byte stream. Generally works only with formats that - * can resync. This includes headerless formats like MPEG-TS/TS but should also - * work with NUT, Ogg and in a limited way AVI for example. - * - * The set of streams, the detected duration, stream parameters and codecs do - * not change when calling this function. If you want a complete reset, it's - * better to open a new AVFormatContext. - * - * This does not flush the AVIOContext (s->pb). If necessary, call - * avio_flush(s->pb) before calling this function. - * - * @param s media file handle - * @return >=0 on success, error code otherwise - */ -int avformat_flush(AVFormatContext *s); - -/** - * Start playing a network-based stream (e.g. RTSP stream) at the - * current position. - */ -int av_read_play(AVFormatContext *s); - -/** - * Pause a network-based stream (e.g. RTSP stream). - * - * Use av_read_play() to resume it. - */ -int av_read_pause(AVFormatContext *s); - -/** - * Close an opened input AVFormatContext. Free it and all its contents - * and set *s to NULL. - */ -void avformat_close_input(AVFormatContext **s); -/** - * @} - */ - -#define AVSEEK_FLAG_BACKWARD 1 ///< seek backward -#define AVSEEK_FLAG_BYTE 2 ///< seeking based on position in bytes -#define AVSEEK_FLAG_ANY 4 ///< seek to any frame, even non-keyframes -#define AVSEEK_FLAG_FRAME 8 ///< seeking based on frame number - -/** - * @addtogroup lavf_encoding - * @{ - */ -/** - * Allocate the stream private data and write the stream header to - * an output media file. - * - * @param s Media file handle, must be allocated with avformat_alloc_context(). - * Its oformat field must be set to the desired output format; - * Its pb field must be set to an already opened AVIOContext. - * @param options An AVDictionary filled with AVFormatContext and muxer-private options. - * On return this parameter will be destroyed and replaced with a dict containing - * options that were not found. May be NULL. - * - * @return 0 on success, negative AVERROR on failure. - * - * @see av_opt_find, av_dict_set, avio_open, av_oformat_next. - */ -av_warn_unused_result -int avformat_write_header(AVFormatContext *s, AVDictionary **options); - -/** - * Write a packet to an output media file. - * - * This function passes the packet directly to the muxer, without any buffering - * or reordering. The caller is responsible for correctly interleaving the - * packets if the format requires it. Callers that want libavformat to handle - * the interleaving should call av_interleaved_write_frame() instead of this - * function. - * - * @param s media file handle - * @param pkt The packet containing the data to be written. Note that unlike - * av_interleaved_write_frame(), this function does not take - * ownership of the packet passed to it (though some muxers may make - * an internal reference to the input packet). - * <br> - * This parameter can be NULL (at any time, not just at the end), in - * order to immediately flush data buffered within the muxer, for - * muxers that buffer up data internally before writing it to the - * output. - * <br> - * Packet's @ref AVPacket.stream_index "stream_index" field must be - * set to the index of the corresponding stream in @ref - * AVFormatContext.streams "s->streams". - * <br> - * The timestamps (@ref AVPacket.pts "pts", @ref AVPacket.dts "dts") - * must be set to correct values in the stream's timebase (unless the - * output format is flagged with the AVFMT_NOTIMESTAMPS flag, then - * they can be set to AV_NOPTS_VALUE). - * The dts for subsequent packets passed to this function must be strictly - * increasing when compared in their respective timebases (unless the - * output format is flagged with the AVFMT_TS_NONSTRICT, then they - * merely have to be nondecreasing). @ref AVPacket.duration - * "duration") should also be set if known. - * @return < 0 on error, = 0 if OK, 1 if flushed and there is no more data to flush - * - * @see av_interleaved_write_frame() - */ -int av_write_frame(AVFormatContext *s, AVPacket *pkt); - -/** - * Write a packet to an output media file ensuring correct interleaving. - * - * This function will buffer the packets internally as needed to make sure the - * packets in the output file are properly interleaved in the order of - * increasing dts. Callers doing their own interleaving should call - * av_write_frame() instead of this function. - * - * @param s media file handle - * @param pkt The packet containing the data to be written. - * <br> - * If the packet is reference-counted, this function will take - * ownership of this reference and unreference it later when it sees - * fit. - * The caller must not access the data through this reference after - * this function returns. If the packet is not reference-counted, - * libavformat will make a copy. - * <br> - * This parameter can be NULL (at any time, not just at the end), to - * flush the interleaving queues. - * <br> - * Packet's @ref AVPacket.stream_index "stream_index" field must be - * set to the index of the corresponding stream in @ref - * AVFormatContext.streams "s->streams". - * <br> - * The timestamps (@ref AVPacket.pts "pts", @ref AVPacket.dts "dts") - * must be set to correct values in the stream's timebase (unless the - * output format is flagged with the AVFMT_NOTIMESTAMPS flag, then - * they can be set to AV_NOPTS_VALUE). - * The dts for subsequent packets in one stream must be strictly - * increasing (unless the output format is flagged with the - * AVFMT_TS_NONSTRICT, then they merely have to be nondecreasing). - * @ref AVPacket.duration "duration") should also be set if known. - * - * @return 0 on success, a negative AVERROR on error. Libavformat will always - * take care of freeing the packet, even if this function fails. - * - * @see av_write_frame(), AVFormatContext.max_interleave_delta - */ -int av_interleaved_write_frame(AVFormatContext *s, AVPacket *pkt); - -/** - * Write a uncoded frame to an output media file. - * - * The frame must be correctly interleaved according to the container - * specification; if not, then av_interleaved_write_frame() must be used. - * - * See av_interleaved_write_frame() for details. - */ -int av_write_uncoded_frame(AVFormatContext *s, int stream_index, - AVFrame *frame); - -/** - * Write a uncoded frame to an output media file. - * - * If the muxer supports it, this function makes it possible to write an AVFrame - * structure directly, without encoding it into a packet. - * It is mostly useful for devices and similar special muxers that use raw - * video or PCM data and will not serialize it into a byte stream. - * - * To test whether it is possible to use it with a given muxer and stream, - * use av_write_uncoded_frame_query(). - * - * The caller gives up ownership of the frame and must not access it - * afterwards. - * - * @return >=0 for success, a negative code on error - */ -int av_interleaved_write_uncoded_frame(AVFormatContext *s, int stream_index, - AVFrame *frame); - -/** - * Test whether a muxer supports uncoded frame. - * - * @return >=0 if an uncoded frame can be written to that muxer and stream, - * <0 if not - */ -int av_write_uncoded_frame_query(AVFormatContext *s, int stream_index); - -/** - * Write the stream trailer to an output media file and free the - * file private data. - * - * May only be called after a successful call to avformat_write_header. - * - * @param s media file handle - * @return 0 if OK, AVERROR_xxx on error - */ -int av_write_trailer(AVFormatContext *s); - -/** - * Return the output format in the list of registered output formats - * which best matches the provided parameters, or return NULL if - * there is no match. - * - * @param short_name if non-NULL checks if short_name matches with the - * names of the registered formats - * @param filename if non-NULL checks if filename terminates with the - * extensions of the registered formats - * @param mime_type if non-NULL checks if mime_type matches with the - * MIME type of the registered formats - */ -AVOutputFormat *av_guess_format(const char *short_name, - const char *filename, - const char *mime_type); - -/** - * Guess the codec ID based upon muxer and filename. - */ -enum AVCodecID av_guess_codec(AVOutputFormat *fmt, const char *short_name, - const char *filename, const char *mime_type, - enum AVMediaType type); - -/** - * Get timing information for the data currently output. - * The exact meaning of "currently output" depends on the format. - * It is mostly relevant for devices that have an internal buffer and/or - * work in real time. - * @param s media file handle - * @param stream stream in the media file - * @param[out] dts DTS of the last packet output for the stream, in stream - * time_base units - * @param[out] wall absolute time when that packet whas output, - * in microsecond - * @return 0 if OK, AVERROR(ENOSYS) if the format does not support it - * Note: some formats or devices may not allow to measure dts and wall - * atomically. - */ -int av_get_output_timestamp(struct AVFormatContext *s, int stream, - int64_t *dts, int64_t *wall); - - -/** - * @} - */ - - -/** - * @defgroup lavf_misc Utility functions - * @ingroup libavf - * @{ - * - * Miscellaneous utility functions related to both muxing and demuxing - * (or neither). - */ - -/** - * Send a nice hexadecimal dump of a buffer to the specified file stream. - * - * @param f The file stream pointer where the dump should be sent to. - * @param buf buffer - * @param size buffer size - * - * @see av_hex_dump_log, av_pkt_dump2, av_pkt_dump_log2 - */ -void av_hex_dump(FILE *f, const uint8_t *buf, int size); - -/** - * Send a nice hexadecimal dump of a buffer to the log. - * - * @param avcl A pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct. - * @param level The importance level of the message, lower values signifying - * higher importance. - * @param buf buffer - * @param size buffer size - * - * @see av_hex_dump, av_pkt_dump2, av_pkt_dump_log2 - */ -void av_hex_dump_log(void *avcl, int level, const uint8_t *buf, int size); - -/** - * Send a nice dump of a packet to the specified file stream. - * - * @param f The file stream pointer where the dump should be sent to. - * @param pkt packet to dump - * @param dump_payload True if the payload must be displayed, too. - * @param st AVStream that the packet belongs to - */ -void av_pkt_dump2(FILE *f, const AVPacket *pkt, int dump_payload, const AVStream *st); - - -/** - * Send a nice dump of a packet to the log. - * - * @param avcl A pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct. - * @param level The importance level of the message, lower values signifying - * higher importance. - * @param pkt packet to dump - * @param dump_payload True if the payload must be displayed, too. - * @param st AVStream that the packet belongs to - */ -void av_pkt_dump_log2(void *avcl, int level, const AVPacket *pkt, int dump_payload, - const AVStream *st); - -/** - * Get the AVCodecID for the given codec tag tag. - * If no codec id is found returns AV_CODEC_ID_NONE. - * - * @param tags list of supported codec_id-codec_tag pairs, as stored - * in AVInputFormat.codec_tag and AVOutputFormat.codec_tag - * @param tag codec tag to match to a codec ID - */ -enum AVCodecID av_codec_get_id(const struct AVCodecTag * const *tags, unsigned int tag); - -/** - * Get the codec tag for the given codec id id. - * If no codec tag is found returns 0. - * - * @param tags list of supported codec_id-codec_tag pairs, as stored - * in AVInputFormat.codec_tag and AVOutputFormat.codec_tag - * @param id codec ID to match to a codec tag - */ -unsigned int av_codec_get_tag(const struct AVCodecTag * const *tags, enum AVCodecID id); - -/** - * Get the codec tag for the given codec id. - * - * @param tags list of supported codec_id - codec_tag pairs, as stored - * in AVInputFormat.codec_tag and AVOutputFormat.codec_tag - * @param id codec id that should be searched for in the list - * @param tag A pointer to the found tag - * @return 0 if id was not found in tags, > 0 if it was found - */ -int av_codec_get_tag2(const struct AVCodecTag * const *tags, enum AVCodecID id, - unsigned int *tag); - -int av_find_default_stream_index(AVFormatContext *s); - -/** - * Get the index for a specific timestamp. - * - * @param st stream that the timestamp belongs to - * @param timestamp timestamp to retrieve the index for - * @param flags if AVSEEK_FLAG_BACKWARD then the returned index will correspond - * to the timestamp which is <= the requested one, if backward - * is 0, then it will be >= - * if AVSEEK_FLAG_ANY seek to any frame, only keyframes otherwise - * @return < 0 if no such timestamp could be found - */ -int av_index_search_timestamp(AVStream *st, int64_t timestamp, int flags); - -/** - * Add an index entry into a sorted list. Update the entry if the list - * already contains it. - * - * @param timestamp timestamp in the time base of the given stream - */ -int av_add_index_entry(AVStream *st, int64_t pos, int64_t timestamp, - int size, int distance, int flags); - - -/** - * Split a URL string into components. - * - * The pointers to buffers for storing individual components may be null, - * in order to ignore that component. Buffers for components not found are - * set to empty strings. If the port is not found, it is set to a negative - * value. - * - * @param proto the buffer for the protocol - * @param proto_size the size of the proto buffer - * @param authorization the buffer for the authorization - * @param authorization_size the size of the authorization buffer - * @param hostname the buffer for the host name - * @param hostname_size the size of the hostname buffer - * @param port_ptr a pointer to store the port number in - * @param path the buffer for the path - * @param path_size the size of the path buffer - * @param url the URL to split - */ -void av_url_split(char *proto, int proto_size, - char *authorization, int authorization_size, - char *hostname, int hostname_size, - int *port_ptr, - char *path, int path_size, - const char *url); - - -/** - * Print detailed information about the input or output format, such as - * duration, bitrate, streams, container, programs, metadata, side data, - * codec and time base. - * - * @param ic the context to analyze - * @param index index of the stream to dump information about - * @param url the URL to print, such as source or destination file - * @param is_output Select whether the specified context is an input(0) or output(1) - */ -void av_dump_format(AVFormatContext *ic, - int index, - const char *url, - int is_output); - -/** - * Return in 'buf' the path with '%d' replaced by a number. - * - * Also handles the '%0nd' format where 'n' is the total number - * of digits and '%%'. - * - * @param buf destination buffer - * @param buf_size destination buffer size - * @param path numbered sequence string - * @param number frame number - * @return 0 if OK, -1 on format error - */ -int av_get_frame_filename(char *buf, int buf_size, - const char *path, int number); - -/** - * Check whether filename actually is a numbered sequence generator. - * - * @param filename possible numbered sequence string - * @return 1 if a valid numbered sequence string, 0 otherwise - */ -int av_filename_number_test(const char *filename); - -/** - * Generate an SDP for an RTP session. - * - * Note, this overwrites the id values of AVStreams in the muxer contexts - * for getting unique dynamic payload types. - * - * @param ac array of AVFormatContexts describing the RTP streams. If the - * array is composed by only one context, such context can contain - * multiple AVStreams (one AVStream per RTP stream). Otherwise, - * all the contexts in the array (an AVCodecContext per RTP stream) - * must contain only one AVStream. - * @param n_files number of AVCodecContexts contained in ac - * @param buf buffer where the SDP will be stored (must be allocated by - * the caller) - * @param size the size of the buffer - * @return 0 if OK, AVERROR_xxx on error - */ -int av_sdp_create(AVFormatContext *ac[], int n_files, char *buf, int size); - -/** - * Return a positive value if the given filename has one of the given - * extensions, 0 otherwise. - * - * @param filename file name to check against the given extensions - * @param extensions a comma-separated list of filename extensions - */ -int av_match_ext(const char *filename, const char *extensions); - -/** - * Test if the given container can store a codec. - * - * @param ofmt container to check for compatibility - * @param codec_id codec to potentially store in container - * @param std_compliance standards compliance level, one of FF_COMPLIANCE_* - * - * @return 1 if codec with ID codec_id can be stored in ofmt, 0 if it cannot. - * A negative number if this information is not available. - */ -int avformat_query_codec(const AVOutputFormat *ofmt, enum AVCodecID codec_id, - int std_compliance); - -/** - * @defgroup riff_fourcc RIFF FourCCs - * @{ - * Get the tables mapping RIFF FourCCs to libavcodec AVCodecIDs. The tables are - * meant to be passed to av_codec_get_id()/av_codec_get_tag() as in the - * following code: - * @code - * uint32_t tag = MKTAG('H', '2', '6', '4'); - * const struct AVCodecTag *table[] = { avformat_get_riff_video_tags(), 0 }; - * enum AVCodecID id = av_codec_get_id(table, tag); - * @endcode - */ -/** - * @return the table mapping RIFF FourCCs for video to libavcodec AVCodecID. - */ -const struct AVCodecTag *avformat_get_riff_video_tags(void); -/** - * @return the table mapping RIFF FourCCs for audio to AVCodecID. - */ -const struct AVCodecTag *avformat_get_riff_audio_tags(void); -/** - * @return the table mapping MOV FourCCs for video to libavcodec AVCodecID. - */ -const struct AVCodecTag *avformat_get_mov_video_tags(void); -/** - * @return the table mapping MOV FourCCs for audio to AVCodecID. - */ -const struct AVCodecTag *avformat_get_mov_audio_tags(void); - -/** - * @} - */ - -/** - * Guess the sample aspect ratio of a frame, based on both the stream and the - * frame aspect ratio. - * - * Since the frame aspect ratio is set by the codec but the stream aspect ratio - * is set by the demuxer, these two may not be equal. This function tries to - * return the value that you should use if you would like to display the frame. - * - * Basic logic is to use the stream aspect ratio if it is set to something sane - * otherwise use the frame aspect ratio. This way a container setting, which is - * usually easy to modify can override the coded value in the frames. - * - * @param format the format context which the stream is part of - * @param stream the stream which the frame is part of - * @param frame the frame with the aspect ratio to be determined - * @return the guessed (valid) sample_aspect_ratio, 0/1 if no idea - */ -AVRational av_guess_sample_aspect_ratio(AVFormatContext *format, AVStream *stream, AVFrame *frame); - -/** - * Guess the frame rate, based on both the container and codec information. - * - * @param ctx the format context which the stream is part of - * @param stream the stream which the frame is part of - * @param frame the frame for which the frame rate should be determined, may be NULL - * @return the guessed (valid) frame rate, 0/1 if no idea - */ -AVRational av_guess_frame_rate(AVFormatContext *ctx, AVStream *stream, AVFrame *frame); - -/** - * Check if the stream st contained in s is matched by the stream specifier - * spec. - * - * See the "stream specifiers" chapter in the documentation for the syntax - * of spec. - * - * @return >0 if st is matched by spec; - * 0 if st is not matched by spec; - * AVERROR code if spec is invalid - * - * @note A stream specifier can match several streams in the format. - */ -int avformat_match_stream_specifier(AVFormatContext *s, AVStream *st, - const char *spec); - -int avformat_queue_attached_pictures(AVFormatContext *s); - -/** - * Apply a list of bitstream filters to a packet. - * - * @param codec AVCodecContext, usually from an AVStream - * @param pkt the packet to apply filters to - * @param bsfc a NULL-terminated list of filters to apply - * @return >=0 on success; - * AVERROR code on failure - */ -int av_apply_bitstream_filters(AVCodecContext *codec, AVPacket *pkt, - AVBitStreamFilterContext *bsfc); - -/** - * @} - */ - -#endif /* AVFORMAT_AVFORMAT_H */ diff --git a/src/main/jni/ffmpeg/include/libavformat/avio.h b/src/main/jni/ffmpeg/include/libavformat/avio.h deleted file mode 100644 index c3c0b73f2..000000000 --- a/src/main/jni/ffmpeg/include/libavformat/avio.h +++ /dev/null @@ -1,727 +0,0 @@ -/* - * copyright (c) 2001 Fabrice Bellard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ -#ifndef AVFORMAT_AVIO_H -#define AVFORMAT_AVIO_H - -/** - * @file - * @ingroup lavf_io - * Buffered I/O operations - */ - -#include <stdint.h> - -#include "libavutil/common.h" -#include "libavutil/dict.h" -#include "libavutil/log.h" - -#include "libavformat/version.h" - -#define AVIO_SEEKABLE_NORMAL 0x0001 /**< Seeking works like for a local file */ - -/** - * Callback for checking whether to abort blocking functions. - * AVERROR_EXIT is returned in this case by the interrupted - * function. During blocking operations, callback is called with - * opaque as parameter. If the callback returns 1, the - * blocking operation will be aborted. - * - * No members can be added to this struct without a major bump, if - * new elements have been added after this struct in AVFormatContext - * or AVIOContext. - */ -typedef struct AVIOInterruptCB { - int (*callback)(void*); - void *opaque; -} AVIOInterruptCB; - -/** - * Directory entry types. - */ -enum AVIODirEntryType { - AVIO_ENTRY_UNKNOWN, - AVIO_ENTRY_BLOCK_DEVICE, - AVIO_ENTRY_CHARACTER_DEVICE, - AVIO_ENTRY_DIRECTORY, - AVIO_ENTRY_NAMED_PIPE, - AVIO_ENTRY_SYMBOLIC_LINK, - AVIO_ENTRY_SOCKET, - AVIO_ENTRY_FILE, - AVIO_ENTRY_SERVER, - AVIO_ENTRY_SHARE, - AVIO_ENTRY_WORKGROUP, -}; - -/** - * Describes single entry of the directory. - * - * Only name and type fields are guaranteed be set. - * Rest of fields are protocol or/and platform dependent and might be unknown. - */ -typedef struct AVIODirEntry { - char *name; /**< Filename */ - int type; /**< Type of the entry */ - int utf8; /**< Set to 1 when name is encoded with UTF-8, 0 otherwise. - Name can be encoded with UTF-8 even though 0 is set. */ - int64_t size; /**< File size in bytes, -1 if unknown. */ - int64_t modification_timestamp; /**< Time of last modification in microseconds since unix - epoch, -1 if unknown. */ - int64_t access_timestamp; /**< Time of last access in microseconds since unix epoch, - -1 if unknown. */ - int64_t status_change_timestamp; /**< Time of last status change in microseconds since unix - epoch, -1 if unknown. */ - int64_t user_id; /**< User ID of owner, -1 if unknown. */ - int64_t group_id; /**< Group ID of owner, -1 if unknown. */ - int64_t filemode; /**< Unix file mode, -1 if unknown. */ -} AVIODirEntry; - -typedef struct AVIODirContext { - struct URLContext *url_context; -} AVIODirContext; - -/** - * Bytestream IO Context. - * New fields can be added to the end with minor version bumps. - * Removal, reordering and changes to existing fields require a major - * version bump. - * sizeof(AVIOContext) must not be used outside libav*. - * - * @note None of the function pointers in AVIOContext should be called - * directly, they should only be set by the client application - * when implementing custom I/O. Normally these are set to the - * function pointers specified in avio_alloc_context() - */ -typedef struct AVIOContext { - /** - * A class for private options. - * - * If this AVIOContext is created by avio_open2(), av_class is set and - * passes the options down to protocols. - * - * If this AVIOContext is manually allocated, then av_class may be set by - * the caller. - * - * warning -- this field can be NULL, be sure to not pass this AVIOContext - * to any av_opt_* functions in that case. - */ - const AVClass *av_class; - - /* - * The following shows the relationship between buffer, buf_ptr, buf_end, buf_size, - * and pos, when reading and when writing (since AVIOContext is used for both): - * - ********************************************************************************** - * READING - ********************************************************************************** - * - * | buffer_size | - * |---------------------------------------| - * | | - * - * buffer buf_ptr buf_end - * +---------------+-----------------------+ - * |/ / / / / / / /|/ / / / / / /| | - * read buffer: |/ / consumed / | to be read /| | - * |/ / / / / / / /|/ / / / / / /| | - * +---------------+-----------------------+ - * - * pos - * +-------------------------------------------+-----------------+ - * input file: | | | - * +-------------------------------------------+-----------------+ - * - * - ********************************************************************************** - * WRITING - ********************************************************************************** - * - * | buffer_size | - * |-------------------------------| - * | | - * - * buffer buf_ptr buf_end - * +-------------------+-----------+ - * |/ / / / / / / / / /| | - * write buffer: | / to be flushed / | | - * |/ / / / / / / / / /| | - * +-------------------+-----------+ - * - * pos - * +--------------------------+-----------------------------------+ - * output file: | | | - * +--------------------------+-----------------------------------+ - * - */ - unsigned char *buffer; /**< Start of the buffer. */ - int buffer_size; /**< Maximum buffer size */ - unsigned char *buf_ptr; /**< Current position in the buffer */ - unsigned char *buf_end; /**< End of the data, may be less than - buffer+buffer_size if the read function returned - less data than requested, e.g. for streams where - no more data has been received yet. */ - void *opaque; /**< A private pointer, passed to the read/write/seek/... - functions. */ - int (*read_packet)(void *opaque, uint8_t *buf, int buf_size); - int (*write_packet)(void *opaque, uint8_t *buf, int buf_size); - int64_t (*seek)(void *opaque, int64_t offset, int whence); - int64_t pos; /**< position in the file of the current buffer */ - int must_flush; /**< true if the next seek should flush */ - int eof_reached; /**< true if eof reached */ - int write_flag; /**< true if open for writing */ - int max_packet_size; - unsigned long checksum; - unsigned char *checksum_ptr; - unsigned long (*update_checksum)(unsigned long checksum, const uint8_t *buf, unsigned int size); - int error; /**< contains the error code or 0 if no error happened */ - /** - * Pause or resume playback for network streaming protocols - e.g. MMS. - */ - int (*read_pause)(void *opaque, int pause); - /** - * Seek to a given timestamp in stream with the specified stream_index. - * Needed for some network streaming protocols which don't support seeking - * to byte position. - */ - int64_t (*read_seek)(void *opaque, int stream_index, - int64_t timestamp, int flags); - /** - * A combination of AVIO_SEEKABLE_ flags or 0 when the stream is not seekable. - */ - int seekable; - - /** - * max filesize, used to limit allocations - * This field is internal to libavformat and access from outside is not allowed. - */ - int64_t maxsize; - - /** - * avio_read and avio_write should if possible be satisfied directly - * instead of going through a buffer, and avio_seek will always - * call the underlying seek function directly. - */ - int direct; - - /** - * Bytes read statistic - * This field is internal to libavformat and access from outside is not allowed. - */ - int64_t bytes_read; - - /** - * seek statistic - * This field is internal to libavformat and access from outside is not allowed. - */ - int seek_count; - - /** - * writeout statistic - * This field is internal to libavformat and access from outside is not allowed. - */ - int writeout_count; - - /** - * Original buffer size - * used internally after probing and ensure seekback to reset the buffer size - * This field is internal to libavformat and access from outside is not allowed. - */ - int orig_buffer_size; - - /** - * Threshold to favor readahead over seek. - * This is current internal only, do not use from outside. - */ - int short_seek_threshold; -} AVIOContext; - -/* unbuffered I/O */ - -/** - * Return the name of the protocol that will handle the passed URL. - * - * NULL is returned if no protocol could be found for the given URL. - * - * @return Name of the protocol or NULL. - */ -const char *avio_find_protocol_name(const char *url); - -/** - * Return AVIO_FLAG_* access flags corresponding to the access permissions - * of the resource in url, or a negative value corresponding to an - * AVERROR code in case of failure. The returned access flags are - * masked by the value in flags. - * - * @note This function is intrinsically unsafe, in the sense that the - * checked resource may change its existence or permission status from - * one call to another. Thus you should not trust the returned value, - * unless you are sure that no other processes are accessing the - * checked resource. - */ -int avio_check(const char *url, int flags); - -/** - * Move or rename a resource. - * - * @note url_src and url_dst should share the same protocol and authority. - * - * @param url_src url to resource to be moved - * @param url_dst new url to resource if the operation succeeded - * @return >=0 on success or negative on error. - */ -int avpriv_io_move(const char *url_src, const char *url_dst); - -/** - * Delete a resource. - * - * @param url resource to be deleted. - * @return >=0 on success or negative on error. - */ -int avpriv_io_delete(const char *url); - -/** - * Open directory for reading. - * - * @param s directory read context. Pointer to a NULL pointer must be passed. - * @param url directory to be listed. - * @param options A dictionary filled with protocol-private options. On return - * this parameter will be destroyed and replaced with a dictionary - * containing options that were not found. May be NULL. - * @return >=0 on success or negative on error. - */ -int avio_open_dir(AVIODirContext **s, const char *url, AVDictionary **options); - -/** - * Get next directory entry. - * - * Returned entry must be freed with avio_free_directory_entry(). In particular - * it may outlive AVIODirContext. - * - * @param s directory read context. - * @param[out] next next entry or NULL when no more entries. - * @return >=0 on success or negative on error. End of list is not considered an - * error. - */ -int avio_read_dir(AVIODirContext *s, AVIODirEntry **next); - -/** - * Close directory. - * - * @note Entries created using avio_read_dir() are not deleted and must be - * freeded with avio_free_directory_entry(). - * - * @param s directory read context. - * @return >=0 on success or negative on error. - */ -int avio_close_dir(AVIODirContext **s); - -/** - * Free entry allocated by avio_read_dir(). - * - * @param entry entry to be freed. - */ -void avio_free_directory_entry(AVIODirEntry **entry); - -/** - * Allocate and initialize an AVIOContext for buffered I/O. It must be later - * freed with av_free(). - * - * @param buffer Memory block for input/output operations via AVIOContext. - * The buffer must be allocated with av_malloc() and friends. - * It may be freed and replaced with a new buffer by libavformat. - * AVIOContext.buffer holds the buffer currently in use, - * which must be later freed with av_free(). - * @param buffer_size The buffer size is very important for performance. - * For protocols with fixed blocksize it should be set to this blocksize. - * For others a typical size is a cache page, e.g. 4kb. - * @param write_flag Set to 1 if the buffer should be writable, 0 otherwise. - * @param opaque An opaque pointer to user-specific data. - * @param read_packet A function for refilling the buffer, may be NULL. - * @param write_packet A function for writing the buffer contents, may be NULL. - * The function may not change the input buffers content. - * @param seek A function for seeking to specified byte position, may be NULL. - * - * @return Allocated AVIOContext or NULL on failure. - */ -AVIOContext *avio_alloc_context( - unsigned char *buffer, - int buffer_size, - int write_flag, - void *opaque, - int (*read_packet)(void *opaque, uint8_t *buf, int buf_size), - int (*write_packet)(void *opaque, uint8_t *buf, int buf_size), - int64_t (*seek)(void *opaque, int64_t offset, int whence)); - -void avio_w8(AVIOContext *s, int b); -void avio_write(AVIOContext *s, const unsigned char *buf, int size); -void avio_wl64(AVIOContext *s, uint64_t val); -void avio_wb64(AVIOContext *s, uint64_t val); -void avio_wl32(AVIOContext *s, unsigned int val); -void avio_wb32(AVIOContext *s, unsigned int val); -void avio_wl24(AVIOContext *s, unsigned int val); -void avio_wb24(AVIOContext *s, unsigned int val); -void avio_wl16(AVIOContext *s, unsigned int val); -void avio_wb16(AVIOContext *s, unsigned int val); - -/** - * Write a NULL-terminated string. - * @return number of bytes written. - */ -int avio_put_str(AVIOContext *s, const char *str); - -/** - * Convert an UTF-8 string to UTF-16LE and write it. - * @param s the AVIOContext - * @param str NULL-terminated UTF-8 string - * - * @return number of bytes written. - */ -int avio_put_str16le(AVIOContext *s, const char *str); - -/** - * Convert an UTF-8 string to UTF-16BE and write it. - * @param s the AVIOContext - * @param str NULL-terminated UTF-8 string - * - * @return number of bytes written. - */ -int avio_put_str16be(AVIOContext *s, const char *str); - -/** - * Passing this as the "whence" parameter to a seek function causes it to - * return the filesize without seeking anywhere. Supporting this is optional. - * If it is not supported then the seek function will return <0. - */ -#define AVSEEK_SIZE 0x10000 - -/** - * Oring this flag as into the "whence" parameter to a seek function causes it to - * seek by any means (like reopening and linear reading) or other normally unreasonable - * means that can be extremely slow. - * This may be ignored by the seek code. - */ -#define AVSEEK_FORCE 0x20000 - -/** - * fseek() equivalent for AVIOContext. - * @return new position or AVERROR. - */ -int64_t avio_seek(AVIOContext *s, int64_t offset, int whence); - -/** - * Skip given number of bytes forward - * @return new position or AVERROR. - */ -int64_t avio_skip(AVIOContext *s, int64_t offset); - -/** - * ftell() equivalent for AVIOContext. - * @return position or AVERROR. - */ -static av_always_inline int64_t avio_tell(AVIOContext *s) -{ - return avio_seek(s, 0, SEEK_CUR); -} - -/** - * Get the filesize. - * @return filesize or AVERROR - */ -int64_t avio_size(AVIOContext *s); - -/** - * feof() equivalent for AVIOContext. - * @return non zero if and only if end of file - */ -int avio_feof(AVIOContext *s); -#if FF_API_URL_FEOF -/** - * @deprecated use avio_feof() - */ -attribute_deprecated -int url_feof(AVIOContext *s); -#endif - -/** @warning Writes up to 4 KiB per call */ -int avio_printf(AVIOContext *s, const char *fmt, ...) av_printf_format(2, 3); - -/** - * Force flushing of buffered data. - * - * For write streams, force the buffered data to be immediately written to the output, - * without to wait to fill the internal buffer. - * - * For read streams, discard all currently buffered data, and advance the - * reported file position to that of the underlying stream. This does not - * read new data, and does not perform any seeks. - */ -void avio_flush(AVIOContext *s); - -/** - * Read size bytes from AVIOContext into buf. - * @return number of bytes read or AVERROR - */ -int avio_read(AVIOContext *s, unsigned char *buf, int size); - -/** - * @name Functions for reading from AVIOContext - * @{ - * - * @note return 0 if EOF, so you cannot use it if EOF handling is - * necessary - */ -int avio_r8 (AVIOContext *s); -unsigned int avio_rl16(AVIOContext *s); -unsigned int avio_rl24(AVIOContext *s); -unsigned int avio_rl32(AVIOContext *s); -uint64_t avio_rl64(AVIOContext *s); -unsigned int avio_rb16(AVIOContext *s); -unsigned int avio_rb24(AVIOContext *s); -unsigned int avio_rb32(AVIOContext *s); -uint64_t avio_rb64(AVIOContext *s); -/** - * @} - */ - -/** - * Read a string from pb into buf. The reading will terminate when either - * a NULL character was encountered, maxlen bytes have been read, or nothing - * more can be read from pb. The result is guaranteed to be NULL-terminated, it - * will be truncated if buf is too small. - * Note that the string is not interpreted or validated in any way, it - * might get truncated in the middle of a sequence for multi-byte encodings. - * - * @return number of bytes read (is always <= maxlen). - * If reading ends on EOF or error, the return value will be one more than - * bytes actually read. - */ -int avio_get_str(AVIOContext *pb, int maxlen, char *buf, int buflen); - -/** - * Read a UTF-16 string from pb and convert it to UTF-8. - * The reading will terminate when either a null or invalid character was - * encountered or maxlen bytes have been read. - * @return number of bytes read (is always <= maxlen) - */ -int avio_get_str16le(AVIOContext *pb, int maxlen, char *buf, int buflen); -int avio_get_str16be(AVIOContext *pb, int maxlen, char *buf, int buflen); - - -/** - * @name URL open modes - * The flags argument to avio_open must be one of the following - * constants, optionally ORed with other flags. - * @{ - */ -#define AVIO_FLAG_READ 1 /**< read-only */ -#define AVIO_FLAG_WRITE 2 /**< write-only */ -#define AVIO_FLAG_READ_WRITE (AVIO_FLAG_READ|AVIO_FLAG_WRITE) /**< read-write pseudo flag */ -/** - * @} - */ - -/** - * Use non-blocking mode. - * If this flag is set, operations on the context will return - * AVERROR(EAGAIN) if they can not be performed immediately. - * If this flag is not set, operations on the context will never return - * AVERROR(EAGAIN). - * Note that this flag does not affect the opening/connecting of the - * context. Connecting a protocol will always block if necessary (e.g. on - * network protocols) but never hang (e.g. on busy devices). - * Warning: non-blocking protocols is work-in-progress; this flag may be - * silently ignored. - */ -#define AVIO_FLAG_NONBLOCK 8 - -/** - * Use direct mode. - * avio_read and avio_write should if possible be satisfied directly - * instead of going through a buffer, and avio_seek will always - * call the underlying seek function directly. - */ -#define AVIO_FLAG_DIRECT 0x8000 - -/** - * Create and initialize a AVIOContext for accessing the - * resource indicated by url. - * @note When the resource indicated by url has been opened in - * read+write mode, the AVIOContext can be used only for writing. - * - * @param s Used to return the pointer to the created AVIOContext. - * In case of failure the pointed to value is set to NULL. - * @param url resource to access - * @param flags flags which control how the resource indicated by url - * is to be opened - * @return >= 0 in case of success, a negative value corresponding to an - * AVERROR code in case of failure - */ -int avio_open(AVIOContext **s, const char *url, int flags); - -/** - * Create and initialize a AVIOContext for accessing the - * resource indicated by url. - * @note When the resource indicated by url has been opened in - * read+write mode, the AVIOContext can be used only for writing. - * - * @param s Used to return the pointer to the created AVIOContext. - * In case of failure the pointed to value is set to NULL. - * @param url resource to access - * @param flags flags which control how the resource indicated by url - * is to be opened - * @param int_cb an interrupt callback to be used at the protocols level - * @param options A dictionary filled with protocol-private options. On return - * this parameter will be destroyed and replaced with a dict containing options - * that were not found. May be NULL. - * @return >= 0 in case of success, a negative value corresponding to an - * AVERROR code in case of failure - */ -int avio_open2(AVIOContext **s, const char *url, int flags, - const AVIOInterruptCB *int_cb, AVDictionary **options); - -/** - * Close the resource accessed by the AVIOContext s and free it. - * This function can only be used if s was opened by avio_open(). - * - * The internal buffer is automatically flushed before closing the - * resource. - * - * @return 0 on success, an AVERROR < 0 on error. - * @see avio_closep - */ -int avio_close(AVIOContext *s); - -/** - * Close the resource accessed by the AVIOContext *s, free it - * and set the pointer pointing to it to NULL. - * This function can only be used if s was opened by avio_open(). - * - * The internal buffer is automatically flushed before closing the - * resource. - * - * @return 0 on success, an AVERROR < 0 on error. - * @see avio_close - */ -int avio_closep(AVIOContext **s); - - -/** - * Open a write only memory stream. - * - * @param s new IO context - * @return zero if no error. - */ -int avio_open_dyn_buf(AVIOContext **s); - -/** - * Return the written size and a pointer to the buffer. The buffer - * must be freed with av_free(). - * Padding of AV_INPUT_BUFFER_PADDING_SIZE is added to the buffer. - * - * @param s IO context - * @param pbuffer pointer to a byte buffer - * @return the length of the byte buffer - */ -int avio_close_dyn_buf(AVIOContext *s, uint8_t **pbuffer); - -/** - * Iterate through names of available protocols. - * - * @param opaque A private pointer representing current protocol. - * It must be a pointer to NULL on first iteration and will - * be updated by successive calls to avio_enum_protocols. - * @param output If set to 1, iterate over output protocols, - * otherwise over input protocols. - * - * @return A static string containing the name of current protocol or NULL - */ -const char *avio_enum_protocols(void **opaque, int output); - -/** - * Pause and resume playing - only meaningful if using a network streaming - * protocol (e.g. MMS). - * - * @param h IO context from which to call the read_pause function pointer - * @param pause 1 for pause, 0 for resume - */ -int avio_pause(AVIOContext *h, int pause); - -/** - * Seek to a given timestamp relative to some component stream. - * Only meaningful if using a network streaming protocol (e.g. MMS.). - * - * @param h IO context from which to call the seek function pointers - * @param stream_index The stream index that the timestamp is relative to. - * If stream_index is (-1) the timestamp should be in AV_TIME_BASE - * units from the beginning of the presentation. - * If a stream_index >= 0 is used and the protocol does not support - * seeking based on component streams, the call will fail. - * @param timestamp timestamp in AVStream.time_base units - * or if there is no stream specified then in AV_TIME_BASE units. - * @param flags Optional combination of AVSEEK_FLAG_BACKWARD, AVSEEK_FLAG_BYTE - * and AVSEEK_FLAG_ANY. The protocol may silently ignore - * AVSEEK_FLAG_BACKWARD and AVSEEK_FLAG_ANY, but AVSEEK_FLAG_BYTE will - * fail if used and not supported. - * @return >= 0 on success - * @see AVInputFormat::read_seek - */ -int64_t avio_seek_time(AVIOContext *h, int stream_index, - int64_t timestamp, int flags); - -/* Avoid a warning. The header can not be included because it breaks c++. */ -struct AVBPrint; - -/** - * Read contents of h into print buffer, up to max_size bytes, or up to EOF. - * - * @return 0 for success (max_size bytes read or EOF reached), negative error - * code otherwise - */ -int avio_read_to_bprint(AVIOContext *h, struct AVBPrint *pb, size_t max_size); - -/** - * Accept and allocate a client context on a server context. - * @param s the server context - * @param c the client context, must be unallocated - * @return >= 0 on success or a negative value corresponding - * to an AVERROR on failure - */ -int avio_accept(AVIOContext *s, AVIOContext **c); - -/** - * Perform one step of the protocol handshake to accept a new client. - * This function must be called on a client returned by avio_accept() before - * using it as a read/write context. - * It is separate from avio_accept() because it may block. - * A step of the handshake is defined by places where the application may - * decide to change the proceedings. - * For example, on a protocol with a request header and a reply header, each - * one can constitute a step because the application may use the parameters - * from the request to change parameters in the reply; or each individual - * chunk of the request can constitute a step. - * If the handshake is already finished, avio_handshake() does nothing and - * returns 0 immediately. - * - * @param c the client context to perform the handshake on - * @return 0 on a complete and successful handshake - * > 0 if the handshake progressed, but is not complete - * < 0 for an AVERROR code - */ -int avio_handshake(AVIOContext *c); -#endif /* AVFORMAT_AVIO_H */ diff --git a/src/main/jni/ffmpeg/include/libavformat/version.h b/src/main/jni/ffmpeg/include/libavformat/version.h deleted file mode 100644 index a57eb12fd..000000000 --- a/src/main/jni/ffmpeg/include/libavformat/version.h +++ /dev/null @@ -1,78 +0,0 @@ -/* - * Version macros. - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVFORMAT_VERSION_H -#define AVFORMAT_VERSION_H - -/** - * @file - * @ingroup libavf - * Libavformat version macros - */ - -#include "libavutil/version.h" - -#define LIBAVFORMAT_VERSION_MAJOR 57 -#define LIBAVFORMAT_VERSION_MINOR 21 -#define LIBAVFORMAT_VERSION_MICRO 100 - -#define LIBAVFORMAT_VERSION_INT AV_VERSION_INT(LIBAVFORMAT_VERSION_MAJOR, \ - LIBAVFORMAT_VERSION_MINOR, \ - LIBAVFORMAT_VERSION_MICRO) -#define LIBAVFORMAT_VERSION AV_VERSION(LIBAVFORMAT_VERSION_MAJOR, \ - LIBAVFORMAT_VERSION_MINOR, \ - LIBAVFORMAT_VERSION_MICRO) -#define LIBAVFORMAT_BUILD LIBAVFORMAT_VERSION_INT - -#define LIBAVFORMAT_IDENT "Lavf" AV_STRINGIFY(LIBAVFORMAT_VERSION) - -/** - * FF_API_* defines may be placed below to indicate public API that will be - * dropped at a future version bump. The defines themselves are not part of - * the public API and may change, break or disappear at any time. - * - * @note, when bumping the major version it is recommended to manually - * disable each FF_API_* in its own commit instead of disabling them all - * at once through the bump. This improves the git bisect-ability of the change. - * - */ -#ifndef FF_API_LAVF_BITEXACT -#define FF_API_LAVF_BITEXACT (LIBAVFORMAT_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_LAVF_FRAC -#define FF_API_LAVF_FRAC (LIBAVFORMAT_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_LAVF_CODEC_TB -#define FF_API_LAVF_CODEC_TB (LIBAVFORMAT_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_URL_FEOF -#define FF_API_URL_FEOF (LIBAVFORMAT_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_LAVF_FMT_RAWPICTURE -#define FF_API_LAVF_FMT_RAWPICTURE (LIBAVFORMAT_VERSION_MAJOR < 58) -#endif -#ifndef FF_API_COMPUTE_PKT_FIELDS2 -#define FF_API_COMPUTE_PKT_FIELDS2 (LIBAVFORMAT_VERSION_MAJOR < 58) -#endif - -#ifndef FF_API_R_FRAME_RATE -#define FF_API_R_FRAME_RATE 1 -#endif -#endif /* AVFORMAT_VERSION_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/adler32.h b/src/main/jni/ffmpeg/include/libavutil/adler32.h deleted file mode 100644 index 0dc69ec0a..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/adler32.h +++ /dev/null @@ -1,55 +0,0 @@ -/* - * copyright (c) 2006 Mans Rullgard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_ADLER32_H -#define AVUTIL_ADLER32_H - -#include <stdint.h> -#include "attributes.h" - -/** - * @file - * Public header for libavutil Adler32 hasher - * - * @defgroup lavu_adler32 Adler32 - * @ingroup lavu_crypto - * @{ - */ - -/** - * Calculate the Adler32 checksum of a buffer. - * - * Passing the return value to a subsequent av_adler32_update() call - * allows the checksum of multiple buffers to be calculated as though - * they were concatenated. - * - * @param adler initial checksum value - * @param buf pointer to input buffer - * @param len size of input buffer - * @return updated checksum - */ -unsigned long av_adler32_update(unsigned long adler, const uint8_t *buf, - unsigned int len) av_pure; - -/** - * @} - */ - -#endif /* AVUTIL_ADLER32_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/aes.h b/src/main/jni/ffmpeg/include/libavutil/aes.h deleted file mode 100644 index 09efbda10..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/aes.h +++ /dev/null @@ -1,65 +0,0 @@ -/* - * copyright (c) 2007 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_AES_H -#define AVUTIL_AES_H - -#include <stdint.h> - -#include "attributes.h" -#include "version.h" - -/** - * @defgroup lavu_aes AES - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_aes_size; - -struct AVAES; - -/** - * Allocate an AVAES context. - */ -struct AVAES *av_aes_alloc(void); - -/** - * Initialize an AVAES context. - * @param key_bits 128, 192 or 256 - * @param decrypt 0 for encryption, 1 for decryption - */ -int av_aes_init(struct AVAES *a, const uint8_t *key, int key_bits, int decrypt); - -/** - * Encrypt or decrypt a buffer using a previously initialized context. - * @param count number of 16 byte blocks - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param iv initialization vector for CBC mode, if NULL then ECB will be used - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_aes_crypt(struct AVAES *a, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt); - -/** - * @} - */ - -#endif /* AVUTIL_AES_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/aes_ctr.h b/src/main/jni/ffmpeg/include/libavutil/aes_ctr.h deleted file mode 100644 index f596fa6a4..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/aes_ctr.h +++ /dev/null @@ -1,83 +0,0 @@ -/* - * AES-CTR cipher - * Copyright (c) 2015 Eran Kornblau <erankor at gmail dot com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_AES_CTR_H -#define AVUTIL_AES_CTR_H - -#include <stdint.h> - -#include "attributes.h" -#include "version.h" - -#define AES_CTR_KEY_SIZE (16) -#define AES_CTR_IV_SIZE (8) - -struct AVAESCTR; - -/** - * Allocate an AVAESCTR context. - */ -struct AVAESCTR *av_aes_ctr_alloc(void); - -/** - * Initialize an AVAESCTR context. - * @param key encryption key, must have a length of AES_CTR_KEY_SIZE - */ -int av_aes_ctr_init(struct AVAESCTR *a, const uint8_t *key); - -/** - * Release an AVAESCTR context. - */ -void av_aes_ctr_free(struct AVAESCTR *a); - -/** - * Process a buffer using a previously initialized context. - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param size the size of src and dst - */ -void av_aes_ctr_crypt(struct AVAESCTR *a, uint8_t *dst, const uint8_t *src, int size); - -/** - * Get the current iv - */ -const uint8_t* av_aes_ctr_get_iv(struct AVAESCTR *a); - -/** - * Generate a random iv - */ -void av_aes_ctr_set_random_iv(struct AVAESCTR *a); - -/** - * Forcefully change the iv - */ -void av_aes_ctr_set_iv(struct AVAESCTR *a, const uint8_t* iv); - -/** - * Increment the top 64 bit of the iv (performed after each frame) - */ -void av_aes_ctr_increment_iv(struct AVAESCTR *a); - -/** - * @} - */ - -#endif /* AVUTIL_AES_CTR_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/attributes.h b/src/main/jni/ffmpeg/include/libavutil/attributes.h deleted file mode 100644 index 5c6b9deec..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/attributes.h +++ /dev/null @@ -1,168 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * Macro definitions for various function/variable attributes - */ - -#ifndef AVUTIL_ATTRIBUTES_H -#define AVUTIL_ATTRIBUTES_H - -#ifdef __GNUC__ -# define AV_GCC_VERSION_AT_LEAST(x,y) (__GNUC__ > (x) || __GNUC__ == (x) && __GNUC_MINOR__ >= (y)) -# define AV_GCC_VERSION_AT_MOST(x,y) (__GNUC__ < (x) || __GNUC__ == (x) && __GNUC_MINOR__ <= (y)) -#else -# define AV_GCC_VERSION_AT_LEAST(x,y) 0 -# define AV_GCC_VERSION_AT_MOST(x,y) 0 -#endif - -#ifndef av_always_inline -#if AV_GCC_VERSION_AT_LEAST(3,1) -# define av_always_inline __attribute__((always_inline)) inline -#elif defined(_MSC_VER) -# define av_always_inline __forceinline -#else -# define av_always_inline inline -#endif -#endif - -#ifndef av_extern_inline -#if defined(__ICL) && __ICL >= 1210 || defined(__GNUC_STDC_INLINE__) -# define av_extern_inline extern inline -#else -# define av_extern_inline inline -#endif -#endif - -#if AV_GCC_VERSION_AT_LEAST(3,4) -# define av_warn_unused_result __attribute__((warn_unused_result)) -#else -# define av_warn_unused_result -#endif - -#if AV_GCC_VERSION_AT_LEAST(3,1) -# define av_noinline __attribute__((noinline)) -#elif defined(_MSC_VER) -# define av_noinline __declspec(noinline) -#else -# define av_noinline -#endif - -#if AV_GCC_VERSION_AT_LEAST(3,1) -# define av_pure __attribute__((pure)) -#else -# define av_pure -#endif - -#if AV_GCC_VERSION_AT_LEAST(2,6) -# define av_const __attribute__((const)) -#else -# define av_const -#endif - -#if AV_GCC_VERSION_AT_LEAST(4,3) -# define av_cold __attribute__((cold)) -#else -# define av_cold -#endif - -#if AV_GCC_VERSION_AT_LEAST(4,1) && !defined(__llvm__) -# define av_flatten __attribute__((flatten)) -#else -# define av_flatten -#endif - -#if AV_GCC_VERSION_AT_LEAST(3,1) -# define attribute_deprecated __attribute__((deprecated)) -#elif defined(_MSC_VER) -# define attribute_deprecated __declspec(deprecated) -#else -# define attribute_deprecated -#endif - -/** - * Disable warnings about deprecated features - * This is useful for sections of code kept for backward compatibility and - * scheduled for removal. - */ -#ifndef AV_NOWARN_DEPRECATED -#if AV_GCC_VERSION_AT_LEAST(4,6) -# define AV_NOWARN_DEPRECATED(code) \ - _Pragma("GCC diagnostic push") \ - _Pragma("GCC diagnostic ignored \"-Wdeprecated-declarations\"") \ - code \ - _Pragma("GCC diagnostic pop") -#elif defined(_MSC_VER) -# define AV_NOWARN_DEPRECATED(code) \ - __pragma(warning(push)) \ - __pragma(warning(disable : 4996)) \ - code; \ - __pragma(warning(pop)) -#else -# define AV_NOWARN_DEPRECATED(code) code -#endif -#endif - - -#if defined(__GNUC__) -# define av_unused __attribute__((unused)) -#else -# define av_unused -#endif - -/** - * Mark a variable as used and prevent the compiler from optimizing it - * away. This is useful for variables accessed only from inline - * assembler without the compiler being aware. - */ -#if AV_GCC_VERSION_AT_LEAST(3,1) -# define av_used __attribute__((used)) -#else -# define av_used -#endif - -#if AV_GCC_VERSION_AT_LEAST(3,3) -# define av_alias __attribute__((may_alias)) -#else -# define av_alias -#endif - -#if defined(__GNUC__) && !defined(__INTEL_COMPILER) && !defined(__clang__) -# define av_uninit(x) x=x -#else -# define av_uninit(x) x -#endif - -#ifdef __GNUC__ -# define av_builtin_constant_p __builtin_constant_p -# define av_printf_format(fmtpos, attrpos) __attribute__((__format__(__printf__, fmtpos, attrpos))) -#else -# define av_builtin_constant_p(x) 0 -# define av_printf_format(fmtpos, attrpos) -#endif - -#if AV_GCC_VERSION_AT_LEAST(2,5) -# define av_noreturn __attribute__((noreturn)) -#else -# define av_noreturn -#endif - -#endif /* AVUTIL_ATTRIBUTES_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/audio_fifo.h b/src/main/jni/ffmpeg/include/libavutil/audio_fifo.h deleted file mode 100644 index 24f91dab7..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/audio_fifo.h +++ /dev/null @@ -1,170 +0,0 @@ -/* - * Audio FIFO - * Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * Audio FIFO Buffer - */ - -#ifndef AVUTIL_AUDIO_FIFO_H -#define AVUTIL_AUDIO_FIFO_H - -#include "avutil.h" -#include "fifo.h" -#include "samplefmt.h" - -/** - * @addtogroup lavu_audio - * @{ - * - * @defgroup lavu_audiofifo Audio FIFO Buffer - * @{ - */ - -/** - * Context for an Audio FIFO Buffer. - * - * - Operates at the sample level rather than the byte level. - * - Supports multiple channels with either planar or packed sample format. - * - Automatic reallocation when writing to a full buffer. - */ -typedef struct AVAudioFifo AVAudioFifo; - -/** - * Free an AVAudioFifo. - * - * @param af AVAudioFifo to free - */ -void av_audio_fifo_free(AVAudioFifo *af); - -/** - * Allocate an AVAudioFifo. - * - * @param sample_fmt sample format - * @param channels number of channels - * @param nb_samples initial allocation size, in samples - * @return newly allocated AVAudioFifo, or NULL on error - */ -AVAudioFifo *av_audio_fifo_alloc(enum AVSampleFormat sample_fmt, int channels, - int nb_samples); - -/** - * Reallocate an AVAudioFifo. - * - * @param af AVAudioFifo to reallocate - * @param nb_samples new allocation size, in samples - * @return 0 if OK, or negative AVERROR code on failure - */ -av_warn_unused_result -int av_audio_fifo_realloc(AVAudioFifo *af, int nb_samples); - -/** - * Write data to an AVAudioFifo. - * - * The AVAudioFifo will be reallocated automatically if the available space - * is less than nb_samples. - * - * @see enum AVSampleFormat - * The documentation for AVSampleFormat describes the data layout. - * - * @param af AVAudioFifo to write to - * @param data audio data plane pointers - * @param nb_samples number of samples to write - * @return number of samples actually written, or negative AVERROR - * code on failure. If successful, the number of samples - * actually written will always be nb_samples. - */ -int av_audio_fifo_write(AVAudioFifo *af, void **data, int nb_samples); - -/** - * Peek data from an AVAudioFifo. - * - * @see enum AVSampleFormat - * The documentation for AVSampleFormat describes the data layout. - * - * @param af AVAudioFifo to read from - * @param data audio data plane pointers - * @param nb_samples number of samples to peek - * @return number of samples actually peek, or negative AVERROR code - * on failure. The number of samples actually peek will not - * be greater than nb_samples, and will only be less than - * nb_samples if av_audio_fifo_size is less than nb_samples. - */ -int av_audio_fifo_peek(AVAudioFifo *af, void **data, int nb_samples); - -/** - * Read data from an AVAudioFifo. - * - * @see enum AVSampleFormat - * The documentation for AVSampleFormat describes the data layout. - * - * @param af AVAudioFifo to read from - * @param data audio data plane pointers - * @param nb_samples number of samples to read - * @return number of samples actually read, or negative AVERROR code - * on failure. The number of samples actually read will not - * be greater than nb_samples, and will only be less than - * nb_samples if av_audio_fifo_size is less than nb_samples. - */ -int av_audio_fifo_read(AVAudioFifo *af, void **data, int nb_samples); - -/** - * Drain data from an AVAudioFifo. - * - * Removes the data without reading it. - * - * @param af AVAudioFifo to drain - * @param nb_samples number of samples to drain - * @return 0 if OK, or negative AVERROR code on failure - */ -int av_audio_fifo_drain(AVAudioFifo *af, int nb_samples); - -/** - * Reset the AVAudioFifo buffer. - * - * This empties all data in the buffer. - * - * @param af AVAudioFifo to reset - */ -void av_audio_fifo_reset(AVAudioFifo *af); - -/** - * Get the current number of samples in the AVAudioFifo available for reading. - * - * @param af the AVAudioFifo to query - * @return number of samples available for reading - */ -int av_audio_fifo_size(AVAudioFifo *af); - -/** - * Get the current number of samples in the AVAudioFifo available for writing. - * - * @param af the AVAudioFifo to query - * @return number of samples available for writing - */ -int av_audio_fifo_space(AVAudioFifo *af); - -/** - * @} - * @} - */ - -#endif /* AVUTIL_AUDIO_FIFO_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/avassert.h b/src/main/jni/ffmpeg/include/libavutil/avassert.h deleted file mode 100644 index 41f5e0eea..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/avassert.h +++ /dev/null @@ -1,66 +0,0 @@ -/* - * copyright (c) 2010 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * simple assert() macros that are a bit more flexible than ISO C assert(). - * @author Michael Niedermayer <michaelni@gmx.at> - */ - -#ifndef AVUTIL_AVASSERT_H -#define AVUTIL_AVASSERT_H - -#include <stdlib.h> -#include "avutil.h" -#include "log.h" - -/** - * assert() equivalent, that is always enabled. - */ -#define av_assert0(cond) do { \ - if (!(cond)) { \ - av_log(NULL, AV_LOG_PANIC, "Assertion %s failed at %s:%d\n", \ - AV_STRINGIFY(cond), __FILE__, __LINE__); \ - abort(); \ - } \ -} while (0) - - -/** - * assert() equivalent, that does not lie in speed critical code. - * These asserts() thus can be enabled without fearing speedloss. - */ -#if defined(ASSERT_LEVEL) && ASSERT_LEVEL > 0 -#define av_assert1(cond) av_assert0(cond) -#else -#define av_assert1(cond) ((void)0) -#endif - - -/** - * assert() equivalent, that does lie in speed critical code. - */ -#if defined(ASSERT_LEVEL) && ASSERT_LEVEL > 1 -#define av_assert2(cond) av_assert0(cond) -#else -#define av_assert2(cond) ((void)0) -#endif - -#endif /* AVUTIL_AVASSERT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/avconfig.h b/src/main/jni/ffmpeg/include/libavutil/avconfig.h deleted file mode 100644 index 123b3b28a..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/avconfig.h +++ /dev/null @@ -1,7 +0,0 @@ -/* Generated by ffconf */ -#ifndef AVUTIL_AVCONFIG_H -#define AVUTIL_AVCONFIG_H -#define AV_HAVE_BIGENDIAN 0 -#define AV_HAVE_FAST_UNALIGNED 0 -#define AV_HAVE_INCOMPATIBLE_LIBAV_ABI 0 -#endif /* AVUTIL_AVCONFIG_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/avstring.h b/src/main/jni/ffmpeg/include/libavutil/avstring.h deleted file mode 100644 index a46d01255..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/avstring.h +++ /dev/null @@ -1,397 +0,0 @@ -/* - * Copyright (c) 2007 Mans Rullgard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_AVSTRING_H -#define AVUTIL_AVSTRING_H - -#include <stddef.h> -#include <stdint.h> -#include "attributes.h" - -/** - * @addtogroup lavu_string - * @{ - */ - -/** - * Return non-zero if pfx is a prefix of str. If it is, *ptr is set to - * the address of the first character in str after the prefix. - * - * @param str input string - * @param pfx prefix to test - * @param ptr updated if the prefix is matched inside str - * @return non-zero if the prefix matches, zero otherwise - */ -int av_strstart(const char *str, const char *pfx, const char **ptr); - -/** - * Return non-zero if pfx is a prefix of str independent of case. If - * it is, *ptr is set to the address of the first character in str - * after the prefix. - * - * @param str input string - * @param pfx prefix to test - * @param ptr updated if the prefix is matched inside str - * @return non-zero if the prefix matches, zero otherwise - */ -int av_stristart(const char *str, const char *pfx, const char **ptr); - -/** - * Locate the first case-independent occurrence in the string haystack - * of the string needle. A zero-length string needle is considered to - * match at the start of haystack. - * - * This function is a case-insensitive version of the standard strstr(). - * - * @param haystack string to search in - * @param needle string to search for - * @return pointer to the located match within haystack - * or a null pointer if no match - */ -char *av_stristr(const char *haystack, const char *needle); - -/** - * Locate the first occurrence of the string needle in the string haystack - * where not more than hay_length characters are searched. A zero-length - * string needle is considered to match at the start of haystack. - * - * This function is a length-limited version of the standard strstr(). - * - * @param haystack string to search in - * @param needle string to search for - * @param hay_length length of string to search in - * @return pointer to the located match within haystack - * or a null pointer if no match - */ -char *av_strnstr(const char *haystack, const char *needle, size_t hay_length); - -/** - * Copy the string src to dst, but no more than size - 1 bytes, and - * null-terminate dst. - * - * This function is the same as BSD strlcpy(). - * - * @param dst destination buffer - * @param src source string - * @param size size of destination buffer - * @return the length of src - * - * @warning since the return value is the length of src, src absolutely - * _must_ be a properly 0-terminated string, otherwise this will read beyond - * the end of the buffer and possibly crash. - */ -size_t av_strlcpy(char *dst, const char *src, size_t size); - -/** - * Append the string src to the string dst, but to a total length of - * no more than size - 1 bytes, and null-terminate dst. - * - * This function is similar to BSD strlcat(), but differs when - * size <= strlen(dst). - * - * @param dst destination buffer - * @param src source string - * @param size size of destination buffer - * @return the total length of src and dst - * - * @warning since the return value use the length of src and dst, these - * absolutely _must_ be a properly 0-terminated strings, otherwise this - * will read beyond the end of the buffer and possibly crash. - */ -size_t av_strlcat(char *dst, const char *src, size_t size); - -/** - * Append output to a string, according to a format. Never write out of - * the destination buffer, and always put a terminating 0 within - * the buffer. - * @param dst destination buffer (string to which the output is - * appended) - * @param size total size of the destination buffer - * @param fmt printf-compatible format string, specifying how the - * following parameters are used - * @return the length of the string that would have been generated - * if enough space had been available - */ -size_t av_strlcatf(char *dst, size_t size, const char *fmt, ...) av_printf_format(3, 4); - -/** - * Get the count of continuous non zero chars starting from the beginning. - * - * @param len maximum number of characters to check in the string, that - * is the maximum value which is returned by the function - */ -static inline size_t av_strnlen(const char *s, size_t len) -{ - size_t i; - for (i = 0; i < len && s[i]; i++) - ; - return i; -} - -/** - * Print arguments following specified format into a large enough auto - * allocated buffer. It is similar to GNU asprintf(). - * @param fmt printf-compatible format string, specifying how the - * following parameters are used. - * @return the allocated string - * @note You have to free the string yourself with av_free(). - */ -char *av_asprintf(const char *fmt, ...) av_printf_format(1, 2); - -/** - * Convert a number to a av_malloced string. - */ -char *av_d2str(double d); - -/** - * Unescape the given string until a non escaped terminating char, - * and return the token corresponding to the unescaped string. - * - * The normal \ and ' escaping is supported. Leading and trailing - * whitespaces are removed, unless they are escaped with '\' or are - * enclosed between ''. - * - * @param buf the buffer to parse, buf will be updated to point to the - * terminating char - * @param term a 0-terminated list of terminating chars - * @return the malloced unescaped string, which must be av_freed by - * the user, NULL in case of allocation failure - */ -char *av_get_token(const char **buf, const char *term); - -/** - * Split the string into several tokens which can be accessed by - * successive calls to av_strtok(). - * - * A token is defined as a sequence of characters not belonging to the - * set specified in delim. - * - * On the first call to av_strtok(), s should point to the string to - * parse, and the value of saveptr is ignored. In subsequent calls, s - * should be NULL, and saveptr should be unchanged since the previous - * call. - * - * This function is similar to strtok_r() defined in POSIX.1. - * - * @param s the string to parse, may be NULL - * @param delim 0-terminated list of token delimiters, must be non-NULL - * @param saveptr user-provided pointer which points to stored - * information necessary for av_strtok() to continue scanning the same - * string. saveptr is updated to point to the next character after the - * first delimiter found, or to NULL if the string was terminated - * @return the found token, or NULL when no token is found - */ -char *av_strtok(char *s, const char *delim, char **saveptr); - -/** - * Locale-independent conversion of ASCII isdigit. - */ -static inline av_const int av_isdigit(int c) -{ - return c >= '0' && c <= '9'; -} - -/** - * Locale-independent conversion of ASCII isgraph. - */ -static inline av_const int av_isgraph(int c) -{ - return c > 32 && c < 127; -} - -/** - * Locale-independent conversion of ASCII isspace. - */ -static inline av_const int av_isspace(int c) -{ - return c == ' ' || c == '\f' || c == '\n' || c == '\r' || c == '\t' || - c == '\v'; -} - -/** - * Locale-independent conversion of ASCII characters to uppercase. - */ -static inline av_const int av_toupper(int c) -{ - if (c >= 'a' && c <= 'z') - c ^= 0x20; - return c; -} - -/** - * Locale-independent conversion of ASCII characters to lowercase. - */ -static inline av_const int av_tolower(int c) -{ - if (c >= 'A' && c <= 'Z') - c ^= 0x20; - return c; -} - -/** - * Locale-independent conversion of ASCII isxdigit. - */ -static inline av_const int av_isxdigit(int c) -{ - c = av_tolower(c); - return av_isdigit(c) || (c >= 'a' && c <= 'f'); -} - -/** - * Locale-independent case-insensitive compare. - * @note This means only ASCII-range characters are case-insensitive - */ -int av_strcasecmp(const char *a, const char *b); - -/** - * Locale-independent case-insensitive compare. - * @note This means only ASCII-range characters are case-insensitive - */ -int av_strncasecmp(const char *a, const char *b, size_t n); - - -/** - * Thread safe basename. - * @param path the path, on DOS both \ and / are considered separators. - * @return pointer to the basename substring. - */ -const char *av_basename(const char *path); - -/** - * Thread safe dirname. - * @param path the path, on DOS both \ and / are considered separators. - * @return the path with the separator replaced by the string terminator or ".". - * @note the function may change the input string. - */ -const char *av_dirname(char *path); - -/** - * Match instances of a name in a comma-separated list of names. - * @param name Name to look for. - * @param names List of names. - * @return 1 on match, 0 otherwise. - */ -int av_match_name(const char *name, const char *names); - -/** - * Append path component to the existing path. - * Path separator '/' is placed between when needed. - * Resulting string have to be freed with av_free(). - * @param path base path - * @param component component to be appended - * @return new path or NULL on error. - */ -char *av_append_path_component(const char *path, const char *component); - -enum AVEscapeMode { - AV_ESCAPE_MODE_AUTO, ///< Use auto-selected escaping mode. - AV_ESCAPE_MODE_BACKSLASH, ///< Use backslash escaping. - AV_ESCAPE_MODE_QUOTE, ///< Use single-quote escaping. -}; - -/** - * Consider spaces special and escape them even in the middle of the - * string. - * - * This is equivalent to adding the whitespace characters to the special - * characters lists, except it is guaranteed to use the exact same list - * of whitespace characters as the rest of libavutil. - */ -#define AV_ESCAPE_FLAG_WHITESPACE (1 << 0) - -/** - * Escape only specified special characters. - * Without this flag, escape also any characters that may be considered - * special by av_get_token(), such as the single quote. - */ -#define AV_ESCAPE_FLAG_STRICT (1 << 1) - -/** - * Escape string in src, and put the escaped string in an allocated - * string in *dst, which must be freed with av_free(). - * - * @param dst pointer where an allocated string is put - * @param src string to escape, must be non-NULL - * @param special_chars string containing the special characters which - * need to be escaped, can be NULL - * @param mode escape mode to employ, see AV_ESCAPE_MODE_* macros. - * Any unknown value for mode will be considered equivalent to - * AV_ESCAPE_MODE_BACKSLASH, but this behaviour can change without - * notice. - * @param flags flags which control how to escape, see AV_ESCAPE_FLAG_ macros - * @return the length of the allocated string, or a negative error code in case of error - * @see av_bprint_escape() - */ -av_warn_unused_result -int av_escape(char **dst, const char *src, const char *special_chars, - enum AVEscapeMode mode, int flags); - -#define AV_UTF8_FLAG_ACCEPT_INVALID_BIG_CODES 1 ///< accept codepoints over 0x10FFFF -#define AV_UTF8_FLAG_ACCEPT_NON_CHARACTERS 2 ///< accept non-characters - 0xFFFE and 0xFFFF -#define AV_UTF8_FLAG_ACCEPT_SURROGATES 4 ///< accept UTF-16 surrogates codes -#define AV_UTF8_FLAG_EXCLUDE_XML_INVALID_CONTROL_CODES 8 ///< exclude control codes not accepted by XML - -#define AV_UTF8_FLAG_ACCEPT_ALL \ - AV_UTF8_FLAG_ACCEPT_INVALID_BIG_CODES|AV_UTF8_FLAG_ACCEPT_NON_CHARACTERS|AV_UTF8_FLAG_ACCEPT_SURROGATES - -/** - * Read and decode a single UTF-8 code point (character) from the - * buffer in *buf, and update *buf to point to the next byte to - * decode. - * - * In case of an invalid byte sequence, the pointer will be updated to - * the next byte after the invalid sequence and the function will - * return an error code. - * - * Depending on the specified flags, the function will also fail in - * case the decoded code point does not belong to a valid range. - * - * @note For speed-relevant code a carefully implemented use of - * GET_UTF8() may be preferred. - * - * @param codep pointer used to return the parsed code in case of success. - * The value in *codep is set even in case the range check fails. - * @param bufp pointer to the address the first byte of the sequence - * to decode, updated by the function to point to the - * byte next after the decoded sequence - * @param buf_end pointer to the end of the buffer, points to the next - * byte past the last in the buffer. This is used to - * avoid buffer overreads (in case of an unfinished - * UTF-8 sequence towards the end of the buffer). - * @param flags a collection of AV_UTF8_FLAG_* flags - * @return >= 0 in case a sequence was successfully read, a negative - * value in case of invalid sequence - */ -av_warn_unused_result -int av_utf8_decode(int32_t *codep, const uint8_t **bufp, const uint8_t *buf_end, - unsigned int flags); - -/** - * Check if a name is in a list. - * @returns 0 if not found, or the 1 based index where it has been found in the - * list. - */ -int av_match_list(const char *name, const char *list, char separator); - -/** - * @} - */ - -#endif /* AVUTIL_AVSTRING_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/avutil.h b/src/main/jni/ffmpeg/include/libavutil/avutil.h deleted file mode 100644 index 9bcf67412..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/avutil.h +++ /dev/null @@ -1,343 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_AVUTIL_H -#define AVUTIL_AVUTIL_H - -/** - * @file - * external API header - */ - -/** - * @mainpage - * - * @section ffmpeg_intro Introduction - * - * This document describes the usage of the different libraries - * provided by FFmpeg. - * - * @li @ref libavc "libavcodec" encoding/decoding library - * @li @ref lavfi "libavfilter" graph-based frame editing library - * @li @ref libavf "libavformat" I/O and muxing/demuxing library - * @li @ref lavd "libavdevice" special devices muxing/demuxing library - * @li @ref lavu "libavutil" common utility library - * @li @ref lswr "libswresample" audio resampling, format conversion and mixing - * @li @ref lpp "libpostproc" post processing library - * @li @ref libsws "libswscale" color conversion and scaling library - * - * @section ffmpeg_versioning Versioning and compatibility - * - * Each of the FFmpeg libraries contains a version.h header, which defines a - * major, minor and micro version number with the - * <em>LIBRARYNAME_VERSION_{MAJOR,MINOR,MICRO}</em> macros. The major version - * number is incremented with backward incompatible changes - e.g. removing - * parts of the public API, reordering public struct members, etc. The minor - * version number is incremented for backward compatible API changes or major - * new features - e.g. adding a new public function or a new decoder. The micro - * version number is incremented for smaller changes that a calling program - * might still want to check for - e.g. changing behavior in a previously - * unspecified situation. - * - * FFmpeg guarantees backward API and ABI compatibility for each library as long - * as its major version number is unchanged. This means that no public symbols - * will be removed or renamed. Types and names of the public struct members and - * values of public macros and enums will remain the same (unless they were - * explicitly declared as not part of the public API). Documented behavior will - * not change. - * - * In other words, any correct program that works with a given FFmpeg snapshot - * should work just as well without any changes with any later snapshot with the - * same major versions. This applies to both rebuilding the program against new - * FFmpeg versions or to replacing the dynamic FFmpeg libraries that a program - * links against. - * - * However, new public symbols may be added and new members may be appended to - * public structs whose size is not part of public ABI (most public structs in - * FFmpeg). New macros and enum values may be added. Behavior in undocumented - * situations may change slightly (and be documented). All those are accompanied - * by an entry in doc/APIchanges and incrementing either the minor or micro - * version number. - */ - -/** - * @defgroup lavu Common utility functions - * - * @brief - * libavutil contains the code shared across all the other FFmpeg - * libraries - * - * @note In order to use the functions provided by avutil you must include - * the specific header. - * - * @{ - * - * @defgroup lavu_crypto Crypto and Hashing - * - * @{ - * @} - * - * @defgroup lavu_math Maths - * @{ - * - * @} - * - * @defgroup lavu_string String Manipulation - * - * @{ - * - * @} - * - * @defgroup lavu_mem Memory Management - * - * @{ - * - * @} - * - * @defgroup lavu_data Data Structures - * @{ - * - * @} - * - * @defgroup lavu_audio Audio related - * - * @{ - * - * @} - * - * @defgroup lavu_error Error Codes - * - * @{ - * - * @} - * - * @defgroup lavu_log Logging Facility - * - * @{ - * - * @} - * - * @defgroup lavu_misc Other - * - * @{ - * - * @defgroup preproc_misc Preprocessor String Macros - * - * @{ - * - * @} - * - * @defgroup version_utils Library Version Macros - * - * @{ - * - * @} - */ - - -/** - * @addtogroup lavu_ver - * @{ - */ - -/** - * Return the LIBAVUTIL_VERSION_INT constant. - */ -unsigned avutil_version(void); - -/** - * Return an informative version string. This usually is the actual release - * version number or a git commit description. This string has no fixed format - * and can change any time. It should never be parsed by code. - */ -const char *av_version_info(void); - -/** - * Return the libavutil build-time configuration. - */ -const char *avutil_configuration(void); - -/** - * Return the libavutil license. - */ -const char *avutil_license(void); - -/** - * @} - */ - -/** - * @addtogroup lavu_media Media Type - * @brief Media Type - */ - -enum AVMediaType { - AVMEDIA_TYPE_UNKNOWN = -1, ///< Usually treated as AVMEDIA_TYPE_DATA - AVMEDIA_TYPE_VIDEO, - AVMEDIA_TYPE_AUDIO, - AVMEDIA_TYPE_DATA, ///< Opaque data information usually continuous - AVMEDIA_TYPE_SUBTITLE, - AVMEDIA_TYPE_ATTACHMENT, ///< Opaque data information usually sparse - AVMEDIA_TYPE_NB -}; - -/** - * Return a string describing the media_type enum, NULL if media_type - * is unknown. - */ -const char *av_get_media_type_string(enum AVMediaType media_type); - -/** - * @defgroup lavu_const Constants - * @{ - * - * @defgroup lavu_enc Encoding specific - * - * @note those definition should move to avcodec - * @{ - */ - -#define FF_LAMBDA_SHIFT 7 -#define FF_LAMBDA_SCALE (1<<FF_LAMBDA_SHIFT) -#define FF_QP2LAMBDA 118 ///< factor to convert from H.263 QP to lambda -#define FF_LAMBDA_MAX (256*128-1) - -#define FF_QUALITY_SCALE FF_LAMBDA_SCALE //FIXME maybe remove - -/** - * @} - * @defgroup lavu_time Timestamp specific - * - * FFmpeg internal timebase and timestamp definitions - * - * @{ - */ - -/** - * @brief Undefined timestamp value - * - * Usually reported by demuxer that work on containers that do not provide - * either pts or dts. - */ - -#define AV_NOPTS_VALUE ((int64_t)UINT64_C(0x8000000000000000)) - -/** - * Internal time base represented as integer - */ - -#define AV_TIME_BASE 1000000 - -/** - * Internal time base represented as fractional value - */ - -#define AV_TIME_BASE_Q (AVRational){1, AV_TIME_BASE} - -/** - * @} - * @} - * @defgroup lavu_picture Image related - * - * AVPicture types, pixel formats and basic image planes manipulation. - * - * @{ - */ - -enum AVPictureType { - AV_PICTURE_TYPE_NONE = 0, ///< Undefined - AV_PICTURE_TYPE_I, ///< Intra - AV_PICTURE_TYPE_P, ///< Predicted - AV_PICTURE_TYPE_B, ///< Bi-dir predicted - AV_PICTURE_TYPE_S, ///< S(GMC)-VOP MPEG4 - AV_PICTURE_TYPE_SI, ///< Switching Intra - AV_PICTURE_TYPE_SP, ///< Switching Predicted - AV_PICTURE_TYPE_BI, ///< BI type -}; - -/** - * Return a single letter to describe the given picture type - * pict_type. - * - * @param[in] pict_type the picture type @return a single character - * representing the picture type, '?' if pict_type is unknown - */ -char av_get_picture_type_char(enum AVPictureType pict_type); - -/** - * @} - */ - -#include "common.h" -#include "error.h" -#include "rational.h" -#include "version.h" -#include "macros.h" -#include "mathematics.h" -#include "log.h" -#include "pixfmt.h" - -/** - * Return x default pointer in case p is NULL. - */ -static inline void *av_x_if_null(const void *p, const void *x) -{ - return (void *)(intptr_t)(p ? p : x); -} - -/** - * Compute the length of an integer list. - * - * @param elsize size in bytes of each list element (only 1, 2, 4 or 8) - * @param term list terminator (usually 0 or -1) - * @param list pointer to the list - * @return length of the list, in elements, not counting the terminator - */ -unsigned av_int_list_length_for_size(unsigned elsize, - const void *list, uint64_t term) av_pure; - -/** - * Compute the length of an integer list. - * - * @param term list terminator (usually 0 or -1) - * @param list pointer to the list - * @return length of the list, in elements, not counting the terminator - */ -#define av_int_list_length(list, term) \ - av_int_list_length_for_size(sizeof(*(list)), list, term) - -/** - * Open a file using a UTF-8 filename. - * The API of this function matches POSIX fopen(), errors are returned through - * errno. - */ -FILE *av_fopen_utf8(const char *path, const char *mode); - -/** - * Return the fractional representation of the internal time base. - */ -AVRational av_get_time_base_q(void); - -/** - * @} - * @} - */ - -#endif /* AVUTIL_AVUTIL_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/base64.h b/src/main/jni/ffmpeg/include/libavutil/base64.h deleted file mode 100644 index 514498eac..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/base64.h +++ /dev/null @@ -1,67 +0,0 @@ -/* - * Copyright (c) 2006 Ryan Martell. (rdm4@martellventures.com) - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_BASE64_H -#define AVUTIL_BASE64_H - -#include <stdint.h> - -/** - * @defgroup lavu_base64 Base64 - * @ingroup lavu_crypto - * @{ - */ - - -/** - * Decode a base64-encoded string. - * - * @param out buffer for decoded data - * @param in null-terminated input string - * @param out_size size in bytes of the out buffer, must be at - * least 3/4 of the length of in - * @return number of bytes written, or a negative value in case of - * invalid input - */ -int av_base64_decode(uint8_t *out, const char *in, int out_size); - -/** - * Encode data to base64 and null-terminate. - * - * @param out buffer for encoded data - * @param out_size size in bytes of the out buffer (including the - * null terminator), must be at least AV_BASE64_SIZE(in_size) - * @param in input buffer containing the data to encode - * @param in_size size in bytes of the in buffer - * @return out or NULL in case of error - */ -char *av_base64_encode(char *out, int out_size, const uint8_t *in, int in_size); - -/** - * Calculate the output size needed to base64-encode x bytes to a - * null-terminated string. - */ -#define AV_BASE64_SIZE(x) (((x)+2) / 3 * 4 + 1) - - /** - * @} - */ - -#endif /* AVUTIL_BASE64_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/blowfish.h b/src/main/jni/ffmpeg/include/libavutil/blowfish.h deleted file mode 100644 index 9e289a40d..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/blowfish.h +++ /dev/null @@ -1,82 +0,0 @@ -/* - * Blowfish algorithm - * Copyright (c) 2012 Samuel Pitoiset - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_BLOWFISH_H -#define AVUTIL_BLOWFISH_H - -#include <stdint.h> - -/** - * @defgroup lavu_blowfish Blowfish - * @ingroup lavu_crypto - * @{ - */ - -#define AV_BF_ROUNDS 16 - -typedef struct AVBlowfish { - uint32_t p[AV_BF_ROUNDS + 2]; - uint32_t s[4][256]; -} AVBlowfish; - -/** - * Allocate an AVBlowfish context. - */ -AVBlowfish *av_blowfish_alloc(void); - -/** - * Initialize an AVBlowfish context. - * - * @param ctx an AVBlowfish context - * @param key a key - * @param key_len length of the key - */ -void av_blowfish_init(struct AVBlowfish *ctx, const uint8_t *key, int key_len); - -/** - * Encrypt or decrypt a buffer using a previously initialized context. - * - * @param ctx an AVBlowfish context - * @param xl left four bytes halves of input to be encrypted - * @param xr right four bytes halves of input to be encrypted - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_blowfish_crypt_ecb(struct AVBlowfish *ctx, uint32_t *xl, uint32_t *xr, - int decrypt); - -/** - * Encrypt or decrypt a buffer using a previously initialized context. - * - * @param ctx an AVBlowfish context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 8 byte blocks - * @param iv initialization vector for CBC mode, if NULL ECB will be used - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_blowfish_crypt(struct AVBlowfish *ctx, uint8_t *dst, const uint8_t *src, - int count, uint8_t *iv, int decrypt); - -/** - * @} - */ - -#endif /* AVUTIL_BLOWFISH_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/bprint.h b/src/main/jni/ffmpeg/include/libavutil/bprint.h deleted file mode 100644 index c09b1ac1e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/bprint.h +++ /dev/null @@ -1,219 +0,0 @@ -/* - * Copyright (c) 2012 Nicolas George - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_BPRINT_H -#define AVUTIL_BPRINT_H - -#include <stdarg.h> - -#include "attributes.h" -#include "avstring.h" - -/** - * Define a structure with extra padding to a fixed size - * This helps ensuring binary compatibility with future versions. - */ - -#define FF_PAD_STRUCTURE(name, size, ...) \ -struct ff_pad_helper_##name { __VA_ARGS__ }; \ -typedef struct name { \ - __VA_ARGS__ \ - char reserved_padding[size - sizeof(struct ff_pad_helper_##name)]; \ -} name; - -/** - * Buffer to print data progressively - * - * The string buffer grows as necessary and is always 0-terminated. - * The content of the string is never accessed, and thus is - * encoding-agnostic and can even hold binary data. - * - * Small buffers are kept in the structure itself, and thus require no - * memory allocation at all (unless the contents of the buffer is needed - * after the structure goes out of scope). This is almost as lightweight as - * declaring a local "char buf[512]". - * - * The length of the string can go beyond the allocated size: the buffer is - * then truncated, but the functions still keep account of the actual total - * length. - * - * In other words, buf->len can be greater than buf->size and records the - * total length of what would have been to the buffer if there had been - * enough memory. - * - * Append operations do not need to be tested for failure: if a memory - * allocation fails, data stop being appended to the buffer, but the length - * is still updated. This situation can be tested with - * av_bprint_is_complete(). - * - * The size_max field determines several possible behaviours: - * - * size_max = -1 (= UINT_MAX) or any large value will let the buffer be - * reallocated as necessary, with an amortized linear cost. - * - * size_max = 0 prevents writing anything to the buffer: only the total - * length is computed. The write operations can then possibly be repeated in - * a buffer with exactly the necessary size - * (using size_init = size_max = len + 1). - * - * size_max = 1 is automatically replaced by the exact size available in the - * structure itself, thus ensuring no dynamic memory allocation. The - * internal buffer is large enough to hold a reasonable paragraph of text, - * such as the current paragraph. - */ - -FF_PAD_STRUCTURE(AVBPrint, 1024, - char *str; /**< string so far */ - unsigned len; /**< length so far */ - unsigned size; /**< allocated memory */ - unsigned size_max; /**< maximum allocated memory */ - char reserved_internal_buffer[1]; -) - -/** - * Convenience macros for special values for av_bprint_init() size_max - * parameter. - */ -#define AV_BPRINT_SIZE_UNLIMITED ((unsigned)-1) -#define AV_BPRINT_SIZE_AUTOMATIC 1 -#define AV_BPRINT_SIZE_COUNT_ONLY 0 - -/** - * Init a print buffer. - * - * @param buf buffer to init - * @param size_init initial size (including the final 0) - * @param size_max maximum size; - * 0 means do not write anything, just count the length; - * 1 is replaced by the maximum value for automatic storage; - * any large value means that the internal buffer will be - * reallocated as needed up to that limit; -1 is converted to - * UINT_MAX, the largest limit possible. - * Check also AV_BPRINT_SIZE_* macros. - */ -void av_bprint_init(AVBPrint *buf, unsigned size_init, unsigned size_max); - -/** - * Init a print buffer using a pre-existing buffer. - * - * The buffer will not be reallocated. - * - * @param buf buffer structure to init - * @param buffer byte buffer to use for the string data - * @param size size of buffer - */ -void av_bprint_init_for_buffer(AVBPrint *buf, char *buffer, unsigned size); - -/** - * Append a formatted string to a print buffer. - */ -void av_bprintf(AVBPrint *buf, const char *fmt, ...) av_printf_format(2, 3); - -/** - * Append a formatted string to a print buffer. - */ -void av_vbprintf(AVBPrint *buf, const char *fmt, va_list vl_arg); - -/** - * Append char c n times to a print buffer. - */ -void av_bprint_chars(AVBPrint *buf, char c, unsigned n); - -/** - * Append data to a print buffer. - * - * param buf bprint buffer to use - * param data pointer to data - * param size size of data - */ -void av_bprint_append_data(AVBPrint *buf, const char *data, unsigned size); - -struct tm; -/** - * Append a formatted date and time to a print buffer. - * - * param buf bprint buffer to use - * param fmt date and time format string, see strftime() - * param tm broken-down time structure to translate - * - * @note due to poor design of the standard strftime function, it may - * produce poor results if the format string expands to a very long text and - * the bprint buffer is near the limit stated by the size_max option. - */ -void av_bprint_strftime(AVBPrint *buf, const char *fmt, const struct tm *tm); - -/** - * Allocate bytes in the buffer for external use. - * - * @param[in] buf buffer structure - * @param[in] size required size - * @param[out] mem pointer to the memory area - * @param[out] actual_size size of the memory area after allocation; - * can be larger or smaller than size - */ -void av_bprint_get_buffer(AVBPrint *buf, unsigned size, - unsigned char **mem, unsigned *actual_size); - -/** - * Reset the string to "" but keep internal allocated data. - */ -void av_bprint_clear(AVBPrint *buf); - -/** - * Test if the print buffer is complete (not truncated). - * - * It may have been truncated due to a memory allocation failure - * or the size_max limit (compare size and size_max if necessary). - */ -static inline int av_bprint_is_complete(const AVBPrint *buf) -{ - return buf->len < buf->size; -} - -/** - * Finalize a print buffer. - * - * The print buffer can no longer be used afterwards, - * but the len and size fields are still valid. - * - * @arg[out] ret_str if not NULL, used to return a permanent copy of the - * buffer contents, or NULL if memory allocation fails; - * if NULL, the buffer is discarded and freed - * @return 0 for success or error code (probably AVERROR(ENOMEM)) - */ -int av_bprint_finalize(AVBPrint *buf, char **ret_str); - -/** - * Escape the content in src and append it to dstbuf. - * - * @param dstbuf already inited destination bprint buffer - * @param src string containing the text to escape - * @param special_chars string containing the special characters which - * need to be escaped, can be NULL - * @param mode escape mode to employ, see AV_ESCAPE_MODE_* macros. - * Any unknown value for mode will be considered equivalent to - * AV_ESCAPE_MODE_BACKSLASH, but this behaviour can change without - * notice. - * @param flags flags which control how to escape, see AV_ESCAPE_FLAG_* macros - */ -void av_bprint_escape(AVBPrint *dstbuf, const char *src, const char *special_chars, - enum AVEscapeMode mode, int flags); - -#endif /* AVUTIL_BPRINT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/bswap.h b/src/main/jni/ffmpeg/include/libavutil/bswap.h deleted file mode 100644 index 91cb79538..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/bswap.h +++ /dev/null @@ -1,109 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * byte swapping routines - */ - -#ifndef AVUTIL_BSWAP_H -#define AVUTIL_BSWAP_H - -#include <stdint.h> -#include "libavutil/avconfig.h" -#include "attributes.h" - -#ifdef HAVE_AV_CONFIG_H - -#include "config.h" - -#if ARCH_AARCH64 -# include "aarch64/bswap.h" -#elif ARCH_ARM -# include "arm/bswap.h" -#elif ARCH_AVR32 -# include "avr32/bswap.h" -#elif ARCH_SH4 -# include "sh4/bswap.h" -#elif ARCH_X86 -# include "x86/bswap.h" -#endif - -#endif /* HAVE_AV_CONFIG_H */ - -#define AV_BSWAP16C(x) (((x) << 8 & 0xff00) | ((x) >> 8 & 0x00ff)) -#define AV_BSWAP32C(x) (AV_BSWAP16C(x) << 16 | AV_BSWAP16C((x) >> 16)) -#define AV_BSWAP64C(x) (AV_BSWAP32C(x) << 32 | AV_BSWAP32C((x) >> 32)) - -#define AV_BSWAPC(s, x) AV_BSWAP##s##C(x) - -#ifndef av_bswap16 -static av_always_inline av_const uint16_t av_bswap16(uint16_t x) -{ - x= (x>>8) | (x<<8); - return x; -} -#endif - -#ifndef av_bswap32 -static av_always_inline av_const uint32_t av_bswap32(uint32_t x) -{ - return AV_BSWAP32C(x); -} -#endif - -#ifndef av_bswap64 -static inline uint64_t av_const av_bswap64(uint64_t x) -{ - return (uint64_t)av_bswap32(x) << 32 | av_bswap32(x >> 32); -} -#endif - -// be2ne ... big-endian to native-endian -// le2ne ... little-endian to native-endian - -#if AV_HAVE_BIGENDIAN -#define av_be2ne16(x) (x) -#define av_be2ne32(x) (x) -#define av_be2ne64(x) (x) -#define av_le2ne16(x) av_bswap16(x) -#define av_le2ne32(x) av_bswap32(x) -#define av_le2ne64(x) av_bswap64(x) -#define AV_BE2NEC(s, x) (x) -#define AV_LE2NEC(s, x) AV_BSWAPC(s, x) -#else -#define av_be2ne16(x) av_bswap16(x) -#define av_be2ne32(x) av_bswap32(x) -#define av_be2ne64(x) av_bswap64(x) -#define av_le2ne16(x) (x) -#define av_le2ne32(x) (x) -#define av_le2ne64(x) (x) -#define AV_BE2NEC(s, x) AV_BSWAPC(s, x) -#define AV_LE2NEC(s, x) (x) -#endif - -#define AV_BE2NE16C(x) AV_BE2NEC(16, x) -#define AV_BE2NE32C(x) AV_BE2NEC(32, x) -#define AV_BE2NE64C(x) AV_BE2NEC(64, x) -#define AV_LE2NE16C(x) AV_LE2NEC(16, x) -#define AV_LE2NE32C(x) AV_LE2NEC(32, x) -#define AV_LE2NE64C(x) AV_LE2NEC(64, x) - -#endif /* AVUTIL_BSWAP_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/buffer.h b/src/main/jni/ffmpeg/include/libavutil/buffer.h deleted file mode 100644 index b4399fd39..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/buffer.h +++ /dev/null @@ -1,274 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * @ingroup lavu_buffer - * refcounted data buffer API - */ - -#ifndef AVUTIL_BUFFER_H -#define AVUTIL_BUFFER_H - -#include <stdint.h> - -/** - * @defgroup lavu_buffer AVBuffer - * @ingroup lavu_data - * - * @{ - * AVBuffer is an API for reference-counted data buffers. - * - * There are two core objects in this API -- AVBuffer and AVBufferRef. AVBuffer - * represents the data buffer itself; it is opaque and not meant to be accessed - * by the caller directly, but only through AVBufferRef. However, the caller may - * e.g. compare two AVBuffer pointers to check whether two different references - * are describing the same data buffer. AVBufferRef represents a single - * reference to an AVBuffer and it is the object that may be manipulated by the - * caller directly. - * - * There are two functions provided for creating a new AVBuffer with a single - * reference -- av_buffer_alloc() to just allocate a new buffer, and - * av_buffer_create() to wrap an existing array in an AVBuffer. From an existing - * reference, additional references may be created with av_buffer_ref(). - * Use av_buffer_unref() to free a reference (this will automatically free the - * data once all the references are freed). - * - * The convention throughout this API and the rest of FFmpeg is such that the - * buffer is considered writable if there exists only one reference to it (and - * it has not been marked as read-only). The av_buffer_is_writable() function is - * provided to check whether this is true and av_buffer_make_writable() will - * automatically create a new writable buffer when necessary. - * Of course nothing prevents the calling code from violating this convention, - * however that is safe only when all the existing references are under its - * control. - * - * @note Referencing and unreferencing the buffers is thread-safe and thus - * may be done from multiple threads simultaneously without any need for - * additional locking. - * - * @note Two different references to the same buffer can point to different - * parts of the buffer (i.e. their AVBufferRef.data will not be equal). - */ - -/** - * A reference counted buffer type. It is opaque and is meant to be used through - * references (AVBufferRef). - */ -typedef struct AVBuffer AVBuffer; - -/** - * A reference to a data buffer. - * - * The size of this struct is not a part of the public ABI and it is not meant - * to be allocated directly. - */ -typedef struct AVBufferRef { - AVBuffer *buffer; - - /** - * The data buffer. It is considered writable if and only if - * this is the only reference to the buffer, in which case - * av_buffer_is_writable() returns 1. - */ - uint8_t *data; - /** - * Size of data in bytes. - */ - int size; -} AVBufferRef; - -/** - * Allocate an AVBuffer of the given size using av_malloc(). - * - * @return an AVBufferRef of given size or NULL when out of memory - */ -AVBufferRef *av_buffer_alloc(int size); - -/** - * Same as av_buffer_alloc(), except the returned buffer will be initialized - * to zero. - */ -AVBufferRef *av_buffer_allocz(int size); - -/** - * Always treat the buffer as read-only, even when it has only one - * reference. - */ -#define AV_BUFFER_FLAG_READONLY (1 << 0) - -/** - * Create an AVBuffer from an existing array. - * - * If this function is successful, data is owned by the AVBuffer. The caller may - * only access data through the returned AVBufferRef and references derived from - * it. - * If this function fails, data is left untouched. - * @param data data array - * @param size size of data in bytes - * @param free a callback for freeing this buffer's data - * @param opaque parameter to be got for processing or passed to free - * @param flags a combination of AV_BUFFER_FLAG_* - * - * @return an AVBufferRef referring to data on success, NULL on failure. - */ -AVBufferRef *av_buffer_create(uint8_t *data, int size, - void (*free)(void *opaque, uint8_t *data), - void *opaque, int flags); - -/** - * Default free callback, which calls av_free() on the buffer data. - * This function is meant to be passed to av_buffer_create(), not called - * directly. - */ -void av_buffer_default_free(void *opaque, uint8_t *data); - -/** - * Create a new reference to an AVBuffer. - * - * @return a new AVBufferRef referring to the same AVBuffer as buf or NULL on - * failure. - */ -AVBufferRef *av_buffer_ref(AVBufferRef *buf); - -/** - * Free a given reference and automatically free the buffer if there are no more - * references to it. - * - * @param buf the reference to be freed. The pointer is set to NULL on return. - */ -void av_buffer_unref(AVBufferRef **buf); - -/** - * @return 1 if the caller may write to the data referred to by buf (which is - * true if and only if buf is the only reference to the underlying AVBuffer). - * Return 0 otherwise. - * A positive answer is valid until av_buffer_ref() is called on buf. - */ -int av_buffer_is_writable(const AVBufferRef *buf); - -/** - * @return the opaque parameter set by av_buffer_create. - */ -void *av_buffer_get_opaque(const AVBufferRef *buf); - -int av_buffer_get_ref_count(const AVBufferRef *buf); - -/** - * Create a writable reference from a given buffer reference, avoiding data copy - * if possible. - * - * @param buf buffer reference to make writable. On success, buf is either left - * untouched, or it is unreferenced and a new writable AVBufferRef is - * written in its place. On failure, buf is left untouched. - * @return 0 on success, a negative AVERROR on failure. - */ -int av_buffer_make_writable(AVBufferRef **buf); - -/** - * Reallocate a given buffer. - * - * @param buf a buffer reference to reallocate. On success, buf will be - * unreferenced and a new reference with the required size will be - * written in its place. On failure buf will be left untouched. *buf - * may be NULL, then a new buffer is allocated. - * @param size required new buffer size. - * @return 0 on success, a negative AVERROR on failure. - * - * @note the buffer is actually reallocated with av_realloc() only if it was - * initially allocated through av_buffer_realloc(NULL) and there is only one - * reference to it (i.e. the one passed to this function). In all other cases - * a new buffer is allocated and the data is copied. - */ -int av_buffer_realloc(AVBufferRef **buf, int size); - -/** - * @} - */ - -/** - * @defgroup lavu_bufferpool AVBufferPool - * @ingroup lavu_data - * - * @{ - * AVBufferPool is an API for a lock-free thread-safe pool of AVBuffers. - * - * Frequently allocating and freeing large buffers may be slow. AVBufferPool is - * meant to solve this in cases when the caller needs a set of buffers of the - * same size (the most obvious use case being buffers for raw video or audio - * frames). - * - * At the beginning, the user must call av_buffer_pool_init() to create the - * buffer pool. Then whenever a buffer is needed, call av_buffer_pool_get() to - * get a reference to a new buffer, similar to av_buffer_alloc(). This new - * reference works in all aspects the same way as the one created by - * av_buffer_alloc(). However, when the last reference to this buffer is - * unreferenced, it is returned to the pool instead of being freed and will be - * reused for subsequent av_buffer_pool_get() calls. - * - * When the caller is done with the pool and no longer needs to allocate any new - * buffers, av_buffer_pool_uninit() must be called to mark the pool as freeable. - * Once all the buffers are released, it will automatically be freed. - * - * Allocating and releasing buffers with this API is thread-safe as long as - * either the default alloc callback is used, or the user-supplied one is - * thread-safe. - */ - -/** - * The buffer pool. This structure is opaque and not meant to be accessed - * directly. It is allocated with av_buffer_pool_init() and freed with - * av_buffer_pool_uninit(). - */ -typedef struct AVBufferPool AVBufferPool; - -/** - * Allocate and initialize a buffer pool. - * - * @param size size of each buffer in this pool - * @param alloc a function that will be used to allocate new buffers when the - * pool is empty. May be NULL, then the default allocator will be used - * (av_buffer_alloc()). - * @return newly created buffer pool on success, NULL on error. - */ -AVBufferPool *av_buffer_pool_init(int size, AVBufferRef* (*alloc)(int size)); - -/** - * Mark the pool as being available for freeing. It will actually be freed only - * once all the allocated buffers associated with the pool are released. Thus it - * is safe to call this function while some of the allocated buffers are still - * in use. - * - * @param pool pointer to the pool to be freed. It will be set to NULL. - * @see av_buffer_pool_can_uninit() - */ -void av_buffer_pool_uninit(AVBufferPool **pool); - -/** - * Allocate a new AVBuffer, reusing an old buffer from the pool when available. - * This function may be called simultaneously from multiple threads. - * - * @return a reference to the new buffer on success, NULL on error. - */ -AVBufferRef *av_buffer_pool_get(AVBufferPool *pool); - -/** - * @} - */ - -#endif /* AVUTIL_BUFFER_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/camellia.h b/src/main/jni/ffmpeg/include/libavutil/camellia.h deleted file mode 100644 index e674c9b9a..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/camellia.h +++ /dev/null @@ -1,70 +0,0 @@ -/* - * An implementation of the CAMELLIA algorithm as mentioned in RFC3713 - * Copyright (c) 2014 Supraja Meedinti - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_CAMELLIA_H -#define AVUTIL_CAMELLIA_H - -#include <stdint.h> - - -/** - * @file - * @brief Public header for libavutil CAMELLIA algorithm - * @defgroup lavu_camellia CAMELLIA - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_camellia_size; - -struct AVCAMELLIA; - -/** - * Allocate an AVCAMELLIA context - * To free the struct: av_free(ptr) - */ -struct AVCAMELLIA *av_camellia_alloc(void); - -/** - * Initialize an AVCAMELLIA context. - * - * @param ctx an AVCAMELLIA context - * @param key a key of 16, 24, 32 bytes used for encryption/decryption - * @param key_bits number of keybits: possible are 128, 192, 256 - */ -int av_camellia_init(struct AVCAMELLIA *ctx, const uint8_t *key, int key_bits); - -/** - * Encrypt or decrypt a buffer using a previously initialized context - * - * @param ctx an AVCAMELLIA context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 16 byte blocks - * @paran iv initialization vector for CBC mode, NULL for ECB mode - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_camellia_crypt(struct AVCAMELLIA *ctx, uint8_t *dst, const uint8_t *src, int count, uint8_t* iv, int decrypt); - -/** - * @} - */ -#endif /* AVUTIL_CAMELLIA_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/cast5.h b/src/main/jni/ffmpeg/include/libavutil/cast5.h deleted file mode 100644 index ad5b347e6..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/cast5.h +++ /dev/null @@ -1,80 +0,0 @@ -/* - * An implementation of the CAST128 algorithm as mentioned in RFC2144 - * Copyright (c) 2014 Supraja Meedinti - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_CAST5_H -#define AVUTIL_CAST5_H - -#include <stdint.h> - - -/** - * @file - * @brief Public header for libavutil CAST5 algorithm - * @defgroup lavu_cast5 CAST5 - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_cast5_size; - -struct AVCAST5; - -/** - * Allocate an AVCAST5 context - * To free the struct: av_free(ptr) - */ -struct AVCAST5 *av_cast5_alloc(void); -/** - * Initialize an AVCAST5 context. - * - * @param ctx an AVCAST5 context - * @param key a key of 5,6,...16 bytes used for encryption/decryption - * @param key_bits number of keybits: possible are 40,48,...,128 - * @return 0 on success, less than 0 on failure - */ -int av_cast5_init(struct AVCAST5 *ctx, const uint8_t *key, int key_bits); - -/** - * Encrypt or decrypt a buffer using a previously initialized context, ECB mode only - * - * @param ctx an AVCAST5 context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 8 byte blocks - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_cast5_crypt(struct AVCAST5 *ctx, uint8_t *dst, const uint8_t *src, int count, int decrypt); - -/** - * Encrypt or decrypt a buffer using a previously initialized context - * - * @param ctx an AVCAST5 context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 8 byte blocks - * @param iv initialization vector for CBC mode, NULL for ECB mode - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_cast5_crypt2(struct AVCAST5 *ctx, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt); -/** - * @} - */ -#endif /* AVUTIL_CAST5_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/channel_layout.h b/src/main/jni/ffmpeg/include/libavutil/channel_layout.h deleted file mode 100644 index ec7effead..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/channel_layout.h +++ /dev/null @@ -1,223 +0,0 @@ -/* - * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * Copyright (c) 2008 Peter Ross - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_CHANNEL_LAYOUT_H -#define AVUTIL_CHANNEL_LAYOUT_H - -#include <stdint.h> - -/** - * @file - * audio channel layout utility functions - */ - -/** - * @addtogroup lavu_audio - * @{ - */ - -/** - * @defgroup channel_masks Audio channel masks - * - * A channel layout is a 64-bits integer with a bit set for every channel. - * The number of bits set must be equal to the number of channels. - * The value 0 means that the channel layout is not known. - * @note this data structure is not powerful enough to handle channels - * combinations that have the same channel multiple times, such as - * dual-mono. - * - * @{ - */ -#define AV_CH_FRONT_LEFT 0x00000001 -#define AV_CH_FRONT_RIGHT 0x00000002 -#define AV_CH_FRONT_CENTER 0x00000004 -#define AV_CH_LOW_FREQUENCY 0x00000008 -#define AV_CH_BACK_LEFT 0x00000010 -#define AV_CH_BACK_RIGHT 0x00000020 -#define AV_CH_FRONT_LEFT_OF_CENTER 0x00000040 -#define AV_CH_FRONT_RIGHT_OF_CENTER 0x00000080 -#define AV_CH_BACK_CENTER 0x00000100 -#define AV_CH_SIDE_LEFT 0x00000200 -#define AV_CH_SIDE_RIGHT 0x00000400 -#define AV_CH_TOP_CENTER 0x00000800 -#define AV_CH_TOP_FRONT_LEFT 0x00001000 -#define AV_CH_TOP_FRONT_CENTER 0x00002000 -#define AV_CH_TOP_FRONT_RIGHT 0x00004000 -#define AV_CH_TOP_BACK_LEFT 0x00008000 -#define AV_CH_TOP_BACK_CENTER 0x00010000 -#define AV_CH_TOP_BACK_RIGHT 0x00020000 -#define AV_CH_STEREO_LEFT 0x20000000 ///< Stereo downmix. -#define AV_CH_STEREO_RIGHT 0x40000000 ///< See AV_CH_STEREO_LEFT. -#define AV_CH_WIDE_LEFT 0x0000000080000000ULL -#define AV_CH_WIDE_RIGHT 0x0000000100000000ULL -#define AV_CH_SURROUND_DIRECT_LEFT 0x0000000200000000ULL -#define AV_CH_SURROUND_DIRECT_RIGHT 0x0000000400000000ULL -#define AV_CH_LOW_FREQUENCY_2 0x0000000800000000ULL - -/** Channel mask value used for AVCodecContext.request_channel_layout - to indicate that the user requests the channel order of the decoder output - to be the native codec channel order. */ -#define AV_CH_LAYOUT_NATIVE 0x8000000000000000ULL - -/** - * @} - * @defgroup channel_mask_c Audio channel layouts - * @{ - * */ -#define AV_CH_LAYOUT_MONO (AV_CH_FRONT_CENTER) -#define AV_CH_LAYOUT_STEREO (AV_CH_FRONT_LEFT|AV_CH_FRONT_RIGHT) -#define AV_CH_LAYOUT_2POINT1 (AV_CH_LAYOUT_STEREO|AV_CH_LOW_FREQUENCY) -#define AV_CH_LAYOUT_2_1 (AV_CH_LAYOUT_STEREO|AV_CH_BACK_CENTER) -#define AV_CH_LAYOUT_SURROUND (AV_CH_LAYOUT_STEREO|AV_CH_FRONT_CENTER) -#define AV_CH_LAYOUT_3POINT1 (AV_CH_LAYOUT_SURROUND|AV_CH_LOW_FREQUENCY) -#define AV_CH_LAYOUT_4POINT0 (AV_CH_LAYOUT_SURROUND|AV_CH_BACK_CENTER) -#define AV_CH_LAYOUT_4POINT1 (AV_CH_LAYOUT_4POINT0|AV_CH_LOW_FREQUENCY) -#define AV_CH_LAYOUT_2_2 (AV_CH_LAYOUT_STEREO|AV_CH_SIDE_LEFT|AV_CH_SIDE_RIGHT) -#define AV_CH_LAYOUT_QUAD (AV_CH_LAYOUT_STEREO|AV_CH_BACK_LEFT|AV_CH_BACK_RIGHT) -#define AV_CH_LAYOUT_5POINT0 (AV_CH_LAYOUT_SURROUND|AV_CH_SIDE_LEFT|AV_CH_SIDE_RIGHT) -#define AV_CH_LAYOUT_5POINT1 (AV_CH_LAYOUT_5POINT0|AV_CH_LOW_FREQUENCY) -#define AV_CH_LAYOUT_5POINT0_BACK (AV_CH_LAYOUT_SURROUND|AV_CH_BACK_LEFT|AV_CH_BACK_RIGHT) -#define AV_CH_LAYOUT_5POINT1_BACK (AV_CH_LAYOUT_5POINT0_BACK|AV_CH_LOW_FREQUENCY) -#define AV_CH_LAYOUT_6POINT0 (AV_CH_LAYOUT_5POINT0|AV_CH_BACK_CENTER) -#define AV_CH_LAYOUT_6POINT0_FRONT (AV_CH_LAYOUT_2_2|AV_CH_FRONT_LEFT_OF_CENTER|AV_CH_FRONT_RIGHT_OF_CENTER) -#define AV_CH_LAYOUT_HEXAGONAL (AV_CH_LAYOUT_5POINT0_BACK|AV_CH_BACK_CENTER) -#define AV_CH_LAYOUT_6POINT1 (AV_CH_LAYOUT_5POINT1|AV_CH_BACK_CENTER) -#define AV_CH_LAYOUT_6POINT1_BACK (AV_CH_LAYOUT_5POINT1_BACK|AV_CH_BACK_CENTER) -#define AV_CH_LAYOUT_6POINT1_FRONT (AV_CH_LAYOUT_6POINT0_FRONT|AV_CH_LOW_FREQUENCY) -#define AV_CH_LAYOUT_7POINT0 (AV_CH_LAYOUT_5POINT0|AV_CH_BACK_LEFT|AV_CH_BACK_RIGHT) -#define AV_CH_LAYOUT_7POINT0_FRONT (AV_CH_LAYOUT_5POINT0|AV_CH_FRONT_LEFT_OF_CENTER|AV_CH_FRONT_RIGHT_OF_CENTER) -#define AV_CH_LAYOUT_7POINT1 (AV_CH_LAYOUT_5POINT1|AV_CH_BACK_LEFT|AV_CH_BACK_RIGHT) -#define AV_CH_LAYOUT_7POINT1_WIDE (AV_CH_LAYOUT_5POINT1|AV_CH_FRONT_LEFT_OF_CENTER|AV_CH_FRONT_RIGHT_OF_CENTER) -#define AV_CH_LAYOUT_7POINT1_WIDE_BACK (AV_CH_LAYOUT_5POINT1_BACK|AV_CH_FRONT_LEFT_OF_CENTER|AV_CH_FRONT_RIGHT_OF_CENTER) -#define AV_CH_LAYOUT_OCTAGONAL (AV_CH_LAYOUT_5POINT0|AV_CH_BACK_LEFT|AV_CH_BACK_CENTER|AV_CH_BACK_RIGHT) -#define AV_CH_LAYOUT_HEXADECAGONAL (AV_CH_LAYOUT_OCTAGONAL|AV_CH_WIDE_LEFT|AV_CH_WIDE_RIGHT|AV_CH_TOP_BACK_LEFT|AV_CH_TOP_BACK_RIGHT|AV_CH_TOP_BACK_CENTER|AV_CH_TOP_FRONT_CENTER|AV_CH_TOP_FRONT_LEFT|AV_CH_TOP_FRONT_RIGHT) -#define AV_CH_LAYOUT_STEREO_DOWNMIX (AV_CH_STEREO_LEFT|AV_CH_STEREO_RIGHT) - -enum AVMatrixEncoding { - AV_MATRIX_ENCODING_NONE, - AV_MATRIX_ENCODING_DOLBY, - AV_MATRIX_ENCODING_DPLII, - AV_MATRIX_ENCODING_DPLIIX, - AV_MATRIX_ENCODING_DPLIIZ, - AV_MATRIX_ENCODING_DOLBYEX, - AV_MATRIX_ENCODING_DOLBYHEADPHONE, - AV_MATRIX_ENCODING_NB -}; - -/** - * Return a channel layout id that matches name, or 0 if no match is found. - * - * name can be one or several of the following notations, - * separated by '+' or '|': - * - the name of an usual channel layout (mono, stereo, 4.0, quad, 5.0, - * 5.0(side), 5.1, 5.1(side), 7.1, 7.1(wide), downmix); - * - the name of a single channel (FL, FR, FC, LFE, BL, BR, FLC, FRC, BC, - * SL, SR, TC, TFL, TFC, TFR, TBL, TBC, TBR, DL, DR); - * - a number of channels, in decimal, optionally followed by 'c', yielding - * the default channel layout for that number of channels (@see - * av_get_default_channel_layout); - * - a channel layout mask, in hexadecimal starting with "0x" (see the - * AV_CH_* macros). - * - * @warning Starting from the next major bump the trailing character - * 'c' to specify a number of channels will be required, while a - * channel layout mask could also be specified as a decimal number - * (if and only if not followed by "c"). - * - * Example: "stereo+FC" = "2c+FC" = "2c+1c" = "0x7" - */ -uint64_t av_get_channel_layout(const char *name); - -/** - * Return a description of a channel layout. - * If nb_channels is <= 0, it is guessed from the channel_layout. - * - * @param buf put here the string containing the channel layout - * @param buf_size size in bytes of the buffer - */ -void av_get_channel_layout_string(char *buf, int buf_size, int nb_channels, uint64_t channel_layout); - -struct AVBPrint; -/** - * Append a description of a channel layout to a bprint buffer. - */ -void av_bprint_channel_layout(struct AVBPrint *bp, int nb_channels, uint64_t channel_layout); - -/** - * Return the number of channels in the channel layout. - */ -int av_get_channel_layout_nb_channels(uint64_t channel_layout); - -/** - * Return default channel layout for a given number of channels. - */ -int64_t av_get_default_channel_layout(int nb_channels); - -/** - * Get the index of a channel in channel_layout. - * - * @param channel a channel layout describing exactly one channel which must be - * present in channel_layout. - * - * @return index of channel in channel_layout on success, a negative AVERROR - * on error. - */ -int av_get_channel_layout_channel_index(uint64_t channel_layout, - uint64_t channel); - -/** - * Get the channel with the given index in channel_layout. - */ -uint64_t av_channel_layout_extract_channel(uint64_t channel_layout, int index); - -/** - * Get the name of a given channel. - * - * @return channel name on success, NULL on error. - */ -const char *av_get_channel_name(uint64_t channel); - -/** - * Get the description of a given channel. - * - * @param channel a channel layout with a single channel - * @return channel description on success, NULL on error - */ -const char *av_get_channel_description(uint64_t channel); - -/** - * Get the value and name of a standard channel layout. - * - * @param[in] index index in an internal list, starting at 0 - * @param[out] layout channel layout mask - * @param[out] name name of the layout - * @return 0 if the layout exists, - * <0 if index is beyond the limits - */ -int av_get_standard_channel_layout(unsigned index, uint64_t *layout, - const char **name); - -/** - * @} - * @} - */ - -#endif /* AVUTIL_CHANNEL_LAYOUT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/common.h b/src/main/jni/ffmpeg/include/libavutil/common.h deleted file mode 100644 index 6f0f5827e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/common.h +++ /dev/null @@ -1,519 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * common internal and external API header - */ - -#ifndef AVUTIL_COMMON_H -#define AVUTIL_COMMON_H - -#if defined(__cplusplus) && !defined(__STDC_CONSTANT_MACROS) && !defined(UINT64_C) -#error missing -D__STDC_CONSTANT_MACROS / #define __STDC_CONSTANT_MACROS -#endif - -#include <errno.h> -#include <inttypes.h> -#include <limits.h> -#include <math.h> -#include <stdint.h> -#include <stdio.h> -#include <stdlib.h> -#include <string.h> - -#include "attributes.h" -#include "version.h" -#include "libavutil/avconfig.h" - -#if AV_HAVE_BIGENDIAN -# define AV_NE(be, le) (be) -#else -# define AV_NE(be, le) (le) -#endif - -//rounded division & shift -#define RSHIFT(a,b) ((a) > 0 ? ((a) + ((1<<(b))>>1))>>(b) : ((a) + ((1<<(b))>>1)-1)>>(b)) -/* assume b>0 */ -#define ROUNDED_DIV(a,b) (((a)>0 ? (a) + ((b)>>1) : (a) - ((b)>>1))/(b)) -/* assume a>0 and b>0 */ -#define FF_CEIL_RSHIFT(a,b) (!av_builtin_constant_p(b) ? -((-(a)) >> (b)) \ - : ((a) + (1<<(b)) - 1) >> (b)) -#define FFUDIV(a,b) (((a)>0 ?(a):(a)-(b)+1) / (b)) -#define FFUMOD(a,b) ((a)-(b)*FFUDIV(a,b)) - -/** - * Absolute value, Note, INT_MIN / INT64_MIN result in undefined behavior as they - * are not representable as absolute values of their type. This is the same - * as with *abs() - * @see FFNABS() - */ -#define FFABS(a) ((a) >= 0 ? (a) : (-(a))) -#define FFSIGN(a) ((a) > 0 ? 1 : -1) - -/** - * Negative Absolute value. - * this works for all integers of all types. - * As with many macros, this evaluates its argument twice, it thus must not have - * a sideeffect, that is FFNABS(x++) has undefined behavior. - */ -#define FFNABS(a) ((a) <= 0 ? (a) : (-(a))) - -/** - * Comparator. - * For two numerical expressions x and y, gives 1 if x > y, -1 if x < y, and 0 - * if x == y. This is useful for instance in a qsort comparator callback. - * Furthermore, compilers are able to optimize this to branchless code, and - * there is no risk of overflow with signed types. - * As with many macros, this evaluates its argument multiple times, it thus - * must not have a side-effect. - */ -#define FFDIFFSIGN(x,y) (((x)>(y)) - ((x)<(y))) - -#define FFMAX(a,b) ((a) > (b) ? (a) : (b)) -#define FFMAX3(a,b,c) FFMAX(FFMAX(a,b),c) -#define FFMIN(a,b) ((a) > (b) ? (b) : (a)) -#define FFMIN3(a,b,c) FFMIN(FFMIN(a,b),c) - -#define FFSWAP(type,a,b) do{type SWAP_tmp= b; b= a; a= SWAP_tmp;}while(0) -#define FF_ARRAY_ELEMS(a) (sizeof(a) / sizeof((a)[0])) -#define FFALIGN(x, a) (((x)+(a)-1)&~((a)-1)) - -/* misc math functions */ - -#ifdef HAVE_AV_CONFIG_H -# include "config.h" -# include "intmath.h" -#endif - -/* Pull in unguarded fallback defines at the end of this file. */ -#include "common.h" - -#ifndef av_log2 -av_const int av_log2(unsigned v); -#endif - -#ifndef av_log2_16bit -av_const int av_log2_16bit(unsigned v); -#endif - -/** - * Clip a signed integer value into the amin-amax range. - * @param a value to clip - * @param amin minimum value of the clip range - * @param amax maximum value of the clip range - * @return clipped value - */ -static av_always_inline av_const int av_clip_c(int a, int amin, int amax) -{ -#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 - if (amin > amax) abort(); -#endif - if (a < amin) return amin; - else if (a > amax) return amax; - else return a; -} - -/** - * Clip a signed 64bit integer value into the amin-amax range. - * @param a value to clip - * @param amin minimum value of the clip range - * @param amax maximum value of the clip range - * @return clipped value - */ -static av_always_inline av_const int64_t av_clip64_c(int64_t a, int64_t amin, int64_t amax) -{ -#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 - if (amin > amax) abort(); -#endif - if (a < amin) return amin; - else if (a > amax) return amax; - else return a; -} - -/** - * Clip a signed integer value into the 0-255 range. - * @param a value to clip - * @return clipped value - */ -static av_always_inline av_const uint8_t av_clip_uint8_c(int a) -{ - if (a&(~0xFF)) return (-a)>>31; - else return a; -} - -/** - * Clip a signed integer value into the -128,127 range. - * @param a value to clip - * @return clipped value - */ -static av_always_inline av_const int8_t av_clip_int8_c(int a) -{ - if ((a+0x80U) & ~0xFF) return (a>>31) ^ 0x7F; - else return a; -} - -/** - * Clip a signed integer value into the 0-65535 range. - * @param a value to clip - * @return clipped value - */ -static av_always_inline av_const uint16_t av_clip_uint16_c(int a) -{ - if (a&(~0xFFFF)) return (-a)>>31; - else return a; -} - -/** - * Clip a signed integer value into the -32768,32767 range. - * @param a value to clip - * @return clipped value - */ -static av_always_inline av_const int16_t av_clip_int16_c(int a) -{ - if ((a+0x8000U) & ~0xFFFF) return (a>>31) ^ 0x7FFF; - else return a; -} - -/** - * Clip a signed 64-bit integer value into the -2147483648,2147483647 range. - * @param a value to clip - * @return clipped value - */ -static av_always_inline av_const int32_t av_clipl_int32_c(int64_t a) -{ - if ((a+0x80000000u) & ~UINT64_C(0xFFFFFFFF)) return (int32_t)((a>>63) ^ 0x7FFFFFFF); - else return (int32_t)a; -} - -/** - * Clip a signed integer into the -(2^p),(2^p-1) range. - * @param a value to clip - * @param p bit position to clip at - * @return clipped value - */ -static av_always_inline av_const int av_clip_intp2_c(int a, int p) -{ - if ((a + (1 << p)) & ~((2 << p) - 1)) - return (a >> 31) ^ ((1 << p) - 1); - else - return a; -} - -/** - * Clip a signed integer to an unsigned power of two range. - * @param a value to clip - * @param p bit position to clip at - * @return clipped value - */ -static av_always_inline av_const unsigned av_clip_uintp2_c(int a, int p) -{ - if (a & ~((1<<p) - 1)) return -a >> 31 & ((1<<p) - 1); - else return a; -} - -/** - * Clear high bits from an unsigned integer starting with specific bit position - * @param a value to clip - * @param p bit position to clip at - * @return clipped value - */ -static av_always_inline av_const unsigned av_mod_uintp2_c(unsigned a, unsigned p) -{ - return a & ((1 << p) - 1); -} - -/** - * Add two signed 32-bit values with saturation. - * - * @param a one value - * @param b another value - * @return sum with signed saturation - */ -static av_always_inline int av_sat_add32_c(int a, int b) -{ - return av_clipl_int32((int64_t)a + b); -} - -/** - * Add a doubled value to another value with saturation at both stages. - * - * @param a first value - * @param b value doubled and added to a - * @return sum with signed saturation - */ -static av_always_inline int av_sat_dadd32_c(int a, int b) -{ - return av_sat_add32(a, av_sat_add32(b, b)); -} - -/** - * Clip a float value into the amin-amax range. - * @param a value to clip - * @param amin minimum value of the clip range - * @param amax maximum value of the clip range - * @return clipped value - */ -static av_always_inline av_const float av_clipf_c(float a, float amin, float amax) -{ -#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 - if (amin > amax) abort(); -#endif - if (a < amin) return amin; - else if (a > amax) return amax; - else return a; -} - -/** - * Clip a double value into the amin-amax range. - * @param a value to clip - * @param amin minimum value of the clip range - * @param amax maximum value of the clip range - * @return clipped value - */ -static av_always_inline av_const double av_clipd_c(double a, double amin, double amax) -{ -#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 - if (amin > amax) abort(); -#endif - if (a < amin) return amin; - else if (a > amax) return amax; - else return a; -} - -/** Compute ceil(log2(x)). - * @param x value used to compute ceil(log2(x)) - * @return computed ceiling of log2(x) - */ -static av_always_inline av_const int av_ceil_log2_c(int x) -{ - return av_log2((x - 1) << 1); -} - -/** - * Count number of bits set to one in x - * @param x value to count bits of - * @return the number of bits set to one in x - */ -static av_always_inline av_const int av_popcount_c(uint32_t x) -{ - x -= (x >> 1) & 0x55555555; - x = (x & 0x33333333) + ((x >> 2) & 0x33333333); - x = (x + (x >> 4)) & 0x0F0F0F0F; - x += x >> 8; - return (x + (x >> 16)) & 0x3F; -} - -/** - * Count number of bits set to one in x - * @param x value to count bits of - * @return the number of bits set to one in x - */ -static av_always_inline av_const int av_popcount64_c(uint64_t x) -{ - return av_popcount((uint32_t)x) + av_popcount((uint32_t)(x >> 32)); -} - -#define MKTAG(a,b,c,d) ((a) | ((b) << 8) | ((c) << 16) | ((unsigned)(d) << 24)) -#define MKBETAG(a,b,c,d) ((d) | ((c) << 8) | ((b) << 16) | ((unsigned)(a) << 24)) - -/** - * Convert a UTF-8 character (up to 4 bytes) to its 32-bit UCS-4 encoded form. - * - * @param val Output value, must be an lvalue of type uint32_t. - * @param GET_BYTE Expression reading one byte from the input. - * Evaluated up to 7 times (4 for the currently - * assigned Unicode range). With a memory buffer - * input, this could be *ptr++. - * @param ERROR Expression to be evaluated on invalid input, - * typically a goto statement. - * - * @warning ERROR should not contain a loop control statement which - * could interact with the internal while loop, and should force an - * exit from the macro code (e.g. through a goto or a return) in order - * to prevent undefined results. - */ -#define GET_UTF8(val, GET_BYTE, ERROR)\ - val= GET_BYTE;\ - {\ - uint32_t top = (val & 128) >> 1;\ - if ((val & 0xc0) == 0x80 || val >= 0xFE)\ - ERROR\ - while (val & top) {\ - int tmp= GET_BYTE - 128;\ - if(tmp>>6)\ - ERROR\ - val= (val<<6) + tmp;\ - top <<= 5;\ - }\ - val &= (top << 1) - 1;\ - } - -/** - * Convert a UTF-16 character (2 or 4 bytes) to its 32-bit UCS-4 encoded form. - * - * @param val Output value, must be an lvalue of type uint32_t. - * @param GET_16BIT Expression returning two bytes of UTF-16 data converted - * to native byte order. Evaluated one or two times. - * @param ERROR Expression to be evaluated on invalid input, - * typically a goto statement. - */ -#define GET_UTF16(val, GET_16BIT, ERROR)\ - val = GET_16BIT;\ - {\ - unsigned int hi = val - 0xD800;\ - if (hi < 0x800) {\ - val = GET_16BIT - 0xDC00;\ - if (val > 0x3FFU || hi > 0x3FFU)\ - ERROR\ - val += (hi<<10) + 0x10000;\ - }\ - }\ - -/** - * @def PUT_UTF8(val, tmp, PUT_BYTE) - * Convert a 32-bit Unicode character to its UTF-8 encoded form (up to 4 bytes long). - * @param val is an input-only argument and should be of type uint32_t. It holds - * a UCS-4 encoded Unicode character that is to be converted to UTF-8. If - * val is given as a function it is executed only once. - * @param tmp is a temporary variable and should be of type uint8_t. It - * represents an intermediate value during conversion that is to be - * output by PUT_BYTE. - * @param PUT_BYTE writes the converted UTF-8 bytes to any proper destination. - * It could be a function or a statement, and uses tmp as the input byte. - * For example, PUT_BYTE could be "*output++ = tmp;" PUT_BYTE will be - * executed up to 4 times for values in the valid UTF-8 range and up to - * 7 times in the general case, depending on the length of the converted - * Unicode character. - */ -#define PUT_UTF8(val, tmp, PUT_BYTE)\ - {\ - int bytes, shift;\ - uint32_t in = val;\ - if (in < 0x80) {\ - tmp = in;\ - PUT_BYTE\ - } else {\ - bytes = (av_log2(in) + 4) / 5;\ - shift = (bytes - 1) * 6;\ - tmp = (256 - (256 >> bytes)) | (in >> shift);\ - PUT_BYTE\ - while (shift >= 6) {\ - shift -= 6;\ - tmp = 0x80 | ((in >> shift) & 0x3f);\ - PUT_BYTE\ - }\ - }\ - } - -/** - * @def PUT_UTF16(val, tmp, PUT_16BIT) - * Convert a 32-bit Unicode character to its UTF-16 encoded form (2 or 4 bytes). - * @param val is an input-only argument and should be of type uint32_t. It holds - * a UCS-4 encoded Unicode character that is to be converted to UTF-16. If - * val is given as a function it is executed only once. - * @param tmp is a temporary variable and should be of type uint16_t. It - * represents an intermediate value during conversion that is to be - * output by PUT_16BIT. - * @param PUT_16BIT writes the converted UTF-16 data to any proper destination - * in desired endianness. It could be a function or a statement, and uses tmp - * as the input byte. For example, PUT_BYTE could be "*output++ = tmp;" - * PUT_BYTE will be executed 1 or 2 times depending on input character. - */ -#define PUT_UTF16(val, tmp, PUT_16BIT)\ - {\ - uint32_t in = val;\ - if (in < 0x10000) {\ - tmp = in;\ - PUT_16BIT\ - } else {\ - tmp = 0xD800 | ((in - 0x10000) >> 10);\ - PUT_16BIT\ - tmp = 0xDC00 | ((in - 0x10000) & 0x3FF);\ - PUT_16BIT\ - }\ - }\ - - - -#include "mem.h" - -#ifdef HAVE_AV_CONFIG_H -# include "internal.h" -#endif /* HAVE_AV_CONFIG_H */ - -#endif /* AVUTIL_COMMON_H */ - -/* - * The following definitions are outside the multiple inclusion guard - * to ensure they are immediately available in intmath.h. - */ - -#ifndef av_ceil_log2 -# define av_ceil_log2 av_ceil_log2_c -#endif -#ifndef av_clip -# define av_clip av_clip_c -#endif -#ifndef av_clip64 -# define av_clip64 av_clip64_c -#endif -#ifndef av_clip_uint8 -# define av_clip_uint8 av_clip_uint8_c -#endif -#ifndef av_clip_int8 -# define av_clip_int8 av_clip_int8_c -#endif -#ifndef av_clip_uint16 -# define av_clip_uint16 av_clip_uint16_c -#endif -#ifndef av_clip_int16 -# define av_clip_int16 av_clip_int16_c -#endif -#ifndef av_clipl_int32 -# define av_clipl_int32 av_clipl_int32_c -#endif -#ifndef av_clip_intp2 -# define av_clip_intp2 av_clip_intp2_c -#endif -#ifndef av_clip_uintp2 -# define av_clip_uintp2 av_clip_uintp2_c -#endif -#ifndef av_mod_uintp2 -# define av_mod_uintp2 av_mod_uintp2_c -#endif -#ifndef av_sat_add32 -# define av_sat_add32 av_sat_add32_c -#endif -#ifndef av_sat_dadd32 -# define av_sat_dadd32 av_sat_dadd32_c -#endif -#ifndef av_clipf -# define av_clipf av_clipf_c -#endif -#ifndef av_clipd -# define av_clipd av_clipd_c -#endif -#ifndef av_popcount -# define av_popcount av_popcount_c -#endif -#ifndef av_popcount64 -# define av_popcount64 av_popcount64_c -#endif diff --git a/src/main/jni/ffmpeg/include/libavutil/cpu.h b/src/main/jni/ffmpeg/include/libavutil/cpu.h deleted file mode 100644 index bff8518cc..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/cpu.h +++ /dev/null @@ -1,116 +0,0 @@ -/* - * Copyright (c) 2000, 2001, 2002 Fabrice Bellard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_CPU_H -#define AVUTIL_CPU_H - -#include "attributes.h" - -#define AV_CPU_FLAG_FORCE 0x80000000 /* force usage of selected flags (OR) */ - - /* lower 16 bits - CPU features */ -#define AV_CPU_FLAG_MMX 0x0001 ///< standard MMX -#define AV_CPU_FLAG_MMXEXT 0x0002 ///< SSE integer functions or AMD MMX ext -#define AV_CPU_FLAG_MMX2 0x0002 ///< SSE integer functions or AMD MMX ext -#define AV_CPU_FLAG_3DNOW 0x0004 ///< AMD 3DNOW -#define AV_CPU_FLAG_SSE 0x0008 ///< SSE functions -#define AV_CPU_FLAG_SSE2 0x0010 ///< PIV SSE2 functions -#define AV_CPU_FLAG_SSE2SLOW 0x40000000 ///< SSE2 supported, but usually not faster - ///< than regular MMX/SSE (e.g. Core1) -#define AV_CPU_FLAG_3DNOWEXT 0x0020 ///< AMD 3DNowExt -#define AV_CPU_FLAG_SSE3 0x0040 ///< Prescott SSE3 functions -#define AV_CPU_FLAG_SSE3SLOW 0x20000000 ///< SSE3 supported, but usually not faster - ///< than regular MMX/SSE (e.g. Core1) -#define AV_CPU_FLAG_SSSE3 0x0080 ///< Conroe SSSE3 functions -#define AV_CPU_FLAG_ATOM 0x10000000 ///< Atom processor, some SSSE3 instructions are slower -#define AV_CPU_FLAG_SSE4 0x0100 ///< Penryn SSE4.1 functions -#define AV_CPU_FLAG_SSE42 0x0200 ///< Nehalem SSE4.2 functions -#define AV_CPU_FLAG_AESNI 0x80000 ///< Advanced Encryption Standard functions -#define AV_CPU_FLAG_AVX 0x4000 ///< AVX functions: requires OS support even if YMM registers aren't used -#define AV_CPU_FLAG_AVXSLOW 0x8000000 ///< AVX supported, but slow when using YMM registers (e.g. Bulldozer) -#define AV_CPU_FLAG_XOP 0x0400 ///< Bulldozer XOP functions -#define AV_CPU_FLAG_FMA4 0x0800 ///< Bulldozer FMA4 functions -#define AV_CPU_FLAG_CMOV 0x1000 ///< supports cmov instruction -#define AV_CPU_FLAG_AVX2 0x8000 ///< AVX2 functions: requires OS support even if YMM registers aren't used -#define AV_CPU_FLAG_FMA3 0x10000 ///< Haswell FMA3 functions -#define AV_CPU_FLAG_BMI1 0x20000 ///< Bit Manipulation Instruction Set 1 -#define AV_CPU_FLAG_BMI2 0x40000 ///< Bit Manipulation Instruction Set 2 - -#define AV_CPU_FLAG_ALTIVEC 0x0001 ///< standard -#define AV_CPU_FLAG_VSX 0x0002 ///< ISA 2.06 -#define AV_CPU_FLAG_POWER8 0x0004 ///< ISA 2.07 - -#define AV_CPU_FLAG_ARMV5TE (1 << 0) -#define AV_CPU_FLAG_ARMV6 (1 << 1) -#define AV_CPU_FLAG_ARMV6T2 (1 << 2) -#define AV_CPU_FLAG_VFP (1 << 3) -#define AV_CPU_FLAG_VFPV3 (1 << 4) -#define AV_CPU_FLAG_NEON (1 << 5) -#define AV_CPU_FLAG_ARMV8 (1 << 6) -#define AV_CPU_FLAG_SETEND (1 <<16) - -/** - * Return the flags which specify extensions supported by the CPU. - * The returned value is affected by av_force_cpu_flags() if that was used - * before. So av_get_cpu_flags() can easily be used in a application to - * detect the enabled cpu flags. - */ -int av_get_cpu_flags(void); - -/** - * Disables cpu detection and forces the specified flags. - * -1 is a special case that disables forcing of specific flags. - */ -void av_force_cpu_flags(int flags); - -/** - * Set a mask on flags returned by av_get_cpu_flags(). - * This function is mainly useful for testing. - * Please use av_force_cpu_flags() and av_get_cpu_flags() instead which are more flexible - * - * @warning this function is not thread safe. - */ -attribute_deprecated void av_set_cpu_flags_mask(int mask); - -/** - * Parse CPU flags from a string. - * - * The returned flags contain the specified flags as well as related unspecified flags. - * - * This function exists only for compatibility with libav. - * Please use av_parse_cpu_caps() when possible. - * @return a combination of AV_CPU_* flags, negative on error. - */ -attribute_deprecated -int av_parse_cpu_flags(const char *s); - -/** - * Parse CPU caps from a string and update the given AV_CPU_* flags based on that. - * - * @return negative on error. - */ -int av_parse_cpu_caps(unsigned *flags, const char *s); - -/** - * @return the number of logical CPU cores present. - */ -int av_cpu_count(void); - -#endif /* AVUTIL_CPU_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/crc.h b/src/main/jni/ffmpeg/include/libavutil/crc.h deleted file mode 100644 index ef8a7137e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/crc.h +++ /dev/null @@ -1,91 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_CRC_H -#define AVUTIL_CRC_H - -#include <stdint.h> -#include <stddef.h> -#include "attributes.h" -#include "version.h" - -/** - * @defgroup lavu_crc32 CRC32 - * @ingroup lavu_crypto - * @{ - */ - -typedef uint32_t AVCRC; - -typedef enum { - AV_CRC_8_ATM, - AV_CRC_16_ANSI, - AV_CRC_16_CCITT, - AV_CRC_32_IEEE, - AV_CRC_32_IEEE_LE, /*< reversed bitorder version of AV_CRC_32_IEEE */ - AV_CRC_16_ANSI_LE, /*< reversed bitorder version of AV_CRC_16_ANSI */ -#if FF_API_CRC_BIG_TABLE - AV_CRC_24_IEEE = 12, -#else - AV_CRC_24_IEEE, -#endif /* FF_API_CRC_BIG_TABLE */ - AV_CRC_MAX, /*< Not part of public API! Do not use outside libavutil. */ -}AVCRCId; - -/** - * Initialize a CRC table. - * @param ctx must be an array of size sizeof(AVCRC)*257 or sizeof(AVCRC)*1024 - * @param le If 1, the lowest bit represents the coefficient for the highest - * exponent of the corresponding polynomial (both for poly and - * actual CRC). - * If 0, you must swap the CRC parameter and the result of av_crc - * if you need the standard representation (can be simplified in - * most cases to e.g. bswap16): - * av_bswap32(crc << (32-bits)) - * @param bits number of bits for the CRC - * @param poly generator polynomial without the x**bits coefficient, in the - * representation as specified by le - * @param ctx_size size of ctx in bytes - * @return <0 on failure - */ -int av_crc_init(AVCRC *ctx, int le, int bits, uint32_t poly, int ctx_size); - -/** - * Get an initialized standard CRC table. - * @param crc_id ID of a standard CRC - * @return a pointer to the CRC table or NULL on failure - */ -const AVCRC *av_crc_get_table(AVCRCId crc_id); - -/** - * Calculate the CRC of a block. - * @param crc CRC of previous blocks if any or initial value for CRC - * @return CRC updated with the data from the given block - * - * @see av_crc_init() "le" parameter - */ -uint32_t av_crc(const AVCRC *ctx, uint32_t crc, - const uint8_t *buffer, size_t length) av_pure; - -/** - * @} - */ - -#endif /* AVUTIL_CRC_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/des.h b/src/main/jni/ffmpeg/include/libavutil/des.h deleted file mode 100644 index 4cf11f5bc..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/des.h +++ /dev/null @@ -1,77 +0,0 @@ -/* - * DES encryption/decryption - * Copyright (c) 2007 Reimar Doeffinger - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_DES_H -#define AVUTIL_DES_H - -#include <stdint.h> - -/** - * @defgroup lavu_des DES - * @ingroup lavu_crypto - * @{ - */ - -typedef struct AVDES { - uint64_t round_keys[3][16]; - int triple_des; -} AVDES; - -/** - * Allocate an AVDES context. - */ -AVDES *av_des_alloc(void); - -/** - * @brief Initializes an AVDES context. - * - * @param key_bits must be 64 or 192 - * @param decrypt 0 for encryption/CBC-MAC, 1 for decryption - * @return zero on success, negative value otherwise - */ -int av_des_init(struct AVDES *d, const uint8_t *key, int key_bits, int decrypt); - -/** - * @brief Encrypts / decrypts using the DES algorithm. - * - * @param count number of 8 byte blocks - * @param dst destination array, can be equal to src, must be 8-byte aligned - * @param src source array, can be equal to dst, must be 8-byte aligned, may be NULL - * @param iv initialization vector for CBC mode, if NULL then ECB will be used, - * must be 8-byte aligned - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_des_crypt(struct AVDES *d, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt); - -/** - * @brief Calculates CBC-MAC using the DES algorithm. - * - * @param count number of 8 byte blocks - * @param dst destination array, can be equal to src, must be 8-byte aligned - * @param src source array, can be equal to dst, must be 8-byte aligned, may be NULL - */ -void av_des_mac(struct AVDES *d, uint8_t *dst, const uint8_t *src, int count); - -/** - * @} - */ - -#endif /* AVUTIL_DES_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/dict.h b/src/main/jni/ffmpeg/include/libavutil/dict.h deleted file mode 100644 index 5b8d00339..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/dict.h +++ /dev/null @@ -1,198 +0,0 @@ -/* - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * Public dictionary API. - * @deprecated - * AVDictionary is provided for compatibility with libav. It is both in - * implementation as well as API inefficient. It does not scale and is - * extremely slow with large dictionaries. - * It is recommended that new code uses our tree container from tree.c/h - * where applicable, which uses AVL trees to achieve O(log n) performance. - */ - -#ifndef AVUTIL_DICT_H -#define AVUTIL_DICT_H - -#include <stdint.h> - -#include "version.h" - -/** - * @addtogroup lavu_dict AVDictionary - * @ingroup lavu_data - * - * @brief Simple key:value store - * - * @{ - * Dictionaries are used for storing key:value pairs. To create - * an AVDictionary, simply pass an address of a NULL pointer to - * av_dict_set(). NULL can be used as an empty dictionary wherever - * a pointer to an AVDictionary is required. - * Use av_dict_get() to retrieve an entry or iterate over all - * entries and finally av_dict_free() to free the dictionary - * and all its contents. - * - @code - AVDictionary *d = NULL; // "create" an empty dictionary - AVDictionaryEntry *t = NULL; - - av_dict_set(&d, "foo", "bar", 0); // add an entry - - char *k = av_strdup("key"); // if your strings are already allocated, - char *v = av_strdup("value"); // you can avoid copying them like this - av_dict_set(&d, k, v, AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL); - - while (t = av_dict_get(d, "", t, AV_DICT_IGNORE_SUFFIX)) { - <....> // iterate over all entries in d - } - av_dict_free(&d); - @endcode - * - */ - -#define AV_DICT_MATCH_CASE 1 /**< Only get an entry with exact-case key match. Only relevant in av_dict_get(). */ -#define AV_DICT_IGNORE_SUFFIX 2 /**< Return first entry in a dictionary whose first part corresponds to the search key, - ignoring the suffix of the found key string. Only relevant in av_dict_get(). */ -#define AV_DICT_DONT_STRDUP_KEY 4 /**< Take ownership of a key that's been - allocated with av_malloc() or another memory allocation function. */ -#define AV_DICT_DONT_STRDUP_VAL 8 /**< Take ownership of a value that's been - allocated with av_malloc() or another memory allocation function. */ -#define AV_DICT_DONT_OVERWRITE 16 ///< Don't overwrite existing entries. -#define AV_DICT_APPEND 32 /**< If the entry already exists, append to it. Note that no - delimiter is added, the strings are simply concatenated. */ - -typedef struct AVDictionaryEntry { - char *key; - char *value; -} AVDictionaryEntry; - -typedef struct AVDictionary AVDictionary; - -/** - * Get a dictionary entry with matching key. - * - * The returned entry key or value must not be changed, or it will - * cause undefined behavior. - * - * To iterate through all the dictionary entries, you can set the matching key - * to the null string "" and set the AV_DICT_IGNORE_SUFFIX flag. - * - * @param prev Set to the previous matching element to find the next. - * If set to NULL the first matching element is returned. - * @param key matching key - * @param flags a collection of AV_DICT_* flags controlling how the entry is retrieved - * @return found entry or NULL in case no matching entry was found in the dictionary - */ -AVDictionaryEntry *av_dict_get(const AVDictionary *m, const char *key, - const AVDictionaryEntry *prev, int flags); - -/** - * Get number of entries in dictionary. - * - * @param m dictionary - * @return number of entries in dictionary - */ -int av_dict_count(const AVDictionary *m); - -/** - * Set the given entry in *pm, overwriting an existing entry. - * - * Note: If AV_DICT_DONT_STRDUP_KEY or AV_DICT_DONT_STRDUP_VAL is set, - * these arguments will be freed on error. - * - * @param pm pointer to a pointer to a dictionary struct. If *pm is NULL - * a dictionary struct is allocated and put in *pm. - * @param key entry key to add to *pm (will be av_strduped depending on flags) - * @param value entry value to add to *pm (will be av_strduped depending on flags). - * Passing a NULL value will cause an existing entry to be deleted. - * @return >= 0 on success otherwise an error code <0 - */ -int av_dict_set(AVDictionary **pm, const char *key, const char *value, int flags); - -/** - * Convenience wrapper for av_dict_set that converts the value to a string - * and stores it. - * - * Note: If AV_DICT_DONT_STRDUP_KEY is set, key will be freed on error. - */ -int av_dict_set_int(AVDictionary **pm, const char *key, int64_t value, int flags); - -/** - * Parse the key/value pairs list and add the parsed entries to a dictionary. - * - * In case of failure, all the successfully set entries are stored in - * *pm. You may need to manually free the created dictionary. - * - * @param key_val_sep a 0-terminated list of characters used to separate - * key from value - * @param pairs_sep a 0-terminated list of characters used to separate - * two pairs from each other - * @param flags flags to use when adding to dictionary. - * AV_DICT_DONT_STRDUP_KEY and AV_DICT_DONT_STRDUP_VAL - * are ignored since the key/value tokens will always - * be duplicated. - * @return 0 on success, negative AVERROR code on failure - */ -int av_dict_parse_string(AVDictionary **pm, const char *str, - const char *key_val_sep, const char *pairs_sep, - int flags); - -/** - * Copy entries from one AVDictionary struct into another. - * @param dst pointer to a pointer to a AVDictionary struct. If *dst is NULL, - * this function will allocate a struct for you and put it in *dst - * @param src pointer to source AVDictionary struct - * @param flags flags to use when setting entries in *dst - * @note metadata is read using the AV_DICT_IGNORE_SUFFIX flag - * @return 0 on success, negative AVERROR code on failure. If dst was allocated - * by this function, callers should free the associated memory. - */ -int av_dict_copy(AVDictionary **dst, const AVDictionary *src, int flags); - -/** - * Free all the memory allocated for an AVDictionary struct - * and all keys and values. - */ -void av_dict_free(AVDictionary **m); - -/** - * Get dictionary entries as a string. - * - * Create a string containing dictionary's entries. - * Such string may be passed back to av_dict_parse_string(). - * @note String is escaped with backslashes ('\'). - * - * @param[in] m dictionary - * @param[out] buffer Pointer to buffer that will be allocated with string containg entries. - * Buffer must be freed by the caller when is no longer needed. - * @param[in] key_val_sep character used to separate key from value - * @param[in] pairs_sep character used to separate two pairs from each other - * @return >= 0 on success, negative on error - * @warning Separators cannot be neither '\\' nor '\0'. They also cannot be the same. - */ -int av_dict_get_string(const AVDictionary *m, char **buffer, - const char key_val_sep, const char pairs_sep); - -/** - * @} - */ - -#endif /* AVUTIL_DICT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/display.h b/src/main/jni/ffmpeg/include/libavutil/display.h deleted file mode 100644 index c0cfee326..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/display.h +++ /dev/null @@ -1,86 +0,0 @@ -/* - * Copyright (c) 2014 Vittorio Giovara <vittorio.giovara@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_DISPLAY_H -#define AVUTIL_DISPLAY_H - -#include <stdint.h> - -/** - * The display transformation matrix specifies an affine transformation that - * should be applied to video frames for correct presentation. It is compatible - * with the matrices stored in the ISO/IEC 14496-12 container format. - * - * The data is a 3x3 matrix represented as a 9-element array: - * - * | a b u | - * (a, b, u, c, d, v, x, y, w) -> | c d v | - * | x y w | - * - * All numbers are stored in native endianness, as 16.16 fixed-point values, - * except for u, v and w, which are stored as 2.30 fixed-point values. - * - * The transformation maps a point (p, q) in the source (pre-transformation) - * frame to the point (p', q') in the destination (post-transformation) frame as - * follows: - * | a b u | - * (p, q, 1) . | c d v | = z * (p', q', 1) - * | x y w | - * - * The transformation can also be more explicitly written in components as - * follows: - * p' = (a * p + c * q + x) / z; - * q' = (b * p + d * q + y) / z; - * z = u * p + v * q + w - */ - -/** - * Extract the rotation component of the transformation matrix. - * - * @param matrix the transformation matrix - * @return the angle (in degrees) by which the transformation rotates the frame - * counterclockwise. The angle will be in range [-180.0, 180.0], - * or NaN if the matrix is singular. - * - * @note floating point numbers are inherently inexact, so callers are - * recommended to round the return value to nearest integer before use. - */ -double av_display_rotation_get(const int32_t matrix[9]); - -/** - * Initialize a transformation matrix describing a pure counterclockwise - * rotation by the specified angle (in degrees). - * - * @param matrix an allocated transformation matrix (will be fully overwritten - * by this function) - * @param angle rotation angle in degrees. - */ -void av_display_rotation_set(int32_t matrix[9], double angle); - -/** - * Flip the input matrix horizontally and/or vertically. - * - * @param matrix an allocated transformation matrix - * @param hflip whether the matrix should be flipped horizontally - * @param vflip whether the matrix should be flipped vertically - */ -void av_display_matrix_flip(int32_t matrix[9], int hflip, int vflip); - -#endif /* AVUTIL_DISPLAY_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/downmix_info.h b/src/main/jni/ffmpeg/include/libavutil/downmix_info.h deleted file mode 100644 index 221cf5bf9..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/downmix_info.h +++ /dev/null @@ -1,115 +0,0 @@ -/* - * Copyright (c) 2014 Tim Walker <tdskywalker@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_DOWNMIX_INFO_H -#define AVUTIL_DOWNMIX_INFO_H - -#include "frame.h" - -/** - * @file - * audio downmix medatata - */ - -/** - * @addtogroup lavu_audio - * @{ - */ - -/** - * @defgroup downmix_info Audio downmix metadata - * @{ - */ - -/** - * Possible downmix types. - */ -enum AVDownmixType { - AV_DOWNMIX_TYPE_UNKNOWN, /**< Not indicated. */ - AV_DOWNMIX_TYPE_LORO, /**< Lo/Ro 2-channel downmix (Stereo). */ - AV_DOWNMIX_TYPE_LTRT, /**< Lt/Rt 2-channel downmix, Dolby Surround compatible. */ - AV_DOWNMIX_TYPE_DPLII, /**< Lt/Rt 2-channel downmix, Dolby Pro Logic II compatible. */ - AV_DOWNMIX_TYPE_NB /**< Number of downmix types. Not part of ABI. */ -}; - -/** - * This structure describes optional metadata relevant to a downmix procedure. - * - * All fields are set by the decoder to the value indicated in the audio - * bitstream (if present), or to a "sane" default otherwise. - */ -typedef struct AVDownmixInfo { - /** - * Type of downmix preferred by the mastering engineer. - */ - enum AVDownmixType preferred_downmix_type; - - /** - * Absolute scale factor representing the nominal level of the center - * channel during a regular downmix. - */ - double center_mix_level; - - /** - * Absolute scale factor representing the nominal level of the center - * channel during an Lt/Rt compatible downmix. - */ - double center_mix_level_ltrt; - - /** - * Absolute scale factor representing the nominal level of the surround - * channels during a regular downmix. - */ - double surround_mix_level; - - /** - * Absolute scale factor representing the nominal level of the surround - * channels during an Lt/Rt compatible downmix. - */ - double surround_mix_level_ltrt; - - /** - * Absolute scale factor representing the level at which the LFE data is - * mixed into L/R channels during downmixing. - */ - double lfe_mix_level; -} AVDownmixInfo; - -/** - * Get a frame's AV_FRAME_DATA_DOWNMIX_INFO side data for editing. - * - * If the side data is absent, it is created and added to the frame. - * - * @param frame the frame for which the side data is to be obtained or created - * - * @return the AVDownmixInfo structure to be edited by the caller, or NULL if - * the structure cannot be allocated. - */ -AVDownmixInfo *av_downmix_info_update_side_data(AVFrame *frame); - -/** - * @} - */ - -/** - * @} - */ - -#endif /* AVUTIL_DOWNMIX_INFO_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/error.h b/src/main/jni/ffmpeg/include/libavutil/error.h deleted file mode 100644 index 71df4da35..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/error.h +++ /dev/null @@ -1,126 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * error code definitions - */ - -#ifndef AVUTIL_ERROR_H -#define AVUTIL_ERROR_H - -#include <errno.h> -#include <stddef.h> - -/** - * @addtogroup lavu_error - * - * @{ - */ - - -/* error handling */ -#if EDOM > 0 -#define AVERROR(e) (-(e)) ///< Returns a negative error code from a POSIX error code, to return from library functions. -#define AVUNERROR(e) (-(e)) ///< Returns a POSIX error code from a library function error return value. -#else -/* Some platforms have E* and errno already negated. */ -#define AVERROR(e) (e) -#define AVUNERROR(e) (e) -#endif - -#define FFERRTAG(a, b, c, d) (-(int)MKTAG(a, b, c, d)) - -#define AVERROR_BSF_NOT_FOUND FFERRTAG(0xF8,'B','S','F') ///< Bitstream filter not found -#define AVERROR_BUG FFERRTAG( 'B','U','G','!') ///< Internal bug, also see AVERROR_BUG2 -#define AVERROR_BUFFER_TOO_SMALL FFERRTAG( 'B','U','F','S') ///< Buffer too small -#define AVERROR_DECODER_NOT_FOUND FFERRTAG(0xF8,'D','E','C') ///< Decoder not found -#define AVERROR_DEMUXER_NOT_FOUND FFERRTAG(0xF8,'D','E','M') ///< Demuxer not found -#define AVERROR_ENCODER_NOT_FOUND FFERRTAG(0xF8,'E','N','C') ///< Encoder not found -#define AVERROR_EOF FFERRTAG( 'E','O','F',' ') ///< End of file -#define AVERROR_EXIT FFERRTAG( 'E','X','I','T') ///< Immediate exit was requested; the called function should not be restarted -#define AVERROR_EXTERNAL FFERRTAG( 'E','X','T',' ') ///< Generic error in an external library -#define AVERROR_FILTER_NOT_FOUND FFERRTAG(0xF8,'F','I','L') ///< Filter not found -#define AVERROR_INVALIDDATA FFERRTAG( 'I','N','D','A') ///< Invalid data found when processing input -#define AVERROR_MUXER_NOT_FOUND FFERRTAG(0xF8,'M','U','X') ///< Muxer not found -#define AVERROR_OPTION_NOT_FOUND FFERRTAG(0xF8,'O','P','T') ///< Option not found -#define AVERROR_PATCHWELCOME FFERRTAG( 'P','A','W','E') ///< Not yet implemented in FFmpeg, patches welcome -#define AVERROR_PROTOCOL_NOT_FOUND FFERRTAG(0xF8,'P','R','O') ///< Protocol not found - -#define AVERROR_STREAM_NOT_FOUND FFERRTAG(0xF8,'S','T','R') ///< Stream not found -/** - * This is semantically identical to AVERROR_BUG - * it has been introduced in Libav after our AVERROR_BUG and with a modified value. - */ -#define AVERROR_BUG2 FFERRTAG( 'B','U','G',' ') -#define AVERROR_UNKNOWN FFERRTAG( 'U','N','K','N') ///< Unknown error, typically from an external library -#define AVERROR_EXPERIMENTAL (-0x2bb2afa8) ///< Requested feature is flagged experimental. Set strict_std_compliance if you really want to use it. -#define AVERROR_INPUT_CHANGED (-0x636e6701) ///< Input changed between calls. Reconfiguration is required. (can be OR-ed with AVERROR_OUTPUT_CHANGED) -#define AVERROR_OUTPUT_CHANGED (-0x636e6702) ///< Output changed between calls. Reconfiguration is required. (can be OR-ed with AVERROR_INPUT_CHANGED) -/* HTTP & RTSP errors */ -#define AVERROR_HTTP_BAD_REQUEST FFERRTAG(0xF8,'4','0','0') -#define AVERROR_HTTP_UNAUTHORIZED FFERRTAG(0xF8,'4','0','1') -#define AVERROR_HTTP_FORBIDDEN FFERRTAG(0xF8,'4','0','3') -#define AVERROR_HTTP_NOT_FOUND FFERRTAG(0xF8,'4','0','4') -#define AVERROR_HTTP_OTHER_4XX FFERRTAG(0xF8,'4','X','X') -#define AVERROR_HTTP_SERVER_ERROR FFERRTAG(0xF8,'5','X','X') - -#define AV_ERROR_MAX_STRING_SIZE 64 - -/** - * Put a description of the AVERROR code errnum in errbuf. - * In case of failure the global variable errno is set to indicate the - * error. Even in case of failure av_strerror() will print a generic - * error message indicating the errnum provided to errbuf. - * - * @param errnum error code to describe - * @param errbuf buffer to which description is written - * @param errbuf_size the size in bytes of errbuf - * @return 0 on success, a negative value if a description for errnum - * cannot be found - */ -int av_strerror(int errnum, char *errbuf, size_t errbuf_size); - -/** - * Fill the provided buffer with a string containing an error string - * corresponding to the AVERROR code errnum. - * - * @param errbuf a buffer - * @param errbuf_size size in bytes of errbuf - * @param errnum error code to describe - * @return the buffer in input, filled with the error description - * @see av_strerror() - */ -static inline char *av_make_error_string(char *errbuf, size_t errbuf_size, int errnum) -{ - av_strerror(errnum, errbuf, errbuf_size); - return errbuf; -} - -/** - * Convenience macro, the return value should be used only directly in - * function arguments but never stand-alone. - */ -#define av_err2str(errnum) \ - av_make_error_string((char[AV_ERROR_MAX_STRING_SIZE]){0}, AV_ERROR_MAX_STRING_SIZE, errnum) - -/** - * @} - */ - -#endif /* AVUTIL_ERROR_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/eval.h b/src/main/jni/ffmpeg/include/libavutil/eval.h deleted file mode 100644 index dacd22b96..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/eval.h +++ /dev/null @@ -1,113 +0,0 @@ -/* - * Copyright (c) 2002 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * simple arithmetic expression evaluator - */ - -#ifndef AVUTIL_EVAL_H -#define AVUTIL_EVAL_H - -#include "avutil.h" - -typedef struct AVExpr AVExpr; - -/** - * Parse and evaluate an expression. - * Note, this is significantly slower than av_expr_eval(). - * - * @param res a pointer to a double where is put the result value of - * the expression, or NAN in case of error - * @param s expression as a zero terminated string, for example "1+2^3+5*5+sin(2/3)" - * @param const_names NULL terminated array of zero terminated strings of constant identifiers, for example {"PI", "E", 0} - * @param const_values a zero terminated array of values for the identifiers from const_names - * @param func1_names NULL terminated array of zero terminated strings of funcs1 identifiers - * @param funcs1 NULL terminated array of function pointers for functions which take 1 argument - * @param func2_names NULL terminated array of zero terminated strings of funcs2 identifiers - * @param funcs2 NULL terminated array of function pointers for functions which take 2 arguments - * @param opaque a pointer which will be passed to all functions from funcs1 and funcs2 - * @param log_ctx parent logging context - * @return >= 0 in case of success, a negative value corresponding to an - * AVERROR code otherwise - */ -int av_expr_parse_and_eval(double *res, const char *s, - const char * const *const_names, const double *const_values, - const char * const *func1_names, double (* const *funcs1)(void *, double), - const char * const *func2_names, double (* const *funcs2)(void *, double, double), - void *opaque, int log_offset, void *log_ctx); - -/** - * Parse an expression. - * - * @param expr a pointer where is put an AVExpr containing the parsed - * value in case of successful parsing, or NULL otherwise. - * The pointed to AVExpr must be freed with av_expr_free() by the user - * when it is not needed anymore. - * @param s expression as a zero terminated string, for example "1+2^3+5*5+sin(2/3)" - * @param const_names NULL terminated array of zero terminated strings of constant identifiers, for example {"PI", "E", 0} - * @param func1_names NULL terminated array of zero terminated strings of funcs1 identifiers - * @param funcs1 NULL terminated array of function pointers for functions which take 1 argument - * @param func2_names NULL terminated array of zero terminated strings of funcs2 identifiers - * @param funcs2 NULL terminated array of function pointers for functions which take 2 arguments - * @param log_ctx parent logging context - * @return >= 0 in case of success, a negative value corresponding to an - * AVERROR code otherwise - */ -int av_expr_parse(AVExpr **expr, const char *s, - const char * const *const_names, - const char * const *func1_names, double (* const *funcs1)(void *, double), - const char * const *func2_names, double (* const *funcs2)(void *, double, double), - int log_offset, void *log_ctx); - -/** - * Evaluate a previously parsed expression. - * - * @param const_values a zero terminated array of values for the identifiers from av_expr_parse() const_names - * @param opaque a pointer which will be passed to all functions from funcs1 and funcs2 - * @return the value of the expression - */ -double av_expr_eval(AVExpr *e, const double *const_values, void *opaque); - -/** - * Free a parsed expression previously created with av_expr_parse(). - */ -void av_expr_free(AVExpr *e); - -/** - * Parse the string in numstr and return its value as a double. If - * the string is empty, contains only whitespaces, or does not contain - * an initial substring that has the expected syntax for a - * floating-point number, no conversion is performed. In this case, - * returns a value of zero and the value returned in tail is the value - * of numstr. - * - * @param numstr a string representing a number, may contain one of - * the International System number postfixes, for example 'K', 'M', - * 'G'. If 'i' is appended after the postfix, powers of 2 are used - * instead of powers of 10. The 'B' postfix multiplies the value by - * 8, and can be appended after another postfix or used alone. This - * allows using for example 'KB', 'MiB', 'G' and 'B' as postfix. - * @param tail if non-NULL puts here the pointer to the char next - * after the last parsed character - */ -double av_strtod(const char *numstr, char **tail); - -#endif /* AVUTIL_EVAL_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/ffversion.h b/src/main/jni/ffmpeg/include/libavutil/ffversion.h deleted file mode 100644 index e84489812..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/ffversion.h +++ /dev/null @@ -1,5 +0,0 @@ -/* Automatically generated by version.sh, do not manually edit! */ -#ifndef AVUTIL_FFVERSION_H -#define AVUTIL_FFVERSION_H -#define FFMPEG_VERSION "2.8.git" -#endif /* AVUTIL_FFVERSION_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/fifo.h b/src/main/jni/ffmpeg/include/libavutil/fifo.h deleted file mode 100644 index dc7bc6f0d..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/fifo.h +++ /dev/null @@ -1,179 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * a very simple circular buffer FIFO implementation - */ - -#ifndef AVUTIL_FIFO_H -#define AVUTIL_FIFO_H - -#include <stdint.h> -#include "avutil.h" -#include "attributes.h" - -typedef struct AVFifoBuffer { - uint8_t *buffer; - uint8_t *rptr, *wptr, *end; - uint32_t rndx, wndx; -} AVFifoBuffer; - -/** - * Initialize an AVFifoBuffer. - * @param size of FIFO - * @return AVFifoBuffer or NULL in case of memory allocation failure - */ -AVFifoBuffer *av_fifo_alloc(unsigned int size); - -/** - * Initialize an AVFifoBuffer. - * @param nmemb number of elements - * @param size size of the single element - * @return AVFifoBuffer or NULL in case of memory allocation failure - */ -AVFifoBuffer *av_fifo_alloc_array(size_t nmemb, size_t size); - -/** - * Free an AVFifoBuffer. - * @param f AVFifoBuffer to free - */ -void av_fifo_free(AVFifoBuffer *f); - -/** - * Free an AVFifoBuffer and reset pointer to NULL. - * @param f AVFifoBuffer to free - */ -void av_fifo_freep(AVFifoBuffer **f); - -/** - * Reset the AVFifoBuffer to the state right after av_fifo_alloc, in particular it is emptied. - * @param f AVFifoBuffer to reset - */ -void av_fifo_reset(AVFifoBuffer *f); - -/** - * Return the amount of data in bytes in the AVFifoBuffer, that is the - * amount of data you can read from it. - * @param f AVFifoBuffer to read from - * @return size - */ -int av_fifo_size(const AVFifoBuffer *f); - -/** - * Return the amount of space in bytes in the AVFifoBuffer, that is the - * amount of data you can write into it. - * @param f AVFifoBuffer to write into - * @return size - */ -int av_fifo_space(const AVFifoBuffer *f); - -/** - * Feed data at specific position from an AVFifoBuffer to a user-supplied callback. - * Similar as av_fifo_gereric_read but without discarding data. - * @param f AVFifoBuffer to read from - * @param offset offset from current read position - * @param buf_size number of bytes to read - * @param func generic read function - * @param dest data destination - */ -int av_fifo_generic_peek_at(AVFifoBuffer *f, void *dest, int offset, int buf_size, void (*func)(void*, void*, int)); - -/** - * Feed data from an AVFifoBuffer to a user-supplied callback. - * Similar as av_fifo_gereric_read but without discarding data. - * @param f AVFifoBuffer to read from - * @param buf_size number of bytes to read - * @param func generic read function - * @param dest data destination - */ -int av_fifo_generic_peek(AVFifoBuffer *f, void *dest, int buf_size, void (*func)(void*, void*, int)); - -/** - * Feed data from an AVFifoBuffer to a user-supplied callback. - * @param f AVFifoBuffer to read from - * @param buf_size number of bytes to read - * @param func generic read function - * @param dest data destination - */ -int av_fifo_generic_read(AVFifoBuffer *f, void *dest, int buf_size, void (*func)(void*, void*, int)); - -/** - * Feed data from a user-supplied callback to an AVFifoBuffer. - * @param f AVFifoBuffer to write to - * @param src data source; non-const since it may be used as a - * modifiable context by the function defined in func - * @param size number of bytes to write - * @param func generic write function; the first parameter is src, - * the second is dest_buf, the third is dest_buf_size. - * func must return the number of bytes written to dest_buf, or <= 0 to - * indicate no more data available to write. - * If func is NULL, src is interpreted as a simple byte array for source data. - * @return the number of bytes written to the FIFO - */ -int av_fifo_generic_write(AVFifoBuffer *f, void *src, int size, int (*func)(void*, void*, int)); - -/** - * Resize an AVFifoBuffer. - * In case of reallocation failure, the old FIFO is kept unchanged. - * - * @param f AVFifoBuffer to resize - * @param size new AVFifoBuffer size in bytes - * @return <0 for failure, >=0 otherwise - */ -int av_fifo_realloc2(AVFifoBuffer *f, unsigned int size); - -/** - * Enlarge an AVFifoBuffer. - * In case of reallocation failure, the old FIFO is kept unchanged. - * The new fifo size may be larger than the requested size. - * - * @param f AVFifoBuffer to resize - * @param additional_space the amount of space in bytes to allocate in addition to av_fifo_size() - * @return <0 for failure, >=0 otherwise - */ -int av_fifo_grow(AVFifoBuffer *f, unsigned int additional_space); - -/** - * Read and discard the specified amount of data from an AVFifoBuffer. - * @param f AVFifoBuffer to read from - * @param size amount of data to read in bytes - */ -void av_fifo_drain(AVFifoBuffer *f, int size); - -/** - * Return a pointer to the data stored in a FIFO buffer at a certain offset. - * The FIFO buffer is not modified. - * - * @param f AVFifoBuffer to peek at, f must be non-NULL - * @param offs an offset in bytes, its absolute value must be less - * than the used buffer size or the returned pointer will - * point outside to the buffer data. - * The used buffer size can be checked with av_fifo_size(). - */ -static inline uint8_t *av_fifo_peek2(const AVFifoBuffer *f, int offs) -{ - uint8_t *ptr = f->rptr + offs; - if (ptr >= f->end) - ptr = f->buffer + (ptr - f->end); - else if (ptr < f->buffer) - ptr = f->end - (f->buffer - ptr); - return ptr; -} - -#endif /* AVUTIL_FIFO_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/file.h b/src/main/jni/ffmpeg/include/libavutil/file.h deleted file mode 100644 index e931be71e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/file.h +++ /dev/null @@ -1,68 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_FILE_H -#define AVUTIL_FILE_H - -#include <stdint.h> - -#include "avutil.h" - -/** - * @file - * Misc file utilities. - */ - -/** - * Read the file with name filename, and put its content in a newly - * allocated buffer or map it with mmap() when available. - * In case of success set *bufptr to the read or mmapped buffer, and - * *size to the size in bytes of the buffer in *bufptr. - * The returned buffer must be released with av_file_unmap(). - * - * @param log_offset loglevel offset used for logging - * @param log_ctx context used for logging - * @return a non negative number in case of success, a negative value - * corresponding to an AVERROR error code in case of failure - */ -av_warn_unused_result -int av_file_map(const char *filename, uint8_t **bufptr, size_t *size, - int log_offset, void *log_ctx); - -/** - * Unmap or free the buffer bufptr created by av_file_map(). - * - * @param size size in bytes of bufptr, must be the same as returned - * by av_file_map() - */ -void av_file_unmap(uint8_t *bufptr, size_t size); - -/** - * Wrapper to work around the lack of mkstemp() on mingw. - * Also, tries to create file in /tmp first, if possible. - * *prefix can be a character constant; *filename will be allocated internally. - * @return file descriptor of opened file (or negative value corresponding to an - * AVERROR code on error) - * and opened file name in **filename. - * @note On very old libcs it is necessary to set a secure umask before - * calling this, av_tempfile() can't call umask itself as it is used in - * libraries and could interfere with the calling application. - */ -int av_tempfile(const char *prefix, char **filename, int log_offset, void *log_ctx); - -#endif /* AVUTIL_FILE_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/frame.h b/src/main/jni/ffmpeg/include/libavutil/frame.h deleted file mode 100644 index 9c6061a11..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/frame.h +++ /dev/null @@ -1,713 +0,0 @@ -/* - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * @ingroup lavu_frame - * reference-counted frame API - */ - -#ifndef AVUTIL_FRAME_H -#define AVUTIL_FRAME_H - -#include <stdint.h> - -#include "avutil.h" -#include "buffer.h" -#include "dict.h" -#include "rational.h" -#include "samplefmt.h" -#include "pixfmt.h" -#include "version.h" - - -/** - * @defgroup lavu_frame AVFrame - * @ingroup lavu_data - * - * @{ - * AVFrame is an abstraction for reference-counted raw multimedia data. - */ - -enum AVFrameSideDataType { - /** - * The data is the AVPanScan struct defined in libavcodec. - */ - AV_FRAME_DATA_PANSCAN, - /** - * ATSC A53 Part 4 Closed Captions. - * A53 CC bitstream is stored as uint8_t in AVFrameSideData.data. - * The number of bytes of CC data is AVFrameSideData.size. - */ - AV_FRAME_DATA_A53_CC, - /** - * Stereoscopic 3d metadata. - * The data is the AVStereo3D struct defined in libavutil/stereo3d.h. - */ - AV_FRAME_DATA_STEREO3D, - /** - * The data is the AVMatrixEncoding enum defined in libavutil/channel_layout.h. - */ - AV_FRAME_DATA_MATRIXENCODING, - /** - * Metadata relevant to a downmix procedure. - * The data is the AVDownmixInfo struct defined in libavutil/downmix_info.h. - */ - AV_FRAME_DATA_DOWNMIX_INFO, - /** - * ReplayGain information in the form of the AVReplayGain struct. - */ - AV_FRAME_DATA_REPLAYGAIN, - /** - * This side data contains a 3x3 transformation matrix describing an affine - * transformation that needs to be applied to the frame for correct - * presentation. - * - * See libavutil/display.h for a detailed description of the data. - */ - AV_FRAME_DATA_DISPLAYMATRIX, - /** - * Active Format Description data consisting of a single byte as specified - * in ETSI TS 101 154 using AVActiveFormatDescription enum. - */ - AV_FRAME_DATA_AFD, - /** - * Motion vectors exported by some codecs (on demand through the export_mvs - * flag set in the libavcodec AVCodecContext flags2 option). - * The data is the AVMotionVector struct defined in - * libavutil/motion_vector.h. - */ - AV_FRAME_DATA_MOTION_VECTORS, - /** - * Recommmends skipping the specified number of samples. This is exported - * only if the "skip_manual" AVOption is set in libavcodec. - * This has the same format as AV_PKT_DATA_SKIP_SAMPLES. - * @code - * u32le number of samples to skip from start of this packet - * u32le number of samples to skip from end of this packet - * u8 reason for start skip - * u8 reason for end skip (0=padding silence, 1=convergence) - * @endcode - */ - AV_FRAME_DATA_SKIP_SAMPLES, - - /** - * This side data must be associated with an audio frame and corresponds to - * enum AVAudioServiceType defined in avcodec.h. - */ - AV_FRAME_DATA_AUDIO_SERVICE_TYPE, -}; - -enum AVActiveFormatDescription { - AV_AFD_SAME = 8, - AV_AFD_4_3 = 9, - AV_AFD_16_9 = 10, - AV_AFD_14_9 = 11, - AV_AFD_4_3_SP_14_9 = 13, - AV_AFD_16_9_SP_14_9 = 14, - AV_AFD_SP_4_3 = 15, -}; - - -/** - * Structure to hold side data for an AVFrame. - * - * sizeof(AVFrameSideData) is not a part of the public ABI, so new fields may be added - * to the end with a minor bump. - */ -typedef struct AVFrameSideData { - enum AVFrameSideDataType type; - uint8_t *data; - int size; - AVDictionary *metadata; - AVBufferRef *buf; -} AVFrameSideData; - -/** - * This structure describes decoded (raw) audio or video data. - * - * AVFrame must be allocated using av_frame_alloc(). Note that this only - * allocates the AVFrame itself, the buffers for the data must be managed - * through other means (see below). - * AVFrame must be freed with av_frame_free(). - * - * AVFrame is typically allocated once and then reused multiple times to hold - * different data (e.g. a single AVFrame to hold frames received from a - * decoder). In such a case, av_frame_unref() will free any references held by - * the frame and reset it to its original clean state before it - * is reused again. - * - * The data described by an AVFrame is usually reference counted through the - * AVBuffer API. The underlying buffer references are stored in AVFrame.buf / - * AVFrame.extended_buf. An AVFrame is considered to be reference counted if at - * least one reference is set, i.e. if AVFrame.buf[0] != NULL. In such a case, - * every single data plane must be contained in one of the buffers in - * AVFrame.buf or AVFrame.extended_buf. - * There may be a single buffer for all the data, or one separate buffer for - * each plane, or anything in between. - * - * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added - * to the end with a minor bump. - * Similarly fields that are marked as to be only accessed by - * av_opt_ptr() can be reordered. This allows 2 forks to add fields - * without breaking compatibility with each other. - */ -typedef struct AVFrame { -#define AV_NUM_DATA_POINTERS 8 - /** - * pointer to the picture/channel planes. - * This might be different from the first allocated byte - * - * Some decoders access areas outside 0,0 - width,height, please - * see avcodec_align_dimensions2(). Some filters and swscale can read - * up to 16 bytes beyond the planes, if these filters are to be used, - * then 16 extra bytes must be allocated. - */ - uint8_t *data[AV_NUM_DATA_POINTERS]; - - /** - * For video, size in bytes of each picture line. - * For audio, size in bytes of each plane. - * - * For audio, only linesize[0] may be set. For planar audio, each channel - * plane must be the same size. - * - * For video the linesizes should be multiples of the CPUs alignment - * preference, this is 16 or 32 for modern desktop CPUs. - * Some code requires such alignment other code can be slower without - * correct alignment, for yet other it makes no difference. - * - * @note The linesize may be larger than the size of usable data -- there - * may be extra padding present for performance reasons. - */ - int linesize[AV_NUM_DATA_POINTERS]; - - /** - * pointers to the data planes/channels. - * - * For video, this should simply point to data[]. - * - * For planar audio, each channel has a separate data pointer, and - * linesize[0] contains the size of each channel buffer. - * For packed audio, there is just one data pointer, and linesize[0] - * contains the total size of the buffer for all channels. - * - * Note: Both data and extended_data should always be set in a valid frame, - * but for planar audio with more channels that can fit in data, - * extended_data must be used in order to access all channels. - */ - uint8_t **extended_data; - - /** - * width and height of the video frame - */ - int width, height; - - /** - * number of audio samples (per channel) described by this frame - */ - int nb_samples; - - /** - * format of the frame, -1 if unknown or unset - * Values correspond to enum AVPixelFormat for video frames, - * enum AVSampleFormat for audio) - */ - int format; - - /** - * 1 -> keyframe, 0-> not - */ - int key_frame; - - /** - * Picture type of the frame. - */ - enum AVPictureType pict_type; - - /** - * Sample aspect ratio for the video frame, 0/1 if unknown/unspecified. - */ - AVRational sample_aspect_ratio; - - /** - * Presentation timestamp in time_base units (time when frame should be shown to user). - */ - int64_t pts; - - /** - * PTS copied from the AVPacket that was decoded to produce this frame. - */ - int64_t pkt_pts; - - /** - * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isn't used) - * This is also the Presentation time of this AVFrame calculated from - * only AVPacket.dts values without pts values. - */ - int64_t pkt_dts; - - /** - * picture number in bitstream order - */ - int coded_picture_number; - /** - * picture number in display order - */ - int display_picture_number; - - /** - * quality (between 1 (good) and FF_LAMBDA_MAX (bad)) - */ - int quality; - - /** - * for some private data of the user - */ - void *opaque; - -#if FF_API_ERROR_FRAME - /** - * @deprecated unused - */ - attribute_deprecated - uint64_t error[AV_NUM_DATA_POINTERS]; -#endif - - /** - * When decoding, this signals how much the picture must be delayed. - * extra_delay = repeat_pict / (2*fps) - */ - int repeat_pict; - - /** - * The content of the picture is interlaced. - */ - int interlaced_frame; - - /** - * If the content is interlaced, is top field displayed first. - */ - int top_field_first; - - /** - * Tell user application that palette has changed from previous frame. - */ - int palette_has_changed; - - /** - * reordered opaque 64bit (generally an integer or a double precision float - * PTS but can be anything). - * The user sets AVCodecContext.reordered_opaque to represent the input at - * that time, - * the decoder reorders values as needed and sets AVFrame.reordered_opaque - * to exactly one of the values provided by the user through AVCodecContext.reordered_opaque - * @deprecated in favor of pkt_pts - */ - int64_t reordered_opaque; - - /** - * Sample rate of the audio data. - */ - int sample_rate; - - /** - * Channel layout of the audio data. - */ - uint64_t channel_layout; - - /** - * AVBuffer references backing the data for this frame. If all elements of - * this array are NULL, then this frame is not reference counted. This array - * must be filled contiguously -- if buf[i] is non-NULL then buf[j] must - * also be non-NULL for all j < i. - * - * There may be at most one AVBuffer per data plane, so for video this array - * always contains all the references. For planar audio with more than - * AV_NUM_DATA_POINTERS channels, there may be more buffers than can fit in - * this array. Then the extra AVBufferRef pointers are stored in the - * extended_buf array. - */ - AVBufferRef *buf[AV_NUM_DATA_POINTERS]; - - /** - * For planar audio which requires more than AV_NUM_DATA_POINTERS - * AVBufferRef pointers, this array will hold all the references which - * cannot fit into AVFrame.buf. - * - * Note that this is different from AVFrame.extended_data, which always - * contains all the pointers. This array only contains the extra pointers, - * which cannot fit into AVFrame.buf. - * - * This array is always allocated using av_malloc() by whoever constructs - * the frame. It is freed in av_frame_unref(). - */ - AVBufferRef **extended_buf; - /** - * Number of elements in extended_buf. - */ - int nb_extended_buf; - - AVFrameSideData **side_data; - int nb_side_data; - -/** - * @defgroup lavu_frame_flags AV_FRAME_FLAGS - * Flags describing additional frame properties. - * - * @{ - */ - -/** - * The frame data may be corrupted, e.g. due to decoding errors. - */ -#define AV_FRAME_FLAG_CORRUPT (1 << 0) -/** - * @} - */ - - /** - * Frame flags, a combination of @ref lavu_frame_flags - */ - int flags; - - /** - * MPEG vs JPEG YUV range. - * It must be accessed using av_frame_get_color_range() and - * av_frame_set_color_range(). - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVColorRange color_range; - - enum AVColorPrimaries color_primaries; - - enum AVColorTransferCharacteristic color_trc; - - /** - * YUV colorspace type. - * It must be accessed using av_frame_get_colorspace() and - * av_frame_set_colorspace(). - * - encoding: Set by user - * - decoding: Set by libavcodec - */ - enum AVColorSpace colorspace; - - enum AVChromaLocation chroma_location; - - /** - * frame timestamp estimated using various heuristics, in stream time base - * Code outside libavutil should access this field using: - * av_frame_get_best_effort_timestamp(frame) - * - encoding: unused - * - decoding: set by libavcodec, read by user. - */ - int64_t best_effort_timestamp; - - /** - * reordered pos from the last AVPacket that has been input into the decoder - * Code outside libavutil should access this field using: - * av_frame_get_pkt_pos(frame) - * - encoding: unused - * - decoding: Read by user. - */ - int64_t pkt_pos; - - /** - * duration of the corresponding packet, expressed in - * AVStream->time_base units, 0 if unknown. - * Code outside libavutil should access this field using: - * av_frame_get_pkt_duration(frame) - * - encoding: unused - * - decoding: Read by user. - */ - int64_t pkt_duration; - - /** - * metadata. - * Code outside libavutil should access this field using: - * av_frame_get_metadata(frame) - * - encoding: Set by user. - * - decoding: Set by libavcodec. - */ - AVDictionary *metadata; - - /** - * decode error flags of the frame, set to a combination of - * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there - * were errors during the decoding. - * Code outside libavutil should access this field using: - * av_frame_get_decode_error_flags(frame) - * - encoding: unused - * - decoding: set by libavcodec, read by user. - */ - int decode_error_flags; -#define FF_DECODE_ERROR_INVALID_BITSTREAM 1 -#define FF_DECODE_ERROR_MISSING_REFERENCE 2 - - /** - * number of audio channels, only used for audio. - * Code outside libavutil should access this field using: - * av_frame_get_channels(frame) - * - encoding: unused - * - decoding: Read by user. - */ - int channels; - - /** - * size of the corresponding packet containing the compressed - * frame. It must be accessed using av_frame_get_pkt_size() and - * av_frame_set_pkt_size(). - * It is set to a negative value if unknown. - * - encoding: unused - * - decoding: set by libavcodec, read by user. - */ - int pkt_size; - -#if FF_API_FRAME_QP - /** - * QP table - * Not to be accessed directly from outside libavutil - */ - attribute_deprecated - int8_t *qscale_table; - /** - * QP store stride - * Not to be accessed directly from outside libavutil - */ - attribute_deprecated - int qstride; - - attribute_deprecated - int qscale_type; - - /** - * Not to be accessed directly from outside libavutil - */ - AVBufferRef *qp_table_buf; -#endif -} AVFrame; - -/** - * Accessors for some AVFrame fields. - * The position of these field in the structure is not part of the ABI, - * they should not be accessed directly outside libavutil. - */ -int64_t av_frame_get_best_effort_timestamp(const AVFrame *frame); -void av_frame_set_best_effort_timestamp(AVFrame *frame, int64_t val); -int64_t av_frame_get_pkt_duration (const AVFrame *frame); -void av_frame_set_pkt_duration (AVFrame *frame, int64_t val); -int64_t av_frame_get_pkt_pos (const AVFrame *frame); -void av_frame_set_pkt_pos (AVFrame *frame, int64_t val); -int64_t av_frame_get_channel_layout (const AVFrame *frame); -void av_frame_set_channel_layout (AVFrame *frame, int64_t val); -int av_frame_get_channels (const AVFrame *frame); -void av_frame_set_channels (AVFrame *frame, int val); -int av_frame_get_sample_rate (const AVFrame *frame); -void av_frame_set_sample_rate (AVFrame *frame, int val); -AVDictionary *av_frame_get_metadata (const AVFrame *frame); -void av_frame_set_metadata (AVFrame *frame, AVDictionary *val); -int av_frame_get_decode_error_flags (const AVFrame *frame); -void av_frame_set_decode_error_flags (AVFrame *frame, int val); -int av_frame_get_pkt_size(const AVFrame *frame); -void av_frame_set_pkt_size(AVFrame *frame, int val); -AVDictionary **avpriv_frame_get_metadatap(AVFrame *frame); -#if FF_API_FRAME_QP -int8_t *av_frame_get_qp_table(AVFrame *f, int *stride, int *type); -int av_frame_set_qp_table(AVFrame *f, AVBufferRef *buf, int stride, int type); -#endif -enum AVColorSpace av_frame_get_colorspace(const AVFrame *frame); -void av_frame_set_colorspace(AVFrame *frame, enum AVColorSpace val); -enum AVColorRange av_frame_get_color_range(const AVFrame *frame); -void av_frame_set_color_range(AVFrame *frame, enum AVColorRange val); - -/** - * Get the name of a colorspace. - * @return a static string identifying the colorspace; can be NULL. - */ -const char *av_get_colorspace_name(enum AVColorSpace val); - -/** - * Allocate an AVFrame and set its fields to default values. The resulting - * struct must be freed using av_frame_free(). - * - * @return An AVFrame filled with default values or NULL on failure. - * - * @note this only allocates the AVFrame itself, not the data buffers. Those - * must be allocated through other means, e.g. with av_frame_get_buffer() or - * manually. - */ -AVFrame *av_frame_alloc(void); - -/** - * Free the frame and any dynamically allocated objects in it, - * e.g. extended_data. If the frame is reference counted, it will be - * unreferenced first. - * - * @param frame frame to be freed. The pointer will be set to NULL. - */ -void av_frame_free(AVFrame **frame); - -/** - * Set up a new reference to the data described by the source frame. - * - * Copy frame properties from src to dst and create a new reference for each - * AVBufferRef from src. - * - * If src is not reference counted, new buffers are allocated and the data is - * copied. - * - * @return 0 on success, a negative AVERROR on error - */ -int av_frame_ref(AVFrame *dst, const AVFrame *src); - -/** - * Create a new frame that references the same data as src. - * - * This is a shortcut for av_frame_alloc()+av_frame_ref(). - * - * @return newly created AVFrame on success, NULL on error. - */ -AVFrame *av_frame_clone(const AVFrame *src); - -/** - * Unreference all the buffers referenced by frame and reset the frame fields. - */ -void av_frame_unref(AVFrame *frame); - -/** - * Move everything contained in src to dst and reset src. - */ -void av_frame_move_ref(AVFrame *dst, AVFrame *src); - -/** - * Allocate new buffer(s) for audio or video data. - * - * The following fields must be set on frame before calling this function: - * - format (pixel format for video, sample format for audio) - * - width and height for video - * - nb_samples and channel_layout for audio - * - * This function will fill AVFrame.data and AVFrame.buf arrays and, if - * necessary, allocate and fill AVFrame.extended_data and AVFrame.extended_buf. - * For planar formats, one buffer will be allocated for each plane. - * - * @param frame frame in which to store the new buffers. - * @param align required buffer size alignment - * - * @return 0 on success, a negative AVERROR on error. - */ -int av_frame_get_buffer(AVFrame *frame, int align); - -/** - * Check if the frame data is writable. - * - * @return A positive value if the frame data is writable (which is true if and - * only if each of the underlying buffers has only one reference, namely the one - * stored in this frame). Return 0 otherwise. - * - * If 1 is returned the answer is valid until av_buffer_ref() is called on any - * of the underlying AVBufferRefs (e.g. through av_frame_ref() or directly). - * - * @see av_frame_make_writable(), av_buffer_is_writable() - */ -int av_frame_is_writable(AVFrame *frame); - -/** - * Ensure that the frame data is writable, avoiding data copy if possible. - * - * Do nothing if the frame is writable, allocate new buffers and copy the data - * if it is not. - * - * @return 0 on success, a negative AVERROR on error. - * - * @see av_frame_is_writable(), av_buffer_is_writable(), - * av_buffer_make_writable() - */ -int av_frame_make_writable(AVFrame *frame); - -/** - * Copy the frame data from src to dst. - * - * This function does not allocate anything, dst must be already initialized and - * allocated with the same parameters as src. - * - * This function only copies the frame data (i.e. the contents of the data / - * extended data arrays), not any other properties. - * - * @return >= 0 on success, a negative AVERROR on error. - */ -int av_frame_copy(AVFrame *dst, const AVFrame *src); - -/** - * Copy only "metadata" fields from src to dst. - * - * Metadata for the purpose of this function are those fields that do not affect - * the data layout in the buffers. E.g. pts, sample rate (for audio) or sample - * aspect ratio (for video), but not width/height or channel layout. - * Side data is also copied. - */ -int av_frame_copy_props(AVFrame *dst, const AVFrame *src); - -/** - * Get the buffer reference a given data plane is stored in. - * - * @param plane index of the data plane of interest in frame->extended_data. - * - * @return the buffer reference that contains the plane or NULL if the input - * frame is not valid. - */ -AVBufferRef *av_frame_get_plane_buffer(AVFrame *frame, int plane); - -/** - * Add a new side data to a frame. - * - * @param frame a frame to which the side data should be added - * @param type type of the added side data - * @param size size of the side data - * - * @return newly added side data on success, NULL on error - */ -AVFrameSideData *av_frame_new_side_data(AVFrame *frame, - enum AVFrameSideDataType type, - int size); - -/** - * @return a pointer to the side data of a given type on success, NULL if there - * is no side data with such type in this frame. - */ -AVFrameSideData *av_frame_get_side_data(const AVFrame *frame, - enum AVFrameSideDataType type); - -/** - * If side data of the supplied type exists in the frame, free it and remove it - * from the frame. - */ -void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type); - -/** - * @return a string identifying the side data type - */ -const char *av_frame_side_data_name(enum AVFrameSideDataType type); - -/** - * @} - */ - -#endif /* AVUTIL_FRAME_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/hash.h b/src/main/jni/ffmpeg/include/libavutil/hash.h deleted file mode 100644 index d4bcbf8cc..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/hash.h +++ /dev/null @@ -1,112 +0,0 @@ -/* - * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_HASH_H -#define AVUTIL_HASH_H - -#include <stdint.h> - -struct AVHashContext; - -/** - * Allocate a hash context for the algorithm specified by name. - * - * @return >= 0 for success, a negative error code for failure - * @note The context is not initialized, you must call av_hash_init(). - */ -int av_hash_alloc(struct AVHashContext **ctx, const char *name); - -/** - * Get the names of available hash algorithms. - * - * This function can be used to enumerate the algorithms. - * - * @param i index of the hash algorithm, starting from 0 - * @return a pointer to a static string or NULL if i is out of range - */ -const char *av_hash_names(int i); - -/** - * Get the name of the algorithm corresponding to the given hash context. - */ -const char *av_hash_get_name(const struct AVHashContext *ctx); - -/** - * Maximum value that av_hash_get_size will currently return. - * - * You can use this if you absolutely want or need to use static allocation - * and are fine with not supporting hashes newly added to libavutil without - * recompilation. - * Note that you still need to check against av_hash_get_size, adding new hashes - * with larger sizes will not be considered an ABI change and should not cause - * your code to overflow a buffer. - */ -#define AV_HASH_MAX_SIZE 64 - -/** - * Get the size of the resulting hash value in bytes. - * - * The pointer passed to av_hash_final have space for at least this many bytes. - */ -int av_hash_get_size(const struct AVHashContext *ctx); - -/** - * Initialize or reset a hash context. - */ -void av_hash_init(struct AVHashContext *ctx); - -/** - * Update a hash context with additional data. - */ -void av_hash_update(struct AVHashContext *ctx, const uint8_t *src, int len); - -/** - * Finalize a hash context and compute the actual hash value. - */ -void av_hash_final(struct AVHashContext *ctx, uint8_t *dst); - -/** - * Finalize a hash context and compute the actual hash value. - * If size is smaller than the hash size, the hash is truncated; - * if size is larger, the buffer is padded with 0. - */ -void av_hash_final_bin(struct AVHashContext *ctx, uint8_t *dst, int size); - -/** - * Finalize a hash context and compute the actual hash value as a hex string. - * The string is always 0-terminated. - * If size is smaller than 2 * hash_size + 1, the hex string is truncated. - */ -void av_hash_final_hex(struct AVHashContext *ctx, uint8_t *dst, int size); - -/** - * Finalize a hash context and compute the actual hash value as a base64 string. - * The string is always 0-terminated. - * If size is smaller than AV_BASE64_SIZE(hash_size), the base64 string is - * truncated. - */ -void av_hash_final_b64(struct AVHashContext *ctx, uint8_t *dst, int size); - -/** - * Free hash context. - */ -void av_hash_freep(struct AVHashContext **ctx); - -#endif /* AVUTIL_HASH_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/hmac.h b/src/main/jni/ffmpeg/include/libavutil/hmac.h deleted file mode 100644 index 576a0a4fb..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/hmac.h +++ /dev/null @@ -1,100 +0,0 @@ -/* - * Copyright (C) 2012 Martin Storsjo - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_HMAC_H -#define AVUTIL_HMAC_H - -#include <stdint.h> - -#include "version.h" -/** - * @defgroup lavu_hmac HMAC - * @ingroup lavu_crypto - * @{ - */ - -enum AVHMACType { - AV_HMAC_MD5, - AV_HMAC_SHA1, - AV_HMAC_SHA224, - AV_HMAC_SHA256, - AV_HMAC_SHA384 = 12, - AV_HMAC_SHA512, -}; - -typedef struct AVHMAC AVHMAC; - -/** - * Allocate an AVHMAC context. - * @param type The hash function used for the HMAC. - */ -AVHMAC *av_hmac_alloc(enum AVHMACType type); - -/** - * Free an AVHMAC context. - * @param ctx The context to free, may be NULL - */ -void av_hmac_free(AVHMAC *ctx); - -/** - * Initialize an AVHMAC context with an authentication key. - * @param ctx The HMAC context - * @param key The authentication key - * @param keylen The length of the key, in bytes - */ -void av_hmac_init(AVHMAC *ctx, const uint8_t *key, unsigned int keylen); - -/** - * Hash data with the HMAC. - * @param ctx The HMAC context - * @param data The data to hash - * @param len The length of the data, in bytes - */ -void av_hmac_update(AVHMAC *ctx, const uint8_t *data, unsigned int len); - -/** - * Finish hashing and output the HMAC digest. - * @param ctx The HMAC context - * @param out The output buffer to write the digest into - * @param outlen The length of the out buffer, in bytes - * @return The number of bytes written to out, or a negative error code. - */ -int av_hmac_final(AVHMAC *ctx, uint8_t *out, unsigned int outlen); - -/** - * Hash an array of data with a key. - * @param ctx The HMAC context - * @param data The data to hash - * @param len The length of the data, in bytes - * @param key The authentication key - * @param keylen The length of the key, in bytes - * @param out The output buffer to write the digest into - * @param outlen The length of the out buffer, in bytes - * @return The number of bytes written to out, or a negative error code. - */ -int av_hmac_calc(AVHMAC *ctx, const uint8_t *data, unsigned int len, - const uint8_t *key, unsigned int keylen, - uint8_t *out, unsigned int outlen); - -/** - * @} - */ - -#endif /* AVUTIL_HMAC_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/imgutils.h b/src/main/jni/ffmpeg/include/libavutil/imgutils.h deleted file mode 100644 index 23282a38f..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/imgutils.h +++ /dev/null @@ -1,213 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_IMGUTILS_H -#define AVUTIL_IMGUTILS_H - -/** - * @file - * misc image utilities - * - * @addtogroup lavu_picture - * @{ - */ - -#include "avutil.h" -#include "pixdesc.h" -#include "rational.h" - -/** - * Compute the max pixel step for each plane of an image with a - * format described by pixdesc. - * - * The pixel step is the distance in bytes between the first byte of - * the group of bytes which describe a pixel component and the first - * byte of the successive group in the same plane for the same - * component. - * - * @param max_pixsteps an array which is filled with the max pixel step - * for each plane. Since a plane may contain different pixel - * components, the computed max_pixsteps[plane] is relative to the - * component in the plane with the max pixel step. - * @param max_pixstep_comps an array which is filled with the component - * for each plane which has the max pixel step. May be NULL. - */ -void av_image_fill_max_pixsteps(int max_pixsteps[4], int max_pixstep_comps[4], - const AVPixFmtDescriptor *pixdesc); - -/** - * Compute the size of an image line with format pix_fmt and width - * width for the plane plane. - * - * @return the computed size in bytes - */ -int av_image_get_linesize(enum AVPixelFormat pix_fmt, int width, int plane); - -/** - * Fill plane linesizes for an image with pixel format pix_fmt and - * width width. - * - * @param linesizes array to be filled with the linesize for each plane - * @return >= 0 in case of success, a negative error code otherwise - */ -int av_image_fill_linesizes(int linesizes[4], enum AVPixelFormat pix_fmt, int width); - -/** - * Fill plane data pointers for an image with pixel format pix_fmt and - * height height. - * - * @param data pointers array to be filled with the pointer for each image plane - * @param ptr the pointer to a buffer which will contain the image - * @param linesizes the array containing the linesize for each - * plane, should be filled by av_image_fill_linesizes() - * @return the size in bytes required for the image buffer, a negative - * error code in case of failure - */ -int av_image_fill_pointers(uint8_t *data[4], enum AVPixelFormat pix_fmt, int height, - uint8_t *ptr, const int linesizes[4]); - -/** - * Allocate an image with size w and h and pixel format pix_fmt, and - * fill pointers and linesizes accordingly. - * The allocated image buffer has to be freed by using - * av_freep(&pointers[0]). - * - * @param align the value to use for buffer size alignment - * @return the size in bytes required for the image buffer, a negative - * error code in case of failure - */ -int av_image_alloc(uint8_t *pointers[4], int linesizes[4], - int w, int h, enum AVPixelFormat pix_fmt, int align); - -/** - * Copy image plane from src to dst. - * That is, copy "height" number of lines of "bytewidth" bytes each. - * The first byte of each successive line is separated by *_linesize - * bytes. - * - * bytewidth must be contained by both absolute values of dst_linesize - * and src_linesize, otherwise the function behavior is undefined. - * - * @param dst_linesize linesize for the image plane in dst - * @param src_linesize linesize for the image plane in src - */ -void av_image_copy_plane(uint8_t *dst, int dst_linesize, - const uint8_t *src, int src_linesize, - int bytewidth, int height); - -/** - * Copy image in src_data to dst_data. - * - * @param dst_linesizes linesizes for the image in dst_data - * @param src_linesizes linesizes for the image in src_data - */ -void av_image_copy(uint8_t *dst_data[4], int dst_linesizes[4], - const uint8_t *src_data[4], const int src_linesizes[4], - enum AVPixelFormat pix_fmt, int width, int height); - -/** - * Setup the data pointers and linesizes based on the specified image - * parameters and the provided array. - * - * The fields of the given image are filled in by using the src - * address which points to the image data buffer. Depending on the - * specified pixel format, one or multiple image data pointers and - * line sizes will be set. If a planar format is specified, several - * pointers will be set pointing to the different picture planes and - * the line sizes of the different planes will be stored in the - * lines_sizes array. Call with src == NULL to get the required - * size for the src buffer. - * - * To allocate the buffer and fill in the dst_data and dst_linesize in - * one call, use av_image_alloc(). - * - * @param dst_data data pointers to be filled in - * @param dst_linesizes linesizes for the image in dst_data to be filled in - * @param src buffer which will contain or contains the actual image data, can be NULL - * @param pix_fmt the pixel format of the image - * @param width the width of the image in pixels - * @param height the height of the image in pixels - * @param align the value used in src for linesize alignment - * @return the size in bytes required for src, a negative error code - * in case of failure - */ -int av_image_fill_arrays(uint8_t *dst_data[4], int dst_linesize[4], - const uint8_t *src, - enum AVPixelFormat pix_fmt, int width, int height, int align); - -/** - * Return the size in bytes of the amount of data required to store an - * image with the given parameters. - * - * @param[in] align the assumed linesize alignment - */ -int av_image_get_buffer_size(enum AVPixelFormat pix_fmt, int width, int height, int align); - -/** - * Copy image data from an image into a buffer. - * - * av_image_get_buffer_size() can be used to compute the required size - * for the buffer to fill. - * - * @param dst a buffer into which picture data will be copied - * @param dst_size the size in bytes of dst - * @param src_data pointers containing the source image data - * @param src_linesizes linesizes for the image in src_data - * @param pix_fmt the pixel format of the source image - * @param width the width of the source image in pixels - * @param height the height of the source image in pixels - * @param align the assumed linesize alignment for dst - * @return the number of bytes written to dst, or a negative value - * (error code) on error - */ -int av_image_copy_to_buffer(uint8_t *dst, int dst_size, - const uint8_t * const src_data[4], const int src_linesize[4], - enum AVPixelFormat pix_fmt, int width, int height, int align); - -/** - * Check if the given dimension of an image is valid, meaning that all - * bytes of the image can be addressed with a signed int. - * - * @param w the width of the picture - * @param h the height of the picture - * @param log_offset the offset to sum to the log level for logging with log_ctx - * @param log_ctx the parent logging context, it may be NULL - * @return >= 0 if valid, a negative error code otherwise - */ -int av_image_check_size(unsigned int w, unsigned int h, int log_offset, void *log_ctx); - -/** - * Check if the given sample aspect ratio of an image is valid. - * - * It is considered invalid if the denominator is 0 or if applying the ratio - * to the image size would make the smaller dimension less than 1. If the - * sar numerator is 0, it is considered unknown and will return as valid. - * - * @param w width of the image - * @param h height of the image - * @param sar sample aspect ratio of the image - * @return 0 if valid, a negative AVERROR code otherwise - */ -int av_image_check_sar(unsigned int w, unsigned int h, AVRational sar); - -/** - * @} - */ - - -#endif /* AVUTIL_IMGUTILS_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/intfloat.h b/src/main/jni/ffmpeg/include/libavutil/intfloat.h deleted file mode 100644 index fe3d7ec4a..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/intfloat.h +++ /dev/null @@ -1,77 +0,0 @@ -/* - * Copyright (c) 2011 Mans Rullgard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_INTFLOAT_H -#define AVUTIL_INTFLOAT_H - -#include <stdint.h> -#include "attributes.h" - -union av_intfloat32 { - uint32_t i; - float f; -}; - -union av_intfloat64 { - uint64_t i; - double f; -}; - -/** - * Reinterpret a 32-bit integer as a float. - */ -static av_always_inline float av_int2float(uint32_t i) -{ - union av_intfloat32 v; - v.i = i; - return v.f; -} - -/** - * Reinterpret a float as a 32-bit integer. - */ -static av_always_inline uint32_t av_float2int(float f) -{ - union av_intfloat32 v; - v.f = f; - return v.i; -} - -/** - * Reinterpret a 64-bit integer as a double. - */ -static av_always_inline double av_int2double(uint64_t i) -{ - union av_intfloat64 v; - v.i = i; - return v.f; -} - -/** - * Reinterpret a double as a 64-bit integer. - */ -static av_always_inline uint64_t av_double2int(double f) -{ - union av_intfloat64 v; - v.f = f; - return v.i; -} - -#endif /* AVUTIL_INTFLOAT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/intreadwrite.h b/src/main/jni/ffmpeg/include/libavutil/intreadwrite.h deleted file mode 100644 index 51fbe30a2..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/intreadwrite.h +++ /dev/null @@ -1,629 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_INTREADWRITE_H -#define AVUTIL_INTREADWRITE_H - -#include <stdint.h> -#include "libavutil/avconfig.h" -#include "attributes.h" -#include "bswap.h" - -typedef union { - uint64_t u64; - uint32_t u32[2]; - uint16_t u16[4]; - uint8_t u8 [8]; - double f64; - float f32[2]; -} av_alias av_alias64; - -typedef union { - uint32_t u32; - uint16_t u16[2]; - uint8_t u8 [4]; - float f32; -} av_alias av_alias32; - -typedef union { - uint16_t u16; - uint8_t u8 [2]; -} av_alias av_alias16; - -/* - * Arch-specific headers can provide any combination of - * AV_[RW][BLN](16|24|32|48|64) and AV_(COPY|SWAP|ZERO)(64|128) macros. - * Preprocessor symbols must be defined, even if these are implemented - * as inline functions. - * - * R/W means read/write, B/L/N means big/little/native endianness. - * The following macros require aligned access, compared to their - * unaligned variants: AV_(COPY|SWAP|ZERO)(64|128), AV_[RW]N[8-64]A. - * Incorrect usage may range from abysmal performance to crash - * depending on the platform. - * - * The unaligned variants are AV_[RW][BLN][8-64] and AV_COPY*U. - */ - -#ifdef HAVE_AV_CONFIG_H - -#include "config.h" - -#if ARCH_ARM -# include "arm/intreadwrite.h" -#elif ARCH_AVR32 -# include "avr32/intreadwrite.h" -#elif ARCH_MIPS -# include "mips/intreadwrite.h" -#elif ARCH_PPC -# include "ppc/intreadwrite.h" -#elif ARCH_TOMI -# include "tomi/intreadwrite.h" -#elif ARCH_X86 -# include "x86/intreadwrite.h" -#endif - -#endif /* HAVE_AV_CONFIG_H */ - -/* - * Map AV_RNXX <-> AV_R[BL]XX for all variants provided by per-arch headers. - */ - -#if AV_HAVE_BIGENDIAN - -# if defined(AV_RN16) && !defined(AV_RB16) -# define AV_RB16(p) AV_RN16(p) -# elif !defined(AV_RN16) && defined(AV_RB16) -# define AV_RN16(p) AV_RB16(p) -# endif - -# if defined(AV_WN16) && !defined(AV_WB16) -# define AV_WB16(p, v) AV_WN16(p, v) -# elif !defined(AV_WN16) && defined(AV_WB16) -# define AV_WN16(p, v) AV_WB16(p, v) -# endif - -# if defined(AV_RN24) && !defined(AV_RB24) -# define AV_RB24(p) AV_RN24(p) -# elif !defined(AV_RN24) && defined(AV_RB24) -# define AV_RN24(p) AV_RB24(p) -# endif - -# if defined(AV_WN24) && !defined(AV_WB24) -# define AV_WB24(p, v) AV_WN24(p, v) -# elif !defined(AV_WN24) && defined(AV_WB24) -# define AV_WN24(p, v) AV_WB24(p, v) -# endif - -# if defined(AV_RN32) && !defined(AV_RB32) -# define AV_RB32(p) AV_RN32(p) -# elif !defined(AV_RN32) && defined(AV_RB32) -# define AV_RN32(p) AV_RB32(p) -# endif - -# if defined(AV_WN32) && !defined(AV_WB32) -# define AV_WB32(p, v) AV_WN32(p, v) -# elif !defined(AV_WN32) && defined(AV_WB32) -# define AV_WN32(p, v) AV_WB32(p, v) -# endif - -# if defined(AV_RN48) && !defined(AV_RB48) -# define AV_RB48(p) AV_RN48(p) -# elif !defined(AV_RN48) && defined(AV_RB48) -# define AV_RN48(p) AV_RB48(p) -# endif - -# if defined(AV_WN48) && !defined(AV_WB48) -# define AV_WB48(p, v) AV_WN48(p, v) -# elif !defined(AV_WN48) && defined(AV_WB48) -# define AV_WN48(p, v) AV_WB48(p, v) -# endif - -# if defined(AV_RN64) && !defined(AV_RB64) -# define AV_RB64(p) AV_RN64(p) -# elif !defined(AV_RN64) && defined(AV_RB64) -# define AV_RN64(p) AV_RB64(p) -# endif - -# if defined(AV_WN64) && !defined(AV_WB64) -# define AV_WB64(p, v) AV_WN64(p, v) -# elif !defined(AV_WN64) && defined(AV_WB64) -# define AV_WN64(p, v) AV_WB64(p, v) -# endif - -#else /* AV_HAVE_BIGENDIAN */ - -# if defined(AV_RN16) && !defined(AV_RL16) -# define AV_RL16(p) AV_RN16(p) -# elif !defined(AV_RN16) && defined(AV_RL16) -# define AV_RN16(p) AV_RL16(p) -# endif - -# if defined(AV_WN16) && !defined(AV_WL16) -# define AV_WL16(p, v) AV_WN16(p, v) -# elif !defined(AV_WN16) && defined(AV_WL16) -# define AV_WN16(p, v) AV_WL16(p, v) -# endif - -# if defined(AV_RN24) && !defined(AV_RL24) -# define AV_RL24(p) AV_RN24(p) -# elif !defined(AV_RN24) && defined(AV_RL24) -# define AV_RN24(p) AV_RL24(p) -# endif - -# if defined(AV_WN24) && !defined(AV_WL24) -# define AV_WL24(p, v) AV_WN24(p, v) -# elif !defined(AV_WN24) && defined(AV_WL24) -# define AV_WN24(p, v) AV_WL24(p, v) -# endif - -# if defined(AV_RN32) && !defined(AV_RL32) -# define AV_RL32(p) AV_RN32(p) -# elif !defined(AV_RN32) && defined(AV_RL32) -# define AV_RN32(p) AV_RL32(p) -# endif - -# if defined(AV_WN32) && !defined(AV_WL32) -# define AV_WL32(p, v) AV_WN32(p, v) -# elif !defined(AV_WN32) && defined(AV_WL32) -# define AV_WN32(p, v) AV_WL32(p, v) -# endif - -# if defined(AV_RN48) && !defined(AV_RL48) -# define AV_RL48(p) AV_RN48(p) -# elif !defined(AV_RN48) && defined(AV_RL48) -# define AV_RN48(p) AV_RL48(p) -# endif - -# if defined(AV_WN48) && !defined(AV_WL48) -# define AV_WL48(p, v) AV_WN48(p, v) -# elif !defined(AV_WN48) && defined(AV_WL48) -# define AV_WN48(p, v) AV_WL48(p, v) -# endif - -# if defined(AV_RN64) && !defined(AV_RL64) -# define AV_RL64(p) AV_RN64(p) -# elif !defined(AV_RN64) && defined(AV_RL64) -# define AV_RN64(p) AV_RL64(p) -# endif - -# if defined(AV_WN64) && !defined(AV_WL64) -# define AV_WL64(p, v) AV_WN64(p, v) -# elif !defined(AV_WN64) && defined(AV_WL64) -# define AV_WN64(p, v) AV_WL64(p, v) -# endif - -#endif /* !AV_HAVE_BIGENDIAN */ - -/* - * Define AV_[RW]N helper macros to simplify definitions not provided - * by per-arch headers. - */ - -#if defined(__GNUC__) && !defined(__TI_COMPILER_VERSION__) - -union unaligned_64 { uint64_t l; } __attribute__((packed)) av_alias; -union unaligned_32 { uint32_t l; } __attribute__((packed)) av_alias; -union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; - -# define AV_RN(s, p) (((const union unaligned_##s *) (p))->l) -# define AV_WN(s, p, v) ((((union unaligned_##s *) (p))->l) = (v)) - -#elif defined(__DECC) - -# define AV_RN(s, p) (*((const __unaligned uint##s##_t*)(p))) -# define AV_WN(s, p, v) (*((__unaligned uint##s##_t*)(p)) = (v)) - -#elif AV_HAVE_FAST_UNALIGNED - -# define AV_RN(s, p) (((const av_alias##s*)(p))->u##s) -# define AV_WN(s, p, v) (((av_alias##s*)(p))->u##s = (v)) - -#else - -#ifndef AV_RB16 -# define AV_RB16(x) \ - ((((const uint8_t*)(x))[0] << 8) | \ - ((const uint8_t*)(x))[1]) -#endif -#ifndef AV_WB16 -# define AV_WB16(p, darg) do { \ - unsigned d = (darg); \ - ((uint8_t*)(p))[1] = (d); \ - ((uint8_t*)(p))[0] = (d)>>8; \ - } while(0) -#endif - -#ifndef AV_RL16 -# define AV_RL16(x) \ - ((((const uint8_t*)(x))[1] << 8) | \ - ((const uint8_t*)(x))[0]) -#endif -#ifndef AV_WL16 -# define AV_WL16(p, darg) do { \ - unsigned d = (darg); \ - ((uint8_t*)(p))[0] = (d); \ - ((uint8_t*)(p))[1] = (d)>>8; \ - } while(0) -#endif - -#ifndef AV_RB32 -# define AV_RB32(x) \ - (((uint32_t)((const uint8_t*)(x))[0] << 24) | \ - (((const uint8_t*)(x))[1] << 16) | \ - (((const uint8_t*)(x))[2] << 8) | \ - ((const uint8_t*)(x))[3]) -#endif -#ifndef AV_WB32 -# define AV_WB32(p, darg) do { \ - unsigned d = (darg); \ - ((uint8_t*)(p))[3] = (d); \ - ((uint8_t*)(p))[2] = (d)>>8; \ - ((uint8_t*)(p))[1] = (d)>>16; \ - ((uint8_t*)(p))[0] = (d)>>24; \ - } while(0) -#endif - -#ifndef AV_RL32 -# define AV_RL32(x) \ - (((uint32_t)((const uint8_t*)(x))[3] << 24) | \ - (((const uint8_t*)(x))[2] << 16) | \ - (((const uint8_t*)(x))[1] << 8) | \ - ((const uint8_t*)(x))[0]) -#endif -#ifndef AV_WL32 -# define AV_WL32(p, darg) do { \ - unsigned d = (darg); \ - ((uint8_t*)(p))[0] = (d); \ - ((uint8_t*)(p))[1] = (d)>>8; \ - ((uint8_t*)(p))[2] = (d)>>16; \ - ((uint8_t*)(p))[3] = (d)>>24; \ - } while(0) -#endif - -#ifndef AV_RB64 -# define AV_RB64(x) \ - (((uint64_t)((const uint8_t*)(x))[0] << 56) | \ - ((uint64_t)((const uint8_t*)(x))[1] << 48) | \ - ((uint64_t)((const uint8_t*)(x))[2] << 40) | \ - ((uint64_t)((const uint8_t*)(x))[3] << 32) | \ - ((uint64_t)((const uint8_t*)(x))[4] << 24) | \ - ((uint64_t)((const uint8_t*)(x))[5] << 16) | \ - ((uint64_t)((const uint8_t*)(x))[6] << 8) | \ - (uint64_t)((const uint8_t*)(x))[7]) -#endif -#ifndef AV_WB64 -# define AV_WB64(p, darg) do { \ - uint64_t d = (darg); \ - ((uint8_t*)(p))[7] = (d); \ - ((uint8_t*)(p))[6] = (d)>>8; \ - ((uint8_t*)(p))[5] = (d)>>16; \ - ((uint8_t*)(p))[4] = (d)>>24; \ - ((uint8_t*)(p))[3] = (d)>>32; \ - ((uint8_t*)(p))[2] = (d)>>40; \ - ((uint8_t*)(p))[1] = (d)>>48; \ - ((uint8_t*)(p))[0] = (d)>>56; \ - } while(0) -#endif - -#ifndef AV_RL64 -# define AV_RL64(x) \ - (((uint64_t)((const uint8_t*)(x))[7] << 56) | \ - ((uint64_t)((const uint8_t*)(x))[6] << 48) | \ - ((uint64_t)((const uint8_t*)(x))[5] << 40) | \ - ((uint64_t)((const uint8_t*)(x))[4] << 32) | \ - ((uint64_t)((const uint8_t*)(x))[3] << 24) | \ - ((uint64_t)((const uint8_t*)(x))[2] << 16) | \ - ((uint64_t)((const uint8_t*)(x))[1] << 8) | \ - (uint64_t)((const uint8_t*)(x))[0]) -#endif -#ifndef AV_WL64 -# define AV_WL64(p, darg) do { \ - uint64_t d = (darg); \ - ((uint8_t*)(p))[0] = (d); \ - ((uint8_t*)(p))[1] = (d)>>8; \ - ((uint8_t*)(p))[2] = (d)>>16; \ - ((uint8_t*)(p))[3] = (d)>>24; \ - ((uint8_t*)(p))[4] = (d)>>32; \ - ((uint8_t*)(p))[5] = (d)>>40; \ - ((uint8_t*)(p))[6] = (d)>>48; \ - ((uint8_t*)(p))[7] = (d)>>56; \ - } while(0) -#endif - -#if AV_HAVE_BIGENDIAN -# define AV_RN(s, p) AV_RB##s(p) -# define AV_WN(s, p, v) AV_WB##s(p, v) -#else -# define AV_RN(s, p) AV_RL##s(p) -# define AV_WN(s, p, v) AV_WL##s(p, v) -#endif - -#endif /* HAVE_FAST_UNALIGNED */ - -#ifndef AV_RN16 -# define AV_RN16(p) AV_RN(16, p) -#endif - -#ifndef AV_RN32 -# define AV_RN32(p) AV_RN(32, p) -#endif - -#ifndef AV_RN64 -# define AV_RN64(p) AV_RN(64, p) -#endif - -#ifndef AV_WN16 -# define AV_WN16(p, v) AV_WN(16, p, v) -#endif - -#ifndef AV_WN32 -# define AV_WN32(p, v) AV_WN(32, p, v) -#endif - -#ifndef AV_WN64 -# define AV_WN64(p, v) AV_WN(64, p, v) -#endif - -#if AV_HAVE_BIGENDIAN -# define AV_RB(s, p) AV_RN##s(p) -# define AV_WB(s, p, v) AV_WN##s(p, v) -# define AV_RL(s, p) av_bswap##s(AV_RN##s(p)) -# define AV_WL(s, p, v) AV_WN##s(p, av_bswap##s(v)) -#else -# define AV_RB(s, p) av_bswap##s(AV_RN##s(p)) -# define AV_WB(s, p, v) AV_WN##s(p, av_bswap##s(v)) -# define AV_RL(s, p) AV_RN##s(p) -# define AV_WL(s, p, v) AV_WN##s(p, v) -#endif - -#define AV_RB8(x) (((const uint8_t*)(x))[0]) -#define AV_WB8(p, d) do { ((uint8_t*)(p))[0] = (d); } while(0) - -#define AV_RL8(x) AV_RB8(x) -#define AV_WL8(p, d) AV_WB8(p, d) - -#ifndef AV_RB16 -# define AV_RB16(p) AV_RB(16, p) -#endif -#ifndef AV_WB16 -# define AV_WB16(p, v) AV_WB(16, p, v) -#endif - -#ifndef AV_RL16 -# define AV_RL16(p) AV_RL(16, p) -#endif -#ifndef AV_WL16 -# define AV_WL16(p, v) AV_WL(16, p, v) -#endif - -#ifndef AV_RB32 -# define AV_RB32(p) AV_RB(32, p) -#endif -#ifndef AV_WB32 -# define AV_WB32(p, v) AV_WB(32, p, v) -#endif - -#ifndef AV_RL32 -# define AV_RL32(p) AV_RL(32, p) -#endif -#ifndef AV_WL32 -# define AV_WL32(p, v) AV_WL(32, p, v) -#endif - -#ifndef AV_RB64 -# define AV_RB64(p) AV_RB(64, p) -#endif -#ifndef AV_WB64 -# define AV_WB64(p, v) AV_WB(64, p, v) -#endif - -#ifndef AV_RL64 -# define AV_RL64(p) AV_RL(64, p) -#endif -#ifndef AV_WL64 -# define AV_WL64(p, v) AV_WL(64, p, v) -#endif - -#ifndef AV_RB24 -# define AV_RB24(x) \ - ((((const uint8_t*)(x))[0] << 16) | \ - (((const uint8_t*)(x))[1] << 8) | \ - ((const uint8_t*)(x))[2]) -#endif -#ifndef AV_WB24 -# define AV_WB24(p, d) do { \ - ((uint8_t*)(p))[2] = (d); \ - ((uint8_t*)(p))[1] = (d)>>8; \ - ((uint8_t*)(p))[0] = (d)>>16; \ - } while(0) -#endif - -#ifndef AV_RL24 -# define AV_RL24(x) \ - ((((const uint8_t*)(x))[2] << 16) | \ - (((const uint8_t*)(x))[1] << 8) | \ - ((const uint8_t*)(x))[0]) -#endif -#ifndef AV_WL24 -# define AV_WL24(p, d) do { \ - ((uint8_t*)(p))[0] = (d); \ - ((uint8_t*)(p))[1] = (d)>>8; \ - ((uint8_t*)(p))[2] = (d)>>16; \ - } while(0) -#endif - -#ifndef AV_RB48 -# define AV_RB48(x) \ - (((uint64_t)((const uint8_t*)(x))[0] << 40) | \ - ((uint64_t)((const uint8_t*)(x))[1] << 32) | \ - ((uint64_t)((const uint8_t*)(x))[2] << 24) | \ - ((uint64_t)((const uint8_t*)(x))[3] << 16) | \ - ((uint64_t)((const uint8_t*)(x))[4] << 8) | \ - (uint64_t)((const uint8_t*)(x))[5]) -#endif -#ifndef AV_WB48 -# define AV_WB48(p, darg) do { \ - uint64_t d = (darg); \ - ((uint8_t*)(p))[5] = (d); \ - ((uint8_t*)(p))[4] = (d)>>8; \ - ((uint8_t*)(p))[3] = (d)>>16; \ - ((uint8_t*)(p))[2] = (d)>>24; \ - ((uint8_t*)(p))[1] = (d)>>32; \ - ((uint8_t*)(p))[0] = (d)>>40; \ - } while(0) -#endif - -#ifndef AV_RL48 -# define AV_RL48(x) \ - (((uint64_t)((const uint8_t*)(x))[5] << 40) | \ - ((uint64_t)((const uint8_t*)(x))[4] << 32) | \ - ((uint64_t)((const uint8_t*)(x))[3] << 24) | \ - ((uint64_t)((const uint8_t*)(x))[2] << 16) | \ - ((uint64_t)((const uint8_t*)(x))[1] << 8) | \ - (uint64_t)((const uint8_t*)(x))[0]) -#endif -#ifndef AV_WL48 -# define AV_WL48(p, darg) do { \ - uint64_t d = (darg); \ - ((uint8_t*)(p))[0] = (d); \ - ((uint8_t*)(p))[1] = (d)>>8; \ - ((uint8_t*)(p))[2] = (d)>>16; \ - ((uint8_t*)(p))[3] = (d)>>24; \ - ((uint8_t*)(p))[4] = (d)>>32; \ - ((uint8_t*)(p))[5] = (d)>>40; \ - } while(0) -#endif - -/* - * The AV_[RW]NA macros access naturally aligned data - * in a type-safe way. - */ - -#define AV_RNA(s, p) (((const av_alias##s*)(p))->u##s) -#define AV_WNA(s, p, v) (((av_alias##s*)(p))->u##s = (v)) - -#ifndef AV_RN16A -# define AV_RN16A(p) AV_RNA(16, p) -#endif - -#ifndef AV_RN32A -# define AV_RN32A(p) AV_RNA(32, p) -#endif - -#ifndef AV_RN64A -# define AV_RN64A(p) AV_RNA(64, p) -#endif - -#ifndef AV_WN16A -# define AV_WN16A(p, v) AV_WNA(16, p, v) -#endif - -#ifndef AV_WN32A -# define AV_WN32A(p, v) AV_WNA(32, p, v) -#endif - -#ifndef AV_WN64A -# define AV_WN64A(p, v) AV_WNA(64, p, v) -#endif - -/* - * The AV_COPYxxU macros are suitable for copying data to/from unaligned - * memory locations. - */ - -#define AV_COPYU(n, d, s) AV_WN##n(d, AV_RN##n(s)); - -#ifndef AV_COPY16U -# define AV_COPY16U(d, s) AV_COPYU(16, d, s) -#endif - -#ifndef AV_COPY32U -# define AV_COPY32U(d, s) AV_COPYU(32, d, s) -#endif - -#ifndef AV_COPY64U -# define AV_COPY64U(d, s) AV_COPYU(64, d, s) -#endif - -#ifndef AV_COPY128U -# define AV_COPY128U(d, s) \ - do { \ - AV_COPY64U(d, s); \ - AV_COPY64U((char *)(d) + 8, (const char *)(s) + 8); \ - } while(0) -#endif - -/* Parameters for AV_COPY*, AV_SWAP*, AV_ZERO* must be - * naturally aligned. They may be implemented using MMX, - * so emms_c() must be called before using any float code - * afterwards. - */ - -#define AV_COPY(n, d, s) \ - (((av_alias##n*)(d))->u##n = ((const av_alias##n*)(s))->u##n) - -#ifndef AV_COPY16 -# define AV_COPY16(d, s) AV_COPY(16, d, s) -#endif - -#ifndef AV_COPY32 -# define AV_COPY32(d, s) AV_COPY(32, d, s) -#endif - -#ifndef AV_COPY64 -# define AV_COPY64(d, s) AV_COPY(64, d, s) -#endif - -#ifndef AV_COPY128 -# define AV_COPY128(d, s) \ - do { \ - AV_COPY64(d, s); \ - AV_COPY64((char*)(d)+8, (char*)(s)+8); \ - } while(0) -#endif - -#define AV_SWAP(n, a, b) FFSWAP(av_alias##n, *(av_alias##n*)(a), *(av_alias##n*)(b)) - -#ifndef AV_SWAP64 -# define AV_SWAP64(a, b) AV_SWAP(64, a, b) -#endif - -#define AV_ZERO(n, d) (((av_alias##n*)(d))->u##n = 0) - -#ifndef AV_ZERO16 -# define AV_ZERO16(d) AV_ZERO(16, d) -#endif - -#ifndef AV_ZERO32 -# define AV_ZERO32(d) AV_ZERO(32, d) -#endif - -#ifndef AV_ZERO64 -# define AV_ZERO64(d) AV_ZERO(64, d) -#endif - -#ifndef AV_ZERO128 -# define AV_ZERO128(d) \ - do { \ - AV_ZERO64(d); \ - AV_ZERO64((char*)(d)+8); \ - } while(0) -#endif - -#endif /* AVUTIL_INTREADWRITE_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/lfg.h b/src/main/jni/ffmpeg/include/libavutil/lfg.h deleted file mode 100644 index ec90562cf..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/lfg.h +++ /dev/null @@ -1,62 +0,0 @@ -/* - * Lagged Fibonacci PRNG - * Copyright (c) 2008 Michael Niedermayer - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_LFG_H -#define AVUTIL_LFG_H - -typedef struct AVLFG { - unsigned int state[64]; - int index; -} AVLFG; - -void av_lfg_init(AVLFG *c, unsigned int seed); - -/** - * Get the next random unsigned 32-bit number using an ALFG. - * - * Please also consider a simple LCG like state= state*1664525+1013904223, - * it may be good enough and faster for your specific use case. - */ -static inline unsigned int av_lfg_get(AVLFG *c){ - c->state[c->index & 63] = c->state[(c->index-24) & 63] + c->state[(c->index-55) & 63]; - return c->state[c->index++ & 63]; -} - -/** - * Get the next random unsigned 32-bit number using a MLFG. - * - * Please also consider av_lfg_get() above, it is faster. - */ -static inline unsigned int av_mlfg_get(AVLFG *c){ - unsigned int a= c->state[(c->index-55) & 63]; - unsigned int b= c->state[(c->index-24) & 63]; - return c->state[c->index++ & 63] = 2*a*b+a+b; -} - -/** - * Get the next two numbers generated by a Box-Muller Gaussian - * generator using the random numbers issued by lfg. - * - * @param out array where the two generated numbers are placed - */ -void av_bmg_get(AVLFG *lfg, double out[2]); - -#endif /* AVUTIL_LFG_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/log.h b/src/main/jni/ffmpeg/include/libavutil/log.h deleted file mode 100644 index 321748cd8..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/log.h +++ /dev/null @@ -1,359 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_LOG_H -#define AVUTIL_LOG_H - -#include <stdarg.h> -#include "avutil.h" -#include "attributes.h" -#include "version.h" - -typedef enum { - AV_CLASS_CATEGORY_NA = 0, - AV_CLASS_CATEGORY_INPUT, - AV_CLASS_CATEGORY_OUTPUT, - AV_CLASS_CATEGORY_MUXER, - AV_CLASS_CATEGORY_DEMUXER, - AV_CLASS_CATEGORY_ENCODER, - AV_CLASS_CATEGORY_DECODER, - AV_CLASS_CATEGORY_FILTER, - AV_CLASS_CATEGORY_BITSTREAM_FILTER, - AV_CLASS_CATEGORY_SWSCALER, - AV_CLASS_CATEGORY_SWRESAMPLER, - AV_CLASS_CATEGORY_DEVICE_VIDEO_OUTPUT = 40, - AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT, - AV_CLASS_CATEGORY_DEVICE_AUDIO_OUTPUT, - AV_CLASS_CATEGORY_DEVICE_AUDIO_INPUT, - AV_CLASS_CATEGORY_DEVICE_OUTPUT, - AV_CLASS_CATEGORY_DEVICE_INPUT, - AV_CLASS_CATEGORY_NB, ///< not part of ABI/API -}AVClassCategory; - -#define AV_IS_INPUT_DEVICE(category) \ - (((category) == AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT) || \ - ((category) == AV_CLASS_CATEGORY_DEVICE_AUDIO_INPUT) || \ - ((category) == AV_CLASS_CATEGORY_DEVICE_INPUT)) - -#define AV_IS_OUTPUT_DEVICE(category) \ - (((category) == AV_CLASS_CATEGORY_DEVICE_VIDEO_OUTPUT) || \ - ((category) == AV_CLASS_CATEGORY_DEVICE_AUDIO_OUTPUT) || \ - ((category) == AV_CLASS_CATEGORY_DEVICE_OUTPUT)) - -struct AVOptionRanges; - -/** - * Describe the class of an AVClass context structure. That is an - * arbitrary struct of which the first field is a pointer to an - * AVClass struct (e.g. AVCodecContext, AVFormatContext etc.). - */ -typedef struct AVClass { - /** - * The name of the class; usually it is the same name as the - * context structure type to which the AVClass is associated. - */ - const char* class_name; - - /** - * A pointer to a function which returns the name of a context - * instance ctx associated with the class. - */ - const char* (*item_name)(void* ctx); - - /** - * a pointer to the first option specified in the class if any or NULL - * - * @see av_set_default_options() - */ - const struct AVOption *option; - - /** - * LIBAVUTIL_VERSION with which this structure was created. - * This is used to allow fields to be added without requiring major - * version bumps everywhere. - */ - - int version; - - /** - * Offset in the structure where log_level_offset is stored. - * 0 means there is no such variable - */ - int log_level_offset_offset; - - /** - * Offset in the structure where a pointer to the parent context for - * logging is stored. For example a decoder could pass its AVCodecContext - * to eval as such a parent context, which an av_log() implementation - * could then leverage to display the parent context. - * The offset can be NULL. - */ - int parent_log_context_offset; - - /** - * Return next AVOptions-enabled child or NULL - */ - void* (*child_next)(void *obj, void *prev); - - /** - * Return an AVClass corresponding to the next potential - * AVOptions-enabled child. - * - * The difference between child_next and this is that - * child_next iterates over _already existing_ objects, while - * child_class_next iterates over _all possible_ children. - */ - const struct AVClass* (*child_class_next)(const struct AVClass *prev); - - /** - * Category used for visualization (like color) - * This is only set if the category is equal for all objects using this class. - * available since version (51 << 16 | 56 << 8 | 100) - */ - AVClassCategory category; - - /** - * Callback to return the category. - * available since version (51 << 16 | 59 << 8 | 100) - */ - AVClassCategory (*get_category)(void* ctx); - - /** - * Callback to return the supported/allowed ranges. - * available since version (52.12) - */ - int (*query_ranges)(struct AVOptionRanges **, void *obj, const char *key, int flags); -} AVClass; - -/** - * @addtogroup lavu_log - * - * @{ - * - * @defgroup lavu_log_constants Logging Constants - * - * @{ - */ - -/** - * Print no output. - */ -#define AV_LOG_QUIET -8 - -/** - * Something went really wrong and we will crash now. - */ -#define AV_LOG_PANIC 0 - -/** - * Something went wrong and recovery is not possible. - * For example, no header was found for a format which depends - * on headers or an illegal combination of parameters is used. - */ -#define AV_LOG_FATAL 8 - -/** - * Something went wrong and cannot losslessly be recovered. - * However, not all future data is affected. - */ -#define AV_LOG_ERROR 16 - -/** - * Something somehow does not look correct. This may or may not - * lead to problems. An example would be the use of '-vstrict -2'. - */ -#define AV_LOG_WARNING 24 - -/** - * Standard information. - */ -#define AV_LOG_INFO 32 - -/** - * Detailed information. - */ -#define AV_LOG_VERBOSE 40 - -/** - * Stuff which is only useful for libav* developers. - */ -#define AV_LOG_DEBUG 48 - -/** - * Extremely verbose debugging, useful for libav* development. - */ -#define AV_LOG_TRACE 56 - -#define AV_LOG_MAX_OFFSET (AV_LOG_TRACE - AV_LOG_QUIET) - -/** - * @} - */ - -/** - * Sets additional colors for extended debugging sessions. - * @code - av_log(ctx, AV_LOG_DEBUG|AV_LOG_C(134), "Message in purple\n"); - @endcode - * Requires 256color terminal support. Uses outside debugging is not - * recommended. - */ -#define AV_LOG_C(x) ((x) << 8) - -/** - * Send the specified message to the log if the level is less than or equal - * to the current av_log_level. By default, all logging messages are sent to - * stderr. This behavior can be altered by setting a different logging callback - * function. - * @see av_log_set_callback - * - * @param avcl A pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct or NULL if general log. - * @param level The importance level of the message expressed using a @ref - * lavu_log_constants "Logging Constant". - * @param fmt The format string (printf-compatible) that specifies how - * subsequent arguments are converted to output. - */ -void av_log(void *avcl, int level, const char *fmt, ...) av_printf_format(3, 4); - - -/** - * Send the specified message to the log if the level is less than or equal - * to the current av_log_level. By default, all logging messages are sent to - * stderr. This behavior can be altered by setting a different logging callback - * function. - * @see av_log_set_callback - * - * @param avcl A pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct. - * @param level The importance level of the message expressed using a @ref - * lavu_log_constants "Logging Constant". - * @param fmt The format string (printf-compatible) that specifies how - * subsequent arguments are converted to output. - * @param vl The arguments referenced by the format string. - */ -void av_vlog(void *avcl, int level, const char *fmt, va_list vl); - -/** - * Get the current log level - * - * @see lavu_log_constants - * - * @return Current log level - */ -int av_log_get_level(void); - -/** - * Set the log level - * - * @see lavu_log_constants - * - * @param level Logging level - */ -void av_log_set_level(int level); - -/** - * Set the logging callback - * - * @note The callback must be thread safe, even if the application does not use - * threads itself as some codecs are multithreaded. - * - * @see av_log_default_callback - * - * @param callback A logging function with a compatible signature. - */ -void av_log_set_callback(void (*callback)(void*, int, const char*, va_list)); - -/** - * Default logging callback - * - * It prints the message to stderr, optionally colorizing it. - * - * @param avcl A pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct. - * @param level The importance level of the message expressed using a @ref - * lavu_log_constants "Logging Constant". - * @param fmt The format string (printf-compatible) that specifies how - * subsequent arguments are converted to output. - * @param vl The arguments referenced by the format string. - */ -void av_log_default_callback(void *avcl, int level, const char *fmt, - va_list vl); - -/** - * Return the context name - * - * @param ctx The AVClass context - * - * @return The AVClass class_name - */ -const char* av_default_item_name(void* ctx); -AVClassCategory av_default_get_category(void *ptr); - -/** - * Format a line of log the same way as the default callback. - * @param line buffer to receive the formatted line - * @param line_size size of the buffer - * @param print_prefix used to store whether the prefix must be printed; - * must point to a persistent integer initially set to 1 - */ -void av_log_format_line(void *ptr, int level, const char *fmt, va_list vl, - char *line, int line_size, int *print_prefix); - -#if FF_API_DLOG -/** - * av_dlog macros - * @deprecated unused - * Useful to print debug messages that shouldn't get compiled in normally. - */ - -#ifdef DEBUG -# define av_dlog(pctx, ...) av_log(pctx, AV_LOG_DEBUG, __VA_ARGS__) -#else -# define av_dlog(pctx, ...) do { if (0) av_log(pctx, AV_LOG_DEBUG, __VA_ARGS__); } while (0) -#endif -#endif /* FF_API_DLOG */ - -/** - * Skip repeated messages, this requires the user app to use av_log() instead of - * (f)printf as the 2 would otherwise interfere and lead to - * "Last message repeated x times" messages below (f)printf messages with some - * bad luck. - * Also to receive the last, "last repeated" line if any, the user app must - * call av_log(NULL, AV_LOG_QUIET, "%s", ""); at the end - */ -#define AV_LOG_SKIP_REPEATED 1 - -/** - * Include the log severity in messages originating from codecs. - * - * Results in messages such as: - * [rawvideo @ 0xDEADBEEF] [error] encode did not produce valid pts - */ -#define AV_LOG_PRINT_LEVEL 2 - -void av_log_set_flags(int arg); -int av_log_get_flags(void); - -/** - * @} - */ - -#endif /* AVUTIL_LOG_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/macros.h b/src/main/jni/ffmpeg/include/libavutil/macros.h deleted file mode 100644 index 446532377..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/macros.h +++ /dev/null @@ -1,48 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * @ingroup lavu - * Utility Preprocessor macros - */ - -#ifndef AVUTIL_MACROS_H -#define AVUTIL_MACROS_H - -/** - * @addtogroup preproc_misc Preprocessor String Macros - * - * String manipulation macros - * - * @{ - */ - -#define AV_STRINGIFY(s) AV_TOSTRING(s) -#define AV_TOSTRING(s) #s - -#define AV_GLUE(a, b) a ## b -#define AV_JOIN(a, b) AV_GLUE(a, b) - -/** - * @} - */ - -#define AV_PRAGMA(s) _Pragma(#s) - -#endif /* AVUTIL_MACROS_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/mathematics.h b/src/main/jni/ffmpeg/include/libavutil/mathematics.h deleted file mode 100644 index 57c44f845..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/mathematics.h +++ /dev/null @@ -1,165 +0,0 @@ -/* - * copyright (c) 2005-2012 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_MATHEMATICS_H -#define AVUTIL_MATHEMATICS_H - -#include <stdint.h> -#include <math.h> -#include "attributes.h" -#include "rational.h" -#include "intfloat.h" - -#ifndef M_E -#define M_E 2.7182818284590452354 /* e */ -#endif -#ifndef M_LN2 -#define M_LN2 0.69314718055994530942 /* log_e 2 */ -#endif -#ifndef M_LN10 -#define M_LN10 2.30258509299404568402 /* log_e 10 */ -#endif -#ifndef M_LOG2_10 -#define M_LOG2_10 3.32192809488736234787 /* log_2 10 */ -#endif -#ifndef M_PHI -#define M_PHI 1.61803398874989484820 /* phi / golden ratio */ -#endif -#ifndef M_PI -#define M_PI 3.14159265358979323846 /* pi */ -#endif -#ifndef M_PI_2 -#define M_PI_2 1.57079632679489661923 /* pi/2 */ -#endif -#ifndef M_SQRT1_2 -#define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */ -#endif -#ifndef M_SQRT2 -#define M_SQRT2 1.41421356237309504880 /* sqrt(2) */ -#endif -#ifndef NAN -#define NAN av_int2float(0x7fc00000) -#endif -#ifndef INFINITY -#define INFINITY av_int2float(0x7f800000) -#endif - -/** - * @addtogroup lavu_math - * @{ - */ - - -enum AVRounding { - AV_ROUND_ZERO = 0, ///< Round toward zero. - AV_ROUND_INF = 1, ///< Round away from zero. - AV_ROUND_DOWN = 2, ///< Round toward -infinity. - AV_ROUND_UP = 3, ///< Round toward +infinity. - AV_ROUND_NEAR_INF = 5, ///< Round to nearest and halfway cases away from zero. - AV_ROUND_PASS_MINMAX = 8192, ///< Flag to pass INT64_MIN/MAX through instead of rescaling, this avoids special cases for AV_NOPTS_VALUE -}; - -/** - * Compute the greatest common divisor of a and b. - * - * @return gcd of a and b up to sign; if a >= 0 and b >= 0, return value is >= 0; - * if a == 0 and b == 0, returns 0. - */ -int64_t av_const av_gcd(int64_t a, int64_t b); - -/** - * Rescale a 64-bit integer with rounding to nearest. - * A simple a*b/c isn't possible as it can overflow. - */ -int64_t av_rescale(int64_t a, int64_t b, int64_t c) av_const; - -/** - * Rescale a 64-bit integer with specified rounding. - * A simple a*b/c isn't possible as it can overflow. - * - * @return rescaled value a, or if AV_ROUND_PASS_MINMAX is set and a is - * INT64_MIN or INT64_MAX then a is passed through unchanged. - */ -int64_t av_rescale_rnd(int64_t a, int64_t b, int64_t c, enum AVRounding) av_const; - -/** - * Rescale a 64-bit integer by 2 rational numbers. - */ -int64_t av_rescale_q(int64_t a, AVRational bq, AVRational cq) av_const; - -/** - * Rescale a 64-bit integer by 2 rational numbers with specified rounding. - * - * @return rescaled value a, or if AV_ROUND_PASS_MINMAX is set and a is - * INT64_MIN or INT64_MAX then a is passed through unchanged. - */ -int64_t av_rescale_q_rnd(int64_t a, AVRational bq, AVRational cq, - enum AVRounding) av_const; - -/** - * Compare 2 timestamps each in its own timebases. - * The result of the function is undefined if one of the timestamps - * is outside the int64_t range when represented in the others timebase. - * @return -1 if ts_a is before ts_b, 1 if ts_a is after ts_b or 0 if they represent the same position - */ -int av_compare_ts(int64_t ts_a, AVRational tb_a, int64_t ts_b, AVRational tb_b); - -/** - * Compare 2 integers modulo mod. - * That is we compare integers a and b for which only the least - * significant log2(mod) bits are known. - * - * @param mod must be a power of 2 - * @return a negative value if a is smaller than b - * a positive value if a is greater than b - * 0 if a equals b - */ -int64_t av_compare_mod(uint64_t a, uint64_t b, uint64_t mod); - -/** - * Rescale a timestamp while preserving known durations. - * - * @param in_ts Input timestamp - * @param in_tb Input timebase - * @param fs_tb Duration and *last timebase - * @param duration duration till the next call - * @param out_tb Output timebase - */ -int64_t av_rescale_delta(AVRational in_tb, int64_t in_ts, AVRational fs_tb, int duration, int64_t *last, AVRational out_tb); - -/** - * Add a value to a timestamp. - * - * This function guarantees that when the same value is repeatly added that - * no accumulation of rounding errors occurs. - * - * @param ts Input timestamp - * @param ts_tb Input timestamp timebase - * @param inc value to add to ts - * @param inc_tb inc timebase - */ -int64_t av_add_stable(AVRational ts_tb, int64_t ts, AVRational inc_tb, int64_t inc); - - - /** - * @} - */ - -#endif /* AVUTIL_MATHEMATICS_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/md5.h b/src/main/jni/ffmpeg/include/libavutil/md5.h deleted file mode 100644 index 79702c88c..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/md5.h +++ /dev/null @@ -1,81 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_MD5_H -#define AVUTIL_MD5_H - -#include <stdint.h> - -#include "attributes.h" -#include "version.h" - -/** - * @defgroup lavu_md5 MD5 - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_md5_size; - -struct AVMD5; - -/** - * Allocate an AVMD5 context. - */ -struct AVMD5 *av_md5_alloc(void); - -/** - * Initialize MD5 hashing. - * - * @param ctx pointer to the function context (of size av_md5_size) - */ -void av_md5_init(struct AVMD5 *ctx); - -/** - * Update hash value. - * - * @param ctx hash function context - * @param src input data to update hash with - * @param len input data length - */ -void av_md5_update(struct AVMD5 *ctx, const uint8_t *src, int len); - -/** - * Finish hashing and output digest value. - * - * @param ctx hash function context - * @param dst buffer where output digest value is stored - */ -void av_md5_final(struct AVMD5 *ctx, uint8_t *dst); - -/** - * Hash an array of data. - * - * @param dst The output buffer to write the digest into - * @param src The data to hash - * @param len The length of the data, in bytes - */ -void av_md5_sum(uint8_t *dst, const uint8_t *src, const int len); - -/** - * @} - */ - -#endif /* AVUTIL_MD5_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/mem.h b/src/main/jni/ffmpeg/include/libavutil/mem.h deleted file mode 100644 index d25b3229b..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/mem.h +++ /dev/null @@ -1,406 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * memory handling functions - */ - -#ifndef AVUTIL_MEM_H -#define AVUTIL_MEM_H - -#include <limits.h> -#include <stdint.h> - -#include "attributes.h" -#include "error.h" -#include "avutil.h" - -/** - * @addtogroup lavu_mem - * @{ - */ - - -#if defined(__INTEL_COMPILER) && __INTEL_COMPILER < 1110 || defined(__SUNPRO_C) - #define DECLARE_ALIGNED(n,t,v) t __attribute__ ((aligned (n))) v - #define DECLARE_ASM_CONST(n,t,v) const t __attribute__ ((aligned (n))) v -#elif defined(__TI_COMPILER_VERSION__) - #define DECLARE_ALIGNED(n,t,v) \ - AV_PRAGMA(DATA_ALIGN(v,n)) \ - t __attribute__((aligned(n))) v - #define DECLARE_ASM_CONST(n,t,v) \ - AV_PRAGMA(DATA_ALIGN(v,n)) \ - static const t __attribute__((aligned(n))) v -#elif defined(__GNUC__) - #define DECLARE_ALIGNED(n,t,v) t __attribute__ ((aligned (n))) v - #define DECLARE_ASM_CONST(n,t,v) static const t av_used __attribute__ ((aligned (n))) v -#elif defined(_MSC_VER) - #define DECLARE_ALIGNED(n,t,v) __declspec(align(n)) t v - #define DECLARE_ASM_CONST(n,t,v) __declspec(align(n)) static const t v -#else - #define DECLARE_ALIGNED(n,t,v) t v - #define DECLARE_ASM_CONST(n,t,v) static const t v -#endif - -#if AV_GCC_VERSION_AT_LEAST(3,1) - #define av_malloc_attrib __attribute__((__malloc__)) -#else - #define av_malloc_attrib -#endif - -#if AV_GCC_VERSION_AT_LEAST(4,3) - #define av_alloc_size(...) __attribute__((alloc_size(__VA_ARGS__))) -#else - #define av_alloc_size(...) -#endif - -/** - * Allocate a block of size bytes with alignment suitable for all - * memory accesses (including vectors if available on the CPU). - * @param size Size in bytes for the memory block to be allocated. - * @return Pointer to the allocated block, NULL if the block cannot - * be allocated. - * @see av_mallocz() - */ -void *av_malloc(size_t size) av_malloc_attrib av_alloc_size(1); - -/** - * Allocate a block of size * nmemb bytes with av_malloc(). - * @param nmemb Number of elements - * @param size Size of the single element - * @return Pointer to the allocated block, NULL if the block cannot - * be allocated. - * @see av_malloc() - */ -av_alloc_size(1, 2) static inline void *av_malloc_array(size_t nmemb, size_t size) -{ - if (!size || nmemb >= INT_MAX / size) - return NULL; - return av_malloc(nmemb * size); -} - -/** - * Allocate or reallocate a block of memory. - * If ptr is NULL and size > 0, allocate a new block. If - * size is zero, free the memory block pointed to by ptr. - * @param ptr Pointer to a memory block already allocated with - * av_realloc() or NULL. - * @param size Size in bytes of the memory block to be allocated or - * reallocated. - * @return Pointer to a newly-reallocated block or NULL if the block - * cannot be reallocated or the function is used to free the memory block. - * @warning Pointers originating from the av_malloc() family of functions must - * not be passed to av_realloc(). The former can be implemented using - * memalign() (or other functions), and there is no guarantee that - * pointers from such functions can be passed to realloc() at all. - * The situation is undefined according to POSIX and may crash with - * some libc implementations. - * @see av_fast_realloc() - */ -void *av_realloc(void *ptr, size_t size) av_alloc_size(2); - -/** - * Allocate or reallocate a block of memory. - * This function does the same thing as av_realloc, except: - * - It takes two arguments and checks the result of the multiplication for - * integer overflow. - * - It frees the input block in case of failure, thus avoiding the memory - * leak with the classic "buf = realloc(buf); if (!buf) return -1;". - */ -void *av_realloc_f(void *ptr, size_t nelem, size_t elsize); - -/** - * Allocate or reallocate a block of memory. - * If *ptr is NULL and size > 0, allocate a new block. If - * size is zero, free the memory block pointed to by ptr. - * @param ptr Pointer to a pointer to a memory block already allocated - * with av_realloc(), or pointer to a pointer to NULL. - * The pointer is updated on success, or freed on failure. - * @param size Size in bytes for the memory block to be allocated or - * reallocated - * @return Zero on success, an AVERROR error code on failure. - * @warning Pointers originating from the av_malloc() family of functions must - * not be passed to av_reallocp(). The former can be implemented using - * memalign() (or other functions), and there is no guarantee that - * pointers from such functions can be passed to realloc() at all. - * The situation is undefined according to POSIX and may crash with - * some libc implementations. - */ -av_warn_unused_result -int av_reallocp(void *ptr, size_t size); - -/** - * Allocate or reallocate an array. - * If ptr is NULL and nmemb > 0, allocate a new block. If - * nmemb is zero, free the memory block pointed to by ptr. - * @param ptr Pointer to a memory block already allocated with - * av_realloc() or NULL. - * @param nmemb Number of elements - * @param size Size of the single element - * @return Pointer to a newly-reallocated block or NULL if the block - * cannot be reallocated or the function is used to free the memory block. - * @warning Pointers originating from the av_malloc() family of functions must - * not be passed to av_realloc(). The former can be implemented using - * memalign() (or other functions), and there is no guarantee that - * pointers from such functions can be passed to realloc() at all. - * The situation is undefined according to POSIX and may crash with - * some libc implementations. - */ -av_alloc_size(2, 3) void *av_realloc_array(void *ptr, size_t nmemb, size_t size); - -/** - * Allocate or reallocate an array through a pointer to a pointer. - * If *ptr is NULL and nmemb > 0, allocate a new block. If - * nmemb is zero, free the memory block pointed to by ptr. - * @param ptr Pointer to a pointer to a memory block already allocated - * with av_realloc(), or pointer to a pointer to NULL. - * The pointer is updated on success, or freed on failure. - * @param nmemb Number of elements - * @param size Size of the single element - * @return Zero on success, an AVERROR error code on failure. - * @warning Pointers originating from the av_malloc() family of functions must - * not be passed to av_realloc(). The former can be implemented using - * memalign() (or other functions), and there is no guarantee that - * pointers from such functions can be passed to realloc() at all. - * The situation is undefined according to POSIX and may crash with - * some libc implementations. - */ -av_alloc_size(2, 3) int av_reallocp_array(void *ptr, size_t nmemb, size_t size); - -/** - * Free a memory block which has been allocated with av_malloc(z)() or - * av_realloc(). - * @param ptr Pointer to the memory block which should be freed. - * @note ptr = NULL is explicitly allowed. - * @note It is recommended that you use av_freep() instead. - * @see av_freep() - */ -void av_free(void *ptr); - -/** - * Allocate a block of size bytes with alignment suitable for all - * memory accesses (including vectors if available on the CPU) and - * zero all the bytes of the block. - * @param size Size in bytes for the memory block to be allocated. - * @return Pointer to the allocated block, NULL if it cannot be allocated. - * @see av_malloc() - */ -void *av_mallocz(size_t size) av_malloc_attrib av_alloc_size(1); - -/** - * Allocate a block of nmemb * size bytes with alignment suitable for all - * memory accesses (including vectors if available on the CPU) and - * zero all the bytes of the block. - * The allocation will fail if nmemb * size is greater than or equal - * to INT_MAX. - * @param nmemb - * @param size - * @return Pointer to the allocated block, NULL if it cannot be allocated. - */ -void *av_calloc(size_t nmemb, size_t size) av_malloc_attrib; - -/** - * Allocate a block of size * nmemb bytes with av_mallocz(). - * @param nmemb Number of elements - * @param size Size of the single element - * @return Pointer to the allocated block, NULL if the block cannot - * be allocated. - * @see av_mallocz() - * @see av_malloc_array() - */ -av_alloc_size(1, 2) static inline void *av_mallocz_array(size_t nmemb, size_t size) -{ - if (!size || nmemb >= INT_MAX / size) - return NULL; - return av_mallocz(nmemb * size); -} - -/** - * Duplicate the string s. - * @param s string to be duplicated - * @return Pointer to a newly-allocated string containing a - * copy of s or NULL if the string cannot be allocated. - */ -char *av_strdup(const char *s) av_malloc_attrib; - -/** - * Duplicate a substring of the string s. - * @param s string to be duplicated - * @param len the maximum length of the resulting string (not counting the - * terminating byte). - * @return Pointer to a newly-allocated string containing a - * copy of s or NULL if the string cannot be allocated. - */ -char *av_strndup(const char *s, size_t len) av_malloc_attrib; - -/** - * Duplicate the buffer p. - * @param p buffer to be duplicated - * @return Pointer to a newly allocated buffer containing a - * copy of p or NULL if the buffer cannot be allocated. - */ -void *av_memdup(const void *p, size_t size); - -/** - * Free a memory block which has been allocated with av_malloc(z)() or - * av_realloc() and set the pointer pointing to it to NULL. - * @param ptr Pointer to the pointer to the memory block which should - * be freed. - * @note passing a pointer to a NULL pointer is safe and leads to no action. - * @see av_free() - */ -void av_freep(void *ptr); - -/** - * Add an element to a dynamic array. - * - * The array to grow is supposed to be an array of pointers to - * structures, and the element to add must be a pointer to an already - * allocated structure. - * - * The array is reallocated when its size reaches powers of 2. - * Therefore, the amortized cost of adding an element is constant. - * - * In case of success, the pointer to the array is updated in order to - * point to the new grown array, and the number pointed to by nb_ptr - * is incremented. - * In case of failure, the array is freed, *tab_ptr is set to NULL and - * *nb_ptr is set to 0. - * - * @param tab_ptr pointer to the array to grow - * @param nb_ptr pointer to the number of elements in the array - * @param elem element to add - * @see av_dynarray_add_nofree(), av_dynarray2_add() - */ -void av_dynarray_add(void *tab_ptr, int *nb_ptr, void *elem); - -/** - * Add an element to a dynamic array. - * - * Function has the same functionality as av_dynarray_add(), - * but it doesn't free memory on fails. It returns error code - * instead and leave current buffer untouched. - * - * @param tab_ptr pointer to the array to grow - * @param nb_ptr pointer to the number of elements in the array - * @param elem element to add - * @return >=0 on success, negative otherwise. - * @see av_dynarray_add(), av_dynarray2_add() - */ -av_warn_unused_result -int av_dynarray_add_nofree(void *tab_ptr, int *nb_ptr, void *elem); - -/** - * Add an element of size elem_size to a dynamic array. - * - * The array is reallocated when its number of elements reaches powers of 2. - * Therefore, the amortized cost of adding an element is constant. - * - * In case of success, the pointer to the array is updated in order to - * point to the new grown array, and the number pointed to by nb_ptr - * is incremented. - * In case of failure, the array is freed, *tab_ptr is set to NULL and - * *nb_ptr is set to 0. - * - * @param tab_ptr pointer to the array to grow - * @param nb_ptr pointer to the number of elements in the array - * @param elem_size size in bytes of the elements in the array - * @param elem_data pointer to the data of the element to add. If NULL, the space of - * the new added element is not filled. - * @return pointer to the data of the element to copy in the new allocated space. - * If NULL, the new allocated space is left uninitialized." - * @see av_dynarray_add(), av_dynarray_add_nofree() - */ -void *av_dynarray2_add(void **tab_ptr, int *nb_ptr, size_t elem_size, - const uint8_t *elem_data); - -/** - * Multiply two size_t values checking for overflow. - * @return 0 if success, AVERROR(EINVAL) if overflow. - */ -static inline int av_size_mult(size_t a, size_t b, size_t *r) -{ - size_t t = a * b; - /* Hack inspired from glibc: only try the division if nelem and elsize - * are both greater than sqrt(SIZE_MAX). */ - if ((a | b) >= ((size_t)1 << (sizeof(size_t) * 4)) && a && t / a != b) - return AVERROR(EINVAL); - *r = t; - return 0; -} - -/** - * Set the maximum size that may me allocated in one block. - */ -void av_max_alloc(size_t max); - -/** - * deliberately overlapping memcpy implementation - * @param dst destination buffer - * @param back how many bytes back we start (the initial size of the overlapping window), must be > 0 - * @param cnt number of bytes to copy, must be >= 0 - * - * cnt > back is valid, this will copy the bytes we just copied, - * thus creating a repeating pattern with a period length of back. - */ -void av_memcpy_backptr(uint8_t *dst, int back, int cnt); - -/** - * Reallocate the given block if it is not large enough, otherwise do nothing. - * - * @see av_realloc - */ -void *av_fast_realloc(void *ptr, unsigned int *size, size_t min_size); - -/** - * Allocate a buffer, reusing the given one if large enough. - * - * Contrary to av_fast_realloc the current buffer contents might not be - * preserved and on error the old buffer is freed, thus no special - * handling to avoid memleaks is necessary. - * - * @param ptr pointer to pointer to already allocated buffer, overwritten with pointer to new buffer - * @param size size of the buffer *ptr points to - * @param min_size minimum size of *ptr buffer after returning, *ptr will be NULL and - * *size 0 if an error occurred. - */ -void av_fast_malloc(void *ptr, unsigned int *size, size_t min_size); - -/** - * Allocate a buffer, reusing the given one if large enough. - * - * All newly allocated space is initially cleared - * Contrary to av_fast_realloc the current buffer contents might not be - * preserved and on error the old buffer is freed, thus no special - * handling to avoid memleaks is necessary. - * - * @param ptr pointer to pointer to already allocated buffer, overwritten with pointer to new buffer - * @param size size of the buffer *ptr points to - * @param min_size minimum size of *ptr buffer after returning, *ptr will be NULL and - * *size 0 if an error occurred. - */ -void av_fast_mallocz(void *ptr, unsigned int *size, size_t min_size); - -/** - * @} - */ - -#endif /* AVUTIL_MEM_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/motion_vector.h b/src/main/jni/ffmpeg/include/libavutil/motion_vector.h deleted file mode 100644 index ec2955638..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/motion_vector.h +++ /dev/null @@ -1,57 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_MOTION_VECTOR_H -#define AVUTIL_MOTION_VECTOR_H - -#include <stdint.h> - -typedef struct AVMotionVector { - /** - * Where the current macroblock comes from; negative value when it comes - * from the past, positive value when it comes from the future. - * XXX: set exact relative ref frame reference instead of a +/- 1 "direction". - */ - int32_t source; - /** - * Width and height of the block. - */ - uint8_t w, h; - /** - * Absolute source position. Can be outside the frame area. - */ - int16_t src_x, src_y; - /** - * Absolute destination position. Can be outside the frame area. - */ - int16_t dst_x, dst_y; - /** - * Extra flag information. - * Currently unused. - */ - uint64_t flags; - /** - * Motion vector - * src_x = dst_x + motion_x / motion_scale - * src_y = dst_y + motion_y / motion_scale - */ - int32_t motion_x, motion_y; - uint16_t motion_scale; -} AVMotionVector; - -#endif /* AVUTIL_MOTION_VECTOR_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/murmur3.h b/src/main/jni/ffmpeg/include/libavutil/murmur3.h deleted file mode 100644 index f29ed973e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/murmur3.h +++ /dev/null @@ -1,32 +0,0 @@ -/* - * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_MURMUR3_H -#define AVUTIL_MURMUR3_H - -#include <stdint.h> - -struct AVMurMur3 *av_murmur3_alloc(void); -void av_murmur3_init_seeded(struct AVMurMur3 *c, uint64_t seed); -void av_murmur3_init(struct AVMurMur3 *c); -void av_murmur3_update(struct AVMurMur3 *c, const uint8_t *src, int len); -void av_murmur3_final(struct AVMurMur3 *c, uint8_t dst[16]); - -#endif /* AVUTIL_MURMUR3_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/opt.h b/src/main/jni/ffmpeg/include/libavutil/opt.h deleted file mode 100644 index 753434d62..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/opt.h +++ /dev/null @@ -1,865 +0,0 @@ -/* - * AVOptions - * copyright (c) 2005 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_OPT_H -#define AVUTIL_OPT_H - -/** - * @file - * AVOptions - */ - -#include "rational.h" -#include "avutil.h" -#include "dict.h" -#include "log.h" -#include "pixfmt.h" -#include "samplefmt.h" -#include "version.h" - -/** - * @defgroup avoptions AVOptions - * @ingroup lavu_data - * @{ - * AVOptions provide a generic system to declare options on arbitrary structs - * ("objects"). An option can have a help text, a type and a range of possible - * values. Options may then be enumerated, read and written to. - * - * @section avoptions_implement Implementing AVOptions - * This section describes how to add AVOptions capabilities to a struct. - * - * All AVOptions-related information is stored in an AVClass. Therefore - * the first member of the struct should be a pointer to an AVClass describing it. - * The option field of the AVClass must be set to a NULL-terminated static array - * of AVOptions. Each AVOption must have a non-empty name, a type, a default - * value and for number-type AVOptions also a range of allowed values. It must - * also declare an offset in bytes from the start of the struct, where the field - * associated with this AVOption is located. Other fields in the AVOption struct - * should also be set when applicable, but are not required. - * - * The following example illustrates an AVOptions-enabled struct: - * @code - * typedef struct test_struct { - * AVClass *class; - * int int_opt; - * char *str_opt; - * uint8_t *bin_opt; - * int bin_len; - * } test_struct; - * - * static const AVOption test_options[] = { - * { "test_int", "This is a test option of int type.", offsetof(test_struct, int_opt), - * AV_OPT_TYPE_INT, { .i64 = -1 }, INT_MIN, INT_MAX }, - * { "test_str", "This is a test option of string type.", offsetof(test_struct, str_opt), - * AV_OPT_TYPE_STRING }, - * { "test_bin", "This is a test option of binary type.", offsetof(test_struct, bin_opt), - * AV_OPT_TYPE_BINARY }, - * { NULL }, - * }; - * - * static const AVClass test_class = { - * .class_name = "test class", - * .item_name = av_default_item_name, - * .option = test_options, - * .version = LIBAVUTIL_VERSION_INT, - * }; - * @endcode - * - * Next, when allocating your struct, you must ensure that the AVClass pointer - * is set to the correct value. Then, av_opt_set_defaults() can be called to - * initialize defaults. After that the struct is ready to be used with the - * AVOptions API. - * - * When cleaning up, you may use the av_opt_free() function to automatically - * free all the allocated string and binary options. - * - * Continuing with the above example: - * - * @code - * test_struct *alloc_test_struct(void) - * { - * test_struct *ret = av_malloc(sizeof(*ret)); - * ret->class = &test_class; - * av_opt_set_defaults(ret); - * return ret; - * } - * void free_test_struct(test_struct **foo) - * { - * av_opt_free(*foo); - * av_freep(foo); - * } - * @endcode - * - * @subsection avoptions_implement_nesting Nesting - * It may happen that an AVOptions-enabled struct contains another - * AVOptions-enabled struct as a member (e.g. AVCodecContext in - * libavcodec exports generic options, while its priv_data field exports - * codec-specific options). In such a case, it is possible to set up the - * parent struct to export a child's options. To do that, simply - * implement AVClass.child_next() and AVClass.child_class_next() in the - * parent struct's AVClass. - * Assuming that the test_struct from above now also contains a - * child_struct field: - * - * @code - * typedef struct child_struct { - * AVClass *class; - * int flags_opt; - * } child_struct; - * static const AVOption child_opts[] = { - * { "test_flags", "This is a test option of flags type.", - * offsetof(child_struct, flags_opt), AV_OPT_TYPE_FLAGS, { .i64 = 0 }, INT_MIN, INT_MAX }, - * { NULL }, - * }; - * static const AVClass child_class = { - * .class_name = "child class", - * .item_name = av_default_item_name, - * .option = child_opts, - * .version = LIBAVUTIL_VERSION_INT, - * }; - * - * void *child_next(void *obj, void *prev) - * { - * test_struct *t = obj; - * if (!prev && t->child_struct) - * return t->child_struct; - * return NULL - * } - * const AVClass child_class_next(const AVClass *prev) - * { - * return prev ? NULL : &child_class; - * } - * @endcode - * Putting child_next() and child_class_next() as defined above into - * test_class will now make child_struct's options accessible through - * test_struct (again, proper setup as described above needs to be done on - * child_struct right after it is created). - * - * From the above example it might not be clear why both child_next() - * and child_class_next() are needed. The distinction is that child_next() - * iterates over actually existing objects, while child_class_next() - * iterates over all possible child classes. E.g. if an AVCodecContext - * was initialized to use a codec which has private options, then its - * child_next() will return AVCodecContext.priv_data and finish - * iterating. OTOH child_class_next() on AVCodecContext.av_class will - * iterate over all available codecs with private options. - * - * @subsection avoptions_implement_named_constants Named constants - * It is possible to create named constants for options. Simply set the unit - * field of the option the constants should apply to a string and - * create the constants themselves as options of type AV_OPT_TYPE_CONST - * with their unit field set to the same string. - * Their default_val field should contain the value of the named - * constant. - * For example, to add some named constants for the test_flags option - * above, put the following into the child_opts array: - * @code - * { "test_flags", "This is a test option of flags type.", - * offsetof(child_struct, flags_opt), AV_OPT_TYPE_FLAGS, { .i64 = 0 }, INT_MIN, INT_MAX, "test_unit" }, - * { "flag1", "This is a flag with value 16", 0, AV_OPT_TYPE_CONST, { .i64 = 16 }, 0, 0, "test_unit" }, - * @endcode - * - * @section avoptions_use Using AVOptions - * This section deals with accessing options in an AVOptions-enabled struct. - * Such structs in FFmpeg are e.g. AVCodecContext in libavcodec or - * AVFormatContext in libavformat. - * - * @subsection avoptions_use_examine Examining AVOptions - * The basic functions for examining options are av_opt_next(), which iterates - * over all options defined for one object, and av_opt_find(), which searches - * for an option with the given name. - * - * The situation is more complicated with nesting. An AVOptions-enabled struct - * may have AVOptions-enabled children. Passing the AV_OPT_SEARCH_CHILDREN flag - * to av_opt_find() will make the function search children recursively. - * - * For enumerating there are basically two cases. The first is when you want to - * get all options that may potentially exist on the struct and its children - * (e.g. when constructing documentation). In that case you should call - * av_opt_child_class_next() recursively on the parent struct's AVClass. The - * second case is when you have an already initialized struct with all its - * children and you want to get all options that can be actually written or read - * from it. In that case you should call av_opt_child_next() recursively (and - * av_opt_next() on each result). - * - * @subsection avoptions_use_get_set Reading and writing AVOptions - * When setting options, you often have a string read directly from the - * user. In such a case, simply passing it to av_opt_set() is enough. For - * non-string type options, av_opt_set() will parse the string according to the - * option type. - * - * Similarly av_opt_get() will read any option type and convert it to a string - * which will be returned. Do not forget that the string is allocated, so you - * have to free it with av_free(). - * - * In some cases it may be more convenient to put all options into an - * AVDictionary and call av_opt_set_dict() on it. A specific case of this - * are the format/codec open functions in lavf/lavc which take a dictionary - * filled with option as a parameter. This makes it possible to set some options - * that cannot be set otherwise, since e.g. the input file format is not known - * before the file is actually opened. - */ - -enum AVOptionType{ - AV_OPT_TYPE_FLAGS, - AV_OPT_TYPE_INT, - AV_OPT_TYPE_INT64, - AV_OPT_TYPE_DOUBLE, - AV_OPT_TYPE_FLOAT, - AV_OPT_TYPE_STRING, - AV_OPT_TYPE_RATIONAL, - AV_OPT_TYPE_BINARY, ///< offset must point to a pointer immediately followed by an int for the length - AV_OPT_TYPE_DICT, - AV_OPT_TYPE_CONST = 128, - AV_OPT_TYPE_IMAGE_SIZE = MKBETAG('S','I','Z','E'), ///< offset must point to two consecutive integers - AV_OPT_TYPE_PIXEL_FMT = MKBETAG('P','F','M','T'), - AV_OPT_TYPE_SAMPLE_FMT = MKBETAG('S','F','M','T'), - AV_OPT_TYPE_VIDEO_RATE = MKBETAG('V','R','A','T'), ///< offset must point to AVRational - AV_OPT_TYPE_DURATION = MKBETAG('D','U','R',' '), - AV_OPT_TYPE_COLOR = MKBETAG('C','O','L','R'), - AV_OPT_TYPE_CHANNEL_LAYOUT = MKBETAG('C','H','L','A'), - AV_OPT_TYPE_BOOL = MKBETAG('B','O','O','L'), -}; - -/** - * AVOption - */ -typedef struct AVOption { - const char *name; - - /** - * short English help text - * @todo What about other languages? - */ - const char *help; - - /** - * The offset relative to the context structure where the option - * value is stored. It should be 0 for named constants. - */ - int offset; - enum AVOptionType type; - - /** - * the default value for scalar options - */ - union { - int64_t i64; - double dbl; - const char *str; - /* TODO those are unused now */ - AVRational q; - } default_val; - double min; ///< minimum valid value for the option - double max; ///< maximum valid value for the option - - int flags; -#define AV_OPT_FLAG_ENCODING_PARAM 1 ///< a generic parameter which can be set by the user for muxing or encoding -#define AV_OPT_FLAG_DECODING_PARAM 2 ///< a generic parameter which can be set by the user for demuxing or decoding -#if FF_API_OPT_TYPE_METADATA -#define AV_OPT_FLAG_METADATA 4 ///< some data extracted or inserted into the file like title, comment, ... -#endif -#define AV_OPT_FLAG_AUDIO_PARAM 8 -#define AV_OPT_FLAG_VIDEO_PARAM 16 -#define AV_OPT_FLAG_SUBTITLE_PARAM 32 -/** - * The option is inteded for exporting values to the caller. - */ -#define AV_OPT_FLAG_EXPORT 64 -/** - * The option may not be set through the AVOptions API, only read. - * This flag only makes sense when AV_OPT_FLAG_EXPORT is also set. - */ -#define AV_OPT_FLAG_READONLY 128 -#define AV_OPT_FLAG_FILTERING_PARAM (1<<16) ///< a generic parameter which can be set by the user for filtering -//FIXME think about enc-audio, ... style flags - - /** - * The logical unit to which the option belongs. Non-constant - * options and corresponding named constants share the same - * unit. May be NULL. - */ - const char *unit; -} AVOption; - -/** - * A single allowed range of values, or a single allowed value. - */ -typedef struct AVOptionRange { - const char *str; - /** - * Value range. - * For string ranges this represents the min/max length. - * For dimensions this represents the min/max pixel count or width/height in multi-component case. - */ - double value_min, value_max; - /** - * Value's component range. - * For string this represents the unicode range for chars, 0-127 limits to ASCII. - */ - double component_min, component_max; - /** - * Range flag. - * If set to 1 the struct encodes a range, if set to 0 a single value. - */ - int is_range; -} AVOptionRange; - -/** - * List of AVOptionRange structs. - */ -typedef struct AVOptionRanges { - /** - * Array of option ranges. - * - * Most of option types use just one component. - * Following describes multi-component option types: - * - * AV_OPT_TYPE_IMAGE_SIZE: - * component index 0: range of pixel count (width * height). - * component index 1: range of width. - * component index 2: range of height. - * - * @note To obtain multi-component version of this structure, user must - * provide AV_OPT_MULTI_COMPONENT_RANGE to av_opt_query_ranges or - * av_opt_query_ranges_default function. - * - * Multi-component range can be read as in following example: - * - * @code - * int range_index, component_index; - * AVOptionRanges *ranges; - * AVOptionRange *range[3]; //may require more than 3 in the future. - * av_opt_query_ranges(&ranges, obj, key, AV_OPT_MULTI_COMPONENT_RANGE); - * for (range_index = 0; range_index < ranges->nb_ranges; range_index++) { - * for (component_index = 0; component_index < ranges->nb_components; component_index++) - * range[component_index] = ranges->range[ranges->nb_ranges * component_index + range_index]; - * //do something with range here. - * } - * av_opt_freep_ranges(&ranges); - * @endcode - */ - AVOptionRange **range; - /** - * Number of ranges per component. - */ - int nb_ranges; - /** - * Number of componentes. - */ - int nb_components; -} AVOptionRanges; - -/** - * Show the obj options. - * - * @param req_flags requested flags for the options to show. Show only the - * options for which it is opt->flags & req_flags. - * @param rej_flags rejected flags for the options to show. Show only the - * options for which it is !(opt->flags & req_flags). - * @param av_log_obj log context to use for showing the options - */ -int av_opt_show2(void *obj, void *av_log_obj, int req_flags, int rej_flags); - -/** - * Set the values of all AVOption fields to their default values. - * - * @param s an AVOption-enabled struct (its first member must be a pointer to AVClass) - */ -void av_opt_set_defaults(void *s); - -/** - * Set the values of all AVOption fields to their default values. Only these - * AVOption fields for which (opt->flags & mask) == flags will have their - * default applied to s. - * - * @param s an AVOption-enabled struct (its first member must be a pointer to AVClass) - * @param mask combination of AV_OPT_FLAG_* - * @param flags combination of AV_OPT_FLAG_* - */ -void av_opt_set_defaults2(void *s, int mask, int flags); - -/** - * Parse the key/value pairs list in opts. For each key/value pair - * found, stores the value in the field in ctx that is named like the - * key. ctx must be an AVClass context, storing is done using - * AVOptions. - * - * @param opts options string to parse, may be NULL - * @param key_val_sep a 0-terminated list of characters used to - * separate key from value - * @param pairs_sep a 0-terminated list of characters used to separate - * two pairs from each other - * @return the number of successfully set key/value pairs, or a negative - * value corresponding to an AVERROR code in case of error: - * AVERROR(EINVAL) if opts cannot be parsed, - * the error code issued by av_opt_set() if a key/value pair - * cannot be set - */ -int av_set_options_string(void *ctx, const char *opts, - const char *key_val_sep, const char *pairs_sep); - -/** - * Parse the key-value pairs list in opts. For each key=value pair found, - * set the value of the corresponding option in ctx. - * - * @param ctx the AVClass object to set options on - * @param opts the options string, key-value pairs separated by a - * delimiter - * @param shorthand a NULL-terminated array of options names for shorthand - * notation: if the first field in opts has no key part, - * the key is taken from the first element of shorthand; - * then again for the second, etc., until either opts is - * finished, shorthand is finished or a named option is - * found; after that, all options must be named - * @param key_val_sep a 0-terminated list of characters used to separate - * key from value, for example '=' - * @param pairs_sep a 0-terminated list of characters used to separate - * two pairs from each other, for example ':' or ',' - * @return the number of successfully set key=value pairs, or a negative - * value corresponding to an AVERROR code in case of error: - * AVERROR(EINVAL) if opts cannot be parsed, - * the error code issued by av_set_string3() if a key/value pair - * cannot be set - * - * Options names must use only the following characters: a-z A-Z 0-9 - . / _ - * Separators must use characters distinct from option names and from each - * other. - */ -int av_opt_set_from_string(void *ctx, const char *opts, - const char *const *shorthand, - const char *key_val_sep, const char *pairs_sep); -/** - * Free all allocated objects in obj. - */ -void av_opt_free(void *obj); - -/** - * Check whether a particular flag is set in a flags field. - * - * @param field_name the name of the flag field option - * @param flag_name the name of the flag to check - * @return non-zero if the flag is set, zero if the flag isn't set, - * isn't of the right type, or the flags field doesn't exist. - */ -int av_opt_flag_is_set(void *obj, const char *field_name, const char *flag_name); - -/** - * Set all the options from a given dictionary on an object. - * - * @param obj a struct whose first element is a pointer to AVClass - * @param options options to process. This dictionary will be freed and replaced - * by a new one containing all options not found in obj. - * Of course this new dictionary needs to be freed by caller - * with av_dict_free(). - * - * @return 0 on success, a negative AVERROR if some option was found in obj, - * but could not be set. - * - * @see av_dict_copy() - */ -int av_opt_set_dict(void *obj, struct AVDictionary **options); - - -/** - * Set all the options from a given dictionary on an object. - * - * @param obj a struct whose first element is a pointer to AVClass - * @param options options to process. This dictionary will be freed and replaced - * by a new one containing all options not found in obj. - * Of course this new dictionary needs to be freed by caller - * with av_dict_free(). - * @param search_flags A combination of AV_OPT_SEARCH_*. - * - * @return 0 on success, a negative AVERROR if some option was found in obj, - * but could not be set. - * - * @see av_dict_copy() - */ -int av_opt_set_dict2(void *obj, struct AVDictionary **options, int search_flags); - -/** - * Extract a key-value pair from the beginning of a string. - * - * @param ropts pointer to the options string, will be updated to - * point to the rest of the string (one of the pairs_sep - * or the final NUL) - * @param key_val_sep a 0-terminated list of characters used to separate - * key from value, for example '=' - * @param pairs_sep a 0-terminated list of characters used to separate - * two pairs from each other, for example ':' or ',' - * @param flags flags; see the AV_OPT_FLAG_* values below - * @param rkey parsed key; must be freed using av_free() - * @param rval parsed value; must be freed using av_free() - * - * @return >=0 for success, or a negative value corresponding to an - * AVERROR code in case of error; in particular: - * AVERROR(EINVAL) if no key is present - * - */ -int av_opt_get_key_value(const char **ropts, - const char *key_val_sep, const char *pairs_sep, - unsigned flags, - char **rkey, char **rval); - -enum { - - /** - * Accept to parse a value without a key; the key will then be returned - * as NULL. - */ - AV_OPT_FLAG_IMPLICIT_KEY = 1, -}; - -/** - * @defgroup opt_eval_funcs Evaluating option strings - * @{ - * This group of functions can be used to evaluate option strings - * and get numbers out of them. They do the same thing as av_opt_set(), - * except the result is written into the caller-supplied pointer. - * - * @param obj a struct whose first element is a pointer to AVClass. - * @param o an option for which the string is to be evaluated. - * @param val string to be evaluated. - * @param *_out value of the string will be written here. - * - * @return 0 on success, a negative number on failure. - */ -int av_opt_eval_flags (void *obj, const AVOption *o, const char *val, int *flags_out); -int av_opt_eval_int (void *obj, const AVOption *o, const char *val, int *int_out); -int av_opt_eval_int64 (void *obj, const AVOption *o, const char *val, int64_t *int64_out); -int av_opt_eval_float (void *obj, const AVOption *o, const char *val, float *float_out); -int av_opt_eval_double(void *obj, const AVOption *o, const char *val, double *double_out); -int av_opt_eval_q (void *obj, const AVOption *o, const char *val, AVRational *q_out); -/** - * @} - */ - -#define AV_OPT_SEARCH_CHILDREN (1 << 0) /**< Search in possible children of the - given object first. */ -/** - * The obj passed to av_opt_find() is fake -- only a double pointer to AVClass - * instead of a required pointer to a struct containing AVClass. This is - * useful for searching for options without needing to allocate the corresponding - * object. - */ -#define AV_OPT_SEARCH_FAKE_OBJ (1 << 1) - -/** - * In av_opt_get, return NULL if the option has a pointer type and is set to NULL, - * rather than returning an empty string. - */ -#define AV_OPT_ALLOW_NULL (1 << 2) - -/** - * Allows av_opt_query_ranges and av_opt_query_ranges_default to return more than - * one component for certain option types. - * @see AVOptionRanges for details. - */ -#define AV_OPT_MULTI_COMPONENT_RANGE (1 << 12) - -/** - * Look for an option in an object. Consider only options which - * have all the specified flags set. - * - * @param[in] obj A pointer to a struct whose first element is a - * pointer to an AVClass. - * Alternatively a double pointer to an AVClass, if - * AV_OPT_SEARCH_FAKE_OBJ search flag is set. - * @param[in] name The name of the option to look for. - * @param[in] unit When searching for named constants, name of the unit - * it belongs to. - * @param opt_flags Find only options with all the specified flags set (AV_OPT_FLAG). - * @param search_flags A combination of AV_OPT_SEARCH_*. - * - * @return A pointer to the option found, or NULL if no option - * was found. - * - * @note Options found with AV_OPT_SEARCH_CHILDREN flag may not be settable - * directly with av_opt_set(). Use special calls which take an options - * AVDictionary (e.g. avformat_open_input()) to set options found with this - * flag. - */ -const AVOption *av_opt_find(void *obj, const char *name, const char *unit, - int opt_flags, int search_flags); - -/** - * Look for an option in an object. Consider only options which - * have all the specified flags set. - * - * @param[in] obj A pointer to a struct whose first element is a - * pointer to an AVClass. - * Alternatively a double pointer to an AVClass, if - * AV_OPT_SEARCH_FAKE_OBJ search flag is set. - * @param[in] name The name of the option to look for. - * @param[in] unit When searching for named constants, name of the unit - * it belongs to. - * @param opt_flags Find only options with all the specified flags set (AV_OPT_FLAG). - * @param search_flags A combination of AV_OPT_SEARCH_*. - * @param[out] target_obj if non-NULL, an object to which the option belongs will be - * written here. It may be different from obj if AV_OPT_SEARCH_CHILDREN is present - * in search_flags. This parameter is ignored if search_flags contain - * AV_OPT_SEARCH_FAKE_OBJ. - * - * @return A pointer to the option found, or NULL if no option - * was found. - */ -const AVOption *av_opt_find2(void *obj, const char *name, const char *unit, - int opt_flags, int search_flags, void **target_obj); - -/** - * Iterate over all AVOptions belonging to obj. - * - * @param obj an AVOptions-enabled struct or a double pointer to an - * AVClass describing it. - * @param prev result of the previous call to av_opt_next() on this object - * or NULL - * @return next AVOption or NULL - */ -const AVOption *av_opt_next(const void *obj, const AVOption *prev); - -/** - * Iterate over AVOptions-enabled children of obj. - * - * @param prev result of a previous call to this function or NULL - * @return next AVOptions-enabled child or NULL - */ -void *av_opt_child_next(void *obj, void *prev); - -/** - * Iterate over potential AVOptions-enabled children of parent. - * - * @param prev result of a previous call to this function or NULL - * @return AVClass corresponding to next potential child or NULL - */ -const AVClass *av_opt_child_class_next(const AVClass *parent, const AVClass *prev); - -/** - * @defgroup opt_set_funcs Option setting functions - * @{ - * Those functions set the field of obj with the given name to value. - * - * @param[in] obj A struct whose first element is a pointer to an AVClass. - * @param[in] name the name of the field to set - * @param[in] val The value to set. In case of av_opt_set() if the field is not - * of a string type, then the given string is parsed. - * SI postfixes and some named scalars are supported. - * If the field is of a numeric type, it has to be a numeric or named - * scalar. Behavior with more than one scalar and +- infix operators - * is undefined. - * If the field is of a flags type, it has to be a sequence of numeric - * scalars or named flags separated by '+' or '-'. Prefixing a flag - * with '+' causes it to be set without affecting the other flags; - * similarly, '-' unsets a flag. - * @param search_flags flags passed to av_opt_find2. I.e. if AV_OPT_SEARCH_CHILDREN - * is passed here, then the option may be set on a child of obj. - * - * @return 0 if the value has been set, or an AVERROR code in case of - * error: - * AVERROR_OPTION_NOT_FOUND if no matching option exists - * AVERROR(ERANGE) if the value is out of range - * AVERROR(EINVAL) if the value is not valid - */ -int av_opt_set (void *obj, const char *name, const char *val, int search_flags); -int av_opt_set_int (void *obj, const char *name, int64_t val, int search_flags); -int av_opt_set_double (void *obj, const char *name, double val, int search_flags); -int av_opt_set_q (void *obj, const char *name, AVRational val, int search_flags); -int av_opt_set_bin (void *obj, const char *name, const uint8_t *val, int size, int search_flags); -int av_opt_set_image_size(void *obj, const char *name, int w, int h, int search_flags); -int av_opt_set_pixel_fmt (void *obj, const char *name, enum AVPixelFormat fmt, int search_flags); -int av_opt_set_sample_fmt(void *obj, const char *name, enum AVSampleFormat fmt, int search_flags); -int av_opt_set_video_rate(void *obj, const char *name, AVRational val, int search_flags); -int av_opt_set_channel_layout(void *obj, const char *name, int64_t ch_layout, int search_flags); -/** - * @note Any old dictionary present is discarded and replaced with a copy of the new one. The - * caller still owns val is and responsible for freeing it. - */ -int av_opt_set_dict_val(void *obj, const char *name, const AVDictionary *val, int search_flags); - -/** - * Set a binary option to an integer list. - * - * @param obj AVClass object to set options on - * @param name name of the binary option - * @param val pointer to an integer list (must have the correct type with - * regard to the contents of the list) - * @param term list terminator (usually 0 or -1) - * @param flags search flags - */ -#define av_opt_set_int_list(obj, name, val, term, flags) \ - (av_int_list_length(val, term) > INT_MAX / sizeof(*(val)) ? \ - AVERROR(EINVAL) : \ - av_opt_set_bin(obj, name, (const uint8_t *)(val), \ - av_int_list_length(val, term) * sizeof(*(val)), flags)) - -/** - * @} - */ - -/** - * @defgroup opt_get_funcs Option getting functions - * @{ - * Those functions get a value of the option with the given name from an object. - * - * @param[in] obj a struct whose first element is a pointer to an AVClass. - * @param[in] name name of the option to get. - * @param[in] search_flags flags passed to av_opt_find2. I.e. if AV_OPT_SEARCH_CHILDREN - * is passed here, then the option may be found in a child of obj. - * @param[out] out_val value of the option will be written here - * @return >=0 on success, a negative error code otherwise - */ -/** - * @note the returned string will be av_malloc()ed and must be av_free()ed by the caller - * - * @note if AV_OPT_ALLOW_NULL is set in search_flags in av_opt_get, and the option has - * AV_OPT_TYPE_STRING or AV_OPT_TYPE_BINARY and is set to NULL, *out_val will be set - * to NULL instead of an allocated empty string. - */ -int av_opt_get (void *obj, const char *name, int search_flags, uint8_t **out_val); -int av_opt_get_int (void *obj, const char *name, int search_flags, int64_t *out_val); -int av_opt_get_double (void *obj, const char *name, int search_flags, double *out_val); -int av_opt_get_q (void *obj, const char *name, int search_flags, AVRational *out_val); -int av_opt_get_image_size(void *obj, const char *name, int search_flags, int *w_out, int *h_out); -int av_opt_get_pixel_fmt (void *obj, const char *name, int search_flags, enum AVPixelFormat *out_fmt); -int av_opt_get_sample_fmt(void *obj, const char *name, int search_flags, enum AVSampleFormat *out_fmt); -int av_opt_get_video_rate(void *obj, const char *name, int search_flags, AVRational *out_val); -int av_opt_get_channel_layout(void *obj, const char *name, int search_flags, int64_t *ch_layout); -/** - * @param[out] out_val The returned dictionary is a copy of the actual value and must - * be freed with av_dict_free() by the caller - */ -int av_opt_get_dict_val(void *obj, const char *name, int search_flags, AVDictionary **out_val); -/** - * @} - */ -/** - * Gets a pointer to the requested field in a struct. - * This function allows accessing a struct even when its fields are moved or - * renamed since the application making the access has been compiled, - * - * @returns a pointer to the field, it can be cast to the correct type and read - * or written to. - */ -void *av_opt_ptr(const AVClass *avclass, void *obj, const char *name); - -/** - * Free an AVOptionRanges struct and set it to NULL. - */ -void av_opt_freep_ranges(AVOptionRanges **ranges); - -/** - * Get a list of allowed ranges for the given option. - * - * The returned list may depend on other fields in obj like for example profile. - * - * @param flags is a bitmask of flags, undefined flags should not be set and should be ignored - * AV_OPT_SEARCH_FAKE_OBJ indicates that the obj is a double pointer to a AVClass instead of a full instance - * AV_OPT_MULTI_COMPONENT_RANGE indicates that function may return more than one component, @see AVOptionRanges - * - * The result must be freed with av_opt_freep_ranges. - * - * @return number of compontents returned on success, a negative errro code otherwise - */ -int av_opt_query_ranges(AVOptionRanges **, void *obj, const char *key, int flags); - -/** - * Copy options from src object into dest object. - * - * Options that require memory allocation (e.g. string or binary) are malloc'ed in dest object. - * Original memory allocated for such options is freed unless both src and dest options points to the same memory. - * - * @param dest Object to copy from - * @param src Object to copy into - * @return 0 on success, negative on error - */ -int av_opt_copy(void *dest, const void *src); - -/** - * Get a default list of allowed ranges for the given option. - * - * This list is constructed without using the AVClass.query_ranges() callback - * and can be used as fallback from within the callback. - * - * @param flags is a bitmask of flags, undefined flags should not be set and should be ignored - * AV_OPT_SEARCH_FAKE_OBJ indicates that the obj is a double pointer to a AVClass instead of a full instance - * AV_OPT_MULTI_COMPONENT_RANGE indicates that function may return more than one component, @see AVOptionRanges - * - * The result must be freed with av_opt_free_ranges. - * - * @return number of compontents returned on success, a negative errro code otherwise - */ -int av_opt_query_ranges_default(AVOptionRanges **, void *obj, const char *key, int flags); - -/** - * Check if given option is set to its default value. - * - * Options o must belong to the obj. This function must not be called to check child's options state. - * @see av_opt_is_set_to_default_by_name(). - * - * @param obj AVClass object to check option on - * @param o option to be checked - * @return >0 when option is set to its default, - * 0 when option is not set its default, - * <0 on error - */ -int av_opt_is_set_to_default(void *obj, const AVOption *o); - -/** - * Check if given option is set to its default value. - * - * @param obj AVClass object to check option on - * @param name option name - * @param search_flags combination of AV_OPT_SEARCH_* - * @return >0 when option is set to its default, - * 0 when option is not set its default, - * <0 on error - */ -int av_opt_is_set_to_default_by_name(void *obj, const char *name, int search_flags); - - -#define AV_OPT_SERIALIZE_SKIP_DEFAULTS 0x00000001 ///< Serialize options that are not set to default values only. -#define AV_OPT_SERIALIZE_OPT_FLAGS_EXACT 0x00000002 ///< Serialize options that exactly match opt_flags only. - -/** - * Serialize object's options. - * - * Create a string containing object's serialized options. - * Such string may be passed back to av_opt_set_from_string() in order to restore option values. - * A key/value or pairs separator occurring in the serialized value or - * name string are escaped through the av_escape() function. - * - * @param[in] obj AVClass object to serialize - * @param[in] opt_flags serialize options with all the specified flags set (AV_OPT_FLAG) - * @param[in] flags combination of AV_OPT_SERIALIZE_* flags - * @param[out] buffer Pointer to buffer that will be allocated with string containg serialized options. - * Buffer must be freed by the caller when is no longer needed. - * @param[in] key_val_sep character used to separate key from value - * @param[in] pairs_sep character used to separate two pairs from each other - * @return >= 0 on success, negative on error - * @warning Separators cannot be neither '\\' nor '\0'. They also cannot be the same. - */ -int av_opt_serialize(void *obj, int opt_flags, int flags, char **buffer, - const char key_val_sep, const char pairs_sep); -/** - * @} - */ - -#endif /* AVUTIL_OPT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/parseutils.h b/src/main/jni/ffmpeg/include/libavutil/parseutils.h deleted file mode 100644 index e66d24b76..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/parseutils.h +++ /dev/null @@ -1,193 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_PARSEUTILS_H -#define AVUTIL_PARSEUTILS_H - -#include <time.h> - -#include "rational.h" - -/** - * @file - * misc parsing utilities - */ - -/** - * Parse str and store the parsed ratio in q. - * - * Note that a ratio with infinite (1/0) or negative value is - * considered valid, so you should check on the returned value if you - * want to exclude those values. - * - * The undefined value can be expressed using the "0:0" string. - * - * @param[in,out] q pointer to the AVRational which will contain the ratio - * @param[in] str the string to parse: it has to be a string in the format - * num:den, a float number or an expression - * @param[in] max the maximum allowed numerator and denominator - * @param[in] log_offset log level offset which is applied to the log - * level of log_ctx - * @param[in] log_ctx parent logging context - * @return >= 0 on success, a negative error code otherwise - */ -int av_parse_ratio(AVRational *q, const char *str, int max, - int log_offset, void *log_ctx); - -#define av_parse_ratio_quiet(rate, str, max) \ - av_parse_ratio(rate, str, max, AV_LOG_MAX_OFFSET, NULL) - -/** - * Parse str and put in width_ptr and height_ptr the detected values. - * - * @param[in,out] width_ptr pointer to the variable which will contain the detected - * width value - * @param[in,out] height_ptr pointer to the variable which will contain the detected - * height value - * @param[in] str the string to parse: it has to be a string in the format - * width x height or a valid video size abbreviation. - * @return >= 0 on success, a negative error code otherwise - */ -int av_parse_video_size(int *width_ptr, int *height_ptr, const char *str); - -/** - * Parse str and store the detected values in *rate. - * - * @param[in,out] rate pointer to the AVRational which will contain the detected - * frame rate - * @param[in] str the string to parse: it has to be a string in the format - * rate_num / rate_den, a float number or a valid video rate abbreviation - * @return >= 0 on success, a negative error code otherwise - */ -int av_parse_video_rate(AVRational *rate, const char *str); - -/** - * Put the RGBA values that correspond to color_string in rgba_color. - * - * @param color_string a string specifying a color. It can be the name of - * a color (case insensitive match) or a [0x|#]RRGGBB[AA] sequence, - * possibly followed by "@" and a string representing the alpha - * component. - * The alpha component may be a string composed by "0x" followed by an - * hexadecimal number or a decimal number between 0.0 and 1.0, which - * represents the opacity value (0x00/0.0 means completely transparent, - * 0xff/1.0 completely opaque). - * If the alpha component is not specified then 0xff is assumed. - * The string "random" will result in a random color. - * @param slen length of the initial part of color_string containing the - * color. It can be set to -1 if color_string is a null terminated string - * containing nothing else than the color. - * @return >= 0 in case of success, a negative value in case of - * failure (for example if color_string cannot be parsed). - */ -int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, - void *log_ctx); - -/** - * Get the name of a color from the internal table of hard-coded named - * colors. - * - * This function is meant to enumerate the color names recognized by - * av_parse_color(). - * - * @param color_idx index of the requested color, starting from 0 - * @param rgbp if not NULL, will point to a 3-elements array with the color value in RGB - * @return the color name string or NULL if color_idx is not in the array - */ -const char *av_get_known_color_name(int color_idx, const uint8_t **rgb); - -/** - * Parse timestr and return in *time a corresponding number of - * microseconds. - * - * @param timeval puts here the number of microseconds corresponding - * to the string in timestr. If the string represents a duration, it - * is the number of microseconds contained in the time interval. If - * the string is a date, is the number of microseconds since 1st of - * January, 1970 up to the time of the parsed date. If timestr cannot - * be successfully parsed, set *time to INT64_MIN. - - * @param timestr a string representing a date or a duration. - * - If a date the syntax is: - * @code - * [{YYYY-MM-DD|YYYYMMDD}[T|t| ]]{{HH:MM:SS[.m...]]]}|{HHMMSS[.m...]]]}}[Z] - * now - * @endcode - * If the value is "now" it takes the current time. - * Time is local time unless Z is appended, in which case it is - * interpreted as UTC. - * If the year-month-day part is not specified it takes the current - * year-month-day. - * - If a duration the syntax is: - * @code - * [-][HH:]MM:SS[.m...] - * [-]S+[.m...] - * @endcode - * @param duration flag which tells how to interpret timestr, if not - * zero timestr is interpreted as a duration, otherwise as a date - * @return >= 0 in case of success, a negative value corresponding to an - * AVERROR code otherwise - */ -int av_parse_time(int64_t *timeval, const char *timestr, int duration); - -/** - * Attempt to find a specific tag in a URL. - * - * syntax: '?tag1=val1&tag2=val2...'. Little URL decoding is done. - * Return 1 if found. - */ -int av_find_info_tag(char *arg, int arg_size, const char *tag1, const char *info); - -/** - * Simplified version of strptime - * - * Parse the input string p according to the format string fmt and - * store its results in the structure dt. - * This implementation supports only a subset of the formats supported - * by the standard strptime(). - * - * The supported input field descriptors are listed below. - * - %H: the hour as a decimal number, using a 24-hour clock, in the - * range '00' through '23' - * - %J: hours as a decimal number, in the range '0' through INT_MAX - * - %M: the minute as a decimal number, using a 24-hour clock, in the - * range '00' through '59' - * - %S: the second as a decimal number, using a 24-hour clock, in the - * range '00' through '59' - * - %Y: the year as a decimal number, using the Gregorian calendar - * - %m: the month as a decimal number, in the range '1' through '12' - * - %d: the day of the month as a decimal number, in the range '1' - * through '31' - * - %T: alias for '%H:%M:%S' - * - %%: a literal '%' - * - * @return a pointer to the first character not processed in this function - * call. In case the input string contains more characters than - * required by the format string the return value points right after - * the last consumed input character. In case the whole input string - * is consumed the return value points to the null byte at the end of - * the string. On failure NULL is returned. - */ -char *av_small_strptime(const char *p, const char *fmt, struct tm *dt); - -/** - * Convert the decomposed UTC time in tm to a time_t value. - */ -time_t av_timegm(struct tm *tm); - -#endif /* AVUTIL_PARSEUTILS_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/pixdesc.h b/src/main/jni/ffmpeg/include/libavutil/pixdesc.h deleted file mode 100644 index a6056fe1c..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/pixdesc.h +++ /dev/null @@ -1,394 +0,0 @@ -/* - * pixel format descriptor - * Copyright (c) 2009 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_PIXDESC_H -#define AVUTIL_PIXDESC_H - -#include <inttypes.h> - -#include "attributes.h" -#include "pixfmt.h" -#include "version.h" - -typedef struct AVComponentDescriptor { - /** - * Which of the 4 planes contains the component. - */ - int plane; - - /** - * Number of elements between 2 horizontally consecutive pixels. - * Elements are bits for bitstream formats, bytes otherwise. - */ - int step; - - /** - * Number of elements before the component of the first pixel. - * Elements are bits for bitstream formats, bytes otherwise. - */ - int offset; - - /** - * Number of least significant bits that must be shifted away - * to get the value. - */ - int shift; - - /** - * Number of bits in the component. - */ - int depth; - -#if FF_API_PLUS1_MINUS1 - /** deprecated, use step instead */ - attribute_deprecated int step_minus1; - - /** deprecated, use depth instead */ - attribute_deprecated int depth_minus1; - - /** deprecated, use offset instead */ - attribute_deprecated int offset_plus1; -#endif -} AVComponentDescriptor; - -/** - * Descriptor that unambiguously describes how the bits of a pixel are - * stored in the up to 4 data planes of an image. It also stores the - * subsampling factors and number of components. - * - * @note This is separate of the colorspace (RGB, YCbCr, YPbPr, JPEG-style YUV - * and all the YUV variants) AVPixFmtDescriptor just stores how values - * are stored not what these values represent. - */ -typedef struct AVPixFmtDescriptor { - const char *name; - uint8_t nb_components; ///< The number of components each pixel has, (1-4) - - /** - * Amount to shift the luma width right to find the chroma width. - * For YV12 this is 1 for example. - * chroma_width = -((-luma_width) >> log2_chroma_w) - * The note above is needed to ensure rounding up. - * This value only refers to the chroma components. - */ - uint8_t log2_chroma_w; ///< chroma_width = -((-luma_width )>>log2_chroma_w) - - /** - * Amount to shift the luma height right to find the chroma height. - * For YV12 this is 1 for example. - * chroma_height= -((-luma_height) >> log2_chroma_h) - * The note above is needed to ensure rounding up. - * This value only refers to the chroma components. - */ - uint8_t log2_chroma_h; - - /** - * Combination of AV_PIX_FMT_FLAG_... flags. - */ - uint64_t flags; - - /** - * Parameters that describe how pixels are packed. - * If the format has 1 or 2 components, then luma is 0. - * If the format has 3 or 4 components: - * if the RGB flag is set then 0 is red, 1 is green and 2 is blue; - * otherwise 0 is luma, 1 is chroma-U and 2 is chroma-V. - * - * If present, the Alpha channel is always the last component. - */ - AVComponentDescriptor comp[4]; - - /** - * Alternative comma-separated names. - */ - const char *alias; -} AVPixFmtDescriptor; - -/** - * Pixel format is big-endian. - */ -#define AV_PIX_FMT_FLAG_BE (1 << 0) -/** - * Pixel format has a palette in data[1], values are indexes in this palette. - */ -#define AV_PIX_FMT_FLAG_PAL (1 << 1) -/** - * All values of a component are bit-wise packed end to end. - */ -#define AV_PIX_FMT_FLAG_BITSTREAM (1 << 2) -/** - * Pixel format is an HW accelerated format. - */ -#define AV_PIX_FMT_FLAG_HWACCEL (1 << 3) -/** - * At least one pixel component is not in the first data plane. - */ -#define AV_PIX_FMT_FLAG_PLANAR (1 << 4) -/** - * The pixel format contains RGB-like data (as opposed to YUV/grayscale). - */ -#define AV_PIX_FMT_FLAG_RGB (1 << 5) - -/** - * The pixel format is "pseudo-paletted". This means that it contains a - * fixed palette in the 2nd plane but the palette is fixed/constant for each - * PIX_FMT. This allows interpreting the data as if it was PAL8, which can - * in some cases be simpler. Or the data can be interpreted purely based on - * the pixel format without using the palette. - * An example of a pseudo-paletted format is AV_PIX_FMT_GRAY8 - */ -#define AV_PIX_FMT_FLAG_PSEUDOPAL (1 << 6) - -/** - * The pixel format has an alpha channel. This is set on all formats that - * support alpha in some way. The exception is AV_PIX_FMT_PAL8, which can - * carry alpha as part of the palette. Details are explained in the - * AVPixelFormat enum, and are also encoded in the corresponding - * AVPixFmtDescriptor. - * - * The alpha is always straight, never pre-multiplied. - * - * If a codec or a filter does not support alpha, it should set all alpha to - * opaque, or use the equivalent pixel formats without alpha component, e.g. - * AV_PIX_FMT_RGB0 (or AV_PIX_FMT_RGB24 etc.) instead of AV_PIX_FMT_RGBA. - */ -#define AV_PIX_FMT_FLAG_ALPHA (1 << 7) - -/** - * Read a line from an image, and write the values of the - * pixel format component c to dst. - * - * @param data the array containing the pointers to the planes of the image - * @param linesize the array containing the linesizes of the image - * @param desc the pixel format descriptor for the image - * @param x the horizontal coordinate of the first pixel to read - * @param y the vertical coordinate of the first pixel to read - * @param w the width of the line to read, that is the number of - * values to write to dst - * @param read_pal_component if not zero and the format is a paletted - * format writes the values corresponding to the palette - * component c in data[1] to dst, rather than the palette indexes in - * data[0]. The behavior is undefined if the format is not paletted. - */ -void av_read_image_line(uint16_t *dst, const uint8_t *data[4], - const int linesize[4], const AVPixFmtDescriptor *desc, - int x, int y, int c, int w, int read_pal_component); - -/** - * Write the values from src to the pixel format component c of an - * image line. - * - * @param src array containing the values to write - * @param data the array containing the pointers to the planes of the - * image to write into. It is supposed to be zeroed. - * @param linesize the array containing the linesizes of the image - * @param desc the pixel format descriptor for the image - * @param x the horizontal coordinate of the first pixel to write - * @param y the vertical coordinate of the first pixel to write - * @param w the width of the line to write, that is the number of - * values to write to the image line - */ -void av_write_image_line(const uint16_t *src, uint8_t *data[4], - const int linesize[4], const AVPixFmtDescriptor *desc, - int x, int y, int c, int w); - -/** - * Return the pixel format corresponding to name. - * - * If there is no pixel format with name name, then looks for a - * pixel format with the name corresponding to the native endian - * format of name. - * For example in a little-endian system, first looks for "gray16", - * then for "gray16le". - * - * Finally if no pixel format has been found, returns AV_PIX_FMT_NONE. - */ -enum AVPixelFormat av_get_pix_fmt(const char *name); - -/** - * Return the short name for a pixel format, NULL in case pix_fmt is - * unknown. - * - * @see av_get_pix_fmt(), av_get_pix_fmt_string() - */ -const char *av_get_pix_fmt_name(enum AVPixelFormat pix_fmt); - -/** - * Print in buf the string corresponding to the pixel format with - * number pix_fmt, or a header if pix_fmt is negative. - * - * @param buf the buffer where to write the string - * @param buf_size the size of buf - * @param pix_fmt the number of the pixel format to print the - * corresponding info string, or a negative value to print the - * corresponding header. - */ -char *av_get_pix_fmt_string(char *buf, int buf_size, - enum AVPixelFormat pix_fmt); - -/** - * Return the number of bits per pixel used by the pixel format - * described by pixdesc. Note that this is not the same as the number - * of bits per sample. - * - * The returned number of bits refers to the number of bits actually - * used for storing the pixel information, that is padding bits are - * not counted. - */ -int av_get_bits_per_pixel(const AVPixFmtDescriptor *pixdesc); - -/** - * Return the number of bits per pixel for the pixel format - * described by pixdesc, including any padding or unused bits. - */ -int av_get_padded_bits_per_pixel(const AVPixFmtDescriptor *pixdesc); - -/** - * @return a pixel format descriptor for provided pixel format or NULL if - * this pixel format is unknown. - */ -const AVPixFmtDescriptor *av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt); - -/** - * Iterate over all pixel format descriptors known to libavutil. - * - * @param prev previous descriptor. NULL to get the first descriptor. - * - * @return next descriptor or NULL after the last descriptor - */ -const AVPixFmtDescriptor *av_pix_fmt_desc_next(const AVPixFmtDescriptor *prev); - -/** - * @return an AVPixelFormat id described by desc, or AV_PIX_FMT_NONE if desc - * is not a valid pointer to a pixel format descriptor. - */ -enum AVPixelFormat av_pix_fmt_desc_get_id(const AVPixFmtDescriptor *desc); - -/** - * Utility function to access log2_chroma_w log2_chroma_h from - * the pixel format AVPixFmtDescriptor. - * - * See av_get_chroma_sub_sample() for a function that asserts a - * valid pixel format instead of returning an error code. - * Its recommended that you use avcodec_get_chroma_sub_sample unless - * you do check the return code! - * - * @param[in] pix_fmt the pixel format - * @param[out] h_shift store log2_chroma_w - * @param[out] v_shift store log2_chroma_h - * - * @return 0 on success, AVERROR(ENOSYS) on invalid or unknown pixel format - */ -int av_pix_fmt_get_chroma_sub_sample(enum AVPixelFormat pix_fmt, - int *h_shift, int *v_shift); - -/** - * @return number of planes in pix_fmt, a negative AVERROR if pix_fmt is not a - * valid pixel format. - */ -int av_pix_fmt_count_planes(enum AVPixelFormat pix_fmt); - -/** - * Utility function to swap the endianness of a pixel format. - * - * @param[in] pix_fmt the pixel format - * - * @return pixel format with swapped endianness if it exists, - * otherwise AV_PIX_FMT_NONE - */ -enum AVPixelFormat av_pix_fmt_swap_endianness(enum AVPixelFormat pix_fmt); - -#define FF_LOSS_RESOLUTION 0x0001 /**< loss due to resolution change */ -#define FF_LOSS_DEPTH 0x0002 /**< loss due to color depth change */ -#define FF_LOSS_COLORSPACE 0x0004 /**< loss due to color space conversion */ -#define FF_LOSS_ALPHA 0x0008 /**< loss of alpha bits */ -#define FF_LOSS_COLORQUANT 0x0010 /**< loss due to color quantization */ -#define FF_LOSS_CHROMA 0x0020 /**< loss of chroma (e.g. RGB to gray conversion) */ - -/** - * Compute what kind of losses will occur when converting from one specific - * pixel format to another. - * When converting from one pixel format to another, information loss may occur. - * For example, when converting from RGB24 to GRAY, the color information will - * be lost. Similarly, other losses occur when converting from some formats to - * other formats. These losses can involve loss of chroma, but also loss of - * resolution, loss of color depth, loss due to the color space conversion, loss - * of the alpha bits or loss due to color quantization. - * av_get_fix_fmt_loss() informs you about the various types of losses - * which will occur when converting from one pixel format to another. - * - * @param[in] dst_pix_fmt destination pixel format - * @param[in] src_pix_fmt source pixel format - * @param[in] has_alpha Whether the source pixel format alpha channel is used. - * @return Combination of flags informing you what kind of losses will occur - * (maximum loss for an invalid dst_pix_fmt). - */ -int av_get_pix_fmt_loss(enum AVPixelFormat dst_pix_fmt, - enum AVPixelFormat src_pix_fmt, - int has_alpha); - -/** - * Compute what kind of losses will occur when converting from one specific - * pixel format to another. - * When converting from one pixel format to another, information loss may occur. - * For example, when converting from RGB24 to GRAY, the color information will - * be lost. Similarly, other losses occur when converting from some formats to - * other formats. These losses can involve loss of chroma, but also loss of - * resolution, loss of color depth, loss due to the color space conversion, loss - * of the alpha bits or loss due to color quantization. - * av_get_fix_fmt_loss() informs you about the various types of losses - * which will occur when converting from one pixel format to another. - * - * @param[in] dst_pix_fmt destination pixel format - * @param[in] src_pix_fmt source pixel format - * @param[in] has_alpha Whether the source pixel format alpha channel is used. - * @return Combination of flags informing you what kind of losses will occur - * (maximum loss for an invalid dst_pix_fmt). - */ -enum AVPixelFormat av_find_best_pix_fmt_of_2(enum AVPixelFormat dst_pix_fmt1, enum AVPixelFormat dst_pix_fmt2, - enum AVPixelFormat src_pix_fmt, int has_alpha, int *loss_ptr); - -/** - * @return the name for provided color range or NULL if unknown. - */ -const char *av_color_range_name(enum AVColorRange range); - -/** - * @return the name for provided color primaries or NULL if unknown. - */ -const char *av_color_primaries_name(enum AVColorPrimaries primaries); - -/** - * @return the name for provided color transfer or NULL if unknown. - */ -const char *av_color_transfer_name(enum AVColorTransferCharacteristic transfer); - -/** - * @return the name for provided color space or NULL if unknown. - */ -const char *av_color_space_name(enum AVColorSpace space); - -/** - * @return the name for provided chroma location or NULL if unknown. - */ -const char *av_chroma_location_name(enum AVChromaLocation location); - -#endif /* AVUTIL_PIXDESC_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/pixelutils.h b/src/main/jni/ffmpeg/include/libavutil/pixelutils.h deleted file mode 100644 index a8dbc157e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/pixelutils.h +++ /dev/null @@ -1,52 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_PIXELUTILS_H -#define AVUTIL_PIXELUTILS_H - -#include <stddef.h> -#include <stdint.h> -#include "common.h" - -/** - * Sum of abs(src1[x] - src2[x]) - */ -typedef int (*av_pixelutils_sad_fn)(const uint8_t *src1, ptrdiff_t stride1, - const uint8_t *src2, ptrdiff_t stride2); - -/** - * Get a potentially optimized pointer to a Sum-of-absolute-differences - * function (see the av_pixelutils_sad_fn prototype). - * - * @param w_bits 1<<w_bits is the requested width of the block size - * @param h_bits 1<<h_bits is the requested height of the block size - * @param aligned If set to 2, the returned sad function will assume src1 and - * src2 addresses are aligned on the block size. - * If set to 1, the returned sad function will assume src1 is - * aligned on the block size. - * If set to 0, the returned sad function assume no particular - * alignment. - * @param log_ctx context used for logging, can be NULL - * - * @return a pointer to the SAD function or NULL in case of error (because of - * invalid parameters) - */ -av_pixelutils_sad_fn av_pixelutils_get_sad_fn(int w_bits, int h_bits, - int aligned, void *log_ctx); - -#endif /* AVUTIL_PIXELUTILS_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/pixfmt.h b/src/main/jni/ffmpeg/include/libavutil/pixfmt.h deleted file mode 100644 index 32044f077..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/pixfmt.h +++ /dev/null @@ -1,469 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_PIXFMT_H -#define AVUTIL_PIXFMT_H - -/** - * @file - * pixel format definitions - * - */ - -#include "libavutil/avconfig.h" -#include "version.h" - -#define AVPALETTE_SIZE 1024 -#define AVPALETTE_COUNT 256 - -/** - * Pixel format. - * - * @note - * AV_PIX_FMT_RGB32 is handled in an endian-specific manner. An RGBA - * color is put together as: - * (A << 24) | (R << 16) | (G << 8) | B - * This is stored as BGRA on little-endian CPU architectures and ARGB on - * big-endian CPUs. - * - * @par - * When the pixel format is palettized RGB32 (AV_PIX_FMT_PAL8), the palettized - * image data is stored in AVFrame.data[0]. The palette is transported in - * AVFrame.data[1], is 1024 bytes long (256 4-byte entries) and is - * formatted the same as in AV_PIX_FMT_RGB32 described above (i.e., it is - * also endian-specific). Note also that the individual RGB32 palette - * components stored in AVFrame.data[1] should be in the range 0..255. - * This is important as many custom PAL8 video codecs that were designed - * to run on the IBM VGA graphics adapter use 6-bit palette components. - * - * @par - * For all the 8bit per pixel formats, an RGB32 palette is in data[1] like - * for pal8. This palette is filled in automatically by the function - * allocating the picture. - */ -enum AVPixelFormat { - AV_PIX_FMT_NONE = -1, - AV_PIX_FMT_YUV420P, ///< planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples) - AV_PIX_FMT_YUYV422, ///< packed YUV 4:2:2, 16bpp, Y0 Cb Y1 Cr - AV_PIX_FMT_RGB24, ///< packed RGB 8:8:8, 24bpp, RGBRGB... - AV_PIX_FMT_BGR24, ///< packed RGB 8:8:8, 24bpp, BGRBGR... - AV_PIX_FMT_YUV422P, ///< planar YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples) - AV_PIX_FMT_YUV444P, ///< planar YUV 4:4:4, 24bpp, (1 Cr & Cb sample per 1x1 Y samples) - AV_PIX_FMT_YUV410P, ///< planar YUV 4:1:0, 9bpp, (1 Cr & Cb sample per 4x4 Y samples) - AV_PIX_FMT_YUV411P, ///< planar YUV 4:1:1, 12bpp, (1 Cr & Cb sample per 4x1 Y samples) - AV_PIX_FMT_GRAY8, ///< Y , 8bpp - AV_PIX_FMT_MONOWHITE, ///< Y , 1bpp, 0 is white, 1 is black, in each byte pixels are ordered from the msb to the lsb - AV_PIX_FMT_MONOBLACK, ///< Y , 1bpp, 0 is black, 1 is white, in each byte pixels are ordered from the msb to the lsb - AV_PIX_FMT_PAL8, ///< 8 bit with AV_PIX_FMT_RGB32 palette - AV_PIX_FMT_YUVJ420P, ///< planar YUV 4:2:0, 12bpp, full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV420P and setting color_range - AV_PIX_FMT_YUVJ422P, ///< planar YUV 4:2:2, 16bpp, full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV422P and setting color_range - AV_PIX_FMT_YUVJ444P, ///< planar YUV 4:4:4, 24bpp, full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV444P and setting color_range -#if FF_API_XVMC - AV_PIX_FMT_XVMC_MPEG2_MC,///< XVideo Motion Acceleration via common packet passing - AV_PIX_FMT_XVMC_MPEG2_IDCT, -#define AV_PIX_FMT_XVMC AV_PIX_FMT_XVMC_MPEG2_IDCT -#endif /* FF_API_XVMC */ - AV_PIX_FMT_UYVY422, ///< packed YUV 4:2:2, 16bpp, Cb Y0 Cr Y1 - AV_PIX_FMT_UYYVYY411, ///< packed YUV 4:1:1, 12bpp, Cb Y0 Y1 Cr Y2 Y3 - AV_PIX_FMT_BGR8, ///< packed RGB 3:3:2, 8bpp, (msb)2B 3G 3R(lsb) - AV_PIX_FMT_BGR4, ///< packed RGB 1:2:1 bitstream, 4bpp, (msb)1B 2G 1R(lsb), a byte contains two pixels, the first pixel in the byte is the one composed by the 4 msb bits - AV_PIX_FMT_BGR4_BYTE, ///< packed RGB 1:2:1, 8bpp, (msb)1B 2G 1R(lsb) - AV_PIX_FMT_RGB8, ///< packed RGB 3:3:2, 8bpp, (msb)2R 3G 3B(lsb) - AV_PIX_FMT_RGB4, ///< packed RGB 1:2:1 bitstream, 4bpp, (msb)1R 2G 1B(lsb), a byte contains two pixels, the first pixel in the byte is the one composed by the 4 msb bits - AV_PIX_FMT_RGB4_BYTE, ///< packed RGB 1:2:1, 8bpp, (msb)1R 2G 1B(lsb) - AV_PIX_FMT_NV12, ///< planar YUV 4:2:0, 12bpp, 1 plane for Y and 1 plane for the UV components, which are interleaved (first byte U and the following byte V) - AV_PIX_FMT_NV21, ///< as above, but U and V bytes are swapped - - AV_PIX_FMT_ARGB, ///< packed ARGB 8:8:8:8, 32bpp, ARGBARGB... - AV_PIX_FMT_RGBA, ///< packed RGBA 8:8:8:8, 32bpp, RGBARGBA... - AV_PIX_FMT_ABGR, ///< packed ABGR 8:8:8:8, 32bpp, ABGRABGR... - AV_PIX_FMT_BGRA, ///< packed BGRA 8:8:8:8, 32bpp, BGRABGRA... - - AV_PIX_FMT_GRAY16BE, ///< Y , 16bpp, big-endian - AV_PIX_FMT_GRAY16LE, ///< Y , 16bpp, little-endian - AV_PIX_FMT_YUV440P, ///< planar YUV 4:4:0 (1 Cr & Cb sample per 1x2 Y samples) - AV_PIX_FMT_YUVJ440P, ///< planar YUV 4:4:0 full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV440P and setting color_range - AV_PIX_FMT_YUVA420P, ///< planar YUV 4:2:0, 20bpp, (1 Cr & Cb sample per 2x2 Y & A samples) -#if FF_API_VDPAU - AV_PIX_FMT_VDPAU_H264,///< H.264 HW decoding with VDPAU, data[0] contains a vdpau_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers - AV_PIX_FMT_VDPAU_MPEG1,///< MPEG-1 HW decoding with VDPAU, data[0] contains a vdpau_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers - AV_PIX_FMT_VDPAU_MPEG2,///< MPEG-2 HW decoding with VDPAU, data[0] contains a vdpau_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers - AV_PIX_FMT_VDPAU_WMV3,///< WMV3 HW decoding with VDPAU, data[0] contains a vdpau_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers - AV_PIX_FMT_VDPAU_VC1, ///< VC-1 HW decoding with VDPAU, data[0] contains a vdpau_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers -#endif - AV_PIX_FMT_RGB48BE, ///< packed RGB 16:16:16, 48bpp, 16R, 16G, 16B, the 2-byte value for each R/G/B component is stored as big-endian - AV_PIX_FMT_RGB48LE, ///< packed RGB 16:16:16, 48bpp, 16R, 16G, 16B, the 2-byte value for each R/G/B component is stored as little-endian - - AV_PIX_FMT_RGB565BE, ///< packed RGB 5:6:5, 16bpp, (msb) 5R 6G 5B(lsb), big-endian - AV_PIX_FMT_RGB565LE, ///< packed RGB 5:6:5, 16bpp, (msb) 5R 6G 5B(lsb), little-endian - AV_PIX_FMT_RGB555BE, ///< packed RGB 5:5:5, 16bpp, (msb)1X 5R 5G 5B(lsb), big-endian , X=unused/undefined - AV_PIX_FMT_RGB555LE, ///< packed RGB 5:5:5, 16bpp, (msb)1X 5R 5G 5B(lsb), little-endian, X=unused/undefined - - AV_PIX_FMT_BGR565BE, ///< packed BGR 5:6:5, 16bpp, (msb) 5B 6G 5R(lsb), big-endian - AV_PIX_FMT_BGR565LE, ///< packed BGR 5:6:5, 16bpp, (msb) 5B 6G 5R(lsb), little-endian - AV_PIX_FMT_BGR555BE, ///< packed BGR 5:5:5, 16bpp, (msb)1X 5B 5G 5R(lsb), big-endian , X=unused/undefined - AV_PIX_FMT_BGR555LE, ///< packed BGR 5:5:5, 16bpp, (msb)1X 5B 5G 5R(lsb), little-endian, X=unused/undefined - -#if FF_API_VAAPI - /** @name Deprecated pixel formats */ - /**@{*/ - AV_PIX_FMT_VAAPI_MOCO, ///< HW acceleration through VA API at motion compensation entry-point, Picture.data[3] contains a vaapi_render_state struct which contains macroblocks as well as various fields extracted from headers - AV_PIX_FMT_VAAPI_IDCT, ///< HW acceleration through VA API at IDCT entry-point, Picture.data[3] contains a vaapi_render_state struct which contains fields extracted from headers - AV_PIX_FMT_VAAPI_VLD, ///< HW decoding through VA API, Picture.data[3] contains a vaapi_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers - /**@}*/ - AV_PIX_FMT_VAAPI = AV_PIX_FMT_VAAPI_VLD, -#else - /** - * Hardware acceleration through VA-API, data[3] contains a - * VASurfaceID. - */ - AV_PIX_FMT_VAAPI, -#endif - - AV_PIX_FMT_YUV420P16LE, ///< planar YUV 4:2:0, 24bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian - AV_PIX_FMT_YUV420P16BE, ///< planar YUV 4:2:0, 24bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian - AV_PIX_FMT_YUV422P16LE, ///< planar YUV 4:2:2, 32bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian - AV_PIX_FMT_YUV422P16BE, ///< planar YUV 4:2:2, 32bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian - AV_PIX_FMT_YUV444P16LE, ///< planar YUV 4:4:4, 48bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian - AV_PIX_FMT_YUV444P16BE, ///< planar YUV 4:4:4, 48bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian -#if FF_API_VDPAU - AV_PIX_FMT_VDPAU_MPEG4, ///< MPEG4 HW decoding with VDPAU, data[0] contains a vdpau_render_state struct which contains the bitstream of the slices as well as various fields extracted from headers -#endif - AV_PIX_FMT_DXVA2_VLD, ///< HW decoding through DXVA2, Picture.data[3] contains a LPDIRECT3DSURFACE9 pointer - - AV_PIX_FMT_RGB444LE, ///< packed RGB 4:4:4, 16bpp, (msb)4X 4R 4G 4B(lsb), little-endian, X=unused/undefined - AV_PIX_FMT_RGB444BE, ///< packed RGB 4:4:4, 16bpp, (msb)4X 4R 4G 4B(lsb), big-endian, X=unused/undefined - AV_PIX_FMT_BGR444LE, ///< packed BGR 4:4:4, 16bpp, (msb)4X 4B 4G 4R(lsb), little-endian, X=unused/undefined - AV_PIX_FMT_BGR444BE, ///< packed BGR 4:4:4, 16bpp, (msb)4X 4B 4G 4R(lsb), big-endian, X=unused/undefined - AV_PIX_FMT_YA8, ///< 8bit gray, 8bit alpha - - AV_PIX_FMT_Y400A = AV_PIX_FMT_YA8, ///< alias for AV_PIX_FMT_YA8 - AV_PIX_FMT_GRAY8A= AV_PIX_FMT_YA8, ///< alias for AV_PIX_FMT_YA8 - - AV_PIX_FMT_BGR48BE, ///< packed RGB 16:16:16, 48bpp, 16B, 16G, 16R, the 2-byte value for each R/G/B component is stored as big-endian - AV_PIX_FMT_BGR48LE, ///< packed RGB 16:16:16, 48bpp, 16B, 16G, 16R, the 2-byte value for each R/G/B component is stored as little-endian - - /** - * The following 12 formats have the disadvantage of needing 1 format for each bit depth. - * Notice that each 9/10 bits sample is stored in 16 bits with extra padding. - * If you want to support multiple bit depths, then using AV_PIX_FMT_YUV420P16* with the bpp stored separately is better. - */ - AV_PIX_FMT_YUV420P9BE, ///< planar YUV 4:2:0, 13.5bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian - AV_PIX_FMT_YUV420P9LE, ///< planar YUV 4:2:0, 13.5bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian - AV_PIX_FMT_YUV420P10BE,///< planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian - AV_PIX_FMT_YUV420P10LE,///< planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian - AV_PIX_FMT_YUV422P10BE,///< planar YUV 4:2:2, 20bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian - AV_PIX_FMT_YUV422P10LE,///< planar YUV 4:2:2, 20bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian - AV_PIX_FMT_YUV444P9BE, ///< planar YUV 4:4:4, 27bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian - AV_PIX_FMT_YUV444P9LE, ///< planar YUV 4:4:4, 27bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian - AV_PIX_FMT_YUV444P10BE,///< planar YUV 4:4:4, 30bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian - AV_PIX_FMT_YUV444P10LE,///< planar YUV 4:4:4, 30bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian - AV_PIX_FMT_YUV422P9BE, ///< planar YUV 4:2:2, 18bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian - AV_PIX_FMT_YUV422P9LE, ///< planar YUV 4:2:2, 18bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian - AV_PIX_FMT_VDA_VLD, ///< hardware decoding through VDA - AV_PIX_FMT_GBRP, ///< planar GBR 4:4:4 24bpp - AV_PIX_FMT_GBRP9BE, ///< planar GBR 4:4:4 27bpp, big-endian - AV_PIX_FMT_GBRP9LE, ///< planar GBR 4:4:4 27bpp, little-endian - AV_PIX_FMT_GBRP10BE, ///< planar GBR 4:4:4 30bpp, big-endian - AV_PIX_FMT_GBRP10LE, ///< planar GBR 4:4:4 30bpp, little-endian - AV_PIX_FMT_GBRP16BE, ///< planar GBR 4:4:4 48bpp, big-endian - AV_PIX_FMT_GBRP16LE, ///< planar GBR 4:4:4 48bpp, little-endian - AV_PIX_FMT_YUVA422P, ///< planar YUV 4:2:2 24bpp, (1 Cr & Cb sample per 2x1 Y & A samples) - AV_PIX_FMT_YUVA444P, ///< planar YUV 4:4:4 32bpp, (1 Cr & Cb sample per 1x1 Y & A samples) - AV_PIX_FMT_YUVA420P9BE, ///< planar YUV 4:2:0 22.5bpp, (1 Cr & Cb sample per 2x2 Y & A samples), big-endian - AV_PIX_FMT_YUVA420P9LE, ///< planar YUV 4:2:0 22.5bpp, (1 Cr & Cb sample per 2x2 Y & A samples), little-endian - AV_PIX_FMT_YUVA422P9BE, ///< planar YUV 4:2:2 27bpp, (1 Cr & Cb sample per 2x1 Y & A samples), big-endian - AV_PIX_FMT_YUVA422P9LE, ///< planar YUV 4:2:2 27bpp, (1 Cr & Cb sample per 2x1 Y & A samples), little-endian - AV_PIX_FMT_YUVA444P9BE, ///< planar YUV 4:4:4 36bpp, (1 Cr & Cb sample per 1x1 Y & A samples), big-endian - AV_PIX_FMT_YUVA444P9LE, ///< planar YUV 4:4:4 36bpp, (1 Cr & Cb sample per 1x1 Y & A samples), little-endian - AV_PIX_FMT_YUVA420P10BE, ///< planar YUV 4:2:0 25bpp, (1 Cr & Cb sample per 2x2 Y & A samples, big-endian) - AV_PIX_FMT_YUVA420P10LE, ///< planar YUV 4:2:0 25bpp, (1 Cr & Cb sample per 2x2 Y & A samples, little-endian) - AV_PIX_FMT_YUVA422P10BE, ///< planar YUV 4:2:2 30bpp, (1 Cr & Cb sample per 2x1 Y & A samples, big-endian) - AV_PIX_FMT_YUVA422P10LE, ///< planar YUV 4:2:2 30bpp, (1 Cr & Cb sample per 2x1 Y & A samples, little-endian) - AV_PIX_FMT_YUVA444P10BE, ///< planar YUV 4:4:4 40bpp, (1 Cr & Cb sample per 1x1 Y & A samples, big-endian) - AV_PIX_FMT_YUVA444P10LE, ///< planar YUV 4:4:4 40bpp, (1 Cr & Cb sample per 1x1 Y & A samples, little-endian) - AV_PIX_FMT_YUVA420P16BE, ///< planar YUV 4:2:0 40bpp, (1 Cr & Cb sample per 2x2 Y & A samples, big-endian) - AV_PIX_FMT_YUVA420P16LE, ///< planar YUV 4:2:0 40bpp, (1 Cr & Cb sample per 2x2 Y & A samples, little-endian) - AV_PIX_FMT_YUVA422P16BE, ///< planar YUV 4:2:2 48bpp, (1 Cr & Cb sample per 2x1 Y & A samples, big-endian) - AV_PIX_FMT_YUVA422P16LE, ///< planar YUV 4:2:2 48bpp, (1 Cr & Cb sample per 2x1 Y & A samples, little-endian) - AV_PIX_FMT_YUVA444P16BE, ///< planar YUV 4:4:4 64bpp, (1 Cr & Cb sample per 1x1 Y & A samples, big-endian) - AV_PIX_FMT_YUVA444P16LE, ///< planar YUV 4:4:4 64bpp, (1 Cr & Cb sample per 1x1 Y & A samples, little-endian) - - AV_PIX_FMT_VDPAU, ///< HW acceleration through VDPAU, Picture.data[3] contains a VdpVideoSurface - - AV_PIX_FMT_XYZ12LE, ///< packed XYZ 4:4:4, 36 bpp, (msb) 12X, 12Y, 12Z (lsb), the 2-byte value for each X/Y/Z is stored as little-endian, the 4 lower bits are set to 0 - AV_PIX_FMT_XYZ12BE, ///< packed XYZ 4:4:4, 36 bpp, (msb) 12X, 12Y, 12Z (lsb), the 2-byte value for each X/Y/Z is stored as big-endian, the 4 lower bits are set to 0 - AV_PIX_FMT_NV16, ///< interleaved chroma YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples) - AV_PIX_FMT_NV20LE, ///< interleaved chroma YUV 4:2:2, 20bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian - AV_PIX_FMT_NV20BE, ///< interleaved chroma YUV 4:2:2, 20bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian - - AV_PIX_FMT_RGBA64BE, ///< packed RGBA 16:16:16:16, 64bpp, 16R, 16G, 16B, 16A, the 2-byte value for each R/G/B/A component is stored as big-endian - AV_PIX_FMT_RGBA64LE, ///< packed RGBA 16:16:16:16, 64bpp, 16R, 16G, 16B, 16A, the 2-byte value for each R/G/B/A component is stored as little-endian - AV_PIX_FMT_BGRA64BE, ///< packed RGBA 16:16:16:16, 64bpp, 16B, 16G, 16R, 16A, the 2-byte value for each R/G/B/A component is stored as big-endian - AV_PIX_FMT_BGRA64LE, ///< packed RGBA 16:16:16:16, 64bpp, 16B, 16G, 16R, 16A, the 2-byte value for each R/G/B/A component is stored as little-endian - - AV_PIX_FMT_YVYU422, ///< packed YUV 4:2:2, 16bpp, Y0 Cr Y1 Cb - - AV_PIX_FMT_VDA, ///< HW acceleration through VDA, data[3] contains a CVPixelBufferRef - - AV_PIX_FMT_YA16BE, ///< 16bit gray, 16bit alpha (big-endian) - AV_PIX_FMT_YA16LE, ///< 16bit gray, 16bit alpha (little-endian) - - AV_PIX_FMT_GBRAP, ///< planar GBRA 4:4:4:4 32bpp - AV_PIX_FMT_GBRAP16BE, ///< planar GBRA 4:4:4:4 64bpp, big-endian - AV_PIX_FMT_GBRAP16LE, ///< planar GBRA 4:4:4:4 64bpp, little-endian - /** - * HW acceleration through QSV, data[3] contains a pointer to the - * mfxFrameSurface1 structure. - */ - AV_PIX_FMT_QSV, - /** - * HW acceleration though MMAL, data[3] contains a pointer to the - * MMAL_BUFFER_HEADER_T structure. - */ - AV_PIX_FMT_MMAL, - - AV_PIX_FMT_D3D11VA_VLD, ///< HW decoding through Direct3D11, Picture.data[3] contains a ID3D11VideoDecoderOutputView pointer - - AV_PIX_FMT_0RGB=0x123+4,///< packed RGB 8:8:8, 32bpp, XRGBXRGB... X=unused/undefined - AV_PIX_FMT_RGB0, ///< packed RGB 8:8:8, 32bpp, RGBXRGBX... X=unused/undefined - AV_PIX_FMT_0BGR, ///< packed BGR 8:8:8, 32bpp, XBGRXBGR... X=unused/undefined - AV_PIX_FMT_BGR0, ///< packed BGR 8:8:8, 32bpp, BGRXBGRX... X=unused/undefined - - AV_PIX_FMT_YUV420P12BE, ///< planar YUV 4:2:0,18bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian - AV_PIX_FMT_YUV420P12LE, ///< planar YUV 4:2:0,18bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian - AV_PIX_FMT_YUV420P14BE, ///< planar YUV 4:2:0,21bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian - AV_PIX_FMT_YUV420P14LE, ///< planar YUV 4:2:0,21bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian - AV_PIX_FMT_YUV422P12BE, ///< planar YUV 4:2:2,24bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian - AV_PIX_FMT_YUV422P12LE, ///< planar YUV 4:2:2,24bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian - AV_PIX_FMT_YUV422P14BE, ///< planar YUV 4:2:2,28bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian - AV_PIX_FMT_YUV422P14LE, ///< planar YUV 4:2:2,28bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian - AV_PIX_FMT_YUV444P12BE, ///< planar YUV 4:4:4,36bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian - AV_PIX_FMT_YUV444P12LE, ///< planar YUV 4:4:4,36bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian - AV_PIX_FMT_YUV444P14BE, ///< planar YUV 4:4:4,42bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian - AV_PIX_FMT_YUV444P14LE, ///< planar YUV 4:4:4,42bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian - AV_PIX_FMT_GBRP12BE, ///< planar GBR 4:4:4 36bpp, big-endian - AV_PIX_FMT_GBRP12LE, ///< planar GBR 4:4:4 36bpp, little-endian - AV_PIX_FMT_GBRP14BE, ///< planar GBR 4:4:4 42bpp, big-endian - AV_PIX_FMT_GBRP14LE, ///< planar GBR 4:4:4 42bpp, little-endian - AV_PIX_FMT_YUVJ411P, ///< planar YUV 4:1:1, 12bpp, (1 Cr & Cb sample per 4x1 Y samples) full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV411P and setting color_range - - AV_PIX_FMT_BAYER_BGGR8, ///< bayer, BGBG..(odd line), GRGR..(even line), 8-bit samples */ - AV_PIX_FMT_BAYER_RGGB8, ///< bayer, RGRG..(odd line), GBGB..(even line), 8-bit samples */ - AV_PIX_FMT_BAYER_GBRG8, ///< bayer, GBGB..(odd line), RGRG..(even line), 8-bit samples */ - AV_PIX_FMT_BAYER_GRBG8, ///< bayer, GRGR..(odd line), BGBG..(even line), 8-bit samples */ - AV_PIX_FMT_BAYER_BGGR16LE, ///< bayer, BGBG..(odd line), GRGR..(even line), 16-bit samples, little-endian */ - AV_PIX_FMT_BAYER_BGGR16BE, ///< bayer, BGBG..(odd line), GRGR..(even line), 16-bit samples, big-endian */ - AV_PIX_FMT_BAYER_RGGB16LE, ///< bayer, RGRG..(odd line), GBGB..(even line), 16-bit samples, little-endian */ - AV_PIX_FMT_BAYER_RGGB16BE, ///< bayer, RGRG..(odd line), GBGB..(even line), 16-bit samples, big-endian */ - AV_PIX_FMT_BAYER_GBRG16LE, ///< bayer, GBGB..(odd line), RGRG..(even line), 16-bit samples, little-endian */ - AV_PIX_FMT_BAYER_GBRG16BE, ///< bayer, GBGB..(odd line), RGRG..(even line), 16-bit samples, big-endian */ - AV_PIX_FMT_BAYER_GRBG16LE, ///< bayer, GRGR..(odd line), BGBG..(even line), 16-bit samples, little-endian */ - AV_PIX_FMT_BAYER_GRBG16BE, ///< bayer, GRGR..(odd line), BGBG..(even line), 16-bit samples, big-endian */ -#if !FF_API_XVMC - AV_PIX_FMT_XVMC,///< XVideo Motion Acceleration via common packet passing -#endif /* !FF_API_XVMC */ - AV_PIX_FMT_YUV440P10LE, ///< planar YUV 4:4:0,20bpp, (1 Cr & Cb sample per 1x2 Y samples), little-endian - AV_PIX_FMT_YUV440P10BE, ///< planar YUV 4:4:0,20bpp, (1 Cr & Cb sample per 1x2 Y samples), big-endian - AV_PIX_FMT_YUV440P12LE, ///< planar YUV 4:4:0,24bpp, (1 Cr & Cb sample per 1x2 Y samples), little-endian - AV_PIX_FMT_YUV440P12BE, ///< planar YUV 4:4:0,24bpp, (1 Cr & Cb sample per 1x2 Y samples), big-endian - AV_PIX_FMT_AYUV64LE, ///< packed AYUV 4:4:4,64bpp (1 Cr & Cb sample per 1x1 Y & A samples), little-endian - AV_PIX_FMT_AYUV64BE, ///< packed AYUV 4:4:4,64bpp (1 Cr & Cb sample per 1x1 Y & A samples), big-endian - - AV_PIX_FMT_VIDEOTOOLBOX, ///< hardware decoding through Videotoolbox - - AV_PIX_FMT_NB, ///< number of pixel formats, DO NOT USE THIS if you want to link with shared libav* because the number of formats might differ between versions -}; - -#define AV_PIX_FMT_Y400A AV_PIX_FMT_GRAY8A -#define AV_PIX_FMT_GBR24P AV_PIX_FMT_GBRP - -#if AV_HAVE_BIGENDIAN -# define AV_PIX_FMT_NE(be, le) AV_PIX_FMT_##be -#else -# define AV_PIX_FMT_NE(be, le) AV_PIX_FMT_##le -#endif - -#define AV_PIX_FMT_RGB32 AV_PIX_FMT_NE(ARGB, BGRA) -#define AV_PIX_FMT_RGB32_1 AV_PIX_FMT_NE(RGBA, ABGR) -#define AV_PIX_FMT_BGR32 AV_PIX_FMT_NE(ABGR, RGBA) -#define AV_PIX_FMT_BGR32_1 AV_PIX_FMT_NE(BGRA, ARGB) -#define AV_PIX_FMT_0RGB32 AV_PIX_FMT_NE(0RGB, BGR0) -#define AV_PIX_FMT_0BGR32 AV_PIX_FMT_NE(0BGR, RGB0) - -#define AV_PIX_FMT_GRAY16 AV_PIX_FMT_NE(GRAY16BE, GRAY16LE) -#define AV_PIX_FMT_YA16 AV_PIX_FMT_NE(YA16BE, YA16LE) -#define AV_PIX_FMT_RGB48 AV_PIX_FMT_NE(RGB48BE, RGB48LE) -#define AV_PIX_FMT_RGB565 AV_PIX_FMT_NE(RGB565BE, RGB565LE) -#define AV_PIX_FMT_RGB555 AV_PIX_FMT_NE(RGB555BE, RGB555LE) -#define AV_PIX_FMT_RGB444 AV_PIX_FMT_NE(RGB444BE, RGB444LE) -#define AV_PIX_FMT_RGBA64 AV_PIX_FMT_NE(RGBA64BE, RGBA64LE) -#define AV_PIX_FMT_BGR48 AV_PIX_FMT_NE(BGR48BE, BGR48LE) -#define AV_PIX_FMT_BGR565 AV_PIX_FMT_NE(BGR565BE, BGR565LE) -#define AV_PIX_FMT_BGR555 AV_PIX_FMT_NE(BGR555BE, BGR555LE) -#define AV_PIX_FMT_BGR444 AV_PIX_FMT_NE(BGR444BE, BGR444LE) -#define AV_PIX_FMT_BGRA64 AV_PIX_FMT_NE(BGRA64BE, BGRA64LE) - -#define AV_PIX_FMT_YUV420P9 AV_PIX_FMT_NE(YUV420P9BE , YUV420P9LE) -#define AV_PIX_FMT_YUV422P9 AV_PIX_FMT_NE(YUV422P9BE , YUV422P9LE) -#define AV_PIX_FMT_YUV444P9 AV_PIX_FMT_NE(YUV444P9BE , YUV444P9LE) -#define AV_PIX_FMT_YUV420P10 AV_PIX_FMT_NE(YUV420P10BE, YUV420P10LE) -#define AV_PIX_FMT_YUV422P10 AV_PIX_FMT_NE(YUV422P10BE, YUV422P10LE) -#define AV_PIX_FMT_YUV440P10 AV_PIX_FMT_NE(YUV440P10BE, YUV440P10LE) -#define AV_PIX_FMT_YUV444P10 AV_PIX_FMT_NE(YUV444P10BE, YUV444P10LE) -#define AV_PIX_FMT_YUV420P12 AV_PIX_FMT_NE(YUV420P12BE, YUV420P12LE) -#define AV_PIX_FMT_YUV422P12 AV_PIX_FMT_NE(YUV422P12BE, YUV422P12LE) -#define AV_PIX_FMT_YUV440P12 AV_PIX_FMT_NE(YUV440P12BE, YUV440P12LE) -#define AV_PIX_FMT_YUV444P12 AV_PIX_FMT_NE(YUV444P12BE, YUV444P12LE) -#define AV_PIX_FMT_YUV420P14 AV_PIX_FMT_NE(YUV420P14BE, YUV420P14LE) -#define AV_PIX_FMT_YUV422P14 AV_PIX_FMT_NE(YUV422P14BE, YUV422P14LE) -#define AV_PIX_FMT_YUV444P14 AV_PIX_FMT_NE(YUV444P14BE, YUV444P14LE) -#define AV_PIX_FMT_YUV420P16 AV_PIX_FMT_NE(YUV420P16BE, YUV420P16LE) -#define AV_PIX_FMT_YUV422P16 AV_PIX_FMT_NE(YUV422P16BE, YUV422P16LE) -#define AV_PIX_FMT_YUV444P16 AV_PIX_FMT_NE(YUV444P16BE, YUV444P16LE) - -#define AV_PIX_FMT_GBRP9 AV_PIX_FMT_NE(GBRP9BE , GBRP9LE) -#define AV_PIX_FMT_GBRP10 AV_PIX_FMT_NE(GBRP10BE, GBRP10LE) -#define AV_PIX_FMT_GBRP12 AV_PIX_FMT_NE(GBRP12BE, GBRP12LE) -#define AV_PIX_FMT_GBRP14 AV_PIX_FMT_NE(GBRP14BE, GBRP14LE) -#define AV_PIX_FMT_GBRP16 AV_PIX_FMT_NE(GBRP16BE, GBRP16LE) -#define AV_PIX_FMT_GBRAP16 AV_PIX_FMT_NE(GBRAP16BE, GBRAP16LE) - -#define AV_PIX_FMT_BAYER_BGGR16 AV_PIX_FMT_NE(BAYER_BGGR16BE, BAYER_BGGR16LE) -#define AV_PIX_FMT_BAYER_RGGB16 AV_PIX_FMT_NE(BAYER_RGGB16BE, BAYER_RGGB16LE) -#define AV_PIX_FMT_BAYER_GBRG16 AV_PIX_FMT_NE(BAYER_GBRG16BE, BAYER_GBRG16LE) -#define AV_PIX_FMT_BAYER_GRBG16 AV_PIX_FMT_NE(BAYER_GRBG16BE, BAYER_GRBG16LE) - - -#define AV_PIX_FMT_YUVA420P9 AV_PIX_FMT_NE(YUVA420P9BE , YUVA420P9LE) -#define AV_PIX_FMT_YUVA422P9 AV_PIX_FMT_NE(YUVA422P9BE , YUVA422P9LE) -#define AV_PIX_FMT_YUVA444P9 AV_PIX_FMT_NE(YUVA444P9BE , YUVA444P9LE) -#define AV_PIX_FMT_YUVA420P10 AV_PIX_FMT_NE(YUVA420P10BE, YUVA420P10LE) -#define AV_PIX_FMT_YUVA422P10 AV_PIX_FMT_NE(YUVA422P10BE, YUVA422P10LE) -#define AV_PIX_FMT_YUVA444P10 AV_PIX_FMT_NE(YUVA444P10BE, YUVA444P10LE) -#define AV_PIX_FMT_YUVA420P16 AV_PIX_FMT_NE(YUVA420P16BE, YUVA420P16LE) -#define AV_PIX_FMT_YUVA422P16 AV_PIX_FMT_NE(YUVA422P16BE, YUVA422P16LE) -#define AV_PIX_FMT_YUVA444P16 AV_PIX_FMT_NE(YUVA444P16BE, YUVA444P16LE) - -#define AV_PIX_FMT_XYZ12 AV_PIX_FMT_NE(XYZ12BE, XYZ12LE) -#define AV_PIX_FMT_NV20 AV_PIX_FMT_NE(NV20BE, NV20LE) -#define AV_PIX_FMT_AYUV64 AV_PIX_FMT_NE(AYUV64BE, AYUV64LE) - -/** - * Chromaticity coordinates of the source primaries. - */ -enum AVColorPrimaries { - AVCOL_PRI_RESERVED0 = 0, - AVCOL_PRI_BT709 = 1, ///< also ITU-R BT1361 / IEC 61966-2-4 / SMPTE RP177 Annex B - AVCOL_PRI_UNSPECIFIED = 2, - AVCOL_PRI_RESERVED = 3, - AVCOL_PRI_BT470M = 4, ///< also FCC Title 47 Code of Federal Regulations 73.682 (a)(20) - - AVCOL_PRI_BT470BG = 5, ///< also ITU-R BT601-6 625 / ITU-R BT1358 625 / ITU-R BT1700 625 PAL & SECAM - AVCOL_PRI_SMPTE170M = 6, ///< also ITU-R BT601-6 525 / ITU-R BT1358 525 / ITU-R BT1700 NTSC - AVCOL_PRI_SMPTE240M = 7, ///< functionally identical to above - AVCOL_PRI_FILM = 8, ///< colour filters using Illuminant C - AVCOL_PRI_BT2020 = 9, ///< ITU-R BT2020 - AVCOL_PRI_SMPTEST428_1= 10, ///< SMPTE ST 428-1 (CIE 1931 XYZ) - AVCOL_PRI_NB, ///< Not part of ABI -}; - -/** - * Color Transfer Characteristic. - */ -enum AVColorTransferCharacteristic { - AVCOL_TRC_RESERVED0 = 0, - AVCOL_TRC_BT709 = 1, ///< also ITU-R BT1361 - AVCOL_TRC_UNSPECIFIED = 2, - AVCOL_TRC_RESERVED = 3, - AVCOL_TRC_GAMMA22 = 4, ///< also ITU-R BT470M / ITU-R BT1700 625 PAL & SECAM - AVCOL_TRC_GAMMA28 = 5, ///< also ITU-R BT470BG - AVCOL_TRC_SMPTE170M = 6, ///< also ITU-R BT601-6 525 or 625 / ITU-R BT1358 525 or 625 / ITU-R BT1700 NTSC - AVCOL_TRC_SMPTE240M = 7, - AVCOL_TRC_LINEAR = 8, ///< "Linear transfer characteristics" - AVCOL_TRC_LOG = 9, ///< "Logarithmic transfer characteristic (100:1 range)" - AVCOL_TRC_LOG_SQRT = 10, ///< "Logarithmic transfer characteristic (100 * Sqrt(10) : 1 range)" - AVCOL_TRC_IEC61966_2_4 = 11, ///< IEC 61966-2-4 - AVCOL_TRC_BT1361_ECG = 12, ///< ITU-R BT1361 Extended Colour Gamut - AVCOL_TRC_IEC61966_2_1 = 13, ///< IEC 61966-2-1 (sRGB or sYCC) - AVCOL_TRC_BT2020_10 = 14, ///< ITU-R BT2020 for 10 bit system - AVCOL_TRC_BT2020_12 = 15, ///< ITU-R BT2020 for 12 bit system - AVCOL_TRC_SMPTEST2084 = 16, ///< SMPTE ST 2084 for 10, 12, 14 and 16 bit systems - AVCOL_TRC_SMPTEST428_1 = 17, ///< SMPTE ST 428-1 - AVCOL_TRC_NB, ///< Not part of ABI -}; - -/** - * YUV colorspace type. - */ -enum AVColorSpace { - AVCOL_SPC_RGB = 0, ///< order of coefficients is actually GBR, also IEC 61966-2-1 (sRGB) - AVCOL_SPC_BT709 = 1, ///< also ITU-R BT1361 / IEC 61966-2-4 xvYCC709 / SMPTE RP177 Annex B - AVCOL_SPC_UNSPECIFIED = 2, - AVCOL_SPC_RESERVED = 3, - AVCOL_SPC_FCC = 4, ///< FCC Title 47 Code of Federal Regulations 73.682 (a)(20) - AVCOL_SPC_BT470BG = 5, ///< also ITU-R BT601-6 625 / ITU-R BT1358 625 / ITU-R BT1700 625 PAL & SECAM / IEC 61966-2-4 xvYCC601 - AVCOL_SPC_SMPTE170M = 6, ///< also ITU-R BT601-6 525 / ITU-R BT1358 525 / ITU-R BT1700 NTSC / functionally identical to above - AVCOL_SPC_SMPTE240M = 7, - AVCOL_SPC_YCOCG = 8, ///< Used by Dirac / VC-2 and H.264 FRext, see ITU-T SG16 - AVCOL_SPC_BT2020_NCL = 9, ///< ITU-R BT2020 non-constant luminance system - AVCOL_SPC_BT2020_CL = 10, ///< ITU-R BT2020 constant luminance system - AVCOL_SPC_NB, ///< Not part of ABI -}; -#define AVCOL_SPC_YCGCO AVCOL_SPC_YCOCG - - -/** - * MPEG vs JPEG YUV range. - */ -enum AVColorRange { - AVCOL_RANGE_UNSPECIFIED = 0, - AVCOL_RANGE_MPEG = 1, ///< the normal 219*2^(n-8) "MPEG" YUV ranges - AVCOL_RANGE_JPEG = 2, ///< the normal 2^n-1 "JPEG" YUV ranges - AVCOL_RANGE_NB, ///< Not part of ABI -}; - -/** - * Location of chroma samples. - * - * Illustration showing the location of the first (top left) chroma sample of the - * image, the left shows only luma, the right - * shows the location of the chroma sample, the 2 could be imagined to overlay - * each other but are drawn separately due to limitations of ASCII - * - * 1st 2nd 1st 2nd horizontal luma sample positions - * v v v v - * ______ ______ - *1st luma line > |X X ... |3 4 X ... X are luma samples, - * | |1 2 1-6 are possible chroma positions - *2nd luma line > |X X ... |5 6 X ... 0 is undefined/unknown position - */ -enum AVChromaLocation { - AVCHROMA_LOC_UNSPECIFIED = 0, - AVCHROMA_LOC_LEFT = 1, ///< mpeg2/4 4:2:0, h264 default for 4:2:0 - AVCHROMA_LOC_CENTER = 2, ///< mpeg1 4:2:0, jpeg 4:2:0, h263 4:2:0 - AVCHROMA_LOC_TOPLEFT = 3, ///< ITU-R 601, SMPTE 274M 296M S314M(DV 4:1:1), mpeg2 4:2:2 - AVCHROMA_LOC_TOP = 4, - AVCHROMA_LOC_BOTTOMLEFT = 5, - AVCHROMA_LOC_BOTTOM = 6, - AVCHROMA_LOC_NB, ///< Not part of ABI -}; - -#endif /* AVUTIL_PIXFMT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/random_seed.h b/src/main/jni/ffmpeg/include/libavutil/random_seed.h deleted file mode 100644 index 0462a048e..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/random_seed.h +++ /dev/null @@ -1,43 +0,0 @@ -/* - * Copyright (c) 2009 Baptiste Coudurier <baptiste.coudurier@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_RANDOM_SEED_H -#define AVUTIL_RANDOM_SEED_H - -#include <stdint.h> -/** - * @addtogroup lavu_crypto - * @{ - */ - -/** - * Get a seed to use in conjunction with random functions. - * This function tries to provide a good seed at a best effort bases. - * Its possible to call this function multiple times if more bits are needed. - * It can be quite slow, which is why it should only be used as seed for a faster - * PRNG. The quality of the seed depends on the platform. - */ -uint32_t av_get_random_seed(void); - -/** - * @} - */ - -#endif /* AVUTIL_RANDOM_SEED_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/rational.h b/src/main/jni/ffmpeg/include/libavutil/rational.h deleted file mode 100644 index 289746968..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/rational.h +++ /dev/null @@ -1,173 +0,0 @@ -/* - * rational numbers - * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * rational numbers - * @author Michael Niedermayer <michaelni@gmx.at> - */ - -#ifndef AVUTIL_RATIONAL_H -#define AVUTIL_RATIONAL_H - -#include <stdint.h> -#include <limits.h> -#include "attributes.h" - -/** - * @addtogroup lavu_math - * @{ - */ - -/** - * rational number numerator/denominator - */ -typedef struct AVRational{ - int num; ///< numerator - int den; ///< denominator -} AVRational; - -/** - * Create a rational. - * Useful for compilers that do not support compound literals. - * @note The return value is not reduced. - */ -static inline AVRational av_make_q(int num, int den) -{ - AVRational r = { num, den }; - return r; -} - -/** - * Compare two rationals. - * @param a first rational - * @param b second rational - * @return 0 if a==b, 1 if a>b, -1 if a<b, and INT_MIN if one of the - * values is of the form 0/0 - */ -static inline int av_cmp_q(AVRational a, AVRational b){ - const int64_t tmp= a.num * (int64_t)b.den - b.num * (int64_t)a.den; - - if(tmp) return (int)((tmp ^ a.den ^ b.den)>>63)|1; - else if(b.den && a.den) return 0; - else if(a.num && b.num) return (a.num>>31) - (b.num>>31); - else return INT_MIN; -} - -/** - * Convert rational to double. - * @param a rational to convert - * @return (double) a - */ -static inline double av_q2d(AVRational a){ - return a.num / (double) a.den; -} - -/** - * Reduce a fraction. - * This is useful for framerate calculations. - * @param dst_num destination numerator - * @param dst_den destination denominator - * @param num source numerator - * @param den source denominator - * @param max the maximum allowed for dst_num & dst_den - * @return 1 if exact, 0 otherwise - */ -int av_reduce(int *dst_num, int *dst_den, int64_t num, int64_t den, int64_t max); - -/** - * Multiply two rationals. - * @param b first rational - * @param c second rational - * @return b*c - */ -AVRational av_mul_q(AVRational b, AVRational c) av_const; - -/** - * Divide one rational by another. - * @param b first rational - * @param c second rational - * @return b/c - */ -AVRational av_div_q(AVRational b, AVRational c) av_const; - -/** - * Add two rationals. - * @param b first rational - * @param c second rational - * @return b+c - */ -AVRational av_add_q(AVRational b, AVRational c) av_const; - -/** - * Subtract one rational from another. - * @param b first rational - * @param c second rational - * @return b-c - */ -AVRational av_sub_q(AVRational b, AVRational c) av_const; - -/** - * Invert a rational. - * @param q value - * @return 1 / q - */ -static av_always_inline AVRational av_inv_q(AVRational q) -{ - AVRational r = { q.den, q.num }; - return r; -} - -/** - * Convert a double precision floating point number to a rational. - * inf is expressed as {1,0} or {-1,0} depending on the sign. - * - * @param d double to convert - * @param max the maximum allowed numerator and denominator - * @return (AVRational) d - */ -AVRational av_d2q(double d, int max) av_const; - -/** - * @return 1 if q1 is nearer to q than q2, -1 if q2 is nearer - * than q1, 0 if they have the same distance. - */ -int av_nearer_q(AVRational q, AVRational q1, AVRational q2); - -/** - * Find the nearest value in q_list to q. - * @param q_list an array of rationals terminated by {0, 0} - * @return the index of the nearest value found in the array - */ -int av_find_nearest_q_idx(AVRational q, const AVRational* q_list); - -/** - * Converts a AVRational to a IEEE 32bit float. - * - * The float is returned in a uint32_t and its value is platform indepenant. - */ -uint32_t av_q2intfloat(AVRational q); - -/** - * @} - */ - -#endif /* AVUTIL_RATIONAL_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/rc4.h b/src/main/jni/ffmpeg/include/libavutil/rc4.h deleted file mode 100644 index 029cd2ad5..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/rc4.h +++ /dev/null @@ -1,66 +0,0 @@ -/* - * RC4 encryption/decryption/pseudo-random number generator - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_RC4_H -#define AVUTIL_RC4_H - -#include <stdint.h> - -/** - * @defgroup lavu_rc4 RC4 - * @ingroup lavu_crypto - * @{ - */ - -typedef struct AVRC4 { - uint8_t state[256]; - int x, y; -} AVRC4; - -/** - * Allocate an AVRC4 context. - */ -AVRC4 *av_rc4_alloc(void); - -/** - * @brief Initializes an AVRC4 context. - * - * @param key_bits must be a multiple of 8 - * @param decrypt 0 for encryption, 1 for decryption, currently has no effect - * @return zero on success, negative value otherwise - */ -int av_rc4_init(struct AVRC4 *d, const uint8_t *key, int key_bits, int decrypt); - -/** - * @brief Encrypts / decrypts using the RC4 algorithm. - * - * @param count number of bytes - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst, may be NULL - * @param iv not (yet) used for RC4, should be NULL - * @param decrypt 0 for encryption, 1 for decryption, not (yet) used - */ -void av_rc4_crypt(struct AVRC4 *d, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt); - -/** - * @} - */ - -#endif /* AVUTIL_RC4_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/replaygain.h b/src/main/jni/ffmpeg/include/libavutil/replaygain.h deleted file mode 100644 index 5c03e1993..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/replaygain.h +++ /dev/null @@ -1,51 +0,0 @@ -/* - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_REPLAYGAIN_H -#define AVUTIL_REPLAYGAIN_H - -#include <stdint.h> - -/** - * ReplayGain information (see - * http://wiki.hydrogenaudio.org/index.php?title=ReplayGain_1.0_specification). - * The size of this struct is a part of the public ABI. - */ -typedef struct AVReplayGain { - /** - * Track replay gain in microbels (divide by 100000 to get the value in dB). - * Should be set to INT32_MIN when unknown. - */ - int32_t track_gain; - /** - * Peak track amplitude, with 100000 representing full scale (but values - * may overflow). 0 when unknown. - */ - uint32_t track_peak; - /** - * Same as track_gain, but for the whole album. - */ - int32_t album_gain; - /** - * Same as track_peak, but for the whole album, - */ - uint32_t album_peak; -} AVReplayGain; - -#endif /* AVUTIL_REPLAYGAIN_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/ripemd.h b/src/main/jni/ffmpeg/include/libavutil/ripemd.h deleted file mode 100644 index 7b0c8bc89..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/ripemd.h +++ /dev/null @@ -1,75 +0,0 @@ -/* - * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> - * Copyright (C) 2013 James Almer <jamrial@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_RIPEMD_H -#define AVUTIL_RIPEMD_H - -#include <stdint.h> - -#include "attributes.h" -#include "version.h" - -/** - * @defgroup lavu_ripemd RIPEMD - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_ripemd_size; - -struct AVRIPEMD; - -/** - * Allocate an AVRIPEMD context. - */ -struct AVRIPEMD *av_ripemd_alloc(void); - -/** - * Initialize RIPEMD hashing. - * - * @param context pointer to the function context (of size av_ripemd_size) - * @param bits number of bits in digest (128, 160, 256 or 320 bits) - * @return zero if initialization succeeded, -1 otherwise - */ -int av_ripemd_init(struct AVRIPEMD* context, int bits); - -/** - * Update hash value. - * - * @param context hash function context - * @param data input data to update hash with - * @param len input data length - */ -void av_ripemd_update(struct AVRIPEMD* context, const uint8_t* data, unsigned int len); - -/** - * Finish hashing and output digest value. - * - * @param context hash function context - * @param digest buffer where output digest value is stored - */ -void av_ripemd_final(struct AVRIPEMD* context, uint8_t *digest); - -/** - * @} - */ - -#endif /* AVUTIL_RIPEMD_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/samplefmt.h b/src/main/jni/ffmpeg/include/libavutil/samplefmt.h deleted file mode 100644 index 6a8a031c0..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/samplefmt.h +++ /dev/null @@ -1,271 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_SAMPLEFMT_H -#define AVUTIL_SAMPLEFMT_H - -#include <stdint.h> - -#include "avutil.h" -#include "attributes.h" - -/** - * @addtogroup lavu_audio - * @{ - * - * @defgroup lavu_sampfmts Audio sample formats - * - * Audio sample format enumeration and related convenience functions. - * @{ - * - */ - -/** - * Audio sample formats - * - * - The data described by the sample format is always in native-endian order. - * Sample values can be expressed by native C types, hence the lack of a signed - * 24-bit sample format even though it is a common raw audio data format. - * - * - The floating-point formats are based on full volume being in the range - * [-1.0, 1.0]. Any values outside this range are beyond full volume level. - * - * - The data layout as used in av_samples_fill_arrays() and elsewhere in FFmpeg - * (such as AVFrame in libavcodec) is as follows: - * - * @par - * For planar sample formats, each audio channel is in a separate data plane, - * and linesize is the buffer size, in bytes, for a single plane. All data - * planes must be the same size. For packed sample formats, only the first data - * plane is used, and samples for each channel are interleaved. In this case, - * linesize is the buffer size, in bytes, for the 1 plane. - * - */ -enum AVSampleFormat { - AV_SAMPLE_FMT_NONE = -1, - AV_SAMPLE_FMT_U8, ///< unsigned 8 bits - AV_SAMPLE_FMT_S16, ///< signed 16 bits - AV_SAMPLE_FMT_S32, ///< signed 32 bits - AV_SAMPLE_FMT_FLT, ///< float - AV_SAMPLE_FMT_DBL, ///< double - - AV_SAMPLE_FMT_U8P, ///< unsigned 8 bits, planar - AV_SAMPLE_FMT_S16P, ///< signed 16 bits, planar - AV_SAMPLE_FMT_S32P, ///< signed 32 bits, planar - AV_SAMPLE_FMT_FLTP, ///< float, planar - AV_SAMPLE_FMT_DBLP, ///< double, planar - - AV_SAMPLE_FMT_NB ///< Number of sample formats. DO NOT USE if linking dynamically -}; - -/** - * Return the name of sample_fmt, or NULL if sample_fmt is not - * recognized. - */ -const char *av_get_sample_fmt_name(enum AVSampleFormat sample_fmt); - -/** - * Return a sample format corresponding to name, or AV_SAMPLE_FMT_NONE - * on error. - */ -enum AVSampleFormat av_get_sample_fmt(const char *name); - -/** - * Return the planar<->packed alternative form of the given sample format, or - * AV_SAMPLE_FMT_NONE on error. If the passed sample_fmt is already in the - * requested planar/packed format, the format returned is the same as the - * input. - */ -enum AVSampleFormat av_get_alt_sample_fmt(enum AVSampleFormat sample_fmt, int planar); - -/** - * Get the packed alternative form of the given sample format. - * - * If the passed sample_fmt is already in packed format, the format returned is - * the same as the input. - * - * @return the packed alternative form of the given sample format or - AV_SAMPLE_FMT_NONE on error. - */ -enum AVSampleFormat av_get_packed_sample_fmt(enum AVSampleFormat sample_fmt); - -/** - * Get the planar alternative form of the given sample format. - * - * If the passed sample_fmt is already in planar format, the format returned is - * the same as the input. - * - * @return the planar alternative form of the given sample format or - AV_SAMPLE_FMT_NONE on error. - */ -enum AVSampleFormat av_get_planar_sample_fmt(enum AVSampleFormat sample_fmt); - -/** - * Generate a string corresponding to the sample format with - * sample_fmt, or a header if sample_fmt is negative. - * - * @param buf the buffer where to write the string - * @param buf_size the size of buf - * @param sample_fmt the number of the sample format to print the - * corresponding info string, or a negative value to print the - * corresponding header. - * @return the pointer to the filled buffer or NULL if sample_fmt is - * unknown or in case of other errors - */ -char *av_get_sample_fmt_string(char *buf, int buf_size, enum AVSampleFormat sample_fmt); - -/** - * Return number of bytes per sample. - * - * @param sample_fmt the sample format - * @return number of bytes per sample or zero if unknown for the given - * sample format - */ -int av_get_bytes_per_sample(enum AVSampleFormat sample_fmt); - -/** - * Check if the sample format is planar. - * - * @param sample_fmt the sample format to inspect - * @return 1 if the sample format is planar, 0 if it is interleaved - */ -int av_sample_fmt_is_planar(enum AVSampleFormat sample_fmt); - -/** - * Get the required buffer size for the given audio parameters. - * - * @param[out] linesize calculated linesize, may be NULL - * @param nb_channels the number of channels - * @param nb_samples the number of samples in a single channel - * @param sample_fmt the sample format - * @param align buffer size alignment (0 = default, 1 = no alignment) - * @return required buffer size, or negative error code on failure - */ -int av_samples_get_buffer_size(int *linesize, int nb_channels, int nb_samples, - enum AVSampleFormat sample_fmt, int align); - -/** - * @} - * - * @defgroup lavu_sampmanip Samples manipulation - * - * Functions that manipulate audio samples - * @{ - */ - -/** - * Fill plane data pointers and linesize for samples with sample - * format sample_fmt. - * - * The audio_data array is filled with the pointers to the samples data planes: - * for planar, set the start point of each channel's data within the buffer, - * for packed, set the start point of the entire buffer only. - * - * The value pointed to by linesize is set to the aligned size of each - * channel's data buffer for planar layout, or to the aligned size of the - * buffer for all channels for packed layout. - * - * The buffer in buf must be big enough to contain all the samples - * (use av_samples_get_buffer_size() to compute its minimum size), - * otherwise the audio_data pointers will point to invalid data. - * - * @see enum AVSampleFormat - * The documentation for AVSampleFormat describes the data layout. - * - * @param[out] audio_data array to be filled with the pointer for each channel - * @param[out] linesize calculated linesize, may be NULL - * @param buf the pointer to a buffer containing the samples - * @param nb_channels the number of channels - * @param nb_samples the number of samples in a single channel - * @param sample_fmt the sample format - * @param align buffer size alignment (0 = default, 1 = no alignment) - * @return >=0 on success or a negative error code on failure - * @todo return minimum size in bytes required for the buffer in case - * of success at the next bump - */ -int av_samples_fill_arrays(uint8_t **audio_data, int *linesize, - const uint8_t *buf, - int nb_channels, int nb_samples, - enum AVSampleFormat sample_fmt, int align); - -/** - * Allocate a samples buffer for nb_samples samples, and fill data pointers and - * linesize accordingly. - * The allocated samples buffer can be freed by using av_freep(&audio_data[0]) - * Allocated data will be initialized to silence. - * - * @see enum AVSampleFormat - * The documentation for AVSampleFormat describes the data layout. - * - * @param[out] audio_data array to be filled with the pointer for each channel - * @param[out] linesize aligned size for audio buffer(s), may be NULL - * @param nb_channels number of audio channels - * @param nb_samples number of samples per channel - * @param align buffer size alignment (0 = default, 1 = no alignment) - * @return >=0 on success or a negative error code on failure - * @todo return the size of the allocated buffer in case of success at the next bump - * @see av_samples_fill_arrays() - * @see av_samples_alloc_array_and_samples() - */ -int av_samples_alloc(uint8_t **audio_data, int *linesize, int nb_channels, - int nb_samples, enum AVSampleFormat sample_fmt, int align); - -/** - * Allocate a data pointers array, samples buffer for nb_samples - * samples, and fill data pointers and linesize accordingly. - * - * This is the same as av_samples_alloc(), but also allocates the data - * pointers array. - * - * @see av_samples_alloc() - */ -int av_samples_alloc_array_and_samples(uint8_t ***audio_data, int *linesize, int nb_channels, - int nb_samples, enum AVSampleFormat sample_fmt, int align); - -/** - * Copy samples from src to dst. - * - * @param dst destination array of pointers to data planes - * @param src source array of pointers to data planes - * @param dst_offset offset in samples at which the data will be written to dst - * @param src_offset offset in samples at which the data will be read from src - * @param nb_samples number of samples to be copied - * @param nb_channels number of audio channels - * @param sample_fmt audio sample format - */ -int av_samples_copy(uint8_t **dst, uint8_t * const *src, int dst_offset, - int src_offset, int nb_samples, int nb_channels, - enum AVSampleFormat sample_fmt); - -/** - * Fill an audio buffer with silence. - * - * @param audio_data array of pointers to data planes - * @param offset offset in samples at which to start filling - * @param nb_samples number of samples to fill - * @param nb_channels number of audio channels - * @param sample_fmt audio sample format - */ -int av_samples_set_silence(uint8_t **audio_data, int offset, int nb_samples, - int nb_channels, enum AVSampleFormat sample_fmt); - -/** - * @} - * @} - */ -#endif /* AVUTIL_SAMPLEFMT_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/sha.h b/src/main/jni/ffmpeg/include/libavutil/sha.h deleted file mode 100644 index bf4377e51..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/sha.h +++ /dev/null @@ -1,74 +0,0 @@ -/* - * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_SHA_H -#define AVUTIL_SHA_H - -#include <stdint.h> - -#include "attributes.h" -#include "version.h" - -/** - * @defgroup lavu_sha SHA - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_sha_size; - -struct AVSHA; - -/** - * Allocate an AVSHA context. - */ -struct AVSHA *av_sha_alloc(void); - -/** - * Initialize SHA-1 or SHA-2 hashing. - * - * @param context pointer to the function context (of size av_sha_size) - * @param bits number of bits in digest (SHA-1 - 160 bits, SHA-2 224 or 256 bits) - * @return zero if initialization succeeded, -1 otherwise - */ -int av_sha_init(struct AVSHA* context, int bits); - -/** - * Update hash value. - * - * @param context hash function context - * @param data input data to update hash with - * @param len input data length - */ -void av_sha_update(struct AVSHA* context, const uint8_t* data, unsigned int len); - -/** - * Finish hashing and output digest value. - * - * @param context hash function context - * @param digest buffer where output digest value is stored - */ -void av_sha_final(struct AVSHA* context, uint8_t *digest); - -/** - * @} - */ - -#endif /* AVUTIL_SHA_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/sha512.h b/src/main/jni/ffmpeg/include/libavutil/sha512.h deleted file mode 100644 index 7b0870147..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/sha512.h +++ /dev/null @@ -1,75 +0,0 @@ -/* - * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> - * Copyright (C) 2013 James Almer <jamrial@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_SHA512_H -#define AVUTIL_SHA512_H - -#include <stdint.h> - -#include "attributes.h" -#include "version.h" - -/** - * @defgroup lavu_sha512 SHA512 - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_sha512_size; - -struct AVSHA512; - -/** - * Allocate an AVSHA512 context. - */ -struct AVSHA512 *av_sha512_alloc(void); - -/** - * Initialize SHA-2 512 hashing. - * - * @param context pointer to the function context (of size av_sha512_size) - * @param bits number of bits in digest (224, 256, 384 or 512 bits) - * @return zero if initialization succeeded, -1 otherwise - */ -int av_sha512_init(struct AVSHA512* context, int bits); - -/** - * Update hash value. - * - * @param context hash function context - * @param data input data to update hash with - * @param len input data length - */ -void av_sha512_update(struct AVSHA512* context, const uint8_t* data, unsigned int len); - -/** - * Finish hashing and output digest value. - * - * @param context hash function context - * @param digest buffer where output digest value is stored - */ -void av_sha512_final(struct AVSHA512* context, uint8_t *digest); - -/** - * @} - */ - -#endif /* AVUTIL_SHA512_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/stereo3d.h b/src/main/jni/ffmpeg/include/libavutil/stereo3d.h deleted file mode 100644 index 1135dc9dd..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/stereo3d.h +++ /dev/null @@ -1,152 +0,0 @@ -/* - * Copyright (c) 2013 Vittorio Giovara <vittorio.giovara@gmail.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_STEREO3D_H -#define AVUTIL_STEREO3D_H - -#include <stdint.h> - -#include "frame.h" - -/** - * List of possible 3D Types - */ -enum AVStereo3DType { - /** - * Video is not stereoscopic (and metadata has to be there). - */ - AV_STEREO3D_2D, - - /** - * Views are next to each other. - * - * LLLLRRRR - * LLLLRRRR - * LLLLRRRR - * ... - */ - AV_STEREO3D_SIDEBYSIDE, - - /** - * Views are on top of each other. - * - * LLLLLLLL - * LLLLLLLL - * RRRRRRRR - * RRRRRRRR - */ - AV_STEREO3D_TOPBOTTOM, - - /** - * Views are alternated temporally. - * - * frame0 frame1 frame2 ... - * LLLLLLLL RRRRRRRR LLLLLLLL - * LLLLLLLL RRRRRRRR LLLLLLLL - * LLLLLLLL RRRRRRRR LLLLLLLL - * ... ... ... - */ - AV_STEREO3D_FRAMESEQUENCE, - - /** - * Views are packed in a checkerboard-like structure per pixel. - * - * LRLRLRLR - * RLRLRLRL - * LRLRLRLR - * ... - */ - AV_STEREO3D_CHECKERBOARD, - - /** - * Views are next to each other, but when upscaling - * apply a checkerboard pattern. - * - * LLLLRRRR L L L L R R R R - * LLLLRRRR => L L L L R R R R - * LLLLRRRR L L L L R R R R - * LLLLRRRR L L L L R R R R - */ - AV_STEREO3D_SIDEBYSIDE_QUINCUNX, - - /** - * Views are packed per line, as if interlaced. - * - * LLLLLLLL - * RRRRRRRR - * LLLLLLLL - * ... - */ - AV_STEREO3D_LINES, - - /** - * Views are packed per column. - * - * LRLRLRLR - * LRLRLRLR - * LRLRLRLR - * ... - */ - AV_STEREO3D_COLUMNS, -}; - - -/** - * Inverted views, Right/Bottom represents the left view. - */ -#define AV_STEREO3D_FLAG_INVERT (1 << 0) - -/** - * Stereo 3D type: this structure describes how two videos are packed - * within a single video surface, with additional information as needed. - * - * @note The struct must be allocated with av_stereo3d_alloc() and - * its size is not a part of the public ABI. - */ -typedef struct AVStereo3D { - /** - * How views are packed within the video. - */ - enum AVStereo3DType type; - - /** - * Additional information about the frame packing. - */ - int flags; -} AVStereo3D; - -/** - * Allocate an AVStereo3D structure and set its fields to default values. - * The resulting struct can be freed using av_freep(). - * - * @return An AVStereo3D filled with default values or NULL on failure. - */ -AVStereo3D *av_stereo3d_alloc(void); - -/** - * Allocate a complete AVFrameSideData and add it to the frame. - * - * @param frame The frame which side data is added to. - * - * @return The AVStereo3D structure to be filled by caller. - */ -AVStereo3D *av_stereo3d_create_side_data(AVFrame *frame); - -#endif /* AVUTIL_STEREO3D_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/tea.h b/src/main/jni/ffmpeg/include/libavutil/tea.h deleted file mode 100644 index dd929bdaf..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/tea.h +++ /dev/null @@ -1,71 +0,0 @@ -/* - * A 32-bit implementation of the TEA algorithm - * Copyright (c) 2015 Vesselin Bontchev - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_TEA_H -#define AVUTIL_TEA_H - -#include <stdint.h> - -/** - * @file - * @brief Public header for libavutil TEA algorithm - * @defgroup lavu_tea TEA - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_tea_size; - -struct AVTEA; - -/** - * Allocate an AVTEA context - * To free the struct: av_free(ptr) - */ -struct AVTEA *av_tea_alloc(void); - -/** - * Initialize an AVTEA context. - * - * @param ctx an AVTEA context - * @param key a key of 16 bytes used for encryption/decryption - * @param rounds the number of rounds in TEA (64 is the "standard") - */ -void av_tea_init(struct AVTEA *ctx, const uint8_t key[16], int rounds); - -/** - * Encrypt or decrypt a buffer using a previously initialized context. - * - * @param ctx an AVTEA context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 8 byte blocks - * @param iv initialization vector for CBC mode, if NULL then ECB will be used - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_tea_crypt(struct AVTEA *ctx, uint8_t *dst, const uint8_t *src, - int count, uint8_t *iv, int decrypt); - -/** - * @} - */ - -#endif /* AVUTIL_TEA_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/threadmessage.h b/src/main/jni/ffmpeg/include/libavutil/threadmessage.h deleted file mode 100644 index e256cae9d..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/threadmessage.h +++ /dev/null @@ -1,107 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public License - * as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public License - * along with FFmpeg; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_THREADMESSAGE_H -#define AVUTIL_THREADMESSAGE_H - -typedef struct AVThreadMessageQueue AVThreadMessageQueue; - -typedef enum AVThreadMessageFlags { - - /** - * Perform non-blocking operation. - * If this flag is set, send and recv operations are non-blocking and - * return AVERROR(EAGAIN) immediately if they can not proceed. - */ - AV_THREAD_MESSAGE_NONBLOCK = 1, - -} AVThreadMessageFlags; - -/** - * Allocate a new message queue. - * - * @param mq pointer to the message queue - * @param nelem maximum number of elements in the queue - * @param elsize size of each element in the queue - * @return >=0 for success; <0 for error, in particular AVERROR(ENOSYS) if - * lavu was built without thread support - */ -int av_thread_message_queue_alloc(AVThreadMessageQueue **mq, - unsigned nelem, - unsigned elsize); - -/** - * Free a message queue. - * - * The message queue must no longer be in use by another thread. - */ -void av_thread_message_queue_free(AVThreadMessageQueue **mq); - -/** - * Send a message on the queue. - */ -int av_thread_message_queue_send(AVThreadMessageQueue *mq, - void *msg, - unsigned flags); - -/** - * Receive a message from the queue. - */ -int av_thread_message_queue_recv(AVThreadMessageQueue *mq, - void *msg, - unsigned flags); - -/** - * Set the sending error code. - * - * If the error code is set to non-zero, av_thread_message_queue_recv() will - * return it immediately when there are no longer available messages. - * Conventional values, such as AVERROR_EOF or AVERROR(EAGAIN), can be used - * to cause the receiving thread to stop or suspend its operation. - */ -void av_thread_message_queue_set_err_send(AVThreadMessageQueue *mq, - int err); - -/** - * Set the receiving error code. - * - * If the error code is set to non-zero, av_thread_message_queue_send() will - * return it immediately. Conventional values, such as AVERROR_EOF or - * AVERROR(EAGAIN), can be used to cause the sending thread to stop or - * suspend its operation. - */ -void av_thread_message_queue_set_err_recv(AVThreadMessageQueue *mq, - int err); - -/** - * Set the optional free message callback function which will be called if an - * operation is removing messages from the queue. - */ -void av_thread_message_queue_set_free_func(AVThreadMessageQueue *mq, - void (*free_func)(void *msg)); - -/** - * Flush the message queue - * - * This function is mostly equivalent to reading and free-ing every message - * except that it will be done in a single operation (no lock/unlock between - * reads). - */ -void av_thread_message_flush(AVThreadMessageQueue *mq); - -#endif /* AVUTIL_THREADMESSAGE_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/time.h b/src/main/jni/ffmpeg/include/libavutil/time.h deleted file mode 100644 index dc169b064..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/time.h +++ /dev/null @@ -1,56 +0,0 @@ -/* - * Copyright (c) 2000-2003 Fabrice Bellard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_TIME_H -#define AVUTIL_TIME_H - -#include <stdint.h> - -/** - * Get the current time in microseconds. - */ -int64_t av_gettime(void); - -/** - * Get the current time in microseconds since some unspecified starting point. - * On platforms that support it, the time comes from a monotonic clock - * This property makes this time source ideal for measuring relative time. - * The returned values may not be monotonic on platforms where a monotonic - * clock is not available. - */ -int64_t av_gettime_relative(void); - -/** - * Indicates with a boolean result if the av_gettime_relative() time source - * is monotonic. - */ -int av_gettime_relative_is_monotonic(void); - -/** - * Sleep for a period of time. Although the duration is expressed in - * microseconds, the actual delay may be rounded to the precision of the - * system timer. - * - * @param usec Number of microseconds to sleep. - * @return zero on success or (negative) error code. - */ -int av_usleep(unsigned usec); - -#endif /* AVUTIL_TIME_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/timecode.h b/src/main/jni/ffmpeg/include/libavutil/timecode.h deleted file mode 100644 index 56e3975fd..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/timecode.h +++ /dev/null @@ -1,140 +0,0 @@ -/* - * Copyright (c) 2006 Smartjog S.A.S, Baptiste Coudurier <baptiste.coudurier@gmail.com> - * Copyright (c) 2011-2012 Smartjog S.A.S, Clément Bœsch <clement.boesch@smartjog.com> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * Timecode helpers header - */ - -#ifndef AVUTIL_TIMECODE_H -#define AVUTIL_TIMECODE_H - -#include <stdint.h> -#include "rational.h" - -#define AV_TIMECODE_STR_SIZE 16 - -enum AVTimecodeFlag { - AV_TIMECODE_FLAG_DROPFRAME = 1<<0, ///< timecode is drop frame - AV_TIMECODE_FLAG_24HOURSMAX = 1<<1, ///< timecode wraps after 24 hours - AV_TIMECODE_FLAG_ALLOWNEGATIVE = 1<<2, ///< negative time values are allowed -}; - -typedef struct { - int start; ///< timecode frame start (first base frame number) - uint32_t flags; ///< flags such as drop frame, +24 hours support, ... - AVRational rate; ///< frame rate in rational form - unsigned fps; ///< frame per second; must be consistent with the rate field -} AVTimecode; - -/** - * Adjust frame number for NTSC drop frame time code. - * - * @param framenum frame number to adjust - * @param fps frame per second, 30 or 60 - * @return adjusted frame number - * @warning adjustment is only valid in NTSC 29.97 and 59.94 - */ -int av_timecode_adjust_ntsc_framenum2(int framenum, int fps); - -/** - * Convert frame number to SMPTE 12M binary representation. - * - * @param tc timecode data correctly initialized - * @param framenum frame number - * @return the SMPTE binary representation - * - * @note Frame number adjustment is automatically done in case of drop timecode, - * you do NOT have to call av_timecode_adjust_ntsc_framenum2(). - * @note The frame number is relative to tc->start. - * @note Color frame (CF), binary group flags (BGF) and biphase mark polarity - * correction (PC) bits are set to zero. - */ -uint32_t av_timecode_get_smpte_from_framenum(const AVTimecode *tc, int framenum); - -/** - * Load timecode string in buf. - * - * @param buf destination buffer, must be at least AV_TIMECODE_STR_SIZE long - * @param tc timecode data correctly initialized - * @param framenum frame number - * @return the buf parameter - * - * @note Timecode representation can be a negative timecode and have more than - * 24 hours, but will only be honored if the flags are correctly set. - * @note The frame number is relative to tc->start. - */ -char *av_timecode_make_string(const AVTimecode *tc, char *buf, int framenum); - -/** - * Get the timecode string from the SMPTE timecode format. - * - * @param buf destination buffer, must be at least AV_TIMECODE_STR_SIZE long - * @param tcsmpte the 32-bit SMPTE timecode - * @param prevent_df prevent the use of a drop flag when it is known the DF bit - * is arbitrary - * @return the buf parameter - */ -char *av_timecode_make_smpte_tc_string(char *buf, uint32_t tcsmpte, int prevent_df); - -/** - * Get the timecode string from the 25-bit timecode format (MPEG GOP format). - * - * @param buf destination buffer, must be at least AV_TIMECODE_STR_SIZE long - * @param tc25bit the 25-bits timecode - * @return the buf parameter - */ -char *av_timecode_make_mpeg_tc_string(char *buf, uint32_t tc25bit); - -/** - * Init a timecode struct with the passed parameters. - * - * @param log_ctx a pointer to an arbitrary struct of which the first field - * is a pointer to an AVClass struct (used for av_log) - * @param tc pointer to an allocated AVTimecode - * @param rate frame rate in rational form - * @param flags miscellaneous flags such as drop frame, +24 hours, ... - * (see AVTimecodeFlag) - * @param frame_start the first frame number - * @return 0 on success, AVERROR otherwise - */ -int av_timecode_init(AVTimecode *tc, AVRational rate, int flags, int frame_start, void *log_ctx); - -/** - * Parse timecode representation (hh:mm:ss[:;.]ff). - * - * @param log_ctx a pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct (used for av_log). - * @param tc pointer to an allocated AVTimecode - * @param rate frame rate in rational form - * @param str timecode string which will determine the frame start - * @return 0 on success, AVERROR otherwise - */ -int av_timecode_init_from_string(AVTimecode *tc, AVRational rate, const char *str, void *log_ctx); - -/** - * Check if the timecode feature is available for the given frame rate - * - * @return 0 if supported, <0 otherwise - */ -int av_timecode_check_frame_rate(AVRational rate); - -#endif /* AVUTIL_TIMECODE_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/timestamp.h b/src/main/jni/ffmpeg/include/libavutil/timestamp.h deleted file mode 100644 index f010a7ee3..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/timestamp.h +++ /dev/null @@ -1,78 +0,0 @@ -/* - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * timestamp utils, mostly useful for debugging/logging purposes - */ - -#ifndef AVUTIL_TIMESTAMP_H -#define AVUTIL_TIMESTAMP_H - -#include "common.h" - -#if defined(__cplusplus) && !defined(__STDC_FORMAT_MACROS) && !defined(PRId64) -#error missing -D__STDC_FORMAT_MACROS / #define __STDC_FORMAT_MACROS -#endif - -#define AV_TS_MAX_STRING_SIZE 32 - -/** - * Fill the provided buffer with a string containing a timestamp - * representation. - * - * @param buf a buffer with size in bytes of at least AV_TS_MAX_STRING_SIZE - * @param ts the timestamp to represent - * @return the buffer in input - */ -static inline char *av_ts_make_string(char *buf, int64_t ts) -{ - if (ts == AV_NOPTS_VALUE) snprintf(buf, AV_TS_MAX_STRING_SIZE, "NOPTS"); - else snprintf(buf, AV_TS_MAX_STRING_SIZE, "%"PRId64, ts); - return buf; -} - -/** - * Convenience macro, the return value should be used only directly in - * function arguments but never stand-alone. - */ -#define av_ts2str(ts) av_ts_make_string((char[AV_TS_MAX_STRING_SIZE]){0}, ts) - -/** - * Fill the provided buffer with a string containing a timestamp time - * representation. - * - * @param buf a buffer with size in bytes of at least AV_TS_MAX_STRING_SIZE - * @param ts the timestamp to represent - * @param tb the timebase of the timestamp - * @return the buffer in input - */ -static inline char *av_ts_make_time_string(char *buf, int64_t ts, AVRational *tb) -{ - if (ts == AV_NOPTS_VALUE) snprintf(buf, AV_TS_MAX_STRING_SIZE, "NOPTS"); - else snprintf(buf, AV_TS_MAX_STRING_SIZE, "%.6g", av_q2d(*tb) * ts); - return buf; -} - -/** - * Convenience macro, the return value should be used only directly in - * function arguments but never stand-alone. - */ -#define av_ts2timestr(ts, tb) av_ts_make_time_string((char[AV_TS_MAX_STRING_SIZE]){0}, ts, tb) - -#endif /* AVUTIL_TIMESTAMP_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/tree.h b/src/main/jni/ffmpeg/include/libavutil/tree.h deleted file mode 100644 index e1aefaa9f..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/tree.h +++ /dev/null @@ -1,138 +0,0 @@ -/* - * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -/** - * @file - * A tree container. - * @author Michael Niedermayer <michaelni@gmx.at> - */ - -#ifndef AVUTIL_TREE_H -#define AVUTIL_TREE_H - -#include "attributes.h" -#include "version.h" - -/** - * @addtogroup lavu_tree AVTree - * @ingroup lavu_data - * - * Low-complexity tree container - * - * Insertion, removal, finding equal, largest which is smaller than and - * smallest which is larger than, all have O(log n) worst-case complexity. - * @{ - */ - - -struct AVTreeNode; -extern const int av_tree_node_size; - -/** - * Allocate an AVTreeNode. - */ -struct AVTreeNode *av_tree_node_alloc(void); - -/** - * Find an element. - * @param root a pointer to the root node of the tree - * @param next If next is not NULL, then next[0] will contain the previous - * element and next[1] the next element. If either does not exist, - * then the corresponding entry in next is unchanged. - * @param cmp compare function used to compare elements in the tree, - * API identical to that of Standard C's qsort - * It is guranteed that the first and only the first argument to cmp() - * will be the key parameter to av_tree_find(), thus it could if the - * user wants, be a different type (like an opaque context). - * @return An element with cmp(key, elem) == 0 or NULL if no such element - * exists in the tree. - */ -void *av_tree_find(const struct AVTreeNode *root, void *key, - int (*cmp)(const void *key, const void *b), void *next[2]); - -/** - * Insert or remove an element. - * - * If *next is NULL, then the supplied element will be removed if it exists. - * If *next is non-NULL, then the supplied element will be inserted, unless - * it already exists in the tree. - * - * @param rootp A pointer to a pointer to the root node of the tree; note that - * the root node can change during insertions, this is required - * to keep the tree balanced. - * @param key pointer to the element key to insert in the tree - * @param next Used to allocate and free AVTreeNodes. For insertion the user - * must set it to an allocated and zeroed object of at least - * av_tree_node_size bytes size. av_tree_insert() will set it to - * NULL if it has been consumed. - * For deleting elements *next is set to NULL by the user and - * av_tree_insert() will set it to the AVTreeNode which was - * used for the removed element. - * This allows the use of flat arrays, which have - * lower overhead compared to many malloced elements. - * You might want to define a function like: - * @code - * void *tree_insert(struct AVTreeNode **rootp, void *key, - * int (*cmp)(void *key, const void *b), - * AVTreeNode **next) - * { - * if (!*next) - * *next = av_mallocz(av_tree_node_size); - * return av_tree_insert(rootp, key, cmp, next); - * } - * void *tree_remove(struct AVTreeNode **rootp, void *key, - * int (*cmp)(void *key, const void *b, AVTreeNode **next)) - * { - * av_freep(next); - * return av_tree_insert(rootp, key, cmp, next); - * } - * @endcode - * @param cmp compare function used to compare elements in the tree, API identical - * to that of Standard C's qsort - * @return If no insertion happened, the found element; if an insertion or - * removal happened, then either key or NULL will be returned. - * Which one it is depends on the tree state and the implementation. You - * should make no assumptions that it's one or the other in the code. - */ -void *av_tree_insert(struct AVTreeNode **rootp, void *key, - int (*cmp)(const void *key, const void *b), - struct AVTreeNode **next); - -void av_tree_destroy(struct AVTreeNode *t); - -/** - * Apply enu(opaque, &elem) to all the elements in the tree in a given range. - * - * @param cmp a comparison function that returns < 0 for a element below the - * range, > 0 for a element above the range and == 0 for a - * element inside the range - * - * @note The cmp function should use the same ordering used to construct the - * tree. - */ -void av_tree_enumerate(struct AVTreeNode *t, void *opaque, - int (*cmp)(void *opaque, void *elem), - int (*enu)(void *opaque, void *elem)); - -/** - * @} - */ - -#endif /* AVUTIL_TREE_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/twofish.h b/src/main/jni/ffmpeg/include/libavutil/twofish.h deleted file mode 100644 index 813cfecdf..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/twofish.h +++ /dev/null @@ -1,70 +0,0 @@ -/* - * An implementation of the TwoFish algorithm - * Copyright (c) 2015 Supraja Meedinti - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_TWOFISH_H -#define AVUTIL_TWOFISH_H - -#include <stdint.h> - - -/** - * @file - * @brief Public header for libavutil TWOFISH algorithm - * @defgroup lavu_twofish TWOFISH - * @ingroup lavu_crypto - * @{ - */ - -extern const int av_twofish_size; - -struct AVTWOFISH; - -/** - * Allocate an AVTWOFISH context - * To free the struct: av_free(ptr) - */ -struct AVTWOFISH *av_twofish_alloc(void); - -/** - * Initialize an AVTWOFISH context. - * - * @param ctx an AVTWOFISH context - * @param key a key of size ranging from 1 to 32 bytes used for encryption/decryption - * @param key_bits number of keybits: 128, 192, 256 If less than the required, padded with zeroes to nearest valid value; return value is 0 if key_bits is 128/192/256, -1 if less than 0, 1 otherwise - */ -int av_twofish_init(struct AVTWOFISH *ctx, const uint8_t *key, int key_bits); - -/** - * Encrypt or decrypt a buffer using a previously initialized context - * - * @param ctx an AVTWOFISH context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 16 byte blocks - * @paran iv initialization vector for CBC mode, NULL for ECB mode - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_twofish_crypt(struct AVTWOFISH *ctx, uint8_t *dst, const uint8_t *src, int count, uint8_t* iv, int decrypt); - -/** - * @} - */ -#endif /* AVUTIL_TWOFISH_H */ diff --git a/src/main/jni/ffmpeg/include/libavutil/version.h b/src/main/jni/ffmpeg/include/libavutil/version.h deleted file mode 100644 index 22cd66b55..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/version.h +++ /dev/null @@ -1,129 +0,0 @@ -/* - * copyright (c) 2003 Fabrice Bellard - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_VERSION_H -#define AVUTIL_VERSION_H - -#include "macros.h" - -/** - * @addtogroup version_utils - * - * Useful to check and match library version in order to maintain - * backward compatibility. - * - * @{ - */ - -#define AV_VERSION_INT(a, b, c) ((a)<<16 | (b)<<8 | (c)) -#define AV_VERSION_DOT(a, b, c) a ##.## b ##.## c -#define AV_VERSION(a, b, c) AV_VERSION_DOT(a, b, c) - -/** - * Extract version components from the full ::AV_VERSION_INT int as returned - * by functions like ::avformat_version() and ::avcodec_version() - */ -#define AV_VERSION_MAJOR(a) ((a) >> 16) -#define AV_VERSION_MINOR(a) (((a) & 0x00FF00) >> 8) -#define AV_VERSION_MICRO(a) ((a) & 0xFF) - -/** - * @} - */ - -/** - * @file - * @ingroup lavu - * Libavutil version macros - */ - -/** - * @defgroup lavu_ver Version and Build diagnostics - * - * Macros and function useful to check at compiletime and at runtime - * which version of libavutil is in use. - * - * @{ - */ - -#define LIBAVUTIL_VERSION_MAJOR 55 -#define LIBAVUTIL_VERSION_MINOR 11 -#define LIBAVUTIL_VERSION_MICRO 100 - -#define LIBAVUTIL_VERSION_INT AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \ - LIBAVUTIL_VERSION_MINOR, \ - LIBAVUTIL_VERSION_MICRO) -#define LIBAVUTIL_VERSION AV_VERSION(LIBAVUTIL_VERSION_MAJOR, \ - LIBAVUTIL_VERSION_MINOR, \ - LIBAVUTIL_VERSION_MICRO) -#define LIBAVUTIL_BUILD LIBAVUTIL_VERSION_INT - -#define LIBAVUTIL_IDENT "Lavu" AV_STRINGIFY(LIBAVUTIL_VERSION) - -/** - * @} - * - * @defgroup depr_guards Deprecation guards - * FF_API_* defines may be placed below to indicate public API that will be - * dropped at a future version bump. The defines themselves are not part of - * the public API and may change, break or disappear at any time. - * - * @note, when bumping the major version it is recommended to manually - * disable each FF_API_* in its own commit instead of disabling them all - * at once through the bump. This improves the git bisect-ability of the change. - * - * @{ - */ - -#ifndef FF_API_VDPAU -#define FF_API_VDPAU (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_XVMC -#define FF_API_XVMC (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_OPT_TYPE_METADATA -#define FF_API_OPT_TYPE_METADATA (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_DLOG -#define FF_API_DLOG (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_VAAPI -#define FF_API_VAAPI (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_FRAME_QP -#define FF_API_FRAME_QP (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_PLUS1_MINUS1 -#define FF_API_PLUS1_MINUS1 (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_ERROR_FRAME -#define FF_API_ERROR_FRAME (LIBAVUTIL_VERSION_MAJOR < 56) -#endif -#ifndef FF_API_CRC_BIG_TABLE -#define FF_API_CRC_BIG_TABLE (LIBAVUTIL_VERSION_MAJOR < 56) -#endif - - -/** - * @} - */ - -#endif /* AVUTIL_VERSION_H */ - diff --git a/src/main/jni/ffmpeg/include/libavutil/xtea.h b/src/main/jni/ffmpeg/include/libavutil/xtea.h deleted file mode 100644 index 735427c10..000000000 --- a/src/main/jni/ffmpeg/include/libavutil/xtea.h +++ /dev/null @@ -1,94 +0,0 @@ -/* - * A 32-bit implementation of the XTEA algorithm - * Copyright (c) 2012 Samuel Pitoiset - * - * This file is part of FFmpeg. - * - * FFmpeg is free software; you can redistribute it and/or - * modify it under the terms of the GNU Lesser General Public - * License as published by the Free Software Foundation; either - * version 2.1 of the License, or (at your option) any later version. - * - * FFmpeg is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - * Lesser General Public License for more details. - * - * You should have received a copy of the GNU Lesser General Public - * License along with FFmpeg; if not, write to the Free Software - * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA - */ - -#ifndef AVUTIL_XTEA_H -#define AVUTIL_XTEA_H - -#include <stdint.h> - -/** - * @file - * @brief Public header for libavutil XTEA algorithm - * @defgroup lavu_xtea XTEA - * @ingroup lavu_crypto - * @{ - */ - -typedef struct AVXTEA { - uint32_t key[16]; -} AVXTEA; - -/** - * Allocate an AVXTEA context. - */ -AVXTEA *av_xtea_alloc(void); - -/** - * Initialize an AVXTEA context. - * - * @param ctx an AVXTEA context - * @param key a key of 16 bytes used for encryption/decryption, - * interpreted as big endian 32 bit numbers - */ -void av_xtea_init(struct AVXTEA *ctx, const uint8_t key[16]); - -/** - * Initialize an AVXTEA context. - * - * @param ctx an AVXTEA context - * @param key a key of 16 bytes used for encryption/decryption, - * interpreted as little endian 32 bit numbers - */ -void av_xtea_le_init(struct AVXTEA *ctx, const uint8_t key[16]); - -/** - * Encrypt or decrypt a buffer using a previously initialized context, - * in big endian format. - * - * @param ctx an AVXTEA context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 8 byte blocks - * @param iv initialization vector for CBC mode, if NULL then ECB will be used - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_xtea_crypt(struct AVXTEA *ctx, uint8_t *dst, const uint8_t *src, - int count, uint8_t *iv, int decrypt); - -/** - * Encrypt or decrypt a buffer using a previously initialized context, - * in little endian format. - * - * @param ctx an AVXTEA context - * @param dst destination array, can be equal to src - * @param src source array, can be equal to dst - * @param count number of 8 byte blocks - * @param iv initialization vector for CBC mode, if NULL then ECB will be used - * @param decrypt 0 for encryption, 1 for decryption - */ -void av_xtea_le_crypt(struct AVXTEA *ctx, uint8_t *dst, const uint8_t *src, - int count, uint8_t *iv, int decrypt); - -/** - * @} - */ - -#endif /* AVUTIL_XTEA_H */ diff --git a/src/main/jni/jni.c b/src/main/jni/jni.c deleted file mode 100644 index b8bb4713a..000000000 --- a/src/main/jni/jni.c +++ /dev/null @@ -1,40 +0,0 @@ -#include <stdio.h> -#include <string.h> -#include <jni.h> -#include <sys/types.h> -#include <inttypes.h> -#include <stdlib.h> -#include <openssl/aes.h> -#include <unistd.h> -#include "utils.h" -#include "image.h" - -int registerNativeTgNetFunctions(JavaVM *vm, JNIEnv *env); -int gifvideoOnJNILoad(JavaVM *vm, JNIEnv *env); - -jint JNI_OnLoad(JavaVM *vm, void *reserved) { - JNIEnv *env = 0; - srand(time(NULL)); - - if ((*vm)->GetEnv(vm, (void **) &env, JNI_VERSION_1_6) != JNI_OK) { - return -1; - } - - if (imageOnJNILoad(vm, reserved, env) == -1) { - return -1; - } - - if (gifvideoOnJNILoad(vm, env) == -1) { - return -1; - } - - if (registerNativeTgNetFunctions(vm, env) != JNI_TRUE) { - return -1; - } - - return JNI_VERSION_1_6; -} - -void JNI_OnUnload(JavaVM *vm, void *reserved) { - -} diff --git a/src/main/jni/libyuv/include/libyuv.h b/src/main/jni/libyuv/include/libyuv.h deleted file mode 100644 index 3bebe642c..000000000 --- a/src/main/jni/libyuv/include/libyuv.h +++ /dev/null @@ -1,33 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_H_ // NOLINT -#define INCLUDE_LIBYUV_H_ - -#include "libyuv/basic_types.h" -#include "libyuv/compare.h" -#include "libyuv/convert.h" -#include "libyuv/convert_argb.h" -#include "libyuv/convert_from.h" -#include "libyuv/convert_from_argb.h" -#include "libyuv/cpu_id.h" -#include "libyuv/format_conversion.h" -#include "libyuv/mjpeg_decoder.h" -#include "libyuv/planar_functions.h" -#include "libyuv/rotate.h" -#include "libyuv/rotate_argb.h" -#include "libyuv/row.h" -#include "libyuv/scale.h" -#include "libyuv/scale_argb.h" -#include "libyuv/scale_row.h" -#include "libyuv/version.h" -#include "libyuv/video_common.h" - -#endif // INCLUDE_LIBYUV_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/basic_types.h b/src/main/jni/libyuv/include/libyuv/basic_types.h deleted file mode 100644 index beb750ba6..000000000 --- a/src/main/jni/libyuv/include/libyuv/basic_types.h +++ /dev/null @@ -1,118 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_BASIC_TYPES_H_ // NOLINT -#define INCLUDE_LIBYUV_BASIC_TYPES_H_ - -#include <stddef.h> // for NULL, size_t - -#if defined(__ANDROID__) || (defined(_MSC_VER) && (_MSC_VER < 1600)) -#include <sys/types.h> // for uintptr_t on x86 -#else -#include <stdint.h> // for uintptr_t -#endif - -#ifndef GG_LONGLONG -#ifndef INT_TYPES_DEFINED -#define INT_TYPES_DEFINED -#ifdef COMPILER_MSVC -typedef unsigned __int64 uint64; -typedef __int64 int64; -#ifndef INT64_C -#define INT64_C(x) x ## I64 -#endif -#ifndef UINT64_C -#define UINT64_C(x) x ## UI64 -#endif -#define INT64_F "I64" -#else // COMPILER_MSVC -#if defined(__LP64__) && !defined(__OpenBSD__) && !defined(__APPLE__) -typedef unsigned long uint64; // NOLINT -typedef long int64; // NOLINT -#ifndef INT64_C -#define INT64_C(x) x ## L -#endif -#ifndef UINT64_C -#define UINT64_C(x) x ## UL -#endif -#define INT64_F "l" -#else // defined(__LP64__) && !defined(__OpenBSD__) && !defined(__APPLE__) -typedef unsigned long long uint64; // NOLINT -typedef long long int64; // NOLINT -#ifndef INT64_C -#define INT64_C(x) x ## LL -#endif -#ifndef UINT64_C -#define UINT64_C(x) x ## ULL -#endif -#define INT64_F "ll" -#endif // __LP64__ -#endif // COMPILER_MSVC -typedef unsigned int uint32; -typedef int int32; -typedef unsigned short uint16; // NOLINT -typedef short int16; // NOLINT -typedef unsigned char uint8; -typedef signed char int8; -#endif // INT_TYPES_DEFINED -#endif // GG_LONGLONG - -// Detect compiler is for x86 or x64. -#if defined(__x86_64__) || defined(_M_X64) || \ - defined(__i386__) || defined(_M_IX86) -#define CPU_X86 1 -#endif -// Detect compiler is for ARM. -#if defined(__arm__) || defined(_M_ARM) -#define CPU_ARM 1 -#endif - -#ifndef ALIGNP -#ifdef __cplusplus -#define ALIGNP(p, t) \ - (reinterpret_cast<uint8*>(((reinterpret_cast<uintptr_t>(p) + \ - ((t) - 1)) & ~((t) - 1)))) -#else -#define ALIGNP(p, t) \ - ((uint8*)((((uintptr_t)(p) + ((t) - 1)) & ~((t) - 1)))) /* NOLINT */ -#endif -#endif - -#if !defined(LIBYUV_API) -#if defined(_WIN32) || defined(__CYGWIN__) -#if defined(LIBYUV_BUILDING_SHARED_LIBRARY) -#define LIBYUV_API __declspec(dllexport) -#elif defined(LIBYUV_USING_SHARED_LIBRARY) -#define LIBYUV_API __declspec(dllimport) -#else -#define LIBYUV_API -#endif // LIBYUV_BUILDING_SHARED_LIBRARY -#elif defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__APPLE__) && \ - (defined(LIBYUV_BUILDING_SHARED_LIBRARY) || \ - defined(LIBYUV_USING_SHARED_LIBRARY)) -#define LIBYUV_API __attribute__ ((visibility ("default"))) -#else -#define LIBYUV_API -#endif // __GNUC__ -#endif // LIBYUV_API - -#define LIBYUV_BOOL int -#define LIBYUV_FALSE 0 -#define LIBYUV_TRUE 1 - -// Visual C x86 or GCC little endian. -#if defined(__x86_64__) || defined(_M_X64) || \ - defined(__i386__) || defined(_M_IX86) || \ - defined(__arm__) || defined(_M_ARM) || \ - (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) -#define LIBYUV_LITTLE_ENDIAN -#endif - -#endif // INCLUDE_LIBYUV_BASIC_TYPES_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/compare.h b/src/main/jni/libyuv/include/libyuv/compare.h deleted file mode 100644 index 5dfac7c86..000000000 --- a/src/main/jni/libyuv/include/libyuv/compare.h +++ /dev/null @@ -1,73 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_COMPARE_H_ // NOLINT -#define INCLUDE_LIBYUV_COMPARE_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Compute a hash for specified memory. Seed of 5381 recommended. -LIBYUV_API -uint32 HashDjb2(const uint8* src, uint64 count, uint32 seed); - -// Sum Square Error - used to compute Mean Square Error or PSNR. -LIBYUV_API -uint64 ComputeSumSquareError(const uint8* src_a, - const uint8* src_b, int count); - -LIBYUV_API -uint64 ComputeSumSquareErrorPlane(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b, - int width, int height); - -static const int kMaxPsnr = 128; - -LIBYUV_API -double SumSquareErrorToPsnr(uint64 sse, uint64 count); - -LIBYUV_API -double CalcFramePsnr(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b, - int width, int height); - -LIBYUV_API -double I420Psnr(const uint8* src_y_a, int stride_y_a, - const uint8* src_u_a, int stride_u_a, - const uint8* src_v_a, int stride_v_a, - const uint8* src_y_b, int stride_y_b, - const uint8* src_u_b, int stride_u_b, - const uint8* src_v_b, int stride_v_b, - int width, int height); - -LIBYUV_API -double CalcFrameSsim(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b, - int width, int height); - -LIBYUV_API -double I420Ssim(const uint8* src_y_a, int stride_y_a, - const uint8* src_u_a, int stride_u_a, - const uint8* src_v_a, int stride_v_a, - const uint8* src_y_b, int stride_y_b, - const uint8* src_u_b, int stride_u_b, - const uint8* src_v_b, int stride_v_b, - int width, int height); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_COMPARE_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/convert.h b/src/main/jni/libyuv/include/libyuv/convert.h deleted file mode 100644 index 1bd45c837..000000000 --- a/src/main/jni/libyuv/include/libyuv/convert.h +++ /dev/null @@ -1,254 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_CONVERT_H_ // NOLINT -#define INCLUDE_LIBYUV_CONVERT_H_ - -#include "libyuv/basic_types.h" -// TODO(fbarchard): Remove the following headers includes. -#include "libyuv/convert_from.h" -#include "libyuv/planar_functions.h" -#include "libyuv/rotate.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Convert I444 to I420. -LIBYUV_API -int I444ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert I422 to I420. -LIBYUV_API -int I422ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert I411 to I420. -LIBYUV_API -int I411ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Copy I420 to I420. -#define I420ToI420 I420Copy -LIBYUV_API -int I420Copy(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert I400 (grey) to I420. -LIBYUV_API -int I400ToI420(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert NV12 to I420. -LIBYUV_API -int NV12ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert NV21 to I420. -LIBYUV_API -int NV21ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_vu, int src_stride_vu, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert YUY2 to I420. -LIBYUV_API -int YUY2ToI420(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert UYVY to I420. -LIBYUV_API -int UYVYToI420(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert M420 to I420. -LIBYUV_API -int M420ToI420(const uint8* src_m420, int src_stride_m420, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert Q420 to I420. -LIBYUV_API -int Q420ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// ARGB little endian (bgra in memory) to I420. -LIBYUV_API -int ARGBToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// BGRA little endian (argb in memory) to I420. -LIBYUV_API -int BGRAToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// ABGR little endian (rgba in memory) to I420. -LIBYUV_API -int ABGRToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// RGBA little endian (abgr in memory) to I420. -LIBYUV_API -int RGBAToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// RGB little endian (bgr in memory) to I420. -LIBYUV_API -int RGB24ToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// RGB big endian (rgb in memory) to I420. -LIBYUV_API -int RAWToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// RGB16 (RGBP fourcc) little endian to I420. -LIBYUV_API -int RGB565ToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// RGB15 (RGBO fourcc) little endian to I420. -LIBYUV_API -int ARGB1555ToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// RGB12 (R444 fourcc) little endian to I420. -LIBYUV_API -int ARGB4444ToI420(const uint8* src_frame, int src_stride_frame, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -#ifdef HAVE_JPEG -// src_width/height provided by capture. -// dst_width/height for clipping determine final size. -LIBYUV_API -int MJPGToI420(const uint8* sample, size_t sample_size, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int src_width, int src_height, - int dst_width, int dst_height); - -// Query size of MJPG in pixels. -LIBYUV_API -int MJPGSize(const uint8* sample, size_t sample_size, - int* width, int* height); -#endif - -// Note Bayer formats (BGGR) To I420 are in format_conversion.h - -// Convert camera sample to I420 with cropping, rotation and vertical flip. -// "src_size" is needed to parse MJPG. -// "dst_stride_y" number of bytes in a row of the dst_y plane. -// Normally this would be the same as dst_width, with recommended alignment -// to 16 bytes for better efficiency. -// If rotation of 90 or 270 is used, stride is affected. The caller should -// allocate the I420 buffer according to rotation. -// "dst_stride_u" number of bytes in a row of the dst_u plane. -// Normally this would be the same as (dst_width + 1) / 2, with -// recommended alignment to 16 bytes for better efficiency. -// If rotation of 90 or 270 is used, stride is affected. -// "crop_x" and "crop_y" are starting position for cropping. -// To center, crop_x = (src_width - dst_width) / 2 -// crop_y = (src_height - dst_height) / 2 -// "src_width" / "src_height" is size of src_frame in pixels. -// "src_height" can be negative indicating a vertically flipped image source. -// "crop_width" / "crop_height" is the size to crop the src to. -// Must be less than or equal to src_width/src_height -// Cropping parameters are pre-rotation. -// "rotation" can be 0, 90, 180 or 270. -// "format" is a fourcc. ie 'I420', 'YUY2' -// Returns 0 for successful; -1 for invalid parameter. Non-zero for failure. -LIBYUV_API -int ConvertToI420(const uint8* src_frame, size_t src_size, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int crop_x, int crop_y, - int src_width, int src_height, - int crop_width, int crop_height, - enum RotationMode rotation, - uint32 format); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_CONVERT_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/convert_argb.h b/src/main/jni/libyuv/include/libyuv/convert_argb.h deleted file mode 100644 index a18014ca2..000000000 --- a/src/main/jni/libyuv/include/libyuv/convert_argb.h +++ /dev/null @@ -1,225 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_CONVERT_ARGB_H_ // NOLINT -#define INCLUDE_LIBYUV_CONVERT_ARGB_H_ - -#include "libyuv/basic_types.h" -// TODO(fbarchard): Remove the following headers includes -#include "libyuv/convert_from.h" -#include "libyuv/planar_functions.h" -#include "libyuv/rotate.h" - -// TODO(fbarchard): This set of functions should exactly match convert.h -// Add missing Q420. -// TODO(fbarchard): Add tests. Create random content of right size and convert -// with C vs Opt and or to I420 and compare. -// TODO(fbarchard): Some of these functions lack parameter setting. - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Alias. -#define ARGBToARGB ARGBCopy - -// Copy ARGB to ARGB. -LIBYUV_API -int ARGBCopy(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert I420 to ARGB. -LIBYUV_API -int I420ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert I422 to ARGB. -LIBYUV_API -int I422ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert I444 to ARGB. -LIBYUV_API -int I444ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert I411 to ARGB. -LIBYUV_API -int I411ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert I400 (grey) to ARGB. -LIBYUV_API -int I400ToARGB(const uint8* src_y, int src_stride_y, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Alias. -#define YToARGB I400ToARGB_Reference - -// Convert I400 to ARGB. Reverse of ARGBToI400. -LIBYUV_API -int I400ToARGB_Reference(const uint8* src_y, int src_stride_y, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert NV12 to ARGB. -LIBYUV_API -int NV12ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert NV21 to ARGB. -LIBYUV_API -int NV21ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_vu, int src_stride_vu, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert M420 to ARGB. -LIBYUV_API -int M420ToARGB(const uint8* src_m420, int src_stride_m420, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// TODO(fbarchard): Convert Q420 to ARGB. -// LIBYUV_API -// int Q420ToARGB(const uint8* src_y, int src_stride_y, -// const uint8* src_yuy2, int src_stride_yuy2, -// uint8* dst_argb, int dst_stride_argb, -// int width, int height); - -// Convert YUY2 to ARGB. -LIBYUV_API -int YUY2ToARGB(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert UYVY to ARGB. -LIBYUV_API -int UYVYToARGB(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// BGRA little endian (argb in memory) to ARGB. -LIBYUV_API -int BGRAToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// ABGR little endian (rgba in memory) to ARGB. -LIBYUV_API -int ABGRToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// RGBA little endian (abgr in memory) to ARGB. -LIBYUV_API -int RGBAToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Deprecated function name. -#define BG24ToARGB RGB24ToARGB - -// RGB little endian (bgr in memory) to ARGB. -LIBYUV_API -int RGB24ToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// RGB big endian (rgb in memory) to ARGB. -LIBYUV_API -int RAWToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// RGB16 (RGBP fourcc) little endian to ARGB. -LIBYUV_API -int RGB565ToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// RGB15 (RGBO fourcc) little endian to ARGB. -LIBYUV_API -int ARGB1555ToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// RGB12 (R444 fourcc) little endian to ARGB. -LIBYUV_API -int ARGB4444ToARGB(const uint8* src_frame, int src_stride_frame, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -#ifdef HAVE_JPEG -// src_width/height provided by capture -// dst_width/height for clipping determine final size. -LIBYUV_API -int MJPGToARGB(const uint8* sample, size_t sample_size, - uint8* dst_argb, int dst_stride_argb, - int src_width, int src_height, - int dst_width, int dst_height); -#endif - -// Note Bayer formats (BGGR) to ARGB are in format_conversion.h. - -// Convert camera sample to ARGB with cropping, rotation and vertical flip. -// "src_size" is needed to parse MJPG. -// "dst_stride_argb" number of bytes in a row of the dst_argb plane. -// Normally this would be the same as dst_width, with recommended alignment -// to 16 bytes for better efficiency. -// If rotation of 90 or 270 is used, stride is affected. The caller should -// allocate the I420 buffer according to rotation. -// "dst_stride_u" number of bytes in a row of the dst_u plane. -// Normally this would be the same as (dst_width + 1) / 2, with -// recommended alignment to 16 bytes for better efficiency. -// If rotation of 90 or 270 is used, stride is affected. -// "crop_x" and "crop_y" are starting position for cropping. -// To center, crop_x = (src_width - dst_width) / 2 -// crop_y = (src_height - dst_height) / 2 -// "src_width" / "src_height" is size of src_frame in pixels. -// "src_height" can be negative indicating a vertically flipped image source. -// "crop_width" / "crop_height" is the size to crop the src to. -// Must be less than or equal to src_width/src_height -// Cropping parameters are pre-rotation. -// "rotation" can be 0, 90, 180 or 270. -// "format" is a fourcc. ie 'I420', 'YUY2' -// Returns 0 for successful; -1 for invalid parameter. Non-zero for failure. -LIBYUV_API -int ConvertToARGB(const uint8* src_frame, size_t src_size, - uint8* dst_argb, int dst_stride_argb, - int crop_x, int crop_y, - int src_width, int src_height, - int crop_width, int crop_height, - enum RotationMode rotation, - uint32 format); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_CONVERT_ARGB_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/convert_from.h b/src/main/jni/libyuv/include/libyuv/convert_from.h deleted file mode 100644 index b1cf57f7d..000000000 --- a/src/main/jni/libyuv/include/libyuv/convert_from.h +++ /dev/null @@ -1,173 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_CONVERT_FROM_H_ // NOLINT -#define INCLUDE_LIBYUV_CONVERT_FROM_H_ - -#include "libyuv/basic_types.h" -#include "libyuv/rotate.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// See Also convert.h for conversions from formats to I420. - -// I420Copy in convert to I420ToI420. - -LIBYUV_API -int I420ToI422(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -LIBYUV_API -int I420ToI444(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -LIBYUV_API -int I420ToI411(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Copy to I400. Source can be I420, I422, I444, I400, NV12 or NV21. -LIBYUV_API -int I400Copy(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height); - -// TODO(fbarchard): I420ToM420 -// TODO(fbarchard): I420ToQ420 - -LIBYUV_API -int I420ToNV12(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_uv, int dst_stride_uv, - int width, int height); - -LIBYUV_API -int I420ToNV21(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_vu, int dst_stride_vu, - int width, int height); - -LIBYUV_API -int I420ToYUY2(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToUYVY(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -LIBYUV_API -int I420ToBGRA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -LIBYUV_API -int I420ToABGR(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -LIBYUV_API -int I420ToRGBA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_rgba, int dst_stride_rgba, - int width, int height); - -LIBYUV_API -int I420ToRGB24(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToRAW(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToRGB565(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToARGB1555(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToARGB4444(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -// Note Bayer formats (BGGR) To I420 are in format_conversion.h. - -// Convert I420 to specified format. -// "dst_sample_stride" is bytes in a row for the destination. Pass 0 if the -// buffer has contiguous rows. Can be negative. A multiple of 16 is optimal. -LIBYUV_API -int ConvertFromI420(const uint8* y, int y_stride, - const uint8* u, int u_stride, - const uint8* v, int v_stride, - uint8* dst_sample, int dst_sample_stride, - int width, int height, - uint32 format); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_CONVERT_FROM_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/convert_from_argb.h b/src/main/jni/libyuv/include/libyuv/convert_from_argb.h deleted file mode 100644 index 90f43af04..000000000 --- a/src/main/jni/libyuv/include/libyuv/convert_from_argb.h +++ /dev/null @@ -1,166 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_CONVERT_FROM_ARGB_H_ // NOLINT -#define INCLUDE_LIBYUV_CONVERT_FROM_ARGB_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Copy ARGB to ARGB. -#define ARGBToARGB ARGBCopy -LIBYUV_API -int ARGBCopy(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert ARGB To BGRA. -LIBYUV_API -int ARGBToBGRA(const uint8* src_argb, int src_stride_argb, - uint8* dst_bgra, int dst_stride_bgra, - int width, int height); - -// Convert ARGB To ABGR. -LIBYUV_API -int ARGBToABGR(const uint8* src_argb, int src_stride_argb, - uint8* dst_abgr, int dst_stride_abgr, - int width, int height); - -// Convert ARGB To RGBA. -LIBYUV_API -int ARGBToRGBA(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgba, int dst_stride_rgba, - int width, int height); - -// Convert ARGB To RGB24. -LIBYUV_API -int ARGBToRGB24(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgb24, int dst_stride_rgb24, - int width, int height); - -// Convert ARGB To RAW. -LIBYUV_API -int ARGBToRAW(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgb, int dst_stride_rgb, - int width, int height); - -// Convert ARGB To RGB565. -LIBYUV_API -int ARGBToRGB565(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height); - -// Convert ARGB To ARGB1555. -LIBYUV_API -int ARGBToARGB1555(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb1555, int dst_stride_argb1555, - int width, int height); - -// Convert ARGB To ARGB4444. -LIBYUV_API -int ARGBToARGB4444(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb4444, int dst_stride_argb4444, - int width, int height); - -// Convert ARGB To I444. -LIBYUV_API -int ARGBToI444(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert ARGB To I422. -LIBYUV_API -int ARGBToI422(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert ARGB To I420. (also in convert.h) -LIBYUV_API -int ARGBToI420(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert ARGB to J420. (JPeg full range I420). -LIBYUV_API -int ARGBToJ420(const uint8* src_argb, int src_stride_argb, - uint8* dst_yj, int dst_stride_yj, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert ARGB To I411. -LIBYUV_API -int ARGBToI411(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert ARGB to J400. (JPeg full range). -LIBYUV_API -int ARGBToJ400(const uint8* src_argb, int src_stride_argb, - uint8* dst_yj, int dst_stride_yj, - int width, int height); - -// Convert ARGB to I400. -LIBYUV_API -int ARGBToI400(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - int width, int height); - -// Convert ARGB To NV12. -LIBYUV_API -int ARGBToNV12(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_uv, int dst_stride_uv, - int width, int height); - -// Convert ARGB To NV21. -LIBYUV_API -int ARGBToNV21(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_vu, int dst_stride_vu, - int width, int height); - -// Convert ARGB To NV21. -LIBYUV_API -int ARGBToNV21(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_vu, int dst_stride_vu, - int width, int height); - -// Convert ARGB To YUY2. -LIBYUV_API -int ARGBToYUY2(const uint8* src_argb, int src_stride_argb, - uint8* dst_yuy2, int dst_stride_yuy2, - int width, int height); - -// Convert ARGB To UYVY. -LIBYUV_API -int ARGBToUYVY(const uint8* src_argb, int src_stride_argb, - uint8* dst_uyvy, int dst_stride_uyvy, - int width, int height); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_CONVERT_FROM_ARGB_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/cpu_id.h b/src/main/jni/libyuv/include/libyuv/cpu_id.h deleted file mode 100644 index dc858a814..000000000 --- a/src/main/jni/libyuv/include/libyuv/cpu_id.h +++ /dev/null @@ -1,81 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_CPU_ID_H_ // NOLINT -#define INCLUDE_LIBYUV_CPU_ID_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// TODO(fbarchard): Consider overlapping bits for different architectures. -// Internal flag to indicate cpuid requires initialization. -#define kCpuInit 0x1 - -// These flags are only valid on ARM processors. -static const int kCpuHasARM = 0x2; -static const int kCpuHasNEON = 0x4; -// 0x8 reserved for future ARM flag. - -// These flags are only valid on x86 processors. -static const int kCpuHasX86 = 0x10; -static const int kCpuHasSSE2 = 0x20; -static const int kCpuHasSSSE3 = 0x40; -static const int kCpuHasSSE41 = 0x80; -static const int kCpuHasSSE42 = 0x100; -static const int kCpuHasAVX = 0x200; -static const int kCpuHasAVX2 = 0x400; -static const int kCpuHasERMS = 0x800; -static const int kCpuHasFMA3 = 0x1000; -// 0x2000, 0x4000, 0x8000 reserved for future X86 flags. - -// These flags are only valid on MIPS processors. -static const int kCpuHasMIPS = 0x10000; -static const int kCpuHasMIPS_DSP = 0x20000; -static const int kCpuHasMIPS_DSPR2 = 0x40000; - -// Internal function used to auto-init. -LIBYUV_API -int InitCpuFlags(void); - -// Internal function for parsing /proc/cpuinfo. -LIBYUV_API -int ArmCpuCaps(const char* cpuinfo_name); - -// Detect CPU has SSE2 etc. -// Test_flag parameter should be one of kCpuHas constants above. -// returns non-zero if instruction set is detected -static __inline int TestCpuFlag(int test_flag) { - LIBYUV_API extern int cpu_info_; - return (cpu_info_ == kCpuInit ? InitCpuFlags() : cpu_info_) & test_flag; -} - -// For testing, allow CPU flags to be disabled. -// ie MaskCpuFlags(~kCpuHasSSSE3) to disable SSSE3. -// MaskCpuFlags(-1) to enable all cpu specific optimizations. -// MaskCpuFlags(0) to disable all cpu specific optimizations. -LIBYUV_API -void MaskCpuFlags(int enable_flags); - -// Low level cpuid for X86. Returns zeros on other CPUs. -// eax is the info type that you want. -// ecx is typically the cpu number, and should normally be zero. -LIBYUV_API -void CpuId(uint32 eax, uint32 ecx, uint32* cpu_info); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_CPU_ID_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/format_conversion.h b/src/main/jni/libyuv/include/libyuv/format_conversion.h deleted file mode 100644 index b18bf0534..000000000 --- a/src/main/jni/libyuv/include/libyuv/format_conversion.h +++ /dev/null @@ -1,168 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_FORMATCONVERSION_H_ // NOLINT -#define INCLUDE_LIBYUV_FORMATCONVERSION_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Convert Bayer RGB formats to I420. -LIBYUV_API -int BayerBGGRToI420(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -LIBYUV_API -int BayerGBRGToI420(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -LIBYUV_API -int BayerGRBGToI420(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -LIBYUV_API -int BayerRGGBToI420(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Temporary API mapper. -#define BayerRGBToI420(b, bs, f, y, ys, u, us, v, vs, w, h) \ - BayerToI420(b, bs, y, ys, u, us, v, vs, w, h, f) - -LIBYUV_API -int BayerToI420(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height, - uint32 src_fourcc_bayer); - -// Convert I420 to Bayer RGB formats. -LIBYUV_API -int I420ToBayerBGGR(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToBayerGBRG(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToBayerGRBG(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -LIBYUV_API -int I420ToBayerRGGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -// Temporary API mapper. -#define I420ToBayerRGB(y, ys, u, us, v, vs, b, bs, f, w, h) \ - I420ToBayer(y, ys, u, us, v, vs, b, bs, w, h, f) - -LIBYUV_API -int I420ToBayer(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height, - uint32 dst_fourcc_bayer); - -// Convert Bayer RGB formats to ARGB. -LIBYUV_API -int BayerBGGRToARGB(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -LIBYUV_API -int BayerGBRGToARGB(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -LIBYUV_API -int BayerGRBGToARGB(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -LIBYUV_API -int BayerRGGBToARGB(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Temporary API mapper. -#define BayerRGBToARGB(b, bs, f, a, as, w, h) BayerToARGB(b, bs, a, as, w, h, f) - -LIBYUV_API -int BayerToARGB(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int dst_stride_argb, - int width, int height, - uint32 src_fourcc_bayer); - -// Converts ARGB to Bayer RGB formats. -LIBYUV_API -int ARGBToBayerBGGR(const uint8* src_argb, int src_stride_argb, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height); - -LIBYUV_API -int ARGBToBayerGBRG(const uint8* src_argb, int src_stride_argb, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height); - -LIBYUV_API -int ARGBToBayerGRBG(const uint8* src_argb, int src_stride_argb, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height); - -LIBYUV_API -int ARGBToBayerRGGB(const uint8* src_argb, int src_stride_argb, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height); - -// Temporary API mapper. -#define ARGBToBayerRGB(a, as, b, bs, f, w, h) ARGBToBayer(b, bs, a, as, w, h, f) - -LIBYUV_API -int ARGBToBayer(const uint8* src_argb, int src_stride_argb, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height, - uint32 dst_fourcc_bayer); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_FORMATCONVERSION_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/mjpeg_decoder.h b/src/main/jni/libyuv/include/libyuv/mjpeg_decoder.h deleted file mode 100644 index 8423121d1..000000000 --- a/src/main/jni/libyuv/include/libyuv/mjpeg_decoder.h +++ /dev/null @@ -1,192 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_MJPEG_DECODER_H_ // NOLINT -#define INCLUDE_LIBYUV_MJPEG_DECODER_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -// NOTE: For a simplified public API use convert.h MJPGToI420(). - -struct jpeg_common_struct; -struct jpeg_decompress_struct; -struct jpeg_source_mgr; - -namespace libyuv { - -#ifdef __cplusplus -extern "C" { -#endif - -LIBYUV_BOOL ValidateJpeg(const uint8* sample, size_t sample_size); - -#ifdef __cplusplus -} // extern "C" -#endif - -static const uint32 kUnknownDataSize = 0xFFFFFFFF; - -enum JpegSubsamplingType { - kJpegYuv420, - kJpegYuv422, - kJpegYuv411, - kJpegYuv444, - kJpegYuv400, - kJpegUnknown -}; - -struct Buffer { - const uint8* data; - int len; -}; - -struct BufferVector { - Buffer* buffers; - int len; - int pos; -}; - -struct SetJmpErrorMgr; - -// MJPEG ("Motion JPEG") is a pseudo-standard video codec where the frames are -// simply independent JPEG images with a fixed huffman table (which is omitted). -// It is rarely used in video transmission, but is common as a camera capture -// format, especially in Logitech devices. This class implements a decoder for -// MJPEG frames. -// -// See http://tools.ietf.org/html/rfc2435 -class LIBYUV_API MJpegDecoder { - public: - typedef void (*CallbackFunction)(void* opaque, - const uint8* const* data, - const int* strides, - int rows); - - static const int kColorSpaceUnknown; - static const int kColorSpaceGrayscale; - static const int kColorSpaceRgb; - static const int kColorSpaceYCbCr; - static const int kColorSpaceCMYK; - static const int kColorSpaceYCCK; - - MJpegDecoder(); - ~MJpegDecoder(); - - // Loads a new frame, reads its headers, and determines the uncompressed - // image format. - // Returns LIBYUV_TRUE if image looks valid and format is supported. - // If return value is LIBYUV_TRUE, then the values for all the following - // getters are populated. - // src_len is the size of the compressed mjpeg frame in bytes. - LIBYUV_BOOL LoadFrame(const uint8* src, size_t src_len); - - // Returns width of the last loaded frame in pixels. - int GetWidth(); - - // Returns height of the last loaded frame in pixels. - int GetHeight(); - - // Returns format of the last loaded frame. The return value is one of the - // kColorSpace* constants. - int GetColorSpace(); - - // Number of color components in the color space. - int GetNumComponents(); - - // Sample factors of the n-th component. - int GetHorizSampFactor(int component); - - int GetVertSampFactor(int component); - - int GetHorizSubSampFactor(int component); - - int GetVertSubSampFactor(int component); - - // Public for testability. - int GetImageScanlinesPerImcuRow(); - - // Public for testability. - int GetComponentScanlinesPerImcuRow(int component); - - // Width of a component in bytes. - int GetComponentWidth(int component); - - // Height of a component. - int GetComponentHeight(int component); - - // Width of a component in bytes with padding for DCTSIZE. Public for testing. - int GetComponentStride(int component); - - // Size of a component in bytes. - int GetComponentSize(int component); - - // Call this after LoadFrame() if you decide you don't want to decode it - // after all. - LIBYUV_BOOL UnloadFrame(); - - // Decodes the entire image into a one-buffer-per-color-component format. - // dst_width must match exactly. dst_height must be <= to image height; if - // less, the image is cropped. "planes" must have size equal to at least - // GetNumComponents() and they must point to non-overlapping buffers of size - // at least GetComponentSize(i). The pointers in planes are incremented - // to point to after the end of the written data. - // TODO(fbarchard): Add dst_x, dst_y to allow specific rect to be decoded. - LIBYUV_BOOL DecodeToBuffers(uint8** planes, int dst_width, int dst_height); - - // Decodes the entire image and passes the data via repeated calls to a - // callback function. Each call will get the data for a whole number of - // image scanlines. - // TODO(fbarchard): Add dst_x, dst_y to allow specific rect to be decoded. - LIBYUV_BOOL DecodeToCallback(CallbackFunction fn, void* opaque, - int dst_width, int dst_height); - - // The helper function which recognizes the jpeg sub-sampling type. - static JpegSubsamplingType JpegSubsamplingTypeHelper( - int* subsample_x, int* subsample_y, int number_of_components); - - private: - void AllocOutputBuffers(int num_outbufs); - void DestroyOutputBuffers(); - - LIBYUV_BOOL StartDecode(); - LIBYUV_BOOL FinishDecode(); - - void SetScanlinePointers(uint8** data); - LIBYUV_BOOL DecodeImcuRow(); - - int GetComponentScanlinePadding(int component); - - // A buffer holding the input data for a frame. - Buffer buf_; - BufferVector buf_vec_; - - jpeg_decompress_struct* decompress_struct_; - jpeg_source_mgr* source_mgr_; - SetJmpErrorMgr* error_mgr_; - - // LIBYUV_TRUE iff at least one component has scanline padding. (i.e., - // GetComponentScanlinePadding() != 0.) - LIBYUV_BOOL has_scanline_padding_; - - // Temporaries used to point to scanline outputs. - int num_outbufs_; // Outermost size of all arrays below. - uint8*** scanlines_; - int* scanlines_sizes_; - // Temporary buffer used for decoding when we can't decode directly to the - // output buffers. Large enough for just one iMCU row. - uint8** databuf_; - int* databuf_strides_; -}; - -} // namespace libyuv - -#endif // __cplusplus -#endif // INCLUDE_LIBYUV_MJPEG_DECODER_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/planar_functions.h b/src/main/jni/libyuv/include/libyuv/planar_functions.h deleted file mode 100644 index d10a16985..000000000 --- a/src/main/jni/libyuv/include/libyuv/planar_functions.h +++ /dev/null @@ -1,439 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_PLANAR_FUNCTIONS_H_ // NOLINT -#define INCLUDE_LIBYUV_PLANAR_FUNCTIONS_H_ - -#include "libyuv/basic_types.h" - -// TODO(fbarchard): Remove the following headers includes. -#include "libyuv/convert.h" -#include "libyuv/convert_argb.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Copy a plane of data. -LIBYUV_API -void CopyPlane(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height); - -LIBYUV_API -void CopyPlane_16(const uint16* src_y, int src_stride_y, - uint16* dst_y, int dst_stride_y, - int width, int height); - -// Set a plane of data to a 32 bit value. -LIBYUV_API -void SetPlane(uint8* dst_y, int dst_stride_y, - int width, int height, - uint32 value); - -// Copy I400. Supports inverting. -LIBYUV_API -int I400ToI400(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height); - - -// Copy I422 to I422. -#define I422ToI422 I422Copy -LIBYUV_API -int I422Copy(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Copy I444 to I444. -#define I444ToI444 I444Copy -LIBYUV_API -int I444Copy(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert YUY2 to I422. -LIBYUV_API -int YUY2ToI422(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert UYVY to I422. -LIBYUV_API -int UYVYToI422(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Convert I420 to I400. (calls CopyPlane ignoring u/v). -LIBYUV_API -int I420ToI400(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - int width, int height); - -// Alias -#define I420ToI420Mirror I420Mirror - -// I420 mirror. -LIBYUV_API -int I420Mirror(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height); - -// Alias -#define I400ToI400Mirror I400Mirror - -// I400 mirror. A single plane is mirrored horizontally. -// Pass negative height to achieve 180 degree rotation. -LIBYUV_API -int I400Mirror(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height); - -// Alias -#define ARGBToARGBMirror ARGBMirror - -// ARGB mirror. -LIBYUV_API -int ARGBMirror(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert NV12 to RGB565. -LIBYUV_API -int NV12ToRGB565(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height); - -// Convert NV21 to RGB565. -LIBYUV_API -int NV21ToRGB565(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height); - -// I422ToARGB is in convert_argb.h -// Convert I422 to BGRA. -LIBYUV_API -int I422ToBGRA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_bgra, int dst_stride_bgra, - int width, int height); - -// Convert I422 to ABGR. -LIBYUV_API -int I422ToABGR(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_abgr, int dst_stride_abgr, - int width, int height); - -// Convert I422 to RGBA. -LIBYUV_API -int I422ToRGBA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_rgba, int dst_stride_rgba, - int width, int height); - -// Draw a rectangle into I420. -LIBYUV_API -int I420Rect(uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int x, int y, int width, int height, - int value_y, int value_u, int value_v); - -// Draw a rectangle into ARGB. -LIBYUV_API -int ARGBRect(uint8* dst_argb, int dst_stride_argb, - int x, int y, int width, int height, uint32 value); - -// Convert ARGB to gray scale ARGB. -LIBYUV_API -int ARGBGrayTo(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Make a rectangle of ARGB gray scale. -LIBYUV_API -int ARGBGray(uint8* dst_argb, int dst_stride_argb, - int x, int y, int width, int height); - -// Make a rectangle of ARGB Sepia tone. -LIBYUV_API -int ARGBSepia(uint8* dst_argb, int dst_stride_argb, - int x, int y, int width, int height); - -// Apply a matrix rotation to each ARGB pixel. -// matrix_argb is 4 signed ARGB values. -128 to 127 representing -2 to 2. -// The first 4 coefficients apply to B, G, R, A and produce B of the output. -// The next 4 coefficients apply to B, G, R, A and produce G of the output. -// The next 4 coefficients apply to B, G, R, A and produce R of the output. -// The last 4 coefficients apply to B, G, R, A and produce A of the output. -LIBYUV_API -int ARGBColorMatrix(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - const int8* matrix_argb, - int width, int height); - -// Deprecated. Use ARGBColorMatrix instead. -// Apply a matrix rotation to each ARGB pixel. -// matrix_argb is 3 signed ARGB values. -128 to 127 representing -1 to 1. -// The first 4 coefficients apply to B, G, R, A and produce B of the output. -// The next 4 coefficients apply to B, G, R, A and produce G of the output. -// The last 4 coefficients apply to B, G, R, A and produce R of the output. -LIBYUV_API -int RGBColorMatrix(uint8* dst_argb, int dst_stride_argb, - const int8* matrix_rgb, - int x, int y, int width, int height); - -// Apply a color table each ARGB pixel. -// Table contains 256 ARGB values. -LIBYUV_API -int ARGBColorTable(uint8* dst_argb, int dst_stride_argb, - const uint8* table_argb, - int x, int y, int width, int height); - -// Apply a color table each ARGB pixel but preserve destination alpha. -// Table contains 256 ARGB values. -LIBYUV_API -int RGBColorTable(uint8* dst_argb, int dst_stride_argb, - const uint8* table_argb, - int x, int y, int width, int height); - -// Apply a luma/color table each ARGB pixel but preserve destination alpha. -// Table contains 32768 values indexed by [Y][C] where 7 it 7 bit luma from -// RGB (YJ style) and C is an 8 bit color component (R, G or B). -LIBYUV_API -int ARGBLumaColorTable(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - const uint8* luma_rgb_table, - int width, int height); - -// Apply a 3 term polynomial to ARGB values. -// poly points to a 4x4 matrix. The first row is constants. The 2nd row is -// coefficients for b, g, r and a. The 3rd row is coefficients for b squared, -// g squared, r squared and a squared. The 4rd row is coefficients for b to -// the 3, g to the 3, r to the 3 and a to the 3. The values are summed and -// result clamped to 0 to 255. -// A polynomial approximation can be dirived using software such as 'R'. - -LIBYUV_API -int ARGBPolynomial(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - const float* poly, - int width, int height); - -// Quantize a rectangle of ARGB. Alpha unaffected. -// scale is a 16 bit fractional fixed point scaler between 0 and 65535. -// interval_size should be a value between 1 and 255. -// interval_offset should be a value between 0 and 255. -LIBYUV_API -int ARGBQuantize(uint8* dst_argb, int dst_stride_argb, - int scale, int interval_size, int interval_offset, - int x, int y, int width, int height); - -// Copy ARGB to ARGB. -LIBYUV_API -int ARGBCopy(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Copy ARGB to ARGB. -LIBYUV_API -int ARGBCopyAlpha(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Copy ARGB to ARGB. -LIBYUV_API -int ARGBCopyYToAlpha(const uint8* src_y, int src_stride_y, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -typedef void (*ARGBBlendRow)(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width); - -// Get function to Alpha Blend ARGB pixels and store to destination. -LIBYUV_API -ARGBBlendRow GetARGBBlend(); - -// Alpha Blend ARGB images and store to destination. -// Alpha of destination is set to 255. -LIBYUV_API -int ARGBBlend(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Multiply ARGB image by ARGB image. Shifted down by 8. Saturates to 255. -LIBYUV_API -int ARGBMultiply(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Add ARGB image with ARGB image. Saturates to 255. -LIBYUV_API -int ARGBAdd(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Subtract ARGB image (argb1) from ARGB image (argb0). Saturates to 0. -LIBYUV_API -int ARGBSubtract(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert I422 to YUY2. -LIBYUV_API -int I422ToYUY2(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -// Convert I422 to UYVY. -LIBYUV_API -int I422ToUYVY(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_frame, int dst_stride_frame, - int width, int height); - -// Convert unattentuated ARGB to preattenuated ARGB. -LIBYUV_API -int ARGBAttenuate(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert preattentuated ARGB to unattenuated ARGB. -LIBYUV_API -int ARGBUnattenuate(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Convert MJPG to ARGB. -LIBYUV_API -int MJPGToARGB(const uint8* sample, size_t sample_size, - uint8* argb, int argb_stride, - int w, int h, int dw, int dh); - -// Internal function - do not call directly. -// Computes table of cumulative sum for image where the value is the sum -// of all values above and to the left of the entry. Used by ARGBBlur. -LIBYUV_API -int ARGBComputeCumulativeSum(const uint8* src_argb, int src_stride_argb, - int32* dst_cumsum, int dst_stride32_cumsum, - int width, int height); - -// Blur ARGB image. -// dst_cumsum table of width * (height + 1) * 16 bytes aligned to -// 16 byte boundary. -// dst_stride32_cumsum is number of ints in a row (width * 4). -// radius is number of pixels around the center. e.g. 1 = 3x3. 2=5x5. -// Blur is optimized for radius of 5 (11x11) or less. -LIBYUV_API -int ARGBBlur(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int32* dst_cumsum, int dst_stride32_cumsum, - int width, int height, int radius); - -// Multiply ARGB image by ARGB value. -LIBYUV_API -int ARGBShade(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height, uint32 value); - -// Interpolate between two ARGB images using specified amount of interpolation -// (0 to 255) and store to destination. -// 'interpolation' is specified as 8 bit fraction where 0 means 100% src_argb0 -// and 255 means 1% src_argb0 and 99% src_argb1. -// Internally uses ARGBScale bilinear filtering. -// Caveat: This function will write up to 16 bytes beyond the end of dst_argb. -LIBYUV_API -int ARGBInterpolate(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height, int interpolation); - -#if defined(__pnacl__) || defined(__CLR_VER) || defined(COVERAGE_ENABLED) || \ - defined(TARGET_IPHONE_SIMULATOR) -#define LIBYUV_DISABLE_X86 -#endif - -// Row functions for copying a pixels from a source with a slope to a row -// of destination. Useful for scaling, rotation, mirror, texture mapping. -LIBYUV_API -void ARGBAffineRow_C(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* uv_dudv, int width); -// The following are available on all x86 platforms: -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) -LIBYUV_API -void ARGBAffineRow_SSE2(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* uv_dudv, int width); -#define HAS_ARGBAFFINEROW_SSE2 -#endif // LIBYUV_DISABLE_X86 - -// Shuffle ARGB channel order. e.g. BGRA to ARGB. -// shuffler is 16 bytes and must be aligned. -LIBYUV_API -int ARGBShuffle(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_argb, int dst_stride_argb, - const uint8* shuffler, int width, int height); - -// Sobel ARGB effect with planar output. -LIBYUV_API -int ARGBSobelToPlane(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - int width, int height); - -// Sobel ARGB effect. -LIBYUV_API -int ARGBSobel(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -// Sobel ARGB effect w/ Sobel X, Sobel, Sobel Y in ARGB. -LIBYUV_API -int ARGBSobelXY(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_PLANAR_FUNCTIONS_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/rotate.h b/src/main/jni/libyuv/include/libyuv/rotate.h deleted file mode 100644 index 8af60b895..000000000 --- a/src/main/jni/libyuv/include/libyuv/rotate.h +++ /dev/null @@ -1,117 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_ROTATE_H_ // NOLINT -#define INCLUDE_LIBYUV_ROTATE_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Supported rotation. -typedef enum RotationMode { - kRotate0 = 0, // No rotation. - kRotate90 = 90, // Rotate 90 degrees clockwise. - kRotate180 = 180, // Rotate 180 degrees. - kRotate270 = 270, // Rotate 270 degrees clockwise. - - // Deprecated. - kRotateNone = 0, - kRotateClockwise = 90, - kRotateCounterClockwise = 270, -} RotationModeEnum; - -// Rotate I420 frame. -LIBYUV_API -int I420Rotate(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int src_width, int src_height, enum RotationMode mode); - -// Rotate NV12 input and store in I420. -LIBYUV_API -int NV12ToI420Rotate(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int src_width, int src_height, enum RotationMode mode); - -// Rotate a plane by 0, 90, 180, or 270. -LIBYUV_API -int RotatePlane(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int src_width, int src_height, enum RotationMode mode); - -// Rotate planes by 90, 180, 270. Deprecated. -LIBYUV_API -void RotatePlane90(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height); - -LIBYUV_API -void RotatePlane180(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height); - -LIBYUV_API -void RotatePlane270(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height); - -LIBYUV_API -void RotateUV90(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height); - -// Rotations for when U and V are interleaved. -// These functions take one input pointer and -// split the data into two buffers while -// rotating them. Deprecated. -LIBYUV_API -void RotateUV180(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height); - -LIBYUV_API -void RotateUV270(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height); - -// The 90 and 270 functions are based on transposes. -// Doing a transpose with reversing the read/write -// order will result in a rotation by +- 90 degrees. -// Deprecated. -LIBYUV_API -void TransposePlane(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height); - -LIBYUV_API -void TransposeUV(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_ROTATE_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/rotate_argb.h b/src/main/jni/libyuv/include/libyuv/rotate_argb.h deleted file mode 100644 index 660ff5573..000000000 --- a/src/main/jni/libyuv/include/libyuv/rotate_argb.h +++ /dev/null @@ -1,33 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_ROTATE_ARGB_H_ // NOLINT -#define INCLUDE_LIBYUV_ROTATE_ARGB_H_ - -#include "libyuv/basic_types.h" -#include "libyuv/rotate.h" // For RotationMode. - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Rotate ARGB frame -LIBYUV_API -int ARGBRotate(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int src_width, int src_height, enum RotationMode mode); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_ROTATE_ARGB_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/row.h b/src/main/jni/libyuv/include/libyuv/row.h deleted file mode 100644 index 477b27447..000000000 --- a/src/main/jni/libyuv/include/libyuv/row.h +++ /dev/null @@ -1,1821 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_ROW_H_ // NOLINT -#define INCLUDE_LIBYUV_ROW_H_ - -#include <stdlib.h> // For malloc. - -#include "libyuv/basic_types.h" - -#if defined(__native_client__) -#include "ppapi/c/pp_macros.h" // For PPAPI_RELEASE -#endif - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#define IS_ALIGNED(p, a) (!((uintptr_t)(p) & ((a) - 1))) - -#ifdef __cplusplus -#define align_buffer_64(var, size) \ - uint8* var##_mem = reinterpret_cast<uint8*>(malloc((size) + 63)); \ - uint8* var = reinterpret_cast<uint8*> \ - ((reinterpret_cast<intptr_t>(var##_mem) + 63) & ~63) -#else -#define align_buffer_64(var, size) \ - uint8* var##_mem = (uint8*)(malloc((size) + 63)); /* NOLINT */ \ - uint8* var = (uint8*)(((intptr_t)(var##_mem) + 63) & ~63) /* NOLINT */ -#endif - -#define free_aligned_buffer_64(var) \ - free(var##_mem); \ - var = 0 - -#if defined(__pnacl__) || defined(__CLR_VER) || defined(COVERAGE_ENABLED) || \ - defined(TARGET_IPHONE_SIMULATOR) || \ - (defined(_MSC_VER) && defined(__clang__)) -#define LIBYUV_DISABLE_X86 -#endif -// True if compiling for SSSE3 as a requirement. -#if defined(__SSSE3__) || (defined(_M_IX86_FP) && (_M_IX86_FP >= 3)) -#define LIBYUV_SSSE3_ONLY -#endif - -// Enable for NaCL pepper 33 for bundle and AVX2 support. -#if defined(__native_client__) && PPAPI_RELEASE >= 33 -#define NEW_BINUTILS -#endif -#if defined(__native_client__) && defined(__arm__) && PPAPI_RELEASE < 37 -#define LIBYUV_DISABLE_NEON -#endif - -// The following are available on all x86 platforms: -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) -// Effects: -#define HAS_ARGBADDROW_SSE2 -#define HAS_ARGBAFFINEROW_SSE2 -#define HAS_ARGBATTENUATEROW_SSSE3 -#define HAS_ARGBBLENDROW_SSSE3 -#define HAS_ARGBCOLORMATRIXROW_SSSE3 -#define HAS_ARGBCOLORTABLEROW_X86 -#define HAS_ARGBCOPYALPHAROW_SSE2 -#define HAS_ARGBCOPYYTOALPHAROW_SSE2 -#define HAS_ARGBGRAYROW_SSSE3 -#define HAS_ARGBLUMACOLORTABLEROW_SSSE3 -#define HAS_ARGBMIRRORROW_SSSE3 -#define HAS_ARGBMULTIPLYROW_SSE2 -#define HAS_ARGBPOLYNOMIALROW_SSE2 -#define HAS_ARGBQUANTIZEROW_SSE2 -#define HAS_ARGBSEPIAROW_SSSE3 -#define HAS_ARGBSHADEROW_SSE2 -#define HAS_ARGBSUBTRACTROW_SSE2 -#define HAS_ARGBTOUVROW_SSSE3 -#define HAS_ARGBUNATTENUATEROW_SSE2 -#define HAS_COMPUTECUMULATIVESUMROW_SSE2 -#define HAS_CUMULATIVESUMTOAVERAGEROW_SSE2 -#define HAS_INTERPOLATEROW_SSE2 -#define HAS_INTERPOLATEROW_SSSE3 -#define HAS_RGBCOLORTABLEROW_X86 -#define HAS_SOBELROW_SSE2 -#define HAS_SOBELTOPLANEROW_SSE2 -#define HAS_SOBELXROW_SSE2 -#define HAS_SOBELXYROW_SSE2 -#define HAS_SOBELYROW_SSE2 - -// Conversions: -#define HAS_ABGRTOUVROW_SSSE3 -#define HAS_ABGRTOYROW_SSSE3 -#define HAS_ARGB1555TOARGBROW_SSE2 -#define HAS_ARGB4444TOARGBROW_SSE2 -#define HAS_ARGBSHUFFLEROW_SSE2 -#define HAS_ARGBSHUFFLEROW_SSSE3 -#define HAS_ARGBTOARGB1555ROW_SSE2 -#define HAS_ARGBTOARGB4444ROW_SSE2 -#define HAS_ARGBTOBAYERGGROW_SSE2 -#define HAS_ARGBTOBAYERROW_SSSE3 -#define HAS_ARGBTORAWROW_SSSE3 -#define HAS_ARGBTORGB24ROW_SSSE3 -#define HAS_ARGBTORGB565ROW_SSE2 -#define HAS_ARGBTOUV422ROW_SSSE3 -#define HAS_ARGBTOUV444ROW_SSSE3 -#define HAS_ARGBTOUVJROW_SSSE3 -#define HAS_ARGBTOYJROW_SSSE3 -#define HAS_ARGBTOYROW_SSSE3 -#define HAS_BGRATOUVROW_SSSE3 -#define HAS_BGRATOYROW_SSSE3 -#define HAS_COPYROW_ERMS -#define HAS_COPYROW_SSE2 -#define HAS_COPYROW_X86 -#define HAS_HALFROW_SSE2 -#define HAS_I400TOARGBROW_SSE2 -#define HAS_I411TOARGBROW_SSSE3 -#define HAS_I422TOARGB1555ROW_SSSE3 -#define HAS_I422TOABGRROW_SSSE3 -#define HAS_I422TOARGB1555ROW_SSSE3 -#define HAS_I422TOARGB4444ROW_SSSE3 -#define HAS_I422TOARGBROW_SSSE3 -#define HAS_I422TOBGRAROW_SSSE3 -#define HAS_I422TORAWROW_SSSE3 -#define HAS_I422TORGB24ROW_SSSE3 -#define HAS_I422TORGB565ROW_SSSE3 -#define HAS_I422TORGBAROW_SSSE3 -#define HAS_I422TOUYVYROW_SSE2 -#define HAS_I422TOYUY2ROW_SSE2 -#define HAS_I444TOARGBROW_SSSE3 -#define HAS_MERGEUVROW_SSE2 -#define HAS_MIRRORROW_SSE2 -#define HAS_MIRRORROW_SSSE3 -#define HAS_MIRRORROW_UV_SSSE3 -#define HAS_MIRRORUVROW_SSSE3 -#define HAS_NV12TOARGBROW_SSSE3 -#define HAS_NV12TORGB565ROW_SSSE3 -#define HAS_NV21TOARGBROW_SSSE3 -#define HAS_NV21TORGB565ROW_SSSE3 -#define HAS_RAWTOARGBROW_SSSE3 -#define HAS_RAWTOYROW_SSSE3 -#define HAS_RGB24TOARGBROW_SSSE3 -#define HAS_RGB24TOYROW_SSSE3 -#define HAS_RGB565TOARGBROW_SSE2 -#define HAS_RGBATOUVROW_SSSE3 -#define HAS_RGBATOYROW_SSSE3 -#define HAS_SETROW_X86 -#define HAS_SPLITUVROW_SSE2 -#define HAS_UYVYTOARGBROW_SSSE3 -#define HAS_UYVYTOUV422ROW_SSE2 -#define HAS_UYVYTOUVROW_SSE2 -#define HAS_UYVYTOYROW_SSE2 -#define HAS_YTOARGBROW_SSE2 -#define HAS_YUY2TOARGBROW_SSSE3 -#define HAS_YUY2TOUV422ROW_SSE2 -#define HAS_YUY2TOUVROW_SSE2 -#define HAS_YUY2TOYROW_SSE2 -#endif - -// The following are available on x64 Visual C: -#if !defined(LIBYUV_DISABLE_X86) && defined (_M_X64) -#define HAS_I422TOARGBROW_SSSE3 -#endif - -// GCC >= 4.7.0 required for AVX2. -#if defined(__GNUC__) && (defined(__x86_64__) || defined(__i386__)) -#if (__GNUC__ > 4) || (__GNUC__ == 4 && (__GNUC_MINOR__ >= 7)) -#define GCC_HAS_AVX2 1 -#endif // GNUC >= 4.7 -#endif // __GNUC__ - -// clang >= 3.4.0 required for AVX2. -#if defined(__clang__) && (defined(__x86_64__) || defined(__i386__)) -#if (__clang_major__ > 3) || (__clang_major__ == 3 && (__clang_minor__ >= 4)) -#define CLANG_HAS_AVX2 1 -#endif // clang >= 3.4 -#endif // __clang__ - -// Visual C 2012 required for AVX2. -#if defined(_M_IX86) && defined(_MSC_VER) && _MSC_VER >= 1700 -#define VISUALC_HAS_AVX2 1 -#endif // VisualStudio >= 2012 - -// The following are available on all x86 platforms, but -// require VS2012, clang 3.4 or gcc 4.7. -// The code supports NaCL but requires a new compiler and validator. -#if !defined(LIBYUV_DISABLE_X86) && (defined(VISUALC_HAS_AVX2) || \ - defined(CLANG_HAS_AVX2) || defined(GCC_HAS_AVX2)) -// Effects: -#define HAS_ARGBPOLYNOMIALROW_AVX2 -#define HAS_ARGBSHUFFLEROW_AVX2 -#define HAS_ARGBCOPYALPHAROW_AVX2 -#define HAS_ARGBCOPYYTOALPHAROW_AVX2 -#endif - -// The following are require VS2012. -// TODO(fbarchard): Port to gcc. -#if !defined(LIBYUV_DISABLE_X86) && defined(VISUALC_HAS_AVX2) -#define HAS_ARGBTOUVROW_AVX2 -#define HAS_ARGBTOYJROW_AVX2 -#define HAS_ARGBTOYROW_AVX2 -#define HAS_HALFROW_AVX2 -#define HAS_I422TOARGBROW_AVX2 -#define HAS_INTERPOLATEROW_AVX2 -#define HAS_MERGEUVROW_AVX2 -#define HAS_MIRRORROW_AVX2 -#define HAS_SPLITUVROW_AVX2 -#define HAS_UYVYTOUV422ROW_AVX2 -#define HAS_UYVYTOUVROW_AVX2 -#define HAS_UYVYTOYROW_AVX2 -#define HAS_YUY2TOUV422ROW_AVX2 -#define HAS_YUY2TOUVROW_AVX2 -#define HAS_YUY2TOYROW_AVX2 - -// Effects: -#define HAS_ARGBADDROW_AVX2 -#define HAS_ARGBATTENUATEROW_AVX2 -#define HAS_ARGBMIRRORROW_AVX2 -#define HAS_ARGBMULTIPLYROW_AVX2 -#define HAS_ARGBSUBTRACTROW_AVX2 -#define HAS_ARGBUNATTENUATEROW_AVX2 -#endif // defined(VISUALC_HAS_AVX2) - -// The following are Yasm x86 only: -// TODO(fbarchard): Port AVX2 to inline. -#if !defined(LIBYUV_DISABLE_X86) && defined(HAVE_YASM) - (defined(_M_IX86) || defined(_M_X64) || \ - defined(__x86_64__) || defined(__i386__)) -#define HAS_MERGEUVROW_AVX2 -#define HAS_MERGEUVROW_MMX -#define HAS_SPLITUVROW_AVX2 -#define HAS_SPLITUVROW_MMX -#define HAS_UYVYTOYROW_AVX2 -#define HAS_UYVYTOYROW_MMX -#define HAS_YUY2TOYROW_AVX2 -#define HAS_YUY2TOYROW_MMX -#endif - -// The following are disabled when SSSE3 is available: -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) && \ - !defined(LIBYUV_SSSE3_ONLY) -#define HAS_ARGBBLENDROW_SSE2 -#define HAS_ARGBATTENUATEROW_SSE2 -#define HAS_MIRRORROW_SSE2 -#endif - -// The following are available on arm64 platforms: -#if !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__) -// #define HAS_I444TOARGBROW_NEON -// #define HAS_I422TOARGBROW_NEON -// #define HAS_I411TOARGBROW_NEON -// #define HAS_I422TOBGRAROW_NEON -// #define HAS_I422TOABGRROW_NEON -// #define HAS_I422TORGBAROW_NEON -// #define HAS_I422TORGB24ROW_NEON -// #define HAS_I422TORAWROW_NEON -// #define HAS_I422TORGB565ROW_NEON -// #define HAS_I422TOARGB1555ROW_NEON -// #define HAS_I422TOARGB4444ROW_NEON -// #define HAS_YTOARGBROW_NEON -// #define HAS_I400TOARGBROW_NEON -// #define HAS_NV12TOARGBROW_NEON -// #define HAS_NV21TOARGBROW_NEON -// #define HAS_NV12TORGB565ROW_NEON -// #define HAS_NV21TORGB565ROW_NEON -// #define HAS_YUY2TOARGBROW_NEON -// #define HAS_UYVYTOARGBROW_NEON -#define HAS_SPLITUVROW_NEON -#define HAS_MERGEUVROW_NEON -#define HAS_COPYROW_NEON -#define HAS_SETROW_NEON -#define HAS_ARGBSETROWS_NEON -#define HAS_MIRRORROW_NEON -#define HAS_MIRRORUVROW_NEON -#define HAS_ARGBMIRRORROW_NEON -#define HAS_RGB24TOARGBROW_NEON -#define HAS_RAWTOARGBROW_NEON -// #define HAS_RGB565TOARGBROW_NEON -// #define HAS_ARGB1555TOARGBROW_NEON -// #define HAS_ARGB4444TOARGBROW_NEON -#define HAS_ARGBTORGB24ROW_NEON -#define HAS_ARGBTORAWROW_NEON -#define HAS_YUY2TOYROW_NEON -#define HAS_UYVYTOYROW_NEON -#define HAS_YUY2TOUV422ROW_NEON -#define HAS_UYVYTOUV422ROW_NEON -#define HAS_YUY2TOUVROW_NEON -#define HAS_UYVYTOUVROW_NEON -#define HAS_HALFROW_NEON -#define HAS_ARGBTOBAYERROW_NEON -#define HAS_ARGBTOBAYERGGROW_NEON -#define HAS_ARGBSHUFFLEROW_NEON -#define HAS_I422TOYUY2ROW_NEON -#define HAS_I422TOUYVYROW_NEON -// #define HAS_ARGBTORGB565ROW_NEON -// #define HAS_ARGBTOARGB1555ROW_NEON -// #define HAS_ARGBTOARGB4444ROW_NEON -#define HAS_ARGBTOYROW_NEON -#define HAS_ARGBTOYJROW_NEON -#define HAS_ARGBTOUV444ROW_NEON -#define HAS_ARGBTOUV422ROW_NEON -#define HAS_ARGBTOUV411ROW_NEON -// #define HAS_ARGBTOUVROW_NEON -// #define HAS_ARGBTOUVJROW_NEON -// #define HAS_BGRATOUVROW_NEON -// #define HAS_ABGRTOUVROW_NEON -// #define HAS_RGBATOUVROW_NEON -// #define HAS_RGB24TOUVROW_NEON -// #define HAS_RAWTOUVROW_NEON -// #define HAS_RGB565TOUVROW_NEON -// #define HAS_ARGB1555TOUVROW_NEON -// #define HAS_ARGB4444TOUVROW_NEON -// #define HAS_RGB565TOYROW_NEON -// #define HAS_ARGB1555TOYROW_NEON -// #define HAS_ARGB4444TOYROW_NEON -#define HAS_BGRATOYROW_NEON -#define HAS_ABGRTOYROW_NEON -#define HAS_RGBATOYROW_NEON -#define HAS_RGB24TOYROW_NEON -#define HAS_RAWTOYROW_NEON -#define HAS_INTERPOLATEROW_NEON -#define HAS_ARGBBLENDROW_NEON -#define HAS_ARGBATTENUATEROW_NEON -#define HAS_ARGBQUANTIZEROW_NEON -#define HAS_ARGBSHADEROW_NEON -#define HAS_ARGBGRAYROW_NEON -#define HAS_ARGBSEPIAROW_NEON -#define HAS_ARGBCOLORMATRIXROW_NEON -#define HAS_ARGBMULTIPLYROW_NEON -#define HAS_ARGBADDROW_NEON -#define HAS_ARGBSUBTRACTROW_NEON -#define HAS_SOBELROW_NEON -#define HAS_SOBELTOPLANEROW_NEON -#define HAS_SOBELXYROW_NEON -#define HAS_SOBELXROW_NEON -#define HAS_SOBELYROW_NEON -#endif - -// The following are available on Neon platforms: -#if !defined(LIBYUV_DISABLE_NEON) && \ - (defined(__ARM_NEON__) || defined(LIBYUV_NEON)) -#define HAS_ABGRTOUVROW_NEON -#define HAS_ABGRTOYROW_NEON -#define HAS_ARGB1555TOARGBROW_NEON -#define HAS_ARGB1555TOUVROW_NEON -#define HAS_ARGB1555TOYROW_NEON -#define HAS_ARGB4444TOARGBROW_NEON -#define HAS_ARGB4444TOUVROW_NEON -#define HAS_ARGB4444TOYROW_NEON -#define HAS_ARGBTOARGB1555ROW_NEON -#define HAS_ARGBTOARGB4444ROW_NEON -#define HAS_ARGBTOBAYERROW_NEON -#define HAS_ARGBTOBAYERGGROW_NEON -#define HAS_ARGBTORAWROW_NEON -#define HAS_ARGBTORGB24ROW_NEON -#define HAS_ARGBTORGB565ROW_NEON -#define HAS_ARGBTOUV411ROW_NEON -#define HAS_ARGBTOUV422ROW_NEON -#define HAS_ARGBTOUV444ROW_NEON -#define HAS_ARGBTOUVROW_NEON -#define HAS_ARGBTOUVJROW_NEON -#define HAS_ARGBTOYROW_NEON -#define HAS_ARGBTOYJROW_NEON -#define HAS_BGRATOUVROW_NEON -#define HAS_BGRATOYROW_NEON -#define HAS_COPYROW_NEON -#define HAS_HALFROW_NEON -#define HAS_I400TOARGBROW_NEON -#define HAS_I411TOARGBROW_NEON -#define HAS_I422TOABGRROW_NEON -#define HAS_I422TOARGB1555ROW_NEON -#define HAS_I422TOARGB4444ROW_NEON -#define HAS_I422TOARGBROW_NEON -#define HAS_I422TOBGRAROW_NEON -#define HAS_I422TORAWROW_NEON -#define HAS_I422TORGB24ROW_NEON -#define HAS_I422TORGB565ROW_NEON -#define HAS_I422TORGBAROW_NEON -#define HAS_I422TOUYVYROW_NEON -#define HAS_I422TOYUY2ROW_NEON -#define HAS_I444TOARGBROW_NEON -#define HAS_MERGEUVROW_NEON -#define HAS_MIRRORROW_NEON -#define HAS_MIRRORUVROW_NEON -#define HAS_NV12TOARGBROW_NEON -#define HAS_NV12TORGB565ROW_NEON -#define HAS_NV21TOARGBROW_NEON -#define HAS_NV21TORGB565ROW_NEON -#define HAS_RAWTOARGBROW_NEON -#define HAS_RAWTOUVROW_NEON -#define HAS_RAWTOYROW_NEON -#define HAS_RGB24TOARGBROW_NEON -#define HAS_RGB24TOUVROW_NEON -#define HAS_RGB24TOYROW_NEON -#define HAS_RGB565TOARGBROW_NEON -#define HAS_RGB565TOUVROW_NEON -#define HAS_RGB565TOYROW_NEON -#define HAS_RGBATOUVROW_NEON -#define HAS_RGBATOYROW_NEON -#define HAS_SETROW_NEON -#define HAS_SPLITUVROW_NEON -#define HAS_UYVYTOARGBROW_NEON -#define HAS_UYVYTOUV422ROW_NEON -#define HAS_UYVYTOUVROW_NEON -#define HAS_UYVYTOYROW_NEON -#define HAS_YTOARGBROW_NEON -#define HAS_YUY2TOARGBROW_NEON -#define HAS_YUY2TOUV422ROW_NEON -#define HAS_YUY2TOUVROW_NEON -#define HAS_YUY2TOYROW_NEON - -// Effects: -#define HAS_ARGBADDROW_NEON -#define HAS_ARGBATTENUATEROW_NEON -#define HAS_ARGBBLENDROW_NEON -#define HAS_ARGBGRAYROW_NEON -#define HAS_ARGBMIRRORROW_NEON -#define HAS_ARGBMULTIPLYROW_NEON -#define HAS_ARGBQUANTIZEROW_NEON -#define HAS_ARGBSEPIAROW_NEON -#define HAS_ARGBSHADEROW_NEON -#define HAS_ARGBSUBTRACTROW_NEON -#define HAS_SOBELROW_NEON -#define HAS_SOBELTOPLANEROW_NEON -#define HAS_SOBELXYROW_NEON -#define HAS_SOBELXROW_NEON -#define HAS_SOBELYROW_NEON -#define HAS_INTERPOLATEROW_NEON -// TODO(fbarchard): Investigate neon unittest failure. -// #define HAS_ARGBCOLORMATRIXROW_NEON -#endif - -// The following are available on Mips platforms: -#if !defined(LIBYUV_DISABLE_MIPS) && defined(__mips__) && \ - (_MIPS_SIM == _MIPS_SIM_ABI32) && (__mips_isa_rev < 6) -#define HAS_COPYROW_MIPS -#if defined(__mips_dsp) && (__mips_dsp_rev >= 2) -#define HAS_I422TOABGRROW_MIPS_DSPR2 -#define HAS_I422TOARGBROW_MIPS_DSPR2 -#define HAS_I422TOBGRAROW_MIPS_DSPR2 -#define HAS_INTERPOLATEROWS_MIPS_DSPR2 -#define HAS_MIRRORROW_MIPS_DSPR2 -#define HAS_MIRRORUVROW_MIPS_DSPR2 -#define HAS_SPLITUVROW_MIPS_DSPR2 -#endif -#endif - -#if defined(_MSC_VER) && !defined(__CLR_VER) -#define SIMD_ALIGNED(var) __declspec(align(16)) var -typedef __declspec(align(16)) int16 vec16[8]; -typedef __declspec(align(16)) int32 vec32[4]; -typedef __declspec(align(16)) int8 vec8[16]; -typedef __declspec(align(16)) uint16 uvec16[8]; -typedef __declspec(align(16)) uint32 uvec32[4]; -typedef __declspec(align(16)) uint8 uvec8[16]; -typedef __declspec(align(32)) int16 lvec16[16]; -typedef __declspec(align(32)) int32 lvec32[8]; -typedef __declspec(align(32)) int8 lvec8[32]; -typedef __declspec(align(32)) uint16 ulvec16[16]; -typedef __declspec(align(32)) uint32 ulvec32[8]; -typedef __declspec(align(32)) uint8 ulvec8[32]; - -#elif defined(__GNUC__) -// Caveat GCC 4.2 to 4.7 have a known issue using vectors with const. -#define SIMD_ALIGNED(var) var __attribute__((aligned(16))) -typedef int16 __attribute__((vector_size(16))) vec16; -typedef int32 __attribute__((vector_size(16))) vec32; -typedef int8 __attribute__((vector_size(16))) vec8; -typedef uint16 __attribute__((vector_size(16))) uvec16; -typedef uint32 __attribute__((vector_size(16))) uvec32; -typedef uint8 __attribute__((vector_size(16))) uvec8; -#else -#define SIMD_ALIGNED(var) var -typedef int16 vec16[8]; -typedef int32 vec32[4]; -typedef int8 vec8[16]; -typedef uint16 uvec16[8]; -typedef uint32 uvec32[4]; -typedef uint8 uvec8[16]; -#endif - -#if defined(__APPLE__) || defined(__x86_64__) || defined(__llvm__) -#define OMITFP -#else -#define OMITFP __attribute__((optimize("omit-frame-pointer"))) -#endif - -// NaCL macros for GCC x86 and x64. - -// TODO(nfullagar): When pepper_33 toolchain is distributed, default to -// NEW_BINUTILS and remove all BUNDLEALIGN occurances. -#if defined(__native_client__) -#define LABELALIGN ".p2align 5\n" -#else -#define LABELALIGN ".p2align 2\n" -#endif -#if defined(__native_client__) && defined(__x86_64__) -#if defined(NEW_BINUTILS) -#define BUNDLELOCK ".bundle_lock\n" -#define BUNDLEUNLOCK ".bundle_unlock\n" -#define BUNDLEALIGN "\n" -#else -#define BUNDLELOCK "\n" -#define BUNDLEUNLOCK "\n" -#define BUNDLEALIGN ".p2align 5\n" -#endif -#define MEMACCESS(base) "%%nacl:(%%r15,%q" #base ")" -#define MEMACCESS2(offset, base) "%%nacl:" #offset "(%%r15,%q" #base ")" -#define MEMLEA(offset, base) #offset "(%q" #base ")" -#define MEMLEA3(offset, index, scale) \ - #offset "(,%q" #index "," #scale ")" -#define MEMLEA4(offset, base, index, scale) \ - #offset "(%q" #base ",%q" #index "," #scale ")" -#define MEMMOVESTRING(s, d) "%%nacl:(%q" #s "),%%nacl:(%q" #d "), %%r15" -#define MEMSTORESTRING(reg, d) "%%" #reg ",%%nacl:(%q" #d "), %%r15" -#define MEMOPREG(opcode, offset, base, index, scale, reg) \ - BUNDLELOCK \ - "lea " #offset "(%q" #base ",%q" #index "," #scale "),%%r14d\n" \ - #opcode " (%%r15,%%r14),%%" #reg "\n" \ - BUNDLEUNLOCK -#define MEMOPMEM(opcode, reg, offset, base, index, scale) \ - BUNDLELOCK \ - "lea " #offset "(%q" #base ",%q" #index "," #scale "),%%r14d\n" \ - #opcode " %%" #reg ",(%%r15,%%r14)\n" \ - BUNDLEUNLOCK -#define MEMOPARG(opcode, offset, base, index, scale, arg) \ - BUNDLELOCK \ - "lea " #offset "(%q" #base ",%q" #index "," #scale "),%%r14d\n" \ - #opcode " (%%r15,%%r14),%" #arg "\n" \ - BUNDLEUNLOCK -#else // defined(__native_client__) && defined(__x86_64__) -#define BUNDLEALIGN "\n" -#define MEMACCESS(base) "(%" #base ")" -#define MEMACCESS2(offset, base) #offset "(%" #base ")" -#define MEMLEA(offset, base) #offset "(%" #base ")" -#define MEMLEA3(offset, index, scale) \ - #offset "(,%" #index "," #scale ")" -#define MEMLEA4(offset, base, index, scale) \ - #offset "(%" #base ",%" #index "," #scale ")" -#define MEMMOVESTRING(s, d) -#define MEMSTORESTRING(reg, d) -#define MEMOPREG(opcode, offset, base, index, scale, reg) \ - #opcode " " #offset "(%" #base ",%" #index "," #scale "),%%" #reg "\n" -#define MEMOPMEM(opcode, reg, offset, base, index, scale) \ - #opcode " %%" #reg ","#offset "(%" #base ",%" #index "," #scale ")\n" -#define MEMOPARG(opcode, offset, base, index, scale, arg) \ - #opcode " " #offset "(%" #base ",%" #index "," #scale "),%" #arg "\n" -#endif // defined(__native_client__) && defined(__x86_64__) - -#if defined(__arm__) || defined(__aarch64__) -#undef MEMACCESS -#if defined(__native_client__) -#define MEMACCESS(base) ".p2align 3\nbic %" #base ", #0xc0000000\n" -#else -#define MEMACCESS(base) "\n" -#endif -#endif - -void I444ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I411ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToBGRARow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width); -void I422ToABGRRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width); -void I422ToRGBARow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToRGB24Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb24, - int width); -void I422ToRAWRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_raw, - int width); -void I422ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb565, - int width); -void I422ToARGB1555Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb1555, - int width); -void I422ToARGB4444Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb4444, - int width); -void NV12ToARGBRow_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToARGBRow_NEON(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void NV12ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_rgb565, - int width); -void NV21ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_vu, - uint8* dst_rgb565, - int width); -void YUY2ToARGBRow_NEON(const uint8* src_yuy2, - uint8* dst_argb, - int width); -void UYVYToARGBRow_NEON(const uint8* src_uyvy, - uint8* dst_argb, - int width); - -void ARGBToYRow_AVX2(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYRow_Any_AVX2(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_AVX2(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_Any_AVX2(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix); -void BGRAToYRow_SSSE3(const uint8* src_bgra, uint8* dst_y, int pix); -void ABGRToYRow_SSSE3(const uint8* src_abgr, uint8* dst_y, int pix); -void RGBAToYRow_SSSE3(const uint8* src_rgba, uint8* dst_y, int pix); -void RGB24ToYRow_SSSE3(const uint8* src_rgb24, uint8* dst_y, int pix); -void RAWToYRow_SSSE3(const uint8* src_raw, uint8* dst_y, int pix); -void ARGBToYRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix); -void BGRAToYRow_Unaligned_SSSE3(const uint8* src_bgra, uint8* dst_y, int pix); -void ABGRToYRow_Unaligned_SSSE3(const uint8* src_abgr, uint8* dst_y, int pix); -void RGBAToYRow_Unaligned_SSSE3(const uint8* src_rgba, uint8* dst_y, int pix); -void RGB24ToYRow_Unaligned_SSSE3(const uint8* src_rgb24, uint8* dst_y, int pix); -void RAWToYRow_Unaligned_SSSE3(const uint8* src_raw, uint8* dst_y, int pix); -void ARGBToYRow_NEON(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_NEON(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToUV444Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix); -void ARGBToUV422Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix); -void ARGBToUV411Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix); -void ARGBToUVRow_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix); -void ARGBToUVJRow_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix); -void BGRAToUVRow_NEON(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int pix); -void ABGRToUVRow_NEON(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int pix); -void RGBAToUVRow_NEON(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int pix); -void RGB24ToUVRow_NEON(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_u, uint8* dst_v, int pix); -void RAWToUVRow_NEON(const uint8* src_raw, int src_stride_raw, - uint8* dst_u, uint8* dst_v, int pix); -void RGB565ToUVRow_NEON(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int pix); -void ARGB1555ToUVRow_NEON(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int pix); -void ARGB4444ToUVRow_NEON(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int pix); -void BGRAToYRow_NEON(const uint8* src_bgra, uint8* dst_y, int pix); -void ABGRToYRow_NEON(const uint8* src_abgr, uint8* dst_y, int pix); -void RGBAToYRow_NEON(const uint8* src_rgba, uint8* dst_y, int pix); -void RGB24ToYRow_NEON(const uint8* src_rgb24, uint8* dst_y, int pix); -void RAWToYRow_NEON(const uint8* src_raw, uint8* dst_y, int pix); -void RGB565ToYRow_NEON(const uint8* src_rgb565, uint8* dst_y, int pix); -void ARGB1555ToYRow_NEON(const uint8* src_argb1555, uint8* dst_y, int pix); -void ARGB4444ToYRow_NEON(const uint8* src_argb4444, uint8* dst_y, int pix); -void ARGBToYRow_C(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_C(const uint8* src_argb, uint8* dst_y, int pix); -void BGRAToYRow_C(const uint8* src_bgra, uint8* dst_y, int pix); -void ABGRToYRow_C(const uint8* src_abgr, uint8* dst_y, int pix); -void RGBAToYRow_C(const uint8* src_rgba, uint8* dst_y, int pix); -void RGB24ToYRow_C(const uint8* src_rgb24, uint8* dst_y, int pix); -void RAWToYRow_C(const uint8* src_raw, uint8* dst_y, int pix); -void RGB565ToYRow_C(const uint8* src_rgb565, uint8* dst_y, int pix); -void ARGB1555ToYRow_C(const uint8* src_argb1555, uint8* dst_y, int pix); -void ARGB4444ToYRow_C(const uint8* src_argb4444, uint8* dst_y, int pix); -void ARGBToYRow_Any_SSSE3(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_Any_SSSE3(const uint8* src_argb, uint8* dst_y, int pix); -void BGRAToYRow_Any_SSSE3(const uint8* src_bgra, uint8* dst_y, int pix); -void ABGRToYRow_Any_SSSE3(const uint8* src_abgr, uint8* dst_y, int pix); -void RGBAToYRow_Any_SSSE3(const uint8* src_rgba, uint8* dst_y, int pix); -void RGB24ToYRow_Any_SSSE3(const uint8* src_rgb24, uint8* dst_y, int pix); -void RAWToYRow_Any_SSSE3(const uint8* src_raw, uint8* dst_y, int pix); -void ARGBToYRow_Any_NEON(const uint8* src_argb, uint8* dst_y, int pix); -void ARGBToYJRow_Any_NEON(const uint8* src_argb, uint8* dst_y, int pix); -void BGRAToYRow_Any_NEON(const uint8* src_bgra, uint8* dst_y, int pix); -void ABGRToYRow_Any_NEON(const uint8* src_abgr, uint8* dst_y, int pix); -void RGBAToYRow_Any_NEON(const uint8* src_rgba, uint8* dst_y, int pix); -void RGB24ToYRow_Any_NEON(const uint8* src_rgb24, uint8* dst_y, int pix); -void RAWToYRow_Any_NEON(const uint8* src_raw, uint8* dst_y, int pix); -void RGB565ToYRow_Any_NEON(const uint8* src_rgb565, uint8* dst_y, int pix); -void ARGB1555ToYRow_Any_NEON(const uint8* src_argb1555, uint8* dst_y, int pix); -void ARGB4444ToYRow_Any_NEON(const uint8* src_argb4444, uint8* dst_y, int pix); - -void ARGBToUVRow_AVX2(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVRow_Any_AVX2(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVRow_SSSE3(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVJRow_SSSE3(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void BGRAToUVRow_SSSE3(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width); -void ABGRToUVRow_SSSE3(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width); -void RGBAToUVRow_SSSE3(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVRow_Unaligned_SSSE3(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVJRow_Unaligned_SSSE3(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void BGRAToUVRow_Unaligned_SSSE3(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width); -void ABGRToUVRow_Unaligned_SSSE3(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width); -void RGBAToUVRow_Unaligned_SSSE3(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVRow_Any_SSSE3(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVJRow_Any_SSSE3(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void BGRAToUVRow_Any_SSSE3(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width); -void ABGRToUVRow_Any_SSSE3(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width); -void RGBAToUVRow_Any_SSSE3(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV444Row_Any_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix); -void ARGBToUV422Row_Any_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix); -void ARGBToUV411Row_Any_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix); -void ARGBToUVRow_Any_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix); -void ARGBToUVJRow_Any_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix); -void BGRAToUVRow_Any_NEON(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int pix); -void ABGRToUVRow_Any_NEON(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int pix); -void RGBAToUVRow_Any_NEON(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int pix); -void RGB24ToUVRow_Any_NEON(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_u, uint8* dst_v, int pix); -void RAWToUVRow_Any_NEON(const uint8* src_raw, int src_stride_raw, - uint8* dst_u, uint8* dst_v, int pix); -void RGB565ToUVRow_Any_NEON(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int pix); -void ARGB1555ToUVRow_Any_NEON(const uint8* src_argb1555, - int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int pix); -void ARGB4444ToUVRow_Any_NEON(const uint8* src_argb4444, - int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int pix); -void ARGBToUVRow_C(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUVJRow_C(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width); -void BGRAToUVRow_C(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width); -void ABGRToUVRow_C(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width); -void RGBAToUVRow_C(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width); -void RGB24ToUVRow_C(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_u, uint8* dst_v, int width); -void RAWToUVRow_C(const uint8* src_raw, int src_stride_raw, - uint8* dst_u, uint8* dst_v, int width); -void RGB565ToUVRow_C(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int width); -void ARGB1555ToUVRow_C(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int width); -void ARGB4444ToUVRow_C(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int width); - -void ARGBToUV444Row_SSSE3(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV444Row_Unaligned_SSSE3(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV444Row_Any_SSSE3(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); - -void ARGBToUV422Row_SSSE3(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV422Row_Unaligned_SSSE3(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV422Row_Any_SSSE3(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); - -void ARGBToUV444Row_C(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV422Row_C(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); -void ARGBToUV411Row_C(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width); - -void MirrorRow_AVX2(const uint8* src, uint8* dst, int width); -void MirrorRow_SSSE3(const uint8* src, uint8* dst, int width); -void MirrorRow_SSE2(const uint8* src, uint8* dst, int width); -void MirrorRow_NEON(const uint8* src, uint8* dst, int width); -void MirrorRow_MIPS_DSPR2(const uint8* src, uint8* dst, int width); -void MirrorRow_C(const uint8* src, uint8* dst, int width); - -void MirrorUVRow_SSSE3(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width); -void MirrorUVRow_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width); -void MirrorUVRow_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width); -void MirrorUVRow_C(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width); - -void ARGBMirrorRow_AVX2(const uint8* src, uint8* dst, int width); -void ARGBMirrorRow_SSSE3(const uint8* src, uint8* dst, int width); -void ARGBMirrorRow_NEON(const uint8* src, uint8* dst, int width); -void ARGBMirrorRow_C(const uint8* src, uint8* dst, int width); - -void SplitUVRow_C(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix); -void SplitUVRow_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix); -void SplitUVRow_AVX2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix); -void SplitUVRow_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix); -void SplitUVRow_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix); -void SplitUVRow_Unaligned_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix); -void SplitUVRow_Unaligned_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, - uint8* dst_v, int pix); -void SplitUVRow_Any_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix); -void SplitUVRow_Any_AVX2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix); -void SplitUVRow_Any_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix); -void SplitUVRow_Any_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix); - -void MergeUVRow_C(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); -void MergeUVRow_SSE2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); -void MergeUVRow_AVX2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); -void MergeUVRow_NEON(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); -void MergeUVRow_Unaligned_SSE2(const uint8* src_u, const uint8* src_v, - uint8* dst_uv, int width); -void MergeUVRow_Any_SSE2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); -void MergeUVRow_Any_AVX2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); -void MergeUVRow_Any_NEON(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width); - -void CopyRow_SSE2(const uint8* src, uint8* dst, int count); -void CopyRow_ERMS(const uint8* src, uint8* dst, int count); -void CopyRow_X86(const uint8* src, uint8* dst, int count); -void CopyRow_NEON(const uint8* src, uint8* dst, int count); -void CopyRow_MIPS(const uint8* src, uint8* dst, int count); -void CopyRow_C(const uint8* src, uint8* dst, int count); - -void CopyRow_16_C(const uint16* src, uint16* dst, int count); - -void ARGBCopyAlphaRow_C(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBCopyAlphaRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBCopyAlphaRow_AVX2(const uint8* src_argb, uint8* dst_argb, int width); - -void ARGBCopyYToAlphaRow_C(const uint8* src_y, uint8* dst_argb, int width); -void ARGBCopyYToAlphaRow_SSE2(const uint8* src_y, uint8* dst_argb, int width); -void ARGBCopyYToAlphaRow_AVX2(const uint8* src_y, uint8* dst_argb, int width); - -void SetRow_X86(uint8* dst, uint32 v32, int count); -void ARGBSetRows_X86(uint8* dst, uint32 v32, int width, - int dst_stride, int height); -void SetRow_NEON(uint8* dst, uint32 v32, int count); -void ARGBSetRows_NEON(uint8* dst, uint32 v32, int width, - int dst_stride, int height); -void SetRow_C(uint8* dst, uint32 v32, int count); -void ARGBSetRows_C(uint8* dst, uint32 v32, int width, int dst_stride, - int height); - -// ARGBShufflers for BGRAToARGB etc. -void ARGBShuffleRow_C(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_SSE2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_AVX2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_NEON(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_Any_SSE2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_Any_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_Any_AVX2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); -void ARGBShuffleRow_Any_NEON(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix); - -void RGB24ToARGBRow_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int pix); -void RAWToARGBRow_SSSE3(const uint8* src_raw, uint8* dst_argb, int pix); -void RGB565ToARGBRow_SSE2(const uint8* src_rgb565, uint8* dst_argb, int pix); -void ARGB1555ToARGBRow_SSE2(const uint8* src_argb1555, uint8* dst_argb, - int pix); -void ARGB4444ToARGBRow_SSE2(const uint8* src_argb4444, uint8* dst_argb, - int pix); - -void RGB24ToARGBRow_NEON(const uint8* src_rgb24, uint8* dst_argb, int pix); -void RAWToARGBRow_NEON(const uint8* src_raw, uint8* dst_argb, int pix); -void RGB565ToARGBRow_NEON(const uint8* src_rgb565, uint8* dst_argb, int pix); -void ARGB1555ToARGBRow_NEON(const uint8* src_argb1555, uint8* dst_argb, - int pix); -void ARGB4444ToARGBRow_NEON(const uint8* src_argb4444, uint8* dst_argb, - int pix); -void RGB24ToARGBRow_C(const uint8* src_rgb24, uint8* dst_argb, int pix); -void RAWToARGBRow_C(const uint8* src_raw, uint8* dst_argb, int pix); -void RGB565ToARGBRow_C(const uint8* src_rgb, uint8* dst_argb, int pix); -void ARGB1555ToARGBRow_C(const uint8* src_argb, uint8* dst_argb, int pix); -void ARGB4444ToARGBRow_C(const uint8* src_argb, uint8* dst_argb, int pix); -void RGB24ToARGBRow_Any_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int pix); -void RAWToARGBRow_Any_SSSE3(const uint8* src_raw, uint8* dst_argb, int pix); -void RGB565ToARGBRow_Any_SSE2(const uint8* src_rgb565, uint8* dst_argb, - int pix); -void ARGB1555ToARGBRow_Any_SSE2(const uint8* src_argb1555, uint8* dst_argb, - int pix); -void ARGB4444ToARGBRow_Any_SSE2(const uint8* src_argb4444, uint8* dst_argb, - int pix); -void RGB24ToARGBRow_Any_NEON(const uint8* src_rgb24, uint8* dst_argb, int pix); -void RAWToARGBRow_Any_NEON(const uint8* src_raw, uint8* dst_argb, int pix); -void RGB565ToARGBRow_Any_NEON(const uint8* src_rgb565, uint8* dst_argb, - int pix); -void ARGB1555ToARGBRow_Any_NEON(const uint8* src_argb1555, uint8* dst_argb, - int pix); -void ARGB4444ToARGBRow_Any_NEON(const uint8* src_argb4444, uint8* dst_argb, - int pix); - -void ARGBToRGB24Row_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRAWRow_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRGB565Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB1555Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB4444Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix); - -void ARGBToRGB24Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRAWRow_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRGB565Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB1555Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB4444Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); - -void ARGBToRGBARow_C(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRGB24Row_C(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRAWRow_C(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRGB565Row_C(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB1555Row_C(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB4444Row_C(const uint8* src_argb, uint8* dst_rgb, int pix); - -void I400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int pix); -void I400ToARGBRow_Unaligned_SSE2(const uint8* src_y, uint8* dst_argb, int pix); -void I400ToARGBRow_NEON(const uint8* src_y, uint8* dst_argb, int pix); -void I400ToARGBRow_C(const uint8* src_y, uint8* dst_argb, int pix); -void I400ToARGBRow_Any_SSE2(const uint8* src_y, uint8* dst_argb, int pix); -void I400ToARGBRow_Any_NEON(const uint8* src_y, uint8* dst_argb, int pix); - -void I444ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I411ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void NV12ToARGBRow_C(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToRGB565Row_C(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void NV12ToRGB565Row_C(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToARGBRow_C(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void YUY2ToARGBRow_C(const uint8* src_yuy2, - uint8* dst_argb, - int width); -void UYVYToARGBRow_C(const uint8* src_uyvy, - uint8* dst_argb, - int width); -void I422ToBGRARow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width); -void I422ToABGRRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width); -void I422ToRGBARow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToRGB24Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb24, - int width); -void I422ToRAWRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_raw, - int width); -void I422ToARGB4444Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb4444, - int width); -void I422ToARGB1555Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb4444, - int width); -void I422ToRGB565Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb565, - int width); -void YToARGBRow_C(const uint8* src_y, - uint8* dst_argb, - int width); -void I422ToARGBRow_AVX2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I444ToARGBRow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I411ToARGBRow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void NV12ToARGBRow_SSSE3(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToARGBRow_SSSE3(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void NV12ToRGB565Row_SSSE3(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToRGB565Row_SSSE3(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void YUY2ToARGBRow_SSSE3(const uint8* src_yuy2, - uint8* dst_argb, - int width); -void UYVYToARGBRow_SSSE3(const uint8* src_uyvy, - uint8* dst_argb, - int width); -void I422ToBGRARow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width); -void I422ToABGRRow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width); -void I422ToRGBARow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToARGB4444Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGB1555Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToRGB565Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -// RGB24/RAW are unaligned. -void I422ToRGB24Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb24, - int width); -void I422ToRAWRow_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_raw, - int width); - -void I444ToARGBRow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I411ToARGBRow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void NV12ToARGBRow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToARGBRow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void YUY2ToARGBRow_Unaligned_SSSE3(const uint8* src_yuy2, - uint8* dst_argb, - int width); -void UYVYToARGBRow_Unaligned_SSSE3(const uint8* src_uyvy, - uint8* dst_argb, - int width); -void I422ToBGRARow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width); -void I422ToABGRRow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width); -void I422ToRGBARow_Unaligned_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToARGBRow_Any_AVX2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I444ToARGBRow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I411ToARGBRow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void NV12ToARGBRow_Any_SSSE3(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToARGBRow_Any_SSSE3(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void NV12ToRGB565Row_Any_SSSE3(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToRGB565Row_Any_SSSE3(const uint8* src_y, - const uint8* src_vu, - uint8* dst_argb, - int width); -void YUY2ToARGBRow_Any_SSSE3(const uint8* src_yuy2, - uint8* dst_argb, - int width); -void UYVYToARGBRow_Any_SSSE3(const uint8* src_uyvy, - uint8* dst_argb, - int width); -void I422ToBGRARow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width); -void I422ToABGRRow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width); -void I422ToRGBARow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToARGB4444Row_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToARGB1555Row_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -void I422ToRGB565Row_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width); -// RGB24/RAW are unaligned. -void I422ToRGB24Row_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToRAWRow_Any_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void YToARGBRow_SSE2(const uint8* src_y, - uint8* dst_argb, - int width); -void YToARGBRow_NEON(const uint8* src_y, - uint8* dst_argb, - int width); -void YToARGBRow_Any_SSE2(const uint8* src_y, - uint8* dst_argb, - int width); -void YToARGBRow_Any_NEON(const uint8* src_y, - uint8* dst_argb, - int width); - -// ARGB preattenuated alpha blend. -void ARGBBlendRow_SSSE3(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBBlendRow_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBBlendRow_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBBlendRow_C(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); - -// ARGB multiply images. Same API as Blend, but these require -// pointer and width alignment for SSE2. -void ARGBMultiplyRow_C(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBMultiplyRow_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBMultiplyRow_Any_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBMultiplyRow_AVX2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBMultiplyRow_Any_AVX2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBMultiplyRow_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBMultiplyRow_Any_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); - -// ARGB add images. -void ARGBAddRow_C(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBAddRow_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBAddRow_Any_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBAddRow_AVX2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBAddRow_Any_AVX2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBAddRow_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBAddRow_Any_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); - -// ARGB subtract images. Same API as Blend, but these require -// pointer and width alignment for SSE2. -void ARGBSubtractRow_C(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBSubtractRow_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBSubtractRow_Any_SSE2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBSubtractRow_AVX2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBSubtractRow_Any_AVX2(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBSubtractRow_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); -void ARGBSubtractRow_Any_NEON(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width); - -void ARGBToRGB24Row_Any_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRAWRow_Any_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRGB565Row_Any_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB1555Row_Any_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB4444Row_Any_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix); - -void ARGBToRGB24Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRAWRow_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToRGB565Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB1555Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); -void ARGBToARGB4444Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix); - -void I444ToARGBRow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I411ToARGBRow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToBGRARow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToABGRRow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToRGBARow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToRGB24Row_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToRAWRow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGB4444Row_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGB1555Row_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToRGB565Row_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void NV12ToARGBRow_Any_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToARGBRow_Any_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV12ToRGB565Row_Any_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void NV21ToRGB565Row_Any_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width); -void YUY2ToARGBRow_Any_NEON(const uint8* src_yuy2, - uint8* dst_argb, - int width); -void UYVYToARGBRow_Any_NEON(const uint8* src_uyvy, - uint8* dst_argb, - int width); -void I422ToARGBRow_MIPS_DSPR2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToBGRARow_MIPS_DSPR2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToABGRRow_MIPS_DSPR2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToARGBRow_MIPS_DSPR2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToBGRARow_MIPS_DSPR2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); -void I422ToABGRRow_MIPS_DSPR2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width); - -void YUY2ToYRow_AVX2(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_AVX2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_AVX2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_SSE2(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_SSE2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_Unaligned_SSE2(const uint8* src_yuy2, - uint8* dst_y, int pix); -void YUY2ToUVRow_Unaligned_SSE2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_Unaligned_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_NEON(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_NEON(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_NEON(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_C(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_C(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_C(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_Any_AVX2(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_Any_AVX2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_Any_AVX2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_Any_SSE2(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_Any_SSE2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_Any_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToYRow_Any_NEON(const uint8* src_yuy2, uint8* dst_y, int pix); -void YUY2ToUVRow_Any_NEON(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void YUY2ToUV422Row_Any_NEON(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_AVX2(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_AVX2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_AVX2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_SSE2(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_Unaligned_SSE2(const uint8* src_uyvy, - uint8* dst_y, int pix); -void UYVYToUVRow_Unaligned_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_Unaligned_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_AVX2(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_AVX2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_AVX2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_NEON(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_NEON(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_NEON(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); - -void UYVYToYRow_C(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_C(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_C(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_Any_AVX2(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_Any_AVX2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_Any_AVX2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_Any_SSE2(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_Any_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_Any_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToYRow_Any_NEON(const uint8* src_uyvy, uint8* dst_y, int pix); -void UYVYToUVRow_Any_NEON(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix); -void UYVYToUV422Row_Any_NEON(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix); - -void HalfRow_C(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix); -void HalfRow_SSE2(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix); -void HalfRow_AVX2(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix); -void HalfRow_NEON(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix); - -void HalfRow_16_C(const uint16* src_uv, int src_uv_stride, - uint16* dst_uv, int pix); - -void ARGBToBayerRow_C(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix); -void ARGBToBayerRow_SSSE3(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix); -void ARGBToBayerRow_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix); -void ARGBToBayerRow_Any_SSSE3(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix); -void ARGBToBayerRow_Any_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix); -void ARGBToBayerGGRow_C(const uint8* src_argb, uint8* dst_bayer, - uint32 /* selector */, int pix); -void ARGBToBayerGGRow_SSE2(const uint8* src_argb, uint8* dst_bayer, - uint32 /* selector */, int pix); -void ARGBToBayerGGRow_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 /* selector */, int pix); -void ARGBToBayerGGRow_Any_SSE2(const uint8* src_argb, uint8* dst_bayer, - uint32 /* selector */, int pix); -void ARGBToBayerGGRow_Any_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 /* selector */, int pix); - -void I422ToYUY2Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width); -void I422ToUYVYRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width); -void I422ToYUY2Row_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width); -void I422ToUYVYRow_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width); -void I422ToYUY2Row_Any_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width); -void I422ToUYVYRow_Any_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width); -void I422ToYUY2Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width); -void I422ToUYVYRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width); -void I422ToYUY2Row_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width); -void I422ToUYVYRow_Any_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width); - -// Effects related row functions. -void ARGBAttenuateRow_C(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBAttenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBAttenuateRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBAttenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBAttenuateRow_NEON(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBAttenuateRow_Any_SSE2(const uint8* src_argb, uint8* dst_argb, - int width); -void ARGBAttenuateRow_Any_SSSE3(const uint8* src_argb, uint8* dst_argb, - int width); -void ARGBAttenuateRow_Any_AVX2(const uint8* src_argb, uint8* dst_argb, - int width); -void ARGBAttenuateRow_Any_NEON(const uint8* src_argb, uint8* dst_argb, - int width); - -// Inverse table for unattenuate, shared by C and SSE2. -extern const uint32 fixed_invtbl8[256]; -void ARGBUnattenuateRow_C(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBUnattenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBUnattenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBUnattenuateRow_Any_SSE2(const uint8* src_argb, uint8* dst_argb, - int width); -void ARGBUnattenuateRow_Any_AVX2(const uint8* src_argb, uint8* dst_argb, - int width); - -void ARGBGrayRow_C(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBGrayRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width); -void ARGBGrayRow_NEON(const uint8* src_argb, uint8* dst_argb, int width); - -void ARGBSepiaRow_C(uint8* dst_argb, int width); -void ARGBSepiaRow_SSSE3(uint8* dst_argb, int width); -void ARGBSepiaRow_NEON(uint8* dst_argb, int width); - -void ARGBColorMatrixRow_C(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width); -void ARGBColorMatrixRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width); -void ARGBColorMatrixRow_NEON(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width); - -void ARGBColorTableRow_C(uint8* dst_argb, const uint8* table_argb, int width); -void ARGBColorTableRow_X86(uint8* dst_argb, const uint8* table_argb, int width); - -void RGBColorTableRow_C(uint8* dst_argb, const uint8* table_argb, int width); -void RGBColorTableRow_X86(uint8* dst_argb, const uint8* table_argb, int width); - -void ARGBQuantizeRow_C(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width); -void ARGBQuantizeRow_SSE2(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width); -void ARGBQuantizeRow_NEON(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width); - -void ARGBShadeRow_C(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value); -void ARGBShadeRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value); -void ARGBShadeRow_NEON(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value); - -// Used for blur. -void CumulativeSumToAverageRow_SSE2(const int32* topleft, const int32* botleft, - int width, int area, uint8* dst, int count); -void ComputeCumulativeSumRow_SSE2(const uint8* row, int32* cumsum, - const int32* previous_cumsum, int width); - -void CumulativeSumToAverageRow_C(const int32* topleft, const int32* botleft, - int width, int area, uint8* dst, int count); -void ComputeCumulativeSumRow_C(const uint8* row, int32* cumsum, - const int32* previous_cumsum, int width); - -LIBYUV_API -void ARGBAffineRow_C(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* uv_dudv, int width); -LIBYUV_API -void ARGBAffineRow_SSE2(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* uv_dudv, int width); - -// Used for I420Scale, ARGBScale, and ARGBInterpolate. -void InterpolateRow_C(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, - int width, int source_y_fraction); -void InterpolateRow_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_AVX2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_NEON(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRows_MIPS_DSPR2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_Unaligned_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_Unaligned_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_Any_NEON(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_Any_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_Any_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRow_Any_AVX2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); -void InterpolateRows_Any_MIPS_DSPR2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride_ptr, int width, - int source_y_fraction); - -void InterpolateRow_16_C(uint16* dst_ptr, const uint16* src_ptr, - ptrdiff_t src_stride_ptr, - int width, int source_y_fraction); - -// Sobel images. -void SobelXRow_C(const uint8* src_y0, const uint8* src_y1, const uint8* src_y2, - uint8* dst_sobelx, int width); -void SobelXRow_SSE2(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobelx, int width); -void SobelXRow_NEON(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobelx, int width); -void SobelYRow_C(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width); -void SobelYRow_SSE2(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width); -void SobelYRow_NEON(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width); -void SobelRow_C(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width); -void SobelRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width); -void SobelRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width); -void SobelToPlaneRow_C(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width); -void SobelToPlaneRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width); -void SobelToPlaneRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width); -void SobelXYRow_C(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width); -void SobelXYRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width); -void SobelXYRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width); - -void ARGBPolynomialRow_C(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width); -void ARGBPolynomialRow_SSE2(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width); -void ARGBPolynomialRow_AVX2(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width); - -void ARGBLumaColorTableRow_C(const uint8* src_argb, uint8* dst_argb, int width, - const uint8* luma, uint32 lumacoeff); -void ARGBLumaColorTableRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - int width, - const uint8* luma, uint32 lumacoeff); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_ROW_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/scale.h b/src/main/jni/libyuv/include/libyuv/scale.h deleted file mode 100644 index a3bc07e0f..000000000 --- a/src/main/jni/libyuv/include/libyuv/scale.h +++ /dev/null @@ -1,102 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_SCALE_H_ // NOLINT -#define INCLUDE_LIBYUV_SCALE_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Supported filtering. -typedef enum FilterMode { - kFilterNone = 0, // Point sample; Fastest. - kFilterLinear = 1, // Filter horizontally only. - kFilterBilinear = 2, // Faster than box, but lower quality scaling down. - kFilterBox = 3 // Highest quality. -} FilterModeEnum; - -// Scale a YUV plane. -LIBYUV_API -void ScalePlane(const uint8* src, int src_stride, - int src_width, int src_height, - uint8* dst, int dst_stride, - int dst_width, int dst_height, - enum FilterMode filtering); - -void ScalePlane_16(const uint16* src, int src_stride, - int src_width, int src_height, - uint16* dst, int dst_stride, - int dst_width, int dst_height, - enum FilterMode filtering); - -// Scales a YUV 4:2:0 image from the src width and height to the -// dst width and height. -// If filtering is kFilterNone, a simple nearest-neighbor algorithm is -// used. This produces basic (blocky) quality at the fastest speed. -// If filtering is kFilterBilinear, interpolation is used to produce a better -// quality image, at the expense of speed. -// If filtering is kFilterBox, averaging is used to produce ever better -// quality image, at further expense of speed. -// Returns 0 if successful. - -LIBYUV_API -int I420Scale(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - int src_width, int src_height, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int dst_width, int dst_height, - enum FilterMode filtering); - -LIBYUV_API -int I420Scale_16(const uint16* src_y, int src_stride_y, - const uint16* src_u, int src_stride_u, - const uint16* src_v, int src_stride_v, - int src_width, int src_height, - uint16* dst_y, int dst_stride_y, - uint16* dst_u, int dst_stride_u, - uint16* dst_v, int dst_stride_v, - int dst_width, int dst_height, - enum FilterMode filtering); - -#ifdef __cplusplus -// Legacy API. Deprecated. -LIBYUV_API -int Scale(const uint8* src_y, const uint8* src_u, const uint8* src_v, - int src_stride_y, int src_stride_u, int src_stride_v, - int src_width, int src_height, - uint8* dst_y, uint8* dst_u, uint8* dst_v, - int dst_stride_y, int dst_stride_u, int dst_stride_v, - int dst_width, int dst_height, - LIBYUV_BOOL interpolate); - -// Legacy API. Deprecated. -LIBYUV_API -int ScaleOffset(const uint8* src_i420, int src_width, int src_height, - uint8* dst_i420, int dst_width, int dst_height, int dst_yoffset, - LIBYUV_BOOL interpolate); - -// For testing, allow disabling of specialized scalers. -LIBYUV_API -void SetUseReferenceImpl(LIBYUV_BOOL use); -#endif // __cplusplus - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_SCALE_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/scale_argb.h b/src/main/jni/libyuv/include/libyuv/scale_argb.h deleted file mode 100644 index 0c9b36257..000000000 --- a/src/main/jni/libyuv/include/libyuv/scale_argb.h +++ /dev/null @@ -1,57 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_SCALE_ARGB_H_ // NOLINT -#define INCLUDE_LIBYUV_SCALE_ARGB_H_ - -#include "libyuv/basic_types.h" -#include "libyuv/scale.h" // For FilterMode - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -LIBYUV_API -int ARGBScale(const uint8* src_argb, int src_stride_argb, - int src_width, int src_height, - uint8* dst_argb, int dst_stride_argb, - int dst_width, int dst_height, - enum FilterMode filtering); - -// Clipped scale takes destination rectangle coordinates for clip values. -LIBYUV_API -int ARGBScaleClip(const uint8* src_argb, int src_stride_argb, - int src_width, int src_height, - uint8* dst_argb, int dst_stride_argb, - int dst_width, int dst_height, - int clip_x, int clip_y, int clip_width, int clip_height, - enum FilterMode filtering); - -// TODO(fbarchard): Implement this. -// Scale with YUV conversion to ARGB and clipping. -LIBYUV_API -int YUVToARGBScaleClip(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint32 src_fourcc, - int src_width, int src_height, - uint8* dst_argb, int dst_stride_argb, - uint32 dst_fourcc, - int dst_width, int dst_height, - int clip_x, int clip_y, int clip_width, int clip_height, - enum FilterMode filtering); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_SCALE_ARGB_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/scale_row.h b/src/main/jni/libyuv/include/libyuv/scale_row.h deleted file mode 100644 index 70e6bc55b..000000000 --- a/src/main/jni/libyuv/include/libyuv/scale_row.h +++ /dev/null @@ -1,349 +0,0 @@ -/* - * Copyright 2013 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_SCALE_ROW_H_ // NOLINT -#define INCLUDE_LIBYUV_SCALE_ROW_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if defined(__pnacl__) || defined(__CLR_VER) || defined(COVERAGE_ENABLED) || \ - defined(TARGET_IPHONE_SIMULATOR) -#define LIBYUV_DISABLE_X86 -#endif - -// The following are available on all x86 platforms: -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) -#define HAS_SCALEROWDOWN2_SSE2 -#define HAS_SCALEROWDOWN4_SSE2 -#define HAS_SCALEROWDOWN34_SSSE3 -#define HAS_SCALEROWDOWN38_SSSE3 -#define HAS_SCALEADDROWS_SSE2 -#define HAS_SCALEFILTERCOLS_SSSE3 -#define HAS_SCALECOLSUP2_SSE2 -#define HAS_SCALEARGBROWDOWN2_SSE2 -#define HAS_SCALEARGBROWDOWNEVEN_SSE2 -#define HAS_SCALEARGBCOLS_SSE2 -#define HAS_SCALEARGBFILTERCOLS_SSSE3 -#define HAS_SCALEARGBCOLSUP2_SSE2 -#define HAS_FIXEDDIV_X86 -#define HAS_FIXEDDIV1_X86 -#endif - -// The following are available on Neon platforms: -#if !defined(LIBYUV_DISABLE_NEON) && !defined(__native_client__) && \ - (defined(__ARM_NEON__) || defined(LIBYUV_NEON)) -#define HAS_SCALEROWDOWN2_NEON -#define HAS_SCALEROWDOWN4_NEON -#define HAS_SCALEROWDOWN34_NEON -#define HAS_SCALEROWDOWN38_NEON -#define HAS_SCALEARGBROWDOWNEVEN_NEON -#define HAS_SCALEARGBROWDOWN2_NEON -#elif !defined(LIBYUV_DISABLE_NEON) && !defined(__native_client__) && \ - (defined(__aarch64__) || defined(LIBYUV_NEON)) -#define HAS_SCALEROWDOWN2_NEON -#define HAS_SCALEROWDOWN4_NEON -#define HAS_SCALEROWDOWN34_NEON -#define HAS_SCALEROWDOWN38_NEON -#define HAS_SCALEARGBROWDOWN2_NEON -#define HAS_SCALEARGBROWDOWNEVEN_NEON -#endif - -// The following are available on Mips platforms: -#if !defined(LIBYUV_DISABLE_MIPS) && !defined(__native_client__) && \ - defined(__mips__) && defined(__mips_dsp) && (__mips_dsp_rev >= 2) -#define HAS_SCALEROWDOWN2_MIPS_DSPR2 -#define HAS_SCALEROWDOWN4_MIPS_DSPR2 -#define HAS_SCALEROWDOWN34_MIPS_DSPR2 -#define HAS_SCALEROWDOWN38_MIPS_DSPR2 -#endif - -// Scale ARGB vertically with bilinear interpolation. -void ScalePlaneVertical(int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int y, int dy, - int bpp, enum FilterMode filtering); - -void ScalePlaneVertical_16(int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_argb, uint16* dst_argb, - int x, int y, int dy, - int wpp, enum FilterMode filtering); - -// Simplify the filtering based on scale factors. -enum FilterMode ScaleFilterReduce(int src_width, int src_height, - int dst_width, int dst_height, - enum FilterMode filtering); - -// Divide num by div and return as 16.16 fixed point result. -int FixedDiv_C(int num, int div); -int FixedDiv_X86(int num, int div); -// Divide num - 1 by div - 1 and return as 16.16 fixed point result. -int FixedDiv1_C(int num, int div); -int FixedDiv1_X86(int num, int div); -#ifdef HAS_FIXEDDIV_X86 -#define FixedDiv FixedDiv_X86 -#define FixedDiv1 FixedDiv1_X86 -#else -#define FixedDiv FixedDiv_C -#define FixedDiv1 FixedDiv1_C -#endif - -// Compute slope values for stepping. -void ScaleSlope(int src_width, int src_height, - int dst_width, int dst_height, - enum FilterMode filtering, - int* x, int* y, int* dx, int* dy); - -void ScaleRowDown2_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown2_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown2Linear_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown2Linear_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown2Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown2Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown4_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown4_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown4Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown4Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown34_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown34_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown34_0_Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width); -void ScaleRowDown34_0_Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* d, int dst_width); -void ScaleRowDown34_1_Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width); -void ScaleRowDown34_1_Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* d, int dst_width); -void ScaleCols_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx); -void ScaleCols_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx); -void ScaleColsUp2_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int, int); -void ScaleColsUp2_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int, int); -void ScaleFilterCols_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx); -void ScaleFilterCols_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx); -void ScaleFilterCols64_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx); -void ScaleFilterCols64_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx); -void ScaleRowDown38_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown38_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width); -void ScaleRowDown38_3_Box_C(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown38_3_Box_16_C(const uint16* src_ptr, - ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width); -void ScaleRowDown38_2_Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown38_2_Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width); -void ScaleAddRows_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int src_width, int src_height); -void ScaleAddRows_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint32* dst_ptr, int src_width, int src_height); -void ScaleARGBRowDown2_C(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDown2Linear_C(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDown2Box_C(const uint8* src_argb, ptrdiff_t src_stride, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDownEven_C(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDownEvenBox_C(const uint8* src_argb, - ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width); -void ScaleARGBCols_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); -void ScaleARGBCols64_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); -void ScaleARGBColsUp2_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int, int); -void ScaleARGBFilterCols_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); -void ScaleARGBFilterCols64_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); - -void ScaleRowDown2_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown2Linear_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown2Box_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown2_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown2Linear_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown2Box_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown4_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown4Box_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown34_SSSE3(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown34_1_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown34_0_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown38_SSSE3(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown38_3_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown38_2_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleAddRows_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int src_width, - int src_height); -void ScaleFilterCols_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx); -void ScaleColsUp2_SSE2(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx); -void ScaleARGBRowDown2_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDown2Linear_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDown2Box_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDownEven_SSE2(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDownEvenBox_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width); -void ScaleARGBCols_SSE2(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); -void ScaleARGBFilterCols_SSSE3(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); -void ScaleARGBColsUp2_SSE2(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx); -// Row functions. -void ScaleARGBRowDownEven_NEON(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDownEvenBox_NEON(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width); -void ScaleARGBRowDown2_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleARGBRowDown2Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); - -// ScaleRowDown2Box also used by planar functions -// NEON downscalers with interpolation. - -// Note - not static due to reuse in convert for 444 to 420. -void ScaleRowDown2_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); - -void ScaleRowDown2Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); - -void ScaleRowDown4_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown4Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - -// Down scale from 4 to 3 pixels. Use the neon multilane read/write -// to load up the every 4th pixel into a 4 different registers. -// Point samples 32 pixels to 24 pixels. -void ScaleRowDown34_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown34_0_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown34_1_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - -// 32 -> 12 -void ScaleRowDown38_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -// 32x3 -> 12x1 -void ScaleRowDown38_3_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -// 32x2 -> 12x1 -void ScaleRowDown38_2_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - -void ScaleRowDown2_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown2Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown4_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown4Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown34_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown34_0_Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width); -void ScaleRowDown34_1_Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width); -void ScaleRowDown38_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width); -void ScaleRowDown38_2_Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); -void ScaleRowDown38_3_Box_MIPS_DSPR2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_SCALE_ROW_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/version.h b/src/main/jni/libyuv/include/libyuv/version.h deleted file mode 100644 index c6952040b..000000000 --- a/src/main/jni/libyuv/include/libyuv/version.h +++ /dev/null @@ -1,16 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#ifndef INCLUDE_LIBYUV_VERSION_H_ // NOLINT -#define INCLUDE_LIBYUV_VERSION_H_ - -#define LIBYUV_VERSION 1074 - -#endif // INCLUDE_LIBYUV_VERSION_H_ NOLINT diff --git a/src/main/jni/libyuv/include/libyuv/video_common.h b/src/main/jni/libyuv/include/libyuv/video_common.h deleted file mode 100644 index 91acc2ffc..000000000 --- a/src/main/jni/libyuv/include/libyuv/video_common.h +++ /dev/null @@ -1,182 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -// Common definitions for video, including fourcc and VideoFormat. - -#ifndef INCLUDE_LIBYUV_VIDEO_COMMON_H_ // NOLINT -#define INCLUDE_LIBYUV_VIDEO_COMMON_H_ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -////////////////////////////////////////////////////////////////////////////// -// Definition of FourCC codes -////////////////////////////////////////////////////////////////////////////// - -// Convert four characters to a FourCC code. -// Needs to be a macro otherwise the OS X compiler complains when the kFormat* -// constants are used in a switch. -#ifdef __cplusplus -#define FOURCC(a, b, c, d) ( \ - (static_cast<uint32>(a)) | (static_cast<uint32>(b) << 8) | \ - (static_cast<uint32>(c) << 16) | (static_cast<uint32>(d) << 24)) -#else -#define FOURCC(a, b, c, d) ( \ - ((uint32)(a)) | ((uint32)(b) << 8) | /* NOLINT */ \ - ((uint32)(c) << 16) | ((uint32)(d) << 24)) /* NOLINT */ -#endif - -// Some pages discussing FourCC codes: -// http://www.fourcc.org/yuv.php -// http://v4l2spec.bytesex.org/spec/book1.htm -// http://developer.apple.com/quicktime/icefloe/dispatch020.html -// http://msdn.microsoft.com/library/windows/desktop/dd206750.aspx#nv12 -// http://people.xiph.org/~xiphmont/containers/nut/nut4cc.txt - -// FourCC codes grouped according to implementation efficiency. -// Primary formats should convert in 1 efficient step. -// Secondary formats are converted in 2 steps. -// Auxilliary formats call primary converters. -enum FourCC { - // 9 Primary YUV formats: 5 planar, 2 biplanar, 2 packed. - FOURCC_I420 = FOURCC('I', '4', '2', '0'), - FOURCC_I422 = FOURCC('I', '4', '2', '2'), - FOURCC_I444 = FOURCC('I', '4', '4', '4'), - FOURCC_I411 = FOURCC('I', '4', '1', '1'), - FOURCC_I400 = FOURCC('I', '4', '0', '0'), - FOURCC_NV21 = FOURCC('N', 'V', '2', '1'), - FOURCC_NV12 = FOURCC('N', 'V', '1', '2'), - FOURCC_YUY2 = FOURCC('Y', 'U', 'Y', '2'), - FOURCC_UYVY = FOURCC('U', 'Y', 'V', 'Y'), - - // 2 Secondary YUV formats: row biplanar. - FOURCC_M420 = FOURCC('M', '4', '2', '0'), - FOURCC_Q420 = FOURCC('Q', '4', '2', '0'), - - // 9 Primary RGB formats: 4 32 bpp, 2 24 bpp, 3 16 bpp. - FOURCC_ARGB = FOURCC('A', 'R', 'G', 'B'), - FOURCC_BGRA = FOURCC('B', 'G', 'R', 'A'), - FOURCC_ABGR = FOURCC('A', 'B', 'G', 'R'), - FOURCC_24BG = FOURCC('2', '4', 'B', 'G'), - FOURCC_RAW = FOURCC('r', 'a', 'w', ' '), - FOURCC_RGBA = FOURCC('R', 'G', 'B', 'A'), - FOURCC_RGBP = FOURCC('R', 'G', 'B', 'P'), // rgb565 LE. - FOURCC_RGBO = FOURCC('R', 'G', 'B', 'O'), // argb1555 LE. - FOURCC_R444 = FOURCC('R', '4', '4', '4'), // argb4444 LE. - - // 4 Secondary RGB formats: 4 Bayer Patterns. - FOURCC_RGGB = FOURCC('R', 'G', 'G', 'B'), - FOURCC_BGGR = FOURCC('B', 'G', 'G', 'R'), - FOURCC_GRBG = FOURCC('G', 'R', 'B', 'G'), - FOURCC_GBRG = FOURCC('G', 'B', 'R', 'G'), - - // 1 Primary Compressed YUV format. - FOURCC_MJPG = FOURCC('M', 'J', 'P', 'G'), - - // 5 Auxiliary YUV variations: 3 with U and V planes are swapped, 1 Alias. - FOURCC_YV12 = FOURCC('Y', 'V', '1', '2'), - FOURCC_YV16 = FOURCC('Y', 'V', '1', '6'), - FOURCC_YV24 = FOURCC('Y', 'V', '2', '4'), - FOURCC_YU12 = FOURCC('Y', 'U', '1', '2'), // Linux version of I420. - FOURCC_J420 = FOURCC('J', '4', '2', '0'), - FOURCC_J400 = FOURCC('J', '4', '0', '0'), - - // 14 Auxiliary aliases. CanonicalFourCC() maps these to canonical fourcc. - FOURCC_IYUV = FOURCC('I', 'Y', 'U', 'V'), // Alias for I420. - FOURCC_YU16 = FOURCC('Y', 'U', '1', '6'), // Alias for I422. - FOURCC_YU24 = FOURCC('Y', 'U', '2', '4'), // Alias for I444. - FOURCC_YUYV = FOURCC('Y', 'U', 'Y', 'V'), // Alias for YUY2. - FOURCC_YUVS = FOURCC('y', 'u', 'v', 's'), // Alias for YUY2 on Mac. - FOURCC_HDYC = FOURCC('H', 'D', 'Y', 'C'), // Alias for UYVY. - FOURCC_2VUY = FOURCC('2', 'v', 'u', 'y'), // Alias for UYVY on Mac. - FOURCC_JPEG = FOURCC('J', 'P', 'E', 'G'), // Alias for MJPG. - FOURCC_DMB1 = FOURCC('d', 'm', 'b', '1'), // Alias for MJPG on Mac. - FOURCC_BA81 = FOURCC('B', 'A', '8', '1'), // Alias for BGGR. - FOURCC_RGB3 = FOURCC('R', 'G', 'B', '3'), // Alias for RAW. - FOURCC_BGR3 = FOURCC('B', 'G', 'R', '3'), // Alias for 24BG. - FOURCC_CM32 = FOURCC(0, 0, 0, 32), // Alias for BGRA kCMPixelFormat_32ARGB - FOURCC_CM24 = FOURCC(0, 0, 0, 24), // Alias for RAW kCMPixelFormat_24RGB - FOURCC_L555 = FOURCC('L', '5', '5', '5'), // Alias for RGBO. - FOURCC_L565 = FOURCC('L', '5', '6', '5'), // Alias for RGBP. - FOURCC_5551 = FOURCC('5', '5', '5', '1'), // Alias for RGBO. - - // 1 Auxiliary compressed YUV format set aside for capturer. - FOURCC_H264 = FOURCC('H', '2', '6', '4'), - - // Match any fourcc. - FOURCC_ANY = -1, -}; - -enum FourCCBpp { - // Canonical fourcc codes used in our code. - FOURCC_BPP_I420 = 12, - FOURCC_BPP_I422 = 16, - FOURCC_BPP_I444 = 24, - FOURCC_BPP_I411 = 12, - FOURCC_BPP_I400 = 8, - FOURCC_BPP_NV21 = 12, - FOURCC_BPP_NV12 = 12, - FOURCC_BPP_YUY2 = 16, - FOURCC_BPP_UYVY = 16, - FOURCC_BPP_M420 = 12, - FOURCC_BPP_Q420 = 12, - FOURCC_BPP_ARGB = 32, - FOURCC_BPP_BGRA = 32, - FOURCC_BPP_ABGR = 32, - FOURCC_BPP_RGBA = 32, - FOURCC_BPP_24BG = 24, - FOURCC_BPP_RAW = 24, - FOURCC_BPP_RGBP = 16, - FOURCC_BPP_RGBO = 16, - FOURCC_BPP_R444 = 16, - FOURCC_BPP_RGGB = 8, - FOURCC_BPP_BGGR = 8, - FOURCC_BPP_GRBG = 8, - FOURCC_BPP_GBRG = 8, - FOURCC_BPP_YV12 = 12, - FOURCC_BPP_YV16 = 16, - FOURCC_BPP_YV24 = 24, - FOURCC_BPP_YU12 = 12, - FOURCC_BPP_J420 = 12, - FOURCC_BPP_J400 = 8, - FOURCC_BPP_MJPG = 0, // 0 means unknown. - FOURCC_BPP_H264 = 0, - FOURCC_BPP_IYUV = 12, - FOURCC_BPP_YU16 = 16, - FOURCC_BPP_YU24 = 24, - FOURCC_BPP_YUYV = 16, - FOURCC_BPP_YUVS = 16, - FOURCC_BPP_HDYC = 16, - FOURCC_BPP_2VUY = 16, - FOURCC_BPP_JPEG = 1, - FOURCC_BPP_DMB1 = 1, - FOURCC_BPP_BA81 = 8, - FOURCC_BPP_RGB3 = 24, - FOURCC_BPP_BGR3 = 24, - FOURCC_BPP_CM32 = 32, - FOURCC_BPP_CM24 = 24, - - // Match any fourcc. - FOURCC_BPP_ANY = 0, // 0 means unknown. -}; - -// Converts fourcc aliases into canonical ones. -LIBYUV_API uint32 CanonicalFourCC(uint32 fourcc); - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - -#endif // INCLUDE_LIBYUV_VIDEO_COMMON_H_ NOLINT diff --git a/src/main/jni/libyuv/source/compare.cc b/src/main/jni/libyuv/source/compare.cc deleted file mode 100644 index dc715e019..000000000 --- a/src/main/jni/libyuv/source/compare.cc +++ /dev/null @@ -1,325 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/compare.h" - -#include <float.h> -#include <math.h> -#ifdef _OPENMP -#include <omp.h> -#endif - -#include "libyuv/basic_types.h" -#include "libyuv/cpu_id.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// hash seed of 5381 recommended. -// Internal C version of HashDjb2 with int sized count for efficiency. -uint32 HashDjb2_C(const uint8* src, int count, uint32 seed); - -// This module is for Visual C x86 -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || \ - (defined(__x86_64__) || (defined(__i386__) && !defined(__pic__)))) -#define HAS_HASHDJB2_SSE41 -uint32 HashDjb2_SSE41(const uint8* src, int count, uint32 seed); - -#if _MSC_VER >= 1700 -#define HAS_HASHDJB2_AVX2 -uint32 HashDjb2_AVX2(const uint8* src, int count, uint32 seed); -#endif - -#endif // HAS_HASHDJB2_SSE41 - -// hash seed of 5381 recommended. -LIBYUV_API -uint32 HashDjb2(const uint8* src, uint64 count, uint32 seed) { - const int kBlockSize = 1 << 15; // 32768; - int remainder; - uint32 (*HashDjb2_SSE)(const uint8* src, int count, uint32 seed) = HashDjb2_C; -#if defined(HAS_HASHDJB2_SSE41) - if (TestCpuFlag(kCpuHasSSE41)) { - HashDjb2_SSE = HashDjb2_SSE41; - } -#endif -#if defined(HAS_HASHDJB2_AVX2) - if (TestCpuFlag(kCpuHasAVX2)) { - HashDjb2_SSE = HashDjb2_AVX2; - } -#endif - - while (count >= (uint64)(kBlockSize)) { - seed = HashDjb2_SSE(src, kBlockSize, seed); - src += kBlockSize; - count -= kBlockSize; - } - remainder = (int)(count) & ~15; - if (remainder) { - seed = HashDjb2_SSE(src, remainder, seed); - src += remainder; - count -= remainder; - } - remainder = (int)(count) & 15; - if (remainder) { - seed = HashDjb2_C(src, remainder, seed); - } - return seed; -} - -uint32 SumSquareError_C(const uint8* src_a, const uint8* src_b, int count); -#if !defined(LIBYUV_DISABLE_NEON) && \ - (defined(__ARM_NEON__) || defined(LIBYUV_NEON) || defined(__aarch64__)) -#define HAS_SUMSQUAREERROR_NEON -uint32 SumSquareError_NEON(const uint8* src_a, const uint8* src_b, int count); -#endif -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) -#define HAS_SUMSQUAREERROR_SSE2 -uint32 SumSquareError_SSE2(const uint8* src_a, const uint8* src_b, int count); -#endif -// Visual C 2012 required for AVX2. -#if !defined(LIBYUV_DISABLE_X86) && defined(_M_IX86) && _MSC_VER >= 1700 -#define HAS_SUMSQUAREERROR_AVX2 -uint32 SumSquareError_AVX2(const uint8* src_a, const uint8* src_b, int count); -#endif - -// TODO(fbarchard): Refactor into row function. -LIBYUV_API -uint64 ComputeSumSquareError(const uint8* src_a, const uint8* src_b, - int count) { - // SumSquareError returns values 0 to 65535 for each squared difference. - // Up to 65536 of those can be summed and remain within a uint32. - // After each block of 65536 pixels, accumulate into a uint64. - const int kBlockSize = 65536; - int remainder = count & (kBlockSize - 1) & ~31; - uint64 sse = 0; - int i; - uint32 (*SumSquareError)(const uint8* src_a, const uint8* src_b, int count) = - SumSquareError_C; -#if defined(HAS_SUMSQUAREERROR_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - SumSquareError = SumSquareError_NEON; - } -#endif -#if defined(HAS_SUMSQUAREERROR_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && - IS_ALIGNED(src_a, 16) && IS_ALIGNED(src_b, 16)) { - // Note only used for multiples of 16 so count is not checked. - SumSquareError = SumSquareError_SSE2; - } -#endif -#if defined(HAS_SUMSQUAREERROR_AVX2) - if (TestCpuFlag(kCpuHasAVX2)) { - // Note only used for multiples of 32 so count is not checked. - SumSquareError = SumSquareError_AVX2; - } -#endif -#ifdef _OPENMP -#pragma omp parallel for reduction(+: sse) -#endif - for (i = 0; i < (count - (kBlockSize - 1)); i += kBlockSize) { - sse += SumSquareError(src_a + i, src_b + i, kBlockSize); - } - src_a += count & ~(kBlockSize - 1); - src_b += count & ~(kBlockSize - 1); - if (remainder) { - sse += SumSquareError(src_a, src_b, remainder); - src_a += remainder; - src_b += remainder; - } - remainder = count & 31; - if (remainder) { - sse += SumSquareError_C(src_a, src_b, remainder); - } - return sse; -} - -LIBYUV_API -uint64 ComputeSumSquareErrorPlane(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b, - int width, int height) { - uint64 sse = 0; - int h; - // Coalesce rows. - if (stride_a == width && - stride_b == width) { - width *= height; - height = 1; - stride_a = stride_b = 0; - } - for (h = 0; h < height; ++h) { - sse += ComputeSumSquareError(src_a, src_b, width); - src_a += stride_a; - src_b += stride_b; - } - return sse; -} - -LIBYUV_API -double SumSquareErrorToPsnr(uint64 sse, uint64 count) { - double psnr; - if (sse > 0) { - double mse = (double)(count) / (double)(sse); - psnr = 10.0 * log10(255.0 * 255.0 * mse); - } else { - psnr = kMaxPsnr; // Limit to prevent divide by 0 - } - - if (psnr > kMaxPsnr) - psnr = kMaxPsnr; - - return psnr; -} - -LIBYUV_API -double CalcFramePsnr(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b, - int width, int height) { - const uint64 samples = width * height; - const uint64 sse = ComputeSumSquareErrorPlane(src_a, stride_a, - src_b, stride_b, - width, height); - return SumSquareErrorToPsnr(sse, samples); -} - -LIBYUV_API -double I420Psnr(const uint8* src_y_a, int stride_y_a, - const uint8* src_u_a, int stride_u_a, - const uint8* src_v_a, int stride_v_a, - const uint8* src_y_b, int stride_y_b, - const uint8* src_u_b, int stride_u_b, - const uint8* src_v_b, int stride_v_b, - int width, int height) { - const uint64 sse_y = ComputeSumSquareErrorPlane(src_y_a, stride_y_a, - src_y_b, stride_y_b, - width, height); - const int width_uv = (width + 1) >> 1; - const int height_uv = (height + 1) >> 1; - const uint64 sse_u = ComputeSumSquareErrorPlane(src_u_a, stride_u_a, - src_u_b, stride_u_b, - width_uv, height_uv); - const uint64 sse_v = ComputeSumSquareErrorPlane(src_v_a, stride_v_a, - src_v_b, stride_v_b, - width_uv, height_uv); - const uint64 samples = width * height + 2 * (width_uv * height_uv); - const uint64 sse = sse_y + sse_u + sse_v; - return SumSquareErrorToPsnr(sse, samples); -} - -static const int64 cc1 = 26634; // (64^2*(.01*255)^2 -static const int64 cc2 = 239708; // (64^2*(.03*255)^2 - -static double Ssim8x8_C(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b) { - int64 sum_a = 0; - int64 sum_b = 0; - int64 sum_sq_a = 0; - int64 sum_sq_b = 0; - int64 sum_axb = 0; - - int i; - for (i = 0; i < 8; ++i) { - int j; - for (j = 0; j < 8; ++j) { - sum_a += src_a[j]; - sum_b += src_b[j]; - sum_sq_a += src_a[j] * src_a[j]; - sum_sq_b += src_b[j] * src_b[j]; - sum_axb += src_a[j] * src_b[j]; - } - - src_a += stride_a; - src_b += stride_b; - } - - { - const int64 count = 64; - // scale the constants by number of pixels - const int64 c1 = (cc1 * count * count) >> 12; - const int64 c2 = (cc2 * count * count) >> 12; - - const int64 sum_a_x_sum_b = sum_a * sum_b; - - const int64 ssim_n = (2 * sum_a_x_sum_b + c1) * - (2 * count * sum_axb - 2 * sum_a_x_sum_b + c2); - - const int64 sum_a_sq = sum_a*sum_a; - const int64 sum_b_sq = sum_b*sum_b; - - const int64 ssim_d = (sum_a_sq + sum_b_sq + c1) * - (count * sum_sq_a - sum_a_sq + - count * sum_sq_b - sum_b_sq + c2); - - if (ssim_d == 0.0) { - return DBL_MAX; - } - return ssim_n * 1.0 / ssim_d; - } -} - -// We are using a 8x8 moving window with starting location of each 8x8 window -// on the 4x4 pixel grid. Such arrangement allows the windows to overlap -// block boundaries to penalize blocking artifacts. -LIBYUV_API -double CalcFrameSsim(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b, - int width, int height) { - int samples = 0; - double ssim_total = 0; - double (*Ssim8x8)(const uint8* src_a, int stride_a, - const uint8* src_b, int stride_b) = Ssim8x8_C; - - // sample point start with each 4x4 location - int i; - for (i = 0; i < height - 8; i += 4) { - int j; - for (j = 0; j < width - 8; j += 4) { - ssim_total += Ssim8x8(src_a + j, stride_a, src_b + j, stride_b); - samples++; - } - - src_a += stride_a * 4; - src_b += stride_b * 4; - } - - ssim_total /= samples; - return ssim_total; -} - -LIBYUV_API -double I420Ssim(const uint8* src_y_a, int stride_y_a, - const uint8* src_u_a, int stride_u_a, - const uint8* src_v_a, int stride_v_a, - const uint8* src_y_b, int stride_y_b, - const uint8* src_u_b, int stride_u_b, - const uint8* src_v_b, int stride_v_b, - int width, int height) { - const double ssim_y = CalcFrameSsim(src_y_a, stride_y_a, - src_y_b, stride_y_b, width, height); - const int width_uv = (width + 1) >> 1; - const int height_uv = (height + 1) >> 1; - const double ssim_u = CalcFrameSsim(src_u_a, stride_u_a, - src_u_b, stride_u_b, - width_uv, height_uv); - const double ssim_v = CalcFrameSsim(src_v_a, stride_v_a, - src_v_b, stride_v_b, - width_uv, height_uv); - return ssim_y * 0.8 + 0.1 * (ssim_u + ssim_v); -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/compare_common.cc b/src/main/jni/libyuv/source/compare_common.cc deleted file mode 100644 index c546b5182..000000000 --- a/src/main/jni/libyuv/source/compare_common.cc +++ /dev/null @@ -1,42 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -uint32 SumSquareError_C(const uint8* src_a, const uint8* src_b, int count) { - uint32 sse = 0u; - int i; - for (i = 0; i < count; ++i) { - int diff = src_a[i] - src_b[i]; - sse += (uint32)(diff * diff); - } - return sse; -} - -// hash seed of 5381 recommended. -// Internal C version of HashDjb2 with int sized count for efficiency. -uint32 HashDjb2_C(const uint8* src, int count, uint32 seed) { - uint32 hash = seed; - int i; - for (i = 0; i < count; ++i) { - hash += (hash << 5) + src[i]; - } - return hash; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/compare_neon.cc b/src/main/jni/libyuv/source/compare_neon.cc deleted file mode 100644 index 55052c0ee..000000000 --- a/src/main/jni/libyuv/source/compare_neon.cc +++ /dev/null @@ -1,103 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/basic_types.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_NEON) && defined(__ARM_NEON__) - -uint32 SumSquareError_NEON(const uint8* src_a, const uint8* src_b, int count) { - volatile uint32 sse; - asm volatile ( - "vmov.u8 q8, #0 \n" - "vmov.u8 q10, #0 \n" - "vmov.u8 q9, #0 \n" - "vmov.u8 q11, #0 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" - MEMACCESS(1) - "vld1.8 {q1}, [%1]! \n" - "subs %2, %2, #16 \n" - "vsubl.u8 q2, d0, d2 \n" - "vsubl.u8 q3, d1, d3 \n" - "vmlal.s16 q8, d4, d4 \n" - "vmlal.s16 q9, d6, d6 \n" - "vmlal.s16 q10, d5, d5 \n" - "vmlal.s16 q11, d7, d7 \n" - "bgt 1b \n" - - "vadd.u32 q8, q8, q9 \n" - "vadd.u32 q10, q10, q11 \n" - "vadd.u32 q11, q8, q10 \n" - "vpaddl.u32 q1, q11 \n" - "vadd.u64 d0, d2, d3 \n" - "vmov.32 %3, d0[0] \n" - : "+r"(src_a), - "+r"(src_b), - "+r"(count), - "=r"(sse) - : - : "memory", "cc", "q0", "q1", "q2", "q3", "q8", "q9", "q10", "q11"); - return sse; -} - -#elif !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__) - -uint32 SumSquareError_NEON(const uint8* src_a, const uint8* src_b, int count) { - volatile uint32 sse; - asm volatile ( - "eor v16.16b, v16.16b, v16.16b \n" - "eor v18.16b, v18.16b, v18.16b \n" - "eor v17.16b, v17.16b, v17.16b \n" - "eor v19.16b, v19.16b, v19.16b \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], #16 \n" - MEMACCESS(1) - "ld1 {v1.16b}, [%1], #16 \n" - "subs %2, %2, #16 \n" - "usubl v2.8h, v0.8b, v1.8b \n" - "usubl2 v3.8h, v0.16b, v1.16b \n" - "smlal v16.4s, v2.4h, v2.4h \n" - "smlal v17.4s, v3.4h, v3.4h \n" - "smlal2 v18.4s, v2.8h, v2.8h \n" - "smlal2 v19.4s, v3.8h, v3.8h \n" - "bgt 1b \n" - - "add v16.4s, v16.4s, v17.4s \n" - "add v18.4s, v18.4s, v19.4s \n" - "add v19.4s, v16.4s, v18.4s \n" - "addv s0, v19.4s \n" - "fmov %w3, s0 \n" - : "+r"(src_a), - "+r"(src_b), - "+r"(count), - "=r"(sse) - : - : "cc", "v0", "v1", "v2", "v3", "v16", "v17", "v18", "v19"); - return sse; -} - -#endif // __ARM_NEON__ - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/compare_posix.cc b/src/main/jni/libyuv/source/compare_posix.cc deleted file mode 100644 index ac361190e..000000000 --- a/src/main/jni/libyuv/source/compare_posix.cc +++ /dev/null @@ -1,158 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/basic_types.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_X86) && (defined(__x86_64__) || defined(__i386__)) - -uint32 SumSquareError_SSE2(const uint8* src_a, const uint8* src_b, int count) { - uint32 sse; - asm volatile ( // NOLINT - "pxor %%xmm0,%%xmm0 \n" - "pxor %%xmm5,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm1 \n" - "lea " MEMLEA(0x10, 0) ",%0 \n" - "movdqa " MEMACCESS(1) ",%%xmm2 \n" - "lea " MEMLEA(0x10, 1) ",%1 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm1,%%xmm3 \n" - "psubusb %%xmm2,%%xmm1 \n" - "psubusb %%xmm3,%%xmm2 \n" - "por %%xmm2,%%xmm1 \n" - "movdqa %%xmm1,%%xmm2 \n" - "punpcklbw %%xmm5,%%xmm1 \n" - "punpckhbw %%xmm5,%%xmm2 \n" - "pmaddwd %%xmm1,%%xmm1 \n" - "pmaddwd %%xmm2,%%xmm2 \n" - "paddd %%xmm1,%%xmm0 \n" - "paddd %%xmm2,%%xmm0 \n" - "jg 1b \n" - - "pshufd $0xee,%%xmm0,%%xmm1 \n" - "paddd %%xmm1,%%xmm0 \n" - "pshufd $0x1,%%xmm0,%%xmm1 \n" - "paddd %%xmm1,%%xmm0 \n" - "movd %%xmm0,%3 \n" - - : "+r"(src_a), // %0 - "+r"(src_b), // %1 - "+r"(count), // %2 - "=g"(sse) // %3 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); // NOLINT - return sse; -} - -#endif // defined(__x86_64__) || defined(__i386__) - -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(__x86_64__) || (defined(__i386__) && !defined(__pic__))) -#define HAS_HASHDJB2_SSE41 -static uvec32 kHash16x33 = { 0x92d9e201, 0, 0, 0 }; // 33 ^ 16 -static uvec32 kHashMul0 = { - 0x0c3525e1, // 33 ^ 15 - 0xa3476dc1, // 33 ^ 14 - 0x3b4039a1, // 33 ^ 13 - 0x4f5f0981, // 33 ^ 12 -}; -static uvec32 kHashMul1 = { - 0x30f35d61, // 33 ^ 11 - 0x855cb541, // 33 ^ 10 - 0x040a9121, // 33 ^ 9 - 0x747c7101, // 33 ^ 8 -}; -static uvec32 kHashMul2 = { - 0xec41d4e1, // 33 ^ 7 - 0x4cfa3cc1, // 33 ^ 6 - 0x025528a1, // 33 ^ 5 - 0x00121881, // 33 ^ 4 -}; -static uvec32 kHashMul3 = { - 0x00008c61, // 33 ^ 3 - 0x00000441, // 33 ^ 2 - 0x00000021, // 33 ^ 1 - 0x00000001, // 33 ^ 0 -}; - -uint32 HashDjb2_SSE41(const uint8* src, int count, uint32 seed) { - uint32 hash; - asm volatile ( // NOLINT - "movd %2,%%xmm0 \n" - "pxor %%xmm7,%%xmm7 \n" - "movdqa %4,%%xmm6 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm1 \n" - "lea " MEMLEA(0x10, 0) ",%0 \n" - "pmulld %%xmm6,%%xmm0 \n" - "movdqa %5,%%xmm5 \n" - "movdqa %%xmm1,%%xmm2 \n" - "punpcklbw %%xmm7,%%xmm2 \n" - "movdqa %%xmm2,%%xmm3 \n" - "punpcklwd %%xmm7,%%xmm3 \n" - "pmulld %%xmm5,%%xmm3 \n" - "movdqa %6,%%xmm5 \n" - "movdqa %%xmm2,%%xmm4 \n" - "punpckhwd %%xmm7,%%xmm4 \n" - "pmulld %%xmm5,%%xmm4 \n" - "movdqa %7,%%xmm5 \n" - "punpckhbw %%xmm7,%%xmm1 \n" - "movdqa %%xmm1,%%xmm2 \n" - "punpcklwd %%xmm7,%%xmm2 \n" - "pmulld %%xmm5,%%xmm2 \n" - "movdqa %8,%%xmm5 \n" - "punpckhwd %%xmm7,%%xmm1 \n" - "pmulld %%xmm5,%%xmm1 \n" - "paddd %%xmm4,%%xmm3 \n" - "paddd %%xmm2,%%xmm1 \n" - "sub $0x10,%1 \n" - "paddd %%xmm3,%%xmm1 \n" - "pshufd $0xe,%%xmm1,%%xmm2 \n" - "paddd %%xmm2,%%xmm1 \n" - "pshufd $0x1,%%xmm1,%%xmm2 \n" - "paddd %%xmm2,%%xmm1 \n" - "paddd %%xmm1,%%xmm0 \n" - "jg 1b \n" - "movd %%xmm0,%3 \n" - : "+r"(src), // %0 - "+r"(count), // %1 - "+rm"(seed), // %2 - "=g"(hash) // %3 - : "m"(kHash16x33), // %4 - "m"(kHashMul0), // %5 - "m"(kHashMul1), // %6 - "m"(kHashMul2), // %7 - "m"(kHashMul3) // %8 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); // NOLINT - return hash; -} -#endif // defined(__x86_64__) || (defined(__i386__) && !defined(__pic__))) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - diff --git a/src/main/jni/libyuv/source/compare_win.cc b/src/main/jni/libyuv/source/compare_win.cc deleted file mode 100644 index 99831651f..000000000 --- a/src/main/jni/libyuv/source/compare_win.cc +++ /dev/null @@ -1,232 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/basic_types.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_X86) && defined(_M_IX86) && defined(_MSC_VER) - -__declspec(naked) __declspec(align(16)) -uint32 SumSquareError_SSE2(const uint8* src_a, const uint8* src_b, int count) { - __asm { - mov eax, [esp + 4] // src_a - mov edx, [esp + 8] // src_b - mov ecx, [esp + 12] // count - pxor xmm0, xmm0 - pxor xmm5, xmm5 - - align 4 - wloop: - movdqa xmm1, [eax] - lea eax, [eax + 16] - movdqa xmm2, [edx] - lea edx, [edx + 16] - sub ecx, 16 - movdqa xmm3, xmm1 // abs trick - psubusb xmm1, xmm2 - psubusb xmm2, xmm3 - por xmm1, xmm2 - movdqa xmm2, xmm1 - punpcklbw xmm1, xmm5 - punpckhbw xmm2, xmm5 - pmaddwd xmm1, xmm1 - pmaddwd xmm2, xmm2 - paddd xmm0, xmm1 - paddd xmm0, xmm2 - jg wloop - - pshufd xmm1, xmm0, 0xee - paddd xmm0, xmm1 - pshufd xmm1, xmm0, 0x01 - paddd xmm0, xmm1 - movd eax, xmm0 - ret - } -} - -// Visual C 2012 required for AVX2. -#if _MSC_VER >= 1700 -// C4752: found Intel(R) Advanced Vector Extensions; consider using /arch:AVX. -#pragma warning(disable: 4752) -__declspec(naked) __declspec(align(16)) -uint32 SumSquareError_AVX2(const uint8* src_a, const uint8* src_b, int count) { - __asm { - mov eax, [esp + 4] // src_a - mov edx, [esp + 8] // src_b - mov ecx, [esp + 12] // count - vpxor ymm0, ymm0, ymm0 // sum - vpxor ymm5, ymm5, ymm5 // constant 0 for unpck - sub edx, eax - - align 4 - wloop: - vmovdqu ymm1, [eax] - vmovdqu ymm2, [eax + edx] - lea eax, [eax + 32] - sub ecx, 32 - vpsubusb ymm3, ymm1, ymm2 // abs difference trick - vpsubusb ymm2, ymm2, ymm1 - vpor ymm1, ymm2, ymm3 - vpunpcklbw ymm2, ymm1, ymm5 // u16. mutates order. - vpunpckhbw ymm1, ymm1, ymm5 - vpmaddwd ymm2, ymm2, ymm2 // square + hadd to u32. - vpmaddwd ymm1, ymm1, ymm1 - vpaddd ymm0, ymm0, ymm1 - vpaddd ymm0, ymm0, ymm2 - jg wloop - - vpshufd ymm1, ymm0, 0xee // 3, 2 + 1, 0 both lanes. - vpaddd ymm0, ymm0, ymm1 - vpshufd ymm1, ymm0, 0x01 // 1 + 0 both lanes. - vpaddd ymm0, ymm0, ymm1 - vpermq ymm1, ymm0, 0x02 // high + low lane. - vpaddd ymm0, ymm0, ymm1 - vmovd eax, xmm0 - vzeroupper - ret - } -} -#endif // _MSC_VER >= 1700 - -#define HAS_HASHDJB2_SSE41 -static uvec32 kHash16x33 = { 0x92d9e201, 0, 0, 0 }; // 33 ^ 16 -static uvec32 kHashMul0 = { - 0x0c3525e1, // 33 ^ 15 - 0xa3476dc1, // 33 ^ 14 - 0x3b4039a1, // 33 ^ 13 - 0x4f5f0981, // 33 ^ 12 -}; -static uvec32 kHashMul1 = { - 0x30f35d61, // 33 ^ 11 - 0x855cb541, // 33 ^ 10 - 0x040a9121, // 33 ^ 9 - 0x747c7101, // 33 ^ 8 -}; -static uvec32 kHashMul2 = { - 0xec41d4e1, // 33 ^ 7 - 0x4cfa3cc1, // 33 ^ 6 - 0x025528a1, // 33 ^ 5 - 0x00121881, // 33 ^ 4 -}; -static uvec32 kHashMul3 = { - 0x00008c61, // 33 ^ 3 - 0x00000441, // 33 ^ 2 - 0x00000021, // 33 ^ 1 - 0x00000001, // 33 ^ 0 -}; - -// 27: 66 0F 38 40 C6 pmulld xmm0,xmm6 -// 44: 66 0F 38 40 DD pmulld xmm3,xmm5 -// 59: 66 0F 38 40 E5 pmulld xmm4,xmm5 -// 72: 66 0F 38 40 D5 pmulld xmm2,xmm5 -// 83: 66 0F 38 40 CD pmulld xmm1,xmm5 -#define pmulld(reg) _asm _emit 0x66 _asm _emit 0x0F _asm _emit 0x38 \ - _asm _emit 0x40 _asm _emit reg - -__declspec(naked) __declspec(align(16)) -uint32 HashDjb2_SSE41(const uint8* src, int count, uint32 seed) { - __asm { - mov eax, [esp + 4] // src - mov ecx, [esp + 8] // count - movd xmm0, [esp + 12] // seed - - pxor xmm7, xmm7 // constant 0 for unpck - movdqa xmm6, kHash16x33 - - align 4 - wloop: - movdqu xmm1, [eax] // src[0-15] - lea eax, [eax + 16] - pmulld(0xc6) // pmulld xmm0,xmm6 hash *= 33 ^ 16 - movdqa xmm5, kHashMul0 - movdqa xmm2, xmm1 - punpcklbw xmm2, xmm7 // src[0-7] - movdqa xmm3, xmm2 - punpcklwd xmm3, xmm7 // src[0-3] - pmulld(0xdd) // pmulld xmm3, xmm5 - movdqa xmm5, kHashMul1 - movdqa xmm4, xmm2 - punpckhwd xmm4, xmm7 // src[4-7] - pmulld(0xe5) // pmulld xmm4, xmm5 - movdqa xmm5, kHashMul2 - punpckhbw xmm1, xmm7 // src[8-15] - movdqa xmm2, xmm1 - punpcklwd xmm2, xmm7 // src[8-11] - pmulld(0xd5) // pmulld xmm2, xmm5 - movdqa xmm5, kHashMul3 - punpckhwd xmm1, xmm7 // src[12-15] - pmulld(0xcd) // pmulld xmm1, xmm5 - paddd xmm3, xmm4 // add 16 results - paddd xmm1, xmm2 - sub ecx, 16 - paddd xmm1, xmm3 - - pshufd xmm2, xmm1, 0x0e // upper 2 dwords - paddd xmm1, xmm2 - pshufd xmm2, xmm1, 0x01 - paddd xmm1, xmm2 - paddd xmm0, xmm1 - jg wloop - - movd eax, xmm0 // return hash - ret - } -} - -// Visual C 2012 required for AVX2. -#if _MSC_VER >= 1700 -__declspec(naked) __declspec(align(16)) -uint32 HashDjb2_AVX2(const uint8* src, int count, uint32 seed) { - __asm { - mov eax, [esp + 4] // src - mov ecx, [esp + 8] // count - movd xmm0, [esp + 12] // seed - movdqa xmm6, kHash16x33 - - align 4 - wloop: - vpmovzxbd xmm3, dword ptr [eax] // src[0-3] - pmulld xmm0, xmm6 // hash *= 33 ^ 16 - vpmovzxbd xmm4, dword ptr [eax + 4] // src[4-7] - pmulld xmm3, kHashMul0 - vpmovzxbd xmm2, dword ptr [eax + 8] // src[8-11] - pmulld xmm4, kHashMul1 - vpmovzxbd xmm1, dword ptr [eax + 12] // src[12-15] - pmulld xmm2, kHashMul2 - lea eax, [eax + 16] - pmulld xmm1, kHashMul3 - paddd xmm3, xmm4 // add 16 results - paddd xmm1, xmm2 - sub ecx, 16 - paddd xmm1, xmm3 - pshufd xmm2, xmm1, 0x0e // upper 2 dwords - paddd xmm1, xmm2 - pshufd xmm2, xmm1, 0x01 - paddd xmm1, xmm2 - paddd xmm0, xmm1 - jg wloop - - movd eax, xmm0 // return hash - ret - } -} -#endif // _MSC_VER >= 1700 - -#endif // !defined(LIBYUV_DISABLE_X86) && defined(_M_IX86) && defined(_MSC_VER) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert.cc b/src/main/jni/libyuv/source/convert.cc deleted file mode 100644 index c31ecf263..000000000 --- a/src/main/jni/libyuv/source/convert.cc +++ /dev/null @@ -1,1543 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/convert.h" - -#include "libyuv/basic_types.h" -#include "libyuv/cpu_id.h" -#include "libyuv/planar_functions.h" -#include "libyuv/rotate.h" -#include "libyuv/scale.h" // For ScalePlane() -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#define SUBSAMPLE(v, a, s) (v < 0) ? (-((-v + a) >> s)) : ((v + a) >> s) -static __inline int Abs(int v) { - return v >= 0 ? v : -v; -} - -// Any I4xx To I420 format with mirroring. -static int I4xxToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int src_y_width, int src_y_height, - int src_uv_width, int src_uv_height) { - const int dst_y_width = Abs(src_y_width); - const int dst_y_height = Abs(src_y_height); - const int dst_uv_width = SUBSAMPLE(dst_y_width, 1, 1); - const int dst_uv_height = SUBSAMPLE(dst_y_height, 1, 1); - if (src_y_width == 0 || src_y_height == 0 || - src_uv_width == 0 || src_uv_height == 0) { - return -1; - } - ScalePlane(src_y, src_stride_y, src_y_width, src_y_height, - dst_y, dst_stride_y, dst_y_width, dst_y_height, - kFilterBilinear); - ScalePlane(src_u, src_stride_u, src_uv_width, src_uv_height, - dst_u, dst_stride_u, dst_uv_width, dst_uv_height, - kFilterBilinear); - ScalePlane(src_v, src_stride_v, src_uv_width, src_uv_height, - dst_v, dst_stride_v, dst_uv_width, dst_uv_height, - kFilterBilinear); - return 0; -} - -// Copy I420 with optional flipping -// TODO(fbarchard): Use Scale plane which supports mirroring, but ensure -// is does row coalescing. -LIBYUV_API -int I420Copy(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - if (!src_y || !src_u || !src_v || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - src_y = src_y + (height - 1) * src_stride_y; - src_u = src_u + (halfheight - 1) * src_stride_u; - src_v = src_v + (halfheight - 1) * src_stride_v; - src_stride_y = -src_stride_y; - src_stride_u = -src_stride_u; - src_stride_v = -src_stride_v; - } - - if (dst_y) { - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - } - // Copy UV planes. - CopyPlane(src_u, src_stride_u, dst_u, dst_stride_u, halfwidth, halfheight); - CopyPlane(src_v, src_stride_v, dst_v, dst_stride_v, halfwidth, halfheight); - return 0; -} - -// 422 chroma is 1/2 width, 1x height -// 420 chroma is 1/2 width, 1/2 height -LIBYUV_API -int I422ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - const int src_uv_width = SUBSAMPLE(width, 1, 1); - return I4xxToI420(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height, - src_uv_width, height); -} - -// 444 chroma is 1x width, 1x height -// 420 chroma is 1/2 width, 1/2 height -LIBYUV_API -int I444ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - return I4xxToI420(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height, - width, height); -} - -// 411 chroma is 1/4 width, 1x height -// 420 chroma is 1/2 width, 1/2 height -LIBYUV_API -int I411ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - const int src_uv_width = SUBSAMPLE(width, 3, 2); - return I4xxToI420(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height, - src_uv_width, height); -} - -// I400 is greyscale typically used in MJPG -LIBYUV_API -int I400ToI420(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - if (!src_y || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - SetPlane(dst_u, dst_stride_u, halfwidth, halfheight, 128); - SetPlane(dst_v, dst_stride_v, halfwidth, halfheight, 128); - return 0; -} - -static void CopyPlane2(const uint8* src, int src_stride_0, int src_stride_1, - uint8* dst, int dst_stride, - int width, int height) { - int y; - void (*CopyRow)(const uint8* src, uint8* dst, int width) = CopyRow_C; -#if defined(HAS_COPYROW_X86) - if (TestCpuFlag(kCpuHasX86) && IS_ALIGNED(width, 4)) { - CopyRow = CopyRow_X86; - } -#endif -#if defined(HAS_COPYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 32) && - IS_ALIGNED(src, 16) && - IS_ALIGNED(src_stride_0, 16) && IS_ALIGNED(src_stride_1, 16) && - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - CopyRow = CopyRow_SSE2; - } -#endif -#if defined(HAS_COPYROW_ERMS) - if (TestCpuFlag(kCpuHasERMS)) { - CopyRow = CopyRow_ERMS; - } -#endif -#if defined(HAS_COPYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 32)) { - CopyRow = CopyRow_NEON; - } -#endif -#if defined(HAS_COPYROW_MIPS) - if (TestCpuFlag(kCpuHasMIPS)) { - CopyRow = CopyRow_MIPS; - } -#endif - - // Copy plane - for (y = 0; y < height - 1; y += 2) { - CopyRow(src, dst, width); - CopyRow(src + src_stride_0, dst + dst_stride, width); - src += src_stride_0 + src_stride_1; - dst += dst_stride * 2; - } - if (height & 1) { - CopyRow(src, dst, width); - } -} - -// Support converting from FOURCC_M420 -// Useful for bandwidth constrained transports like USB 1.0 and 2.0 and for -// easy conversion to I420. -// M420 format description: -// M420 is row biplanar 420: 2 rows of Y and 1 row of UV. -// Chroma is half width / half height. (420) -// src_stride_m420 is row planar. Normally this will be the width in pixels. -// The UV plane is half width, but 2 values, so src_stride_m420 applies to -// this as well as the two Y planes. -static int X420ToI420(const uint8* src_y, - int src_stride_y0, int src_stride_y1, - const uint8* src_uv, int src_stride_uv, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - void (*SplitUVRow)(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix) = - SplitUVRow_C; - if (!src_y || !src_uv || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - dst_y = dst_y + (height - 1) * dst_stride_y; - dst_u = dst_u + (halfheight - 1) * dst_stride_u; - dst_v = dst_v + (halfheight - 1) * dst_stride_v; - dst_stride_y = -dst_stride_y; - dst_stride_u = -dst_stride_u; - dst_stride_v = -dst_stride_v; - } - // Coalesce rows. - if (src_stride_y0 == width && - src_stride_y1 == width && - dst_stride_y == width) { - width *= height; - height = 1; - src_stride_y0 = src_stride_y1 = dst_stride_y = 0; - } - // Coalesce rows. - if (src_stride_uv == halfwidth * 2 && - dst_stride_u == halfwidth && - dst_stride_v == halfwidth) { - halfwidth *= halfheight; - halfheight = 1; - src_stride_uv = dst_stride_u = dst_stride_v = 0; - } -#if defined(HAS_SPLITUVROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && halfwidth >= 16) { - SplitUVRow = SplitUVRow_Any_SSE2; - if (IS_ALIGNED(halfwidth, 16)) { - SplitUVRow = SplitUVRow_Unaligned_SSE2; - if (IS_ALIGNED(src_uv, 16) && IS_ALIGNED(src_stride_uv, 16) && - IS_ALIGNED(dst_u, 16) && IS_ALIGNED(dst_stride_u, 16) && - IS_ALIGNED(dst_v, 16) && IS_ALIGNED(dst_stride_v, 16)) { - SplitUVRow = SplitUVRow_SSE2; - } - } - } -#endif -#if defined(HAS_SPLITUVROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && halfwidth >= 32) { - SplitUVRow = SplitUVRow_Any_AVX2; - if (IS_ALIGNED(halfwidth, 32)) { - SplitUVRow = SplitUVRow_AVX2; - } - } -#endif -#if defined(HAS_SPLITUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && halfwidth >= 16) { - SplitUVRow = SplitUVRow_Any_NEON; - if (IS_ALIGNED(halfwidth, 16)) { - SplitUVRow = SplitUVRow_NEON; - } - } -#endif -#if defined(HAS_SPLITUVROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && halfwidth >= 16) { - SplitUVRow = SplitUVRow_Any_MIPS_DSPR2; - if (IS_ALIGNED(halfwidth, 16)) { - SplitUVRow = SplitUVRow_Unaligned_MIPS_DSPR2; - if (IS_ALIGNED(src_uv, 4) && IS_ALIGNED(src_stride_uv, 4) && - IS_ALIGNED(dst_u, 4) && IS_ALIGNED(dst_stride_u, 4) && - IS_ALIGNED(dst_v, 4) && IS_ALIGNED(dst_stride_v, 4)) { - SplitUVRow = SplitUVRow_MIPS_DSPR2; - } - } - } -#endif - - if (dst_y) { - if (src_stride_y0 == src_stride_y1) { - CopyPlane(src_y, src_stride_y0, dst_y, dst_stride_y, width, height); - } else { - CopyPlane2(src_y, src_stride_y0, src_stride_y1, dst_y, dst_stride_y, - width, height); - } - } - - for (y = 0; y < halfheight; ++y) { - // Copy a row of UV. - SplitUVRow(src_uv, dst_u, dst_v, halfwidth); - dst_u += dst_stride_u; - dst_v += dst_stride_v; - src_uv += src_stride_uv; - } - return 0; -} - -// Convert NV12 to I420. -LIBYUV_API -int NV12ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - return X420ToI420(src_y, src_stride_y, src_stride_y, - src_uv, src_stride_uv, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height); -} - -// Convert NV21 to I420. Same as NV12 but u and v pointers swapped. -LIBYUV_API -int NV21ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_vu, int src_stride_vu, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - return X420ToI420(src_y, src_stride_y, src_stride_y, - src_vu, src_stride_vu, - dst_y, dst_stride_y, - dst_v, dst_stride_v, - dst_u, dst_stride_u, - width, height); -} - -// Convert M420 to I420. -LIBYUV_API -int M420ToI420(const uint8* src_m420, int src_stride_m420, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - return X420ToI420(src_m420, src_stride_m420, src_stride_m420 * 2, - src_m420 + src_stride_m420 * 2, src_stride_m420 * 3, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height); -} - -// Convert Q420 to I420. -// Format is rows of YY/YUYV -LIBYUV_API -int Q420ToI420(const uint8* src_y, int src_stride_y, - const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - int halfheight; - void (*CopyRow)(const uint8* src, uint8* dst, int width) = CopyRow_C; - void (*YUY2ToUV422Row)(const uint8* src_yuy2, uint8* dst_u, uint8* dst_v, - int pix) = YUY2ToUV422Row_C; - void (*YUY2ToYRow)(const uint8* src_yuy2, uint8* dst_y, int pix) = - YUY2ToYRow_C; - if (!src_y || !src_yuy2 || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - dst_y = dst_y + (height - 1) * dst_stride_y; - dst_u = dst_u + (halfheight - 1) * dst_stride_u; - dst_v = dst_v + (halfheight - 1) * dst_stride_v; - dst_stride_y = -dst_stride_y; - dst_stride_u = -dst_stride_u; - dst_stride_v = -dst_stride_v; - } - // CopyRow for rows of just Y in Q420 copied to Y plane of I420. -#if defined(HAS_COPYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 32)) { - CopyRow = CopyRow_NEON; - } -#endif -#if defined(HAS_COPYROW_X86) - if (IS_ALIGNED(width, 4)) { - CopyRow = CopyRow_X86; - } -#endif -#if defined(HAS_COPYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 32) && - IS_ALIGNED(src_y, 16) && IS_ALIGNED(src_stride_y, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - CopyRow = CopyRow_SSE2; - } -#endif -#if defined(HAS_COPYROW_ERMS) - if (TestCpuFlag(kCpuHasERMS)) { - CopyRow = CopyRow_ERMS; - } -#endif -#if defined(HAS_COPYROW_MIPS) - if (TestCpuFlag(kCpuHasMIPS)) { - CopyRow = CopyRow_MIPS; - } -#endif - -#if defined(HAS_YUY2TOYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - YUY2ToUV422Row = YUY2ToUV422Row_Any_SSE2; - YUY2ToYRow = YUY2ToYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - YUY2ToUV422Row = YUY2ToUV422Row_Unaligned_SSE2; - YUY2ToYRow = YUY2ToYRow_Unaligned_SSE2; - if (IS_ALIGNED(src_yuy2, 16) && IS_ALIGNED(src_stride_yuy2, 16)) { - YUY2ToUV422Row = YUY2ToUV422Row_SSE2; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - YUY2ToYRow = YUY2ToYRow_SSE2; - } - } - } - } -#endif -#if defined(HAS_YUY2TOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - YUY2ToUV422Row = YUY2ToUV422Row_Any_AVX2; - YUY2ToYRow = YUY2ToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - YUY2ToUV422Row = YUY2ToUV422Row_AVX2; - YUY2ToYRow = YUY2ToYRow_AVX2; - } - } -#endif -#if defined(HAS_YUY2TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - YUY2ToYRow = YUY2ToYRow_Any_NEON; - if (width >= 16) { - YUY2ToUV422Row = YUY2ToUV422Row_Any_NEON; - } - if (IS_ALIGNED(width, 16)) { - YUY2ToYRow = YUY2ToYRow_NEON; - YUY2ToUV422Row = YUY2ToUV422Row_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - CopyRow(src_y, dst_y, width); - src_y += src_stride_y; - dst_y += dst_stride_y; - - YUY2ToUV422Row(src_yuy2, dst_u, dst_v, width); - YUY2ToYRow(src_yuy2, dst_y, width); - src_yuy2 += src_stride_yuy2; - dst_y += dst_stride_y; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - CopyRow(src_y, dst_y, width); - YUY2ToUV422Row(src_yuy2, dst_u, dst_v, width); - } - return 0; -} - -// Convert YUY2 to I420. -LIBYUV_API -int YUY2ToI420(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*YUY2ToUVRow)(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) = YUY2ToUVRow_C; - void (*YUY2ToYRow)(const uint8* src_yuy2, - uint8* dst_y, int pix) = YUY2ToYRow_C; - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_yuy2 = src_yuy2 + (height - 1) * src_stride_yuy2; - src_stride_yuy2 = -src_stride_yuy2; - } -#if defined(HAS_YUY2TOYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - YUY2ToUVRow = YUY2ToUVRow_Any_SSE2; - YUY2ToYRow = YUY2ToYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - YUY2ToUVRow = YUY2ToUVRow_Unaligned_SSE2; - YUY2ToYRow = YUY2ToYRow_Unaligned_SSE2; - if (IS_ALIGNED(src_yuy2, 16) && IS_ALIGNED(src_stride_yuy2, 16)) { - YUY2ToUVRow = YUY2ToUVRow_SSE2; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - YUY2ToYRow = YUY2ToYRow_SSE2; - } - } - } - } -#endif -#if defined(HAS_YUY2TOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - YUY2ToUVRow = YUY2ToUVRow_Any_AVX2; - YUY2ToYRow = YUY2ToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - YUY2ToUVRow = YUY2ToUVRow_AVX2; - YUY2ToYRow = YUY2ToYRow_AVX2; - } - } -#endif -#if defined(HAS_YUY2TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - YUY2ToYRow = YUY2ToYRow_Any_NEON; - if (width >= 16) { - YUY2ToUVRow = YUY2ToUVRow_Any_NEON; - } - if (IS_ALIGNED(width, 16)) { - YUY2ToYRow = YUY2ToYRow_NEON; - YUY2ToUVRow = YUY2ToUVRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - YUY2ToUVRow(src_yuy2, src_stride_yuy2, dst_u, dst_v, width); - YUY2ToYRow(src_yuy2, dst_y, width); - YUY2ToYRow(src_yuy2 + src_stride_yuy2, dst_y + dst_stride_y, width); - src_yuy2 += src_stride_yuy2 * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - YUY2ToUVRow(src_yuy2, 0, dst_u, dst_v, width); - YUY2ToYRow(src_yuy2, dst_y, width); - } - return 0; -} - -// Convert UYVY to I420. -LIBYUV_API -int UYVYToI420(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*UYVYToUVRow)(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) = UYVYToUVRow_C; - void (*UYVYToYRow)(const uint8* src_uyvy, - uint8* dst_y, int pix) = UYVYToYRow_C; - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_uyvy = src_uyvy + (height - 1) * src_stride_uyvy; - src_stride_uyvy = -src_stride_uyvy; - } -#if defined(HAS_UYVYTOYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - UYVYToUVRow = UYVYToUVRow_Any_SSE2; - UYVYToYRow = UYVYToYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - UYVYToUVRow = UYVYToUVRow_Unaligned_SSE2; - UYVYToYRow = UYVYToYRow_Unaligned_SSE2; - if (IS_ALIGNED(src_uyvy, 16) && IS_ALIGNED(src_stride_uyvy, 16)) { - UYVYToUVRow = UYVYToUVRow_SSE2; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - UYVYToYRow = UYVYToYRow_SSE2; - } - } - } - } -#endif -#if defined(HAS_UYVYTOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - UYVYToUVRow = UYVYToUVRow_Any_AVX2; - UYVYToYRow = UYVYToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - UYVYToUVRow = UYVYToUVRow_AVX2; - UYVYToYRow = UYVYToYRow_AVX2; - } - } -#endif -#if defined(HAS_UYVYTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - UYVYToYRow = UYVYToYRow_Any_NEON; - if (width >= 16) { - UYVYToUVRow = UYVYToUVRow_Any_NEON; - } - if (IS_ALIGNED(width, 16)) { - UYVYToYRow = UYVYToYRow_NEON; - UYVYToUVRow = UYVYToUVRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - UYVYToUVRow(src_uyvy, src_stride_uyvy, dst_u, dst_v, width); - UYVYToYRow(src_uyvy, dst_y, width); - UYVYToYRow(src_uyvy + src_stride_uyvy, dst_y + dst_stride_y, width); - src_uyvy += src_stride_uyvy * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - UYVYToUVRow(src_uyvy, 0, dst_u, dst_v, width); - UYVYToYRow(src_uyvy, dst_y, width); - } - return 0; -} - -// Convert ARGB to I420. -LIBYUV_API -int ARGBToI420(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - if (!src_argb || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } -#if defined(HAS_ARGBTOYROW_SSSE3) && defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_Unaligned_SSSE3; - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } - } -#endif -#if defined(HAS_ARGBTOYROW_AVX2) && defined(HAS_ARGBTOUVROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - ARGBToUVRow = ARGBToUVRow_Any_AVX2; - ARGBToYRow = ARGBToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - ARGBToUVRow = ARGBToUVRow_AVX2; - ARGBToYRow = ARGBToYRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - ARGBToUVRow(src_argb, src_stride_argb, dst_u, dst_v, width); - ARGBToYRow(src_argb, dst_y, width); - ARGBToYRow(src_argb + src_stride_argb, dst_y + dst_stride_y, width); - src_argb += src_stride_argb * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - ARGBToUVRow(src_argb, 0, dst_u, dst_v, width); - ARGBToYRow(src_argb, dst_y, width); - } - return 0; -} - -// Convert BGRA to I420. -LIBYUV_API -int BGRAToI420(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*BGRAToUVRow)(const uint8* src_bgra0, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width) = BGRAToUVRow_C; - void (*BGRAToYRow)(const uint8* src_bgra, uint8* dst_y, int pix) = - BGRAToYRow_C; - if (!src_bgra || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_bgra = src_bgra + (height - 1) * src_stride_bgra; - src_stride_bgra = -src_stride_bgra; - } -#if defined(HAS_BGRATOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - BGRAToUVRow = BGRAToUVRow_Any_SSSE3; - BGRAToYRow = BGRAToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - BGRAToUVRow = BGRAToUVRow_Unaligned_SSSE3; - BGRAToYRow = BGRAToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_bgra, 16) && IS_ALIGNED(src_stride_bgra, 16)) { - BGRAToUVRow = BGRAToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - BGRAToYRow = BGRAToYRow_SSSE3; - } - } - } - } -#elif defined(HAS_BGRATOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - BGRAToYRow = BGRAToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - BGRAToYRow = BGRAToYRow_NEON; - } - } -#endif -#if defined(HAS_BGRATOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - BGRAToUVRow = BGRAToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - BGRAToUVRow = BGRAToUVRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - BGRAToUVRow(src_bgra, src_stride_bgra, dst_u, dst_v, width); - BGRAToYRow(src_bgra, dst_y, width); - BGRAToYRow(src_bgra + src_stride_bgra, dst_y + dst_stride_y, width); - src_bgra += src_stride_bgra * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - BGRAToUVRow(src_bgra, 0, dst_u, dst_v, width); - BGRAToYRow(src_bgra, dst_y, width); - } - return 0; -} - -// Convert ABGR to I420. -LIBYUV_API -int ABGRToI420(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*ABGRToUVRow)(const uint8* src_abgr0, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width) = ABGRToUVRow_C; - void (*ABGRToYRow)(const uint8* src_abgr, uint8* dst_y, int pix) = - ABGRToYRow_C; - if (!src_abgr || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_abgr = src_abgr + (height - 1) * src_stride_abgr; - src_stride_abgr = -src_stride_abgr; - } -#if defined(HAS_ABGRTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ABGRToUVRow = ABGRToUVRow_Any_SSSE3; - ABGRToYRow = ABGRToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ABGRToUVRow = ABGRToUVRow_Unaligned_SSSE3; - ABGRToYRow = ABGRToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_abgr, 16) && IS_ALIGNED(src_stride_abgr, 16)) { - ABGRToUVRow = ABGRToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ABGRToYRow = ABGRToYRow_SSSE3; - } - } - } - } -#elif defined(HAS_ABGRTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ABGRToYRow = ABGRToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ABGRToYRow = ABGRToYRow_NEON; - } - } -#endif -#if defined(HAS_ABGRTOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ABGRToUVRow = ABGRToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ABGRToUVRow = ABGRToUVRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - ABGRToUVRow(src_abgr, src_stride_abgr, dst_u, dst_v, width); - ABGRToYRow(src_abgr, dst_y, width); - ABGRToYRow(src_abgr + src_stride_abgr, dst_y + dst_stride_y, width); - src_abgr += src_stride_abgr * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - ABGRToUVRow(src_abgr, 0, dst_u, dst_v, width); - ABGRToYRow(src_abgr, dst_y, width); - } - return 0; -} - -// Convert RGBA to I420. -LIBYUV_API -int RGBAToI420(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*RGBAToUVRow)(const uint8* src_rgba0, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width) = RGBAToUVRow_C; - void (*RGBAToYRow)(const uint8* src_rgba, uint8* dst_y, int pix) = - RGBAToYRow_C; - if (!src_rgba || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_rgba = src_rgba + (height - 1) * src_stride_rgba; - src_stride_rgba = -src_stride_rgba; - } -#if defined(HAS_RGBATOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - RGBAToUVRow = RGBAToUVRow_Any_SSSE3; - RGBAToYRow = RGBAToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - RGBAToUVRow = RGBAToUVRow_Unaligned_SSSE3; - RGBAToYRow = RGBAToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_rgba, 16) && IS_ALIGNED(src_stride_rgba, 16)) { - RGBAToUVRow = RGBAToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - RGBAToYRow = RGBAToYRow_SSSE3; - } - } - } - } -#elif defined(HAS_RGBATOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RGBAToYRow = RGBAToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RGBAToYRow = RGBAToYRow_NEON; - } - } -#endif -#if defined(HAS_RGBATOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - RGBAToUVRow = RGBAToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - RGBAToUVRow = RGBAToUVRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - RGBAToUVRow(src_rgba, src_stride_rgba, dst_u, dst_v, width); - RGBAToYRow(src_rgba, dst_y, width); - RGBAToYRow(src_rgba + src_stride_rgba, dst_y + dst_stride_y, width); - src_rgba += src_stride_rgba * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - RGBAToUVRow(src_rgba, 0, dst_u, dst_v, width); - RGBAToYRow(src_rgba, dst_y, width); - } - return 0; -} - -// Convert RGB24 to I420. -LIBYUV_API -int RGB24ToI420(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; -#if defined(HAS_RGB24TOYROW_NEON) - void (*RGB24ToUVRow)(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_u, uint8* dst_v, int width) = RGB24ToUVRow_C; - void (*RGB24ToYRow)(const uint8* src_rgb24, uint8* dst_y, int pix) = - RGB24ToYRow_C; -#else - void (*RGB24ToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - RGB24ToARGBRow_C; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; -#endif - if (!src_rgb24 || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_rgb24 = src_rgb24 + (height - 1) * src_stride_rgb24; - src_stride_rgb24 = -src_stride_rgb24; - } - -#if defined(HAS_RGB24TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RGB24ToYRow = RGB24ToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RGB24ToYRow = RGB24ToYRow_NEON; - } - } -#endif -#if defined(HAS_RGB24TOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - RGB24ToUVRow = RGB24ToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - RGB24ToUVRow = RGB24ToUVRow_NEON; - } - } -#endif -#if defined(HAS_RGB24TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - RGB24ToARGBRow = RGB24ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - RGB24ToARGBRow = RGB24ToARGBRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif // HAS_ARGBTOUVROW_SSSE3 - - { -#if !defined(HAS_RGB24TOYROW_NEON) - // Allocate 2 rows of ARGB. - const int kRowSize = (width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); -#endif - - for (y = 0; y < height - 1; y += 2) { -#if defined(HAS_RGB24TOYROW_NEON) - RGB24ToUVRow(src_rgb24, src_stride_rgb24, dst_u, dst_v, width); - RGB24ToYRow(src_rgb24, dst_y, width); - RGB24ToYRow(src_rgb24 + src_stride_rgb24, dst_y + dst_stride_y, width); -#else - RGB24ToARGBRow(src_rgb24, row, width); - RGB24ToARGBRow(src_rgb24 + src_stride_rgb24, row + kRowSize, width); - ARGBToUVRow(row, kRowSize, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - ARGBToYRow(row + kRowSize, dst_y + dst_stride_y, width); -#endif - src_rgb24 += src_stride_rgb24 * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { -#if defined(HAS_RGB24TOYROW_NEON) - RGB24ToUVRow(src_rgb24, 0, dst_u, dst_v, width); - RGB24ToYRow(src_rgb24, dst_y, width); -#else - RGB24ToARGBRow(src_rgb24, row, width); - ARGBToUVRow(row, 0, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); -#endif - } -#if !defined(HAS_RGB24TOYROW_NEON) - free_aligned_buffer_64(row); -#endif - } - return 0; -} - -// Convert RAW to I420. -LIBYUV_API -int RAWToI420(const uint8* src_raw, int src_stride_raw, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; -#if defined(HAS_RAWTOYROW_NEON) - void (*RAWToUVRow)(const uint8* src_raw, int src_stride_raw, - uint8* dst_u, uint8* dst_v, int width) = RAWToUVRow_C; - void (*RAWToYRow)(const uint8* src_raw, uint8* dst_y, int pix) = - RAWToYRow_C; -#else - void (*RAWToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - RAWToARGBRow_C; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; -#endif - if (!src_raw || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_raw = src_raw + (height - 1) * src_stride_raw; - src_stride_raw = -src_stride_raw; - } - -#if defined(HAS_RAWTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RAWToYRow = RAWToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RAWToYRow = RAWToYRow_NEON; - } - } -#endif -#if defined(HAS_RAWTOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - RAWToUVRow = RAWToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - RAWToUVRow = RAWToUVRow_NEON; - } - } -#endif -#if defined(HAS_RAWTOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - RAWToARGBRow = RAWToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - RAWToARGBRow = RAWToARGBRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif // HAS_ARGBTOUVROW_SSSE3 - - { - // Allocate 2 rows of ARGB. - const int kRowSize = (width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); - - for (y = 0; y < height - 1; y += 2) { - #if defined(HAS_RAWTOYROW_NEON) - RAWToUVRow(src_raw, src_stride_raw, dst_u, dst_v, width); - RAWToYRow(src_raw, dst_y, width); - RAWToYRow(src_raw + src_stride_raw, dst_y + dst_stride_y, width); - #else - RAWToARGBRow(src_raw, row, width); - RAWToARGBRow(src_raw + src_stride_raw, row + kRowSize, width); - ARGBToUVRow(row, kRowSize, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - ARGBToYRow(row + kRowSize, dst_y + dst_stride_y, width); - #endif - src_raw += src_stride_raw * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - #if defined(HAS_RAWTOYROW_NEON) - RAWToUVRow(src_raw, 0, dst_u, dst_v, width); - RAWToYRow(src_raw, dst_y, width); - #else - RAWToARGBRow(src_raw, row, width); - ARGBToUVRow(row, 0, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - #endif - } - #if !defined(HAS_RAWTOYROW_NEON) - free_aligned_buffer_64(row); - #endif - } - return 0; -} - -// Convert RGB565 to I420. -LIBYUV_API -int RGB565ToI420(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; -#if defined(HAS_RGB565TOYROW_NEON) - void (*RGB565ToUVRow)(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int width) = RGB565ToUVRow_C; - void (*RGB565ToYRow)(const uint8* src_rgb565, uint8* dst_y, int pix) = - RGB565ToYRow_C; -#else - void (*RGB565ToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - RGB565ToARGBRow_C; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; -#endif - if (!src_rgb565 || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_rgb565 = src_rgb565 + (height - 1) * src_stride_rgb565; - src_stride_rgb565 = -src_stride_rgb565; - } - -#if defined(HAS_RGB565TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RGB565ToYRow = RGB565ToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RGB565ToYRow = RGB565ToYRow_NEON; - } - if (width >= 16) { - RGB565ToUVRow = RGB565ToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - RGB565ToUVRow = RGB565ToUVRow_NEON; - } - } - } -#else // HAS_RGB565TOYROW_NEON - -#if defined(HAS_RGB565TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8) { - RGB565ToARGBRow = RGB565ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - RGB565ToARGBRow = RGB565ToARGBRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif // HAS_ARGBTOUVROW_SSSE3 -#endif // HAS_RGB565TOYROW_NEON - - { -#if !defined(HAS_RGB565TOYROW_NEON) - // Allocate 2 rows of ARGB. - const int kRowSize = (width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); -#endif - - for (y = 0; y < height - 1; y += 2) { -#if defined(HAS_RGB565TOYROW_NEON) - RGB565ToUVRow(src_rgb565, src_stride_rgb565, dst_u, dst_v, width); - RGB565ToYRow(src_rgb565, dst_y, width); - RGB565ToYRow(src_rgb565 + src_stride_rgb565, dst_y + dst_stride_y, width); -#else - RGB565ToARGBRow(src_rgb565, row, width); - RGB565ToARGBRow(src_rgb565 + src_stride_rgb565, row + kRowSize, width); - ARGBToUVRow(row, kRowSize, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - ARGBToYRow(row + kRowSize, dst_y + dst_stride_y, width); -#endif - src_rgb565 += src_stride_rgb565 * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { -#if defined(HAS_RGB565TOYROW_NEON) - RGB565ToUVRow(src_rgb565, 0, dst_u, dst_v, width); - RGB565ToYRow(src_rgb565, dst_y, width); -#else - RGB565ToARGBRow(src_rgb565, row, width); - ARGBToUVRow(row, 0, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); -#endif - } -#if !defined(HAS_RGB565TOYROW_NEON) - free_aligned_buffer_64(row); -#endif - } - return 0; -} - -// Convert ARGB1555 to I420. -LIBYUV_API -int ARGB1555ToI420(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; -#if defined(HAS_ARGB1555TOYROW_NEON) - void (*ARGB1555ToUVRow)(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int width) = ARGB1555ToUVRow_C; - void (*ARGB1555ToYRow)(const uint8* src_argb1555, uint8* dst_y, int pix) = - ARGB1555ToYRow_C; -#else - void (*ARGB1555ToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - ARGB1555ToARGBRow_C; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; -#endif - if (!src_argb1555 || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb1555 = src_argb1555 + (height - 1) * src_stride_argb1555; - src_stride_argb1555 = -src_stride_argb1555; - } - -#if defined(HAS_ARGB1555TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGB1555ToYRow = ARGB1555ToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGB1555ToYRow = ARGB1555ToYRow_NEON; - } - if (width >= 16) { - ARGB1555ToUVRow = ARGB1555ToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGB1555ToUVRow = ARGB1555ToUVRow_NEON; - } - } - } -#else // HAS_ARGB1555TOYROW_NEON - -#if defined(HAS_ARGB1555TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8) { - ARGB1555ToARGBRow = ARGB1555ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - ARGB1555ToARGBRow = ARGB1555ToARGBRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif // HAS_ARGBTOUVROW_SSSE3 -#endif // HAS_ARGB1555TOYROW_NEON - - { -#if !defined(HAS_ARGB1555TOYROW_NEON) - // Allocate 2 rows of ARGB. - const int kRowSize = (width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); -#endif - for (y = 0; y < height - 1; y += 2) { -#if defined(HAS_ARGB1555TOYROW_NEON) - ARGB1555ToUVRow(src_argb1555, src_stride_argb1555, dst_u, dst_v, width); - ARGB1555ToYRow(src_argb1555, dst_y, width); - ARGB1555ToYRow(src_argb1555 + src_stride_argb1555, dst_y + dst_stride_y, - width); -#else - ARGB1555ToARGBRow(src_argb1555, row, width); - ARGB1555ToARGBRow(src_argb1555 + src_stride_argb1555, row + kRowSize, - width); - ARGBToUVRow(row, kRowSize, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - ARGBToYRow(row + kRowSize, dst_y + dst_stride_y, width); -#endif - src_argb1555 += src_stride_argb1555 * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { -#if defined(HAS_ARGB1555TOYROW_NEON) - ARGB1555ToUVRow(src_argb1555, 0, dst_u, dst_v, width); - ARGB1555ToYRow(src_argb1555, dst_y, width); -#else - ARGB1555ToARGBRow(src_argb1555, row, width); - ARGBToUVRow(row, 0, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); -#endif - } -#if !defined(HAS_ARGB1555TOYROW_NEON) - free_aligned_buffer_64(row); -#endif - } - return 0; -} - -// Convert ARGB4444 to I420. -LIBYUV_API -int ARGB4444ToI420(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; -#if defined(HAS_ARGB4444TOYROW_NEON) - void (*ARGB4444ToUVRow)(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int width) = ARGB4444ToUVRow_C; - void (*ARGB4444ToYRow)(const uint8* src_argb4444, uint8* dst_y, int pix) = - ARGB4444ToYRow_C; -#else - void (*ARGB4444ToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - ARGB4444ToARGBRow_C; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; -#endif - if (!src_argb4444 || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb4444 = src_argb4444 + (height - 1) * src_stride_argb4444; - src_stride_argb4444 = -src_stride_argb4444; - } - -#if defined(HAS_ARGB4444TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGB4444ToYRow = ARGB4444ToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGB4444ToYRow = ARGB4444ToYRow_NEON; - } - if (width >= 16) { - ARGB4444ToUVRow = ARGB4444ToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGB4444ToUVRow = ARGB4444ToUVRow_NEON; - } - } - } -#else // HAS_ARGB4444TOYROW_NEON - -#if defined(HAS_ARGB4444TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8) { - ARGB4444ToARGBRow = ARGB4444ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - ARGB4444ToARGBRow = ARGB4444ToARGBRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif // HAS_ARGBTOUVROW_SSSE3 -#endif // HAS_ARGB4444TOYROW_NEON - - { -#if !defined(HAS_ARGB4444TOYROW_NEON) - // Allocate 2 rows of ARGB. - const int kRowSize = (width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); -#endif - - for (y = 0; y < height - 1; y += 2) { -#if defined(HAS_ARGB4444TOYROW_NEON) - ARGB4444ToUVRow(src_argb4444, src_stride_argb4444, dst_u, dst_v, width); - ARGB4444ToYRow(src_argb4444, dst_y, width); - ARGB4444ToYRow(src_argb4444 + src_stride_argb4444, dst_y + dst_stride_y, - width); -#else - ARGB4444ToARGBRow(src_argb4444, row, width); - ARGB4444ToARGBRow(src_argb4444 + src_stride_argb4444, row + kRowSize, - width); - ARGBToUVRow(row, kRowSize, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - ARGBToYRow(row + kRowSize, dst_y + dst_stride_y, width); -#endif - src_argb4444 += src_stride_argb4444 * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { -#if defined(HAS_ARGB4444TOYROW_NEON) - ARGB4444ToUVRow(src_argb4444, 0, dst_u, dst_v, width); - ARGB4444ToYRow(src_argb4444, dst_y, width); -#else - ARGB4444ToARGBRow(src_argb4444, row, width); - ARGBToUVRow(row, 0, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); -#endif - } -#if !defined(HAS_ARGB4444TOYROW_NEON) - free_aligned_buffer_64(row); -#endif - } - return 0; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert_argb.cc b/src/main/jni/libyuv/source/convert_argb.cc deleted file mode 100644 index ac0bc3d15..000000000 --- a/src/main/jni/libyuv/source/convert_argb.cc +++ /dev/null @@ -1,938 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/convert_argb.h" - -#include "libyuv/cpu_id.h" -#include "libyuv/format_conversion.h" -#ifdef HAVE_JPEG -#include "libyuv/mjpeg_decoder.h" -#endif -#include "libyuv/rotate_argb.h" -#include "libyuv/row.h" -#include "libyuv/video_common.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Copy ARGB with optional flipping -LIBYUV_API -int ARGBCopy(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - if (!src_argb || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - - CopyPlane(src_argb, src_stride_argb, dst_argb, dst_stride_argb, - width * 4, height); - return 0; -} - -// Convert I444 to ARGB. -LIBYUV_API -int I444ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*I444ToARGBRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I444ToARGBRow_C; - if (!src_y || !src_u || !src_v || - !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u == width && - src_stride_v == width && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_argb = 0; - } -#if defined(HAS_I444TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I444ToARGBRow = I444ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I444ToARGBRow = I444ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - I444ToARGBRow = I444ToARGBRow_SSSE3; - } - } - } -#elif defined(HAS_I444TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I444ToARGBRow = I444ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I444ToARGBRow = I444ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I444ToARGBRow(src_y, src_u, src_v, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - } - return 0; -} - -// Convert I422 to ARGB. -LIBYUV_API -int I422ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*I422ToARGBRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToARGBRow_C; - if (!src_y || !src_u || !src_v || - !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 2 == width && - src_stride_v * 2 == width && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_argb = 0; - } -#if defined(HAS_I422TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToARGBRow = I422ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - I422ToARGBRow = I422ToARGBRow_SSSE3; - } - } - } -#endif -#if defined(HAS_I422TOARGBROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 16) { - I422ToARGBRow = I422ToARGBRow_Any_AVX2; - if (IS_ALIGNED(width, 16)) { - I422ToARGBRow = I422ToARGBRow_AVX2; - } - } -#endif -#if defined(HAS_I422TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToARGBRow = I422ToARGBRow_NEON; - } - } -#endif -#if defined(HAS_I422TOARGBROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_y, 4) && IS_ALIGNED(src_stride_y, 4) && - IS_ALIGNED(src_u, 2) && IS_ALIGNED(src_stride_u, 2) && - IS_ALIGNED(src_v, 2) && IS_ALIGNED(src_stride_v, 2) && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride_argb, 4)) { - I422ToARGBRow = I422ToARGBRow_MIPS_DSPR2; - } -#endif - - for (y = 0; y < height; ++y) { - I422ToARGBRow(src_y, src_u, src_v, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - } - return 0; -} - -// Convert I411 to ARGB. -LIBYUV_API -int I411ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*I411ToARGBRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I411ToARGBRow_C; - if (!src_y || !src_u || !src_v || - !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 4 == width && - src_stride_v * 4 == width && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_argb = 0; - } -#if defined(HAS_I411TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I411ToARGBRow = I411ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I411ToARGBRow = I411ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - I411ToARGBRow = I411ToARGBRow_SSSE3; - } - } - } -#elif defined(HAS_I411TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I411ToARGBRow = I411ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I411ToARGBRow = I411ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I411ToARGBRow(src_y, src_u, src_v, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - } - return 0; -} - -// Convert I400 to ARGB. -LIBYUV_API -int I400ToARGB_Reference(const uint8* src_y, int src_stride_y, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*YToARGBRow)(const uint8* y_buf, - uint8* rgb_buf, - int width) = YToARGBRow_C; - if (!src_y || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_y == width && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_y = dst_stride_argb = 0; - } -#if defined(HAS_YTOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8 && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - YToARGBRow = YToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - YToARGBRow = YToARGBRow_SSE2; - } - } -#elif defined(HAS_YTOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - YToARGBRow = YToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - YToARGBRow = YToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - YToARGBRow(src_y, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - } - return 0; -} - -// Convert I400 to ARGB. -LIBYUV_API -int I400ToARGB(const uint8* src_y, int src_stride_y, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*I400ToARGBRow)(const uint8* src_y, uint8* dst_argb, int pix) = - I400ToARGBRow_C; - if (!src_y || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - // Coalesce rows. - if (src_stride_y == width && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_y = dst_stride_argb = 0; - } -#if defined(HAS_I400TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8) { - I400ToARGBRow = I400ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - I400ToARGBRow = I400ToARGBRow_Unaligned_SSE2; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - I400ToARGBRow = I400ToARGBRow_SSE2; - } - } - } -#elif defined(HAS_I400TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I400ToARGBRow = I400ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I400ToARGBRow = I400ToARGBRow_NEON; - } - } -#endif - for (y = 0; y < height; ++y) { - I400ToARGBRow(src_y, dst_argb, width); - src_y += src_stride_y; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Shuffle table for converting BGRA to ARGB. -static uvec8 kShuffleMaskBGRAToARGB = { - 3u, 2u, 1u, 0u, 7u, 6u, 5u, 4u, 11u, 10u, 9u, 8u, 15u, 14u, 13u, 12u -}; - -// Shuffle table for converting ABGR to ARGB. -static uvec8 kShuffleMaskABGRToARGB = { - 2u, 1u, 0u, 3u, 6u, 5u, 4u, 7u, 10u, 9u, 8u, 11u, 14u, 13u, 12u, 15u -}; - -// Shuffle table for converting RGBA to ARGB. -static uvec8 kShuffleMaskRGBAToARGB = { - 1u, 2u, 3u, 0u, 5u, 6u, 7u, 4u, 9u, 10u, 11u, 8u, 13u, 14u, 15u, 12u -}; - -// Convert BGRA to ARGB. -LIBYUV_API -int BGRAToARGB(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - return ARGBShuffle(src_bgra, src_stride_bgra, - dst_argb, dst_stride_argb, - (const uint8*)(&kShuffleMaskBGRAToARGB), - width, height); -} - -// Convert ARGB to BGRA (same as BGRAToARGB). -LIBYUV_API -int ARGBToBGRA(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - return ARGBShuffle(src_bgra, src_stride_bgra, - dst_argb, dst_stride_argb, - (const uint8*)(&kShuffleMaskBGRAToARGB), - width, height); -} - -// Convert ABGR to ARGB. -LIBYUV_API -int ABGRToARGB(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - return ARGBShuffle(src_abgr, src_stride_abgr, - dst_argb, dst_stride_argb, - (const uint8*)(&kShuffleMaskABGRToARGB), - width, height); -} - -// Convert ARGB to ABGR to (same as ABGRToARGB). -LIBYUV_API -int ARGBToABGR(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - return ARGBShuffle(src_abgr, src_stride_abgr, - dst_argb, dst_stride_argb, - (const uint8*)(&kShuffleMaskABGRToARGB), - width, height); -} - -// Convert RGBA to ARGB. -LIBYUV_API -int RGBAToARGB(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - return ARGBShuffle(src_rgba, src_stride_rgba, - dst_argb, dst_stride_argb, - (const uint8*)(&kShuffleMaskRGBAToARGB), - width, height); -} - -// Convert RGB24 to ARGB. -LIBYUV_API -int RGB24ToARGB(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*RGB24ToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - RGB24ToARGBRow_C; - if (!src_rgb24 || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_rgb24 = src_rgb24 + (height - 1) * src_stride_rgb24; - src_stride_rgb24 = -src_stride_rgb24; - } - // Coalesce rows. - if (src_stride_rgb24 == width * 3 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_rgb24 = dst_stride_argb = 0; - } -#if defined(HAS_RGB24TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16 && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - RGB24ToARGBRow = RGB24ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - RGB24ToARGBRow = RGB24ToARGBRow_SSSE3; - } - } -#elif defined(HAS_RGB24TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RGB24ToARGBRow = RGB24ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RGB24ToARGBRow = RGB24ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - RGB24ToARGBRow(src_rgb24, dst_argb, width); - src_rgb24 += src_stride_rgb24; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert RAW to ARGB. -LIBYUV_API -int RAWToARGB(const uint8* src_raw, int src_stride_raw, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*RAWToARGBRow)(const uint8* src_rgb, uint8* dst_argb, int pix) = - RAWToARGBRow_C; - if (!src_raw || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_raw = src_raw + (height - 1) * src_stride_raw; - src_stride_raw = -src_stride_raw; - } - // Coalesce rows. - if (src_stride_raw == width * 3 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_raw = dst_stride_argb = 0; - } -#if defined(HAS_RAWTOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16 && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - RAWToARGBRow = RAWToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - RAWToARGBRow = RAWToARGBRow_SSSE3; - } - } -#elif defined(HAS_RAWTOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RAWToARGBRow = RAWToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RAWToARGBRow = RAWToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - RAWToARGBRow(src_raw, dst_argb, width); - src_raw += src_stride_raw; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert RGB565 to ARGB. -LIBYUV_API -int RGB565ToARGB(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*RGB565ToARGBRow)(const uint8* src_rgb565, uint8* dst_argb, int pix) = - RGB565ToARGBRow_C; - if (!src_rgb565 || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_rgb565 = src_rgb565 + (height - 1) * src_stride_rgb565; - src_stride_rgb565 = -src_stride_rgb565; - } - // Coalesce rows. - if (src_stride_rgb565 == width * 2 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_rgb565 = dst_stride_argb = 0; - } -#if defined(HAS_RGB565TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8 && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - RGB565ToARGBRow = RGB565ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - RGB565ToARGBRow = RGB565ToARGBRow_SSE2; - } - } -#elif defined(HAS_RGB565TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - RGB565ToARGBRow = RGB565ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - RGB565ToARGBRow = RGB565ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - RGB565ToARGBRow(src_rgb565, dst_argb, width); - src_rgb565 += src_stride_rgb565; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert ARGB1555 to ARGB. -LIBYUV_API -int ARGB1555ToARGB(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGB1555ToARGBRow)(const uint8* src_argb1555, uint8* dst_argb, - int pix) = ARGB1555ToARGBRow_C; - if (!src_argb1555 || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb1555 = src_argb1555 + (height - 1) * src_stride_argb1555; - src_stride_argb1555 = -src_stride_argb1555; - } - // Coalesce rows. - if (src_stride_argb1555 == width * 2 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb1555 = dst_stride_argb = 0; - } -#if defined(HAS_ARGB1555TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8 && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGB1555ToARGBRow = ARGB1555ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - ARGB1555ToARGBRow = ARGB1555ToARGBRow_SSE2; - } - } -#elif defined(HAS_ARGB1555TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGB1555ToARGBRow = ARGB1555ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGB1555ToARGBRow = ARGB1555ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGB1555ToARGBRow(src_argb1555, dst_argb, width); - src_argb1555 += src_stride_argb1555; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert ARGB4444 to ARGB. -LIBYUV_API -int ARGB4444ToARGB(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGB4444ToARGBRow)(const uint8* src_argb4444, uint8* dst_argb, - int pix) = ARGB4444ToARGBRow_C; - if (!src_argb4444 || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb4444 = src_argb4444 + (height - 1) * src_stride_argb4444; - src_stride_argb4444 = -src_stride_argb4444; - } - // Coalesce rows. - if (src_stride_argb4444 == width * 2 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb4444 = dst_stride_argb = 0; - } -#if defined(HAS_ARGB4444TOARGBROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8 && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGB4444ToARGBRow = ARGB4444ToARGBRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - ARGB4444ToARGBRow = ARGB4444ToARGBRow_SSE2; - } - } -#elif defined(HAS_ARGB4444TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGB4444ToARGBRow = ARGB4444ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGB4444ToARGBRow = ARGB4444ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGB4444ToARGBRow(src_argb4444, dst_argb, width); - src_argb4444 += src_stride_argb4444; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert NV12 to ARGB. -LIBYUV_API -int NV12ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*NV12ToARGBRow)(const uint8* y_buf, - const uint8* uv_buf, - uint8* rgb_buf, - int width) = NV12ToARGBRow_C; - if (!src_y || !src_uv || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } -#if defined(HAS_NV12TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - NV12ToARGBRow = NV12ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - NV12ToARGBRow = NV12ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - NV12ToARGBRow = NV12ToARGBRow_SSSE3; - } - } - } -#elif defined(HAS_NV12TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - NV12ToARGBRow = NV12ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - NV12ToARGBRow = NV12ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - NV12ToARGBRow(src_y, src_uv, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - if (y & 1) { - src_uv += src_stride_uv; - } - } - return 0; -} - -// Convert NV21 to ARGB. -LIBYUV_API -int NV21ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*NV21ToARGBRow)(const uint8* y_buf, - const uint8* uv_buf, - uint8* rgb_buf, - int width) = NV21ToARGBRow_C; - if (!src_y || !src_uv || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } -#if defined(HAS_NV21TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - NV21ToARGBRow = NV21ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - NV21ToARGBRow = NV21ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - NV21ToARGBRow = NV21ToARGBRow_SSSE3; - } - } - } -#endif -#if defined(HAS_NV21TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - NV21ToARGBRow = NV21ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - NV21ToARGBRow = NV21ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - NV21ToARGBRow(src_y, src_uv, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - if (y & 1) { - src_uv += src_stride_uv; - } - } - return 0; -} - -// Convert M420 to ARGB. -LIBYUV_API -int M420ToARGB(const uint8* src_m420, int src_stride_m420, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*NV12ToARGBRow)(const uint8* y_buf, - const uint8* uv_buf, - uint8* rgb_buf, - int width) = NV12ToARGBRow_C; - if (!src_m420 || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } -#if defined(HAS_NV12TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - NV12ToARGBRow = NV12ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - NV12ToARGBRow = NV12ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - NV12ToARGBRow = NV12ToARGBRow_SSSE3; - } - } - } -#elif defined(HAS_NV12TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - NV12ToARGBRow = NV12ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - NV12ToARGBRow = NV12ToARGBRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - NV12ToARGBRow(src_m420, src_m420 + src_stride_m420 * 2, dst_argb, width); - NV12ToARGBRow(src_m420 + src_stride_m420, src_m420 + src_stride_m420 * 2, - dst_argb + dst_stride_argb, width); - dst_argb += dst_stride_argb * 2; - src_m420 += src_stride_m420 * 3; - } - if (height & 1) { - NV12ToARGBRow(src_m420, src_m420 + src_stride_m420 * 2, dst_argb, width); - } - return 0; -} - -// Convert YUY2 to ARGB. -LIBYUV_API -int YUY2ToARGB(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*YUY2ToARGBRow)(const uint8* src_yuy2, uint8* dst_argb, int pix) = - YUY2ToARGBRow_C; - if (!src_yuy2 || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_yuy2 = src_yuy2 + (height - 1) * src_stride_yuy2; - src_stride_yuy2 = -src_stride_yuy2; - } - // Coalesce rows. - if (src_stride_yuy2 == width * 2 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_yuy2 = dst_stride_argb = 0; - } -#if defined(HAS_YUY2TOARGBROW_SSSE3) - // Posix is 16, Windows is 8. - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - YUY2ToARGBRow = YUY2ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - YUY2ToARGBRow = YUY2ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_yuy2, 16) && IS_ALIGNED(src_stride_yuy2, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - YUY2ToARGBRow = YUY2ToARGBRow_SSSE3; - } - } - } -#elif defined(HAS_YUY2TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - YUY2ToARGBRow = YUY2ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - YUY2ToARGBRow = YUY2ToARGBRow_NEON; - } - } -#endif - for (y = 0; y < height; ++y) { - YUY2ToARGBRow(src_yuy2, dst_argb, width); - src_yuy2 += src_stride_yuy2; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert UYVY to ARGB. -LIBYUV_API -int UYVYToARGB(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*UYVYToARGBRow)(const uint8* src_uyvy, uint8* dst_argb, int pix) = - UYVYToARGBRow_C; - if (!src_uyvy || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_uyvy = src_uyvy + (height - 1) * src_stride_uyvy; - src_stride_uyvy = -src_stride_uyvy; - } - // Coalesce rows. - if (src_stride_uyvy == width * 2 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_uyvy = dst_stride_argb = 0; - } -#if defined(HAS_UYVYTOARGBROW_SSSE3) - // Posix is 16, Windows is 8. - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - UYVYToARGBRow = UYVYToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - UYVYToARGBRow = UYVYToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_uyvy, 16) && IS_ALIGNED(src_stride_uyvy, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - UYVYToARGBRow = UYVYToARGBRow_SSSE3; - } - } - } -#elif defined(HAS_UYVYTOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - UYVYToARGBRow = UYVYToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - UYVYToARGBRow = UYVYToARGBRow_NEON; - } - } -#endif - for (y = 0; y < height; ++y) { - UYVYToARGBRow(src_uyvy, dst_argb, width); - src_uyvy += src_stride_uyvy; - dst_argb += dst_stride_argb; - } - return 0; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert_from.cc b/src/main/jni/libyuv/source/convert_from.cc deleted file mode 100644 index c1a2f62f0..000000000 --- a/src/main/jni/libyuv/source/convert_from.cc +++ /dev/null @@ -1,1210 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/convert_from.h" - -#include "libyuv/basic_types.h" -#include "libyuv/convert.h" // For I420Copy -#include "libyuv/cpu_id.h" -#include "libyuv/format_conversion.h" -#include "libyuv/planar_functions.h" -#include "libyuv/rotate.h" -#include "libyuv/scale.h" // For ScalePlane() -#include "libyuv/video_common.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#define SUBSAMPLE(v, a, s) (v < 0) ? (-((-v + a) >> s)) : ((v + a) >> s) -static __inline int Abs(int v) { - return v >= 0 ? v : -v; -} - -// I420 To any I4xx YUV format with mirroring. -static int I420ToI4xx(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int src_y_width, int src_y_height, - int dst_uv_width, int dst_uv_height) { - const int dst_y_width = Abs(src_y_width); - const int dst_y_height = Abs(src_y_height); - const int src_uv_width = SUBSAMPLE(src_y_width, 1, 1); - const int src_uv_height = SUBSAMPLE(src_y_height, 1, 1); - if (src_y_width == 0 || src_y_height == 0 || - dst_uv_width <= 0 || dst_uv_height <= 0) { - return -1; - } - ScalePlane(src_y, src_stride_y, src_y_width, src_y_height, - dst_y, dst_stride_y, dst_y_width, dst_y_height, - kFilterBilinear); - ScalePlane(src_u, src_stride_u, src_uv_width, src_uv_height, - dst_u, dst_stride_u, dst_uv_width, dst_uv_height, - kFilterBilinear); - ScalePlane(src_v, src_stride_v, src_uv_width, src_uv_height, - dst_v, dst_stride_v, dst_uv_width, dst_uv_height, - kFilterBilinear); - return 0; -} - -// 420 chroma is 1/2 width, 1/2 height -// 422 chroma is 1/2 width, 1x height -LIBYUV_API -int I420ToI422(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - const int dst_uv_width = (Abs(width) + 1) >> 1; - const int dst_uv_height = Abs(height); - return I420ToI4xx(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height, - dst_uv_width, dst_uv_height); -} - -// 420 chroma is 1/2 width, 1/2 height -// 444 chroma is 1x width, 1x height -LIBYUV_API -int I420ToI444(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - const int dst_uv_width = Abs(width); - const int dst_uv_height = Abs(height); - return I420ToI4xx(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height, - dst_uv_width, dst_uv_height); -} - -// 420 chroma is 1/2 width, 1/2 height -// 411 chroma is 1/4 width, 1x height -LIBYUV_API -int I420ToI411(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - const int dst_uv_width = (Abs(width) + 3) >> 2; - const int dst_uv_height = Abs(height); - return I420ToI4xx(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height, - dst_uv_width, dst_uv_height); -} - -// Copy to I400. Source can be I420,422,444,400,NV12,NV21 -LIBYUV_API -int I400Copy(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height) { - if (!src_y || !dst_y || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - return 0; -} - -LIBYUV_API -int I422ToYUY2(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_yuy2, int dst_stride_yuy2, - int width, int height) { - int y; - void (*I422ToYUY2Row)(const uint8* src_y, const uint8* src_u, - const uint8* src_v, uint8* dst_yuy2, int width) = - I422ToYUY2Row_C; - if (!src_y || !src_u || !src_v || !dst_yuy2 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_yuy2 = dst_yuy2 + (height - 1) * dst_stride_yuy2; - dst_stride_yuy2 = -dst_stride_yuy2; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 2 == width && - src_stride_v * 2 == width && - dst_stride_yuy2 == width * 2) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_yuy2 = 0; - } -#if defined(HAS_I422TOYUY2ROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - I422ToYUY2Row = I422ToYUY2Row_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - I422ToYUY2Row = I422ToYUY2Row_SSE2; - } - } -#elif defined(HAS_I422TOYUY2ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - I422ToYUY2Row = I422ToYUY2Row_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToYUY2Row = I422ToYUY2Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToYUY2Row(src_y, src_u, src_v, dst_yuy2, width); - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - dst_yuy2 += dst_stride_yuy2; - } - return 0; -} - -LIBYUV_API -int I420ToYUY2(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_yuy2, int dst_stride_yuy2, - int width, int height) { - int y; - void (*I422ToYUY2Row)(const uint8* src_y, const uint8* src_u, - const uint8* src_v, uint8* dst_yuy2, int width) = - I422ToYUY2Row_C; - if (!src_y || !src_u || !src_v || !dst_yuy2 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_yuy2 = dst_yuy2 + (height - 1) * dst_stride_yuy2; - dst_stride_yuy2 = -dst_stride_yuy2; - } -#if defined(HAS_I422TOYUY2ROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - I422ToYUY2Row = I422ToYUY2Row_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - I422ToYUY2Row = I422ToYUY2Row_SSE2; - } - } -#elif defined(HAS_I422TOYUY2ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - I422ToYUY2Row = I422ToYUY2Row_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToYUY2Row = I422ToYUY2Row_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - I422ToYUY2Row(src_y, src_u, src_v, dst_yuy2, width); - I422ToYUY2Row(src_y + src_stride_y, src_u, src_v, - dst_yuy2 + dst_stride_yuy2, width); - src_y += src_stride_y * 2; - src_u += src_stride_u; - src_v += src_stride_v; - dst_yuy2 += dst_stride_yuy2 * 2; - } - if (height & 1) { - I422ToYUY2Row(src_y, src_u, src_v, dst_yuy2, width); - } - return 0; -} - -LIBYUV_API -int I422ToUYVY(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_uyvy, int dst_stride_uyvy, - int width, int height) { - int y; - void (*I422ToUYVYRow)(const uint8* src_y, const uint8* src_u, - const uint8* src_v, uint8* dst_uyvy, int width) = - I422ToUYVYRow_C; - if (!src_y || !src_u || !src_v || !dst_uyvy || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_uyvy = dst_uyvy + (height - 1) * dst_stride_uyvy; - dst_stride_uyvy = -dst_stride_uyvy; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 2 == width && - src_stride_v * 2 == width && - dst_stride_uyvy == width * 2) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_uyvy = 0; - } -#if defined(HAS_I422TOUYVYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - I422ToUYVYRow = I422ToUYVYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - I422ToUYVYRow = I422ToUYVYRow_SSE2; - } - } -#elif defined(HAS_I422TOUYVYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - I422ToUYVYRow = I422ToUYVYRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToUYVYRow = I422ToUYVYRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToUYVYRow(src_y, src_u, src_v, dst_uyvy, width); - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - dst_uyvy += dst_stride_uyvy; - } - return 0; -} - -LIBYUV_API -int I420ToUYVY(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_uyvy, int dst_stride_uyvy, - int width, int height) { - int y; - void (*I422ToUYVYRow)(const uint8* src_y, const uint8* src_u, - const uint8* src_v, uint8* dst_uyvy, int width) = - I422ToUYVYRow_C; - if (!src_y || !src_u || !src_v || !dst_uyvy || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_uyvy = dst_uyvy + (height - 1) * dst_stride_uyvy; - dst_stride_uyvy = -dst_stride_uyvy; - } -#if defined(HAS_I422TOUYVYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - I422ToUYVYRow = I422ToUYVYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - I422ToUYVYRow = I422ToUYVYRow_SSE2; - } - } -#elif defined(HAS_I422TOUYVYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - I422ToUYVYRow = I422ToUYVYRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToUYVYRow = I422ToUYVYRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - I422ToUYVYRow(src_y, src_u, src_v, dst_uyvy, width); - I422ToUYVYRow(src_y + src_stride_y, src_u, src_v, - dst_uyvy + dst_stride_uyvy, width); - src_y += src_stride_y * 2; - src_u += src_stride_u; - src_v += src_stride_v; - dst_uyvy += dst_stride_uyvy * 2; - } - if (height & 1) { - I422ToUYVYRow(src_y, src_u, src_v, dst_uyvy, width); - } - return 0; -} - -LIBYUV_API -int I420ToNV12(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_uv, int dst_stride_uv, - int width, int height) { - int y; - void (*MergeUVRow_)(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) = MergeUVRow_C; - // Coalesce rows. - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - if (!src_y || !src_u || !src_v || !dst_y || !dst_uv || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - dst_y = dst_y + (height - 1) * dst_stride_y; - dst_uv = dst_uv + (halfheight - 1) * dst_stride_uv; - dst_stride_y = -dst_stride_y; - dst_stride_uv = -dst_stride_uv; - } - if (src_stride_y == width && - dst_stride_y == width) { - width *= height; - height = 1; - src_stride_y = dst_stride_y = 0; - } - // Coalesce rows. - if (src_stride_u == halfwidth && - src_stride_v == halfwidth && - dst_stride_uv == halfwidth * 2) { - halfwidth *= halfheight; - halfheight = 1; - src_stride_u = src_stride_v = dst_stride_uv = 0; - } -#if defined(HAS_MERGEUVROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && halfwidth >= 16) { - MergeUVRow_ = MergeUVRow_Any_SSE2; - if (IS_ALIGNED(halfwidth, 16)) { - MergeUVRow_ = MergeUVRow_Unaligned_SSE2; - if (IS_ALIGNED(src_u, 16) && IS_ALIGNED(src_stride_u, 16) && - IS_ALIGNED(src_v, 16) && IS_ALIGNED(src_stride_v, 16) && - IS_ALIGNED(dst_uv, 16) && IS_ALIGNED(dst_stride_uv, 16)) { - MergeUVRow_ = MergeUVRow_SSE2; - } - } - } -#endif -#if defined(HAS_MERGEUVROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && halfwidth >= 32) { - MergeUVRow_ = MergeUVRow_Any_AVX2; - if (IS_ALIGNED(halfwidth, 32)) { - MergeUVRow_ = MergeUVRow_AVX2; - } - } -#endif -#if defined(HAS_MERGEUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && halfwidth >= 16) { - MergeUVRow_ = MergeUVRow_Any_NEON; - if (IS_ALIGNED(halfwidth, 16)) { - MergeUVRow_ = MergeUVRow_NEON; - } - } -#endif - - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - for (y = 0; y < halfheight; ++y) { - // Merge a row of U and V into a row of UV. - MergeUVRow_(src_u, src_v, dst_uv, halfwidth); - src_u += src_stride_u; - src_v += src_stride_v; - dst_uv += dst_stride_uv; - } - return 0; -} - -LIBYUV_API -int I420ToNV21(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_vu, int dst_stride_vu, - int width, int height) { - return I420ToNV12(src_y, src_stride_y, - src_v, src_stride_v, - src_u, src_stride_u, - dst_y, src_stride_y, - dst_vu, dst_stride_vu, - width, height); -} - -// Convert I420 to ARGB. -LIBYUV_API -int I420ToARGB(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*I422ToARGBRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToARGBRow_C; - if (!src_y || !src_u || !src_v || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } -#if defined(HAS_I422TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToARGBRow = I422ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - I422ToARGBRow = I422ToARGBRow_SSSE3; - } - } - } -#endif -#if defined(HAS_I422TOARGBROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 16) { - I422ToARGBRow = I422ToARGBRow_Any_AVX2; - if (IS_ALIGNED(width, 16)) { - I422ToARGBRow = I422ToARGBRow_AVX2; - } - } -#endif -#if defined(HAS_I422TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToARGBRow = I422ToARGBRow_NEON; - } - } -#endif -#if defined(HAS_I422TOARGBROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_y, 4) && IS_ALIGNED(src_stride_y, 4) && - IS_ALIGNED(src_u, 2) && IS_ALIGNED(src_stride_u, 2) && - IS_ALIGNED(src_v, 2) && IS_ALIGNED(src_stride_v, 2) && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride_argb, 4)) { - I422ToARGBRow = I422ToARGBRow_MIPS_DSPR2; - } -#endif - - for (y = 0; y < height; ++y) { - I422ToARGBRow(src_y, src_u, src_v, dst_argb, width); - dst_argb += dst_stride_argb; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to BGRA. -LIBYUV_API -int I420ToBGRA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_bgra, int dst_stride_bgra, - int width, int height) { - int y; - void (*I422ToBGRARow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToBGRARow_C; - if (!src_y || !src_u || !src_v || !dst_bgra || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_bgra = dst_bgra + (height - 1) * dst_stride_bgra; - dst_stride_bgra = -dst_stride_bgra; - } -#if defined(HAS_I422TOBGRAROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToBGRARow = I422ToBGRARow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToBGRARow = I422ToBGRARow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_bgra, 16) && IS_ALIGNED(dst_stride_bgra, 16)) { - I422ToBGRARow = I422ToBGRARow_SSSE3; - } - } - } -#elif defined(HAS_I422TOBGRAROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToBGRARow = I422ToBGRARow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToBGRARow = I422ToBGRARow_NEON; - } - } -#elif defined(HAS_I422TOBGRAROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_y, 4) && IS_ALIGNED(src_stride_y, 4) && - IS_ALIGNED(src_u, 2) && IS_ALIGNED(src_stride_u, 2) && - IS_ALIGNED(src_v, 2) && IS_ALIGNED(src_stride_v, 2) && - IS_ALIGNED(dst_bgra, 4) && IS_ALIGNED(dst_stride_bgra, 4)) { - I422ToBGRARow = I422ToBGRARow_MIPS_DSPR2; - } -#endif - - for (y = 0; y < height; ++y) { - I422ToBGRARow(src_y, src_u, src_v, dst_bgra, width); - dst_bgra += dst_stride_bgra; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to ABGR. -LIBYUV_API -int I420ToABGR(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_abgr, int dst_stride_abgr, - int width, int height) { - int y; - void (*I422ToABGRRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToABGRRow_C; - if (!src_y || !src_u || !src_v || !dst_abgr || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_abgr = dst_abgr + (height - 1) * dst_stride_abgr; - dst_stride_abgr = -dst_stride_abgr; - } -#if defined(HAS_I422TOABGRROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToABGRRow = I422ToABGRRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToABGRRow = I422ToABGRRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_abgr, 16) && IS_ALIGNED(dst_stride_abgr, 16)) { - I422ToABGRRow = I422ToABGRRow_SSSE3; - } - } - } -#elif defined(HAS_I422TOABGRROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToABGRRow = I422ToABGRRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToABGRRow = I422ToABGRRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToABGRRow(src_y, src_u, src_v, dst_abgr, width); - dst_abgr += dst_stride_abgr; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to RGBA. -LIBYUV_API -int I420ToRGBA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_rgba, int dst_stride_rgba, - int width, int height) { - int y; - void (*I422ToRGBARow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToRGBARow_C; - if (!src_y || !src_u || !src_v || !dst_rgba || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_rgba = dst_rgba + (height - 1) * dst_stride_rgba; - dst_stride_rgba = -dst_stride_rgba; - } -#if defined(HAS_I422TORGBAROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToRGBARow = I422ToRGBARow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToRGBARow = I422ToRGBARow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_rgba, 16) && IS_ALIGNED(dst_stride_rgba, 16)) { - I422ToRGBARow = I422ToRGBARow_SSSE3; - } - } - } -#elif defined(HAS_I422TORGBAROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToRGBARow = I422ToRGBARow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToRGBARow = I422ToRGBARow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToRGBARow(src_y, src_u, src_v, dst_rgba, width); - dst_rgba += dst_stride_rgba; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to RGB24. -LIBYUV_API -int I420ToRGB24(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_rgb24, int dst_stride_rgb24, - int width, int height) { - int y; - void (*I422ToRGB24Row)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToRGB24Row_C; - if (!src_y || !src_u || !src_v || !dst_rgb24 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_rgb24 = dst_rgb24 + (height - 1) * dst_stride_rgb24; - dst_stride_rgb24 = -dst_stride_rgb24; - } -#if defined(HAS_I422TORGB24ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToRGB24Row = I422ToRGB24Row_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToRGB24Row = I422ToRGB24Row_SSSE3; - } - } -#elif defined(HAS_I422TORGB24ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToRGB24Row = I422ToRGB24Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToRGB24Row = I422ToRGB24Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToRGB24Row(src_y, src_u, src_v, dst_rgb24, width); - dst_rgb24 += dst_stride_rgb24; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to RAW. -LIBYUV_API -int I420ToRAW(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_raw, int dst_stride_raw, - int width, int height) { - int y; - void (*I422ToRAWRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToRAWRow_C; - if (!src_y || !src_u || !src_v || !dst_raw || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_raw = dst_raw + (height - 1) * dst_stride_raw; - dst_stride_raw = -dst_stride_raw; - } -#if defined(HAS_I422TORAWROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToRAWRow = I422ToRAWRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToRAWRow = I422ToRAWRow_SSSE3; - } - } -#elif defined(HAS_I422TORAWROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToRAWRow = I422ToRAWRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToRAWRow = I422ToRAWRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToRAWRow(src_y, src_u, src_v, dst_raw, width); - dst_raw += dst_stride_raw; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to ARGB1555. -LIBYUV_API -int I420ToARGB1555(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb1555, int dst_stride_argb1555, - int width, int height) { - int y; - void (*I422ToARGB1555Row)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToARGB1555Row_C; - if (!src_y || !src_u || !src_v || !dst_argb1555 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb1555 = dst_argb1555 + (height - 1) * dst_stride_argb1555; - dst_stride_argb1555 = -dst_stride_argb1555; - } -#if defined(HAS_I422TOARGB1555ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToARGB1555Row = I422ToARGB1555Row_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToARGB1555Row = I422ToARGB1555Row_SSSE3; - } - } -#elif defined(HAS_I422TOARGB1555ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToARGB1555Row = I422ToARGB1555Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToARGB1555Row = I422ToARGB1555Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToARGB1555Row(src_y, src_u, src_v, dst_argb1555, width); - dst_argb1555 += dst_stride_argb1555; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - - -// Convert I420 to ARGB4444. -LIBYUV_API -int I420ToARGB4444(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_argb4444, int dst_stride_argb4444, - int width, int height) { - int y; - void (*I422ToARGB4444Row)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToARGB4444Row_C; - if (!src_y || !src_u || !src_v || !dst_argb4444 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb4444 = dst_argb4444 + (height - 1) * dst_stride_argb4444; - dst_stride_argb4444 = -dst_stride_argb4444; - } -#if defined(HAS_I422TOARGB4444ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToARGB4444Row = I422ToARGB4444Row_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToARGB4444Row = I422ToARGB4444Row_SSSE3; - } - } -#elif defined(HAS_I422TOARGB4444ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToARGB4444Row = I422ToARGB4444Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToARGB4444Row = I422ToARGB4444Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToARGB4444Row(src_y, src_u, src_v, dst_argb4444, width); - dst_argb4444 += dst_stride_argb4444; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to RGB565. -LIBYUV_API -int I420ToRGB565(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height) { - int y; - void (*I422ToRGB565Row)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToRGB565Row_C; - if (!src_y || !src_u || !src_v || !dst_rgb565 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_rgb565 = dst_rgb565 + (height - 1) * dst_stride_rgb565; - dst_stride_rgb565 = -dst_stride_rgb565; - } -#if defined(HAS_I422TORGB565ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToRGB565Row = I422ToRGB565Row_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToRGB565Row = I422ToRGB565Row_SSSE3; - } - } -#elif defined(HAS_I422TORGB565ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToRGB565Row = I422ToRGB565Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToRGB565Row = I422ToRGB565Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToRGB565Row(src_y, src_u, src_v, dst_rgb565, width); - dst_rgb565 += dst_stride_rgb565; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - return 0; -} - -// Convert I420 to specified format -LIBYUV_API -int ConvertFromI420(const uint8* y, int y_stride, - const uint8* u, int u_stride, - const uint8* v, int v_stride, - uint8* dst_sample, int dst_sample_stride, - int width, int height, - uint32 fourcc) { - uint32 format = CanonicalFourCC(fourcc); - int r = 0; - if (!y || !u|| !v || !dst_sample || - width <= 0 || height == 0) { - return -1; - } - switch (format) { - // Single plane formats - case FOURCC_YUY2: - r = I420ToYUY2(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 2, - width, height); - break; - case FOURCC_UYVY: - r = I420ToUYVY(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 2, - width, height); - break; - case FOURCC_RGBP: - r = I420ToRGB565(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 2, - width, height); - break; - case FOURCC_RGBO: - r = I420ToARGB1555(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 2, - width, height); - break; - case FOURCC_R444: - r = I420ToARGB4444(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 2, - width, height); - break; - case FOURCC_24BG: - r = I420ToRGB24(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 3, - width, height); - break; - case FOURCC_RAW: - r = I420ToRAW(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 3, - width, height); - break; - case FOURCC_ARGB: - r = I420ToARGB(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 4, - width, height); - break; - case FOURCC_BGRA: - r = I420ToBGRA(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 4, - width, height); - break; - case FOURCC_ABGR: - r = I420ToABGR(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 4, - width, height); - break; - case FOURCC_RGBA: - r = I420ToRGBA(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width * 4, - width, height); - break; - case FOURCC_BGGR: - r = I420ToBayerBGGR(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - case FOURCC_GBRG: - r = I420ToBayerGBRG(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - case FOURCC_GRBG: - r = I420ToBayerGRBG(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - case FOURCC_RGGB: - r = I420ToBayerRGGB(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - case FOURCC_I400: - r = I400Copy(y, y_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - case FOURCC_NV12: { - uint8* dst_uv = dst_sample + width * height; - r = I420ToNV12(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - dst_uv, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - } - case FOURCC_NV21: { - uint8* dst_vu = dst_sample + width * height; - r = I420ToNV21(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, - dst_sample_stride ? dst_sample_stride : width, - dst_vu, - dst_sample_stride ? dst_sample_stride : width, - width, height); - break; - } - // TODO(fbarchard): Add M420 and Q420. - // Triplanar formats - // TODO(fbarchard): halfstride instead of halfwidth - case FOURCC_I420: - case FOURCC_YU12: - case FOURCC_YV12: { - int halfwidth = (width + 1) / 2; - int halfheight = (height + 1) / 2; - uint8* dst_u; - uint8* dst_v; - if (format == FOURCC_YV12) { - dst_v = dst_sample + width * height; - dst_u = dst_v + halfwidth * halfheight; - } else { - dst_u = dst_sample + width * height; - dst_v = dst_u + halfwidth * halfheight; - } - r = I420Copy(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, width, - dst_u, halfwidth, - dst_v, halfwidth, - width, height); - break; - } - case FOURCC_I422: - case FOURCC_YV16: { - int halfwidth = (width + 1) / 2; - uint8* dst_u; - uint8* dst_v; - if (format == FOURCC_YV16) { - dst_v = dst_sample + width * height; - dst_u = dst_v + halfwidth * height; - } else { - dst_u = dst_sample + width * height; - dst_v = dst_u + halfwidth * height; - } - r = I420ToI422(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, width, - dst_u, halfwidth, - dst_v, halfwidth, - width, height); - break; - } - case FOURCC_I444: - case FOURCC_YV24: { - uint8* dst_u; - uint8* dst_v; - if (format == FOURCC_YV24) { - dst_v = dst_sample + width * height; - dst_u = dst_v + width * height; - } else { - dst_u = dst_sample + width * height; - dst_v = dst_u + width * height; - } - r = I420ToI444(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, width, - dst_u, width, - dst_v, width, - width, height); - break; - } - case FOURCC_I411: { - int quarterwidth = (width + 3) / 4; - uint8* dst_u = dst_sample + width * height; - uint8* dst_v = dst_u + quarterwidth * height; - r = I420ToI411(y, y_stride, - u, u_stride, - v, v_stride, - dst_sample, width, - dst_u, quarterwidth, - dst_v, quarterwidth, - width, height); - break; - } - - // Formats not supported - MJPG, biplanar, some rgb formats. - default: - return -1; // unknown fourcc - return failure code. - } - return r; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert_from_argb.cc b/src/main/jni/libyuv/source/convert_from_argb.cc deleted file mode 100644 index de461ddb0..000000000 --- a/src/main/jni/libyuv/source/convert_from_argb.cc +++ /dev/null @@ -1,1133 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/convert_from_argb.h" - -#include "libyuv/basic_types.h" -#include "libyuv/cpu_id.h" -#include "libyuv/format_conversion.h" -#include "libyuv/planar_functions.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// ARGB little endian (bgra in memory) to I444 -LIBYUV_API -int ARGBToI444(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - void (*ARGBToUV444Row)(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) = ARGBToUV444Row_C; - if (!src_argb || !dst_y || !dst_u || !dst_v || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_y == width && - dst_stride_u == width && - dst_stride_v == width) { - width *= height; - height = 1; - src_stride_argb = dst_stride_y = dst_stride_u = dst_stride_v = 0; - } -#if defined(HAS_ARGBTOUV444ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUV444Row = ARGBToUV444Row_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUV444Row = ARGBToUV444Row_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUV444Row = ARGBToUV444Row_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOUV444ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToUV444Row = ARGBToUV444Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToUV444Row = ARGBToUV444Row_NEON; - } - } -#endif -#if defined(HAS_ARGBTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } - -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToUV444Row(src_argb, dst_u, dst_v, width); - ARGBToYRow(src_argb, dst_y, width); - src_argb += src_stride_argb; - dst_y += dst_stride_y; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - return 0; -} - -// ARGB little endian (bgra in memory) to I422 -LIBYUV_API -int ARGBToI422(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*ARGBToUV422Row)(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) = ARGBToUV422Row_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - if (!src_argb || !dst_y || !dst_u || !dst_v || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_y == width && - dst_stride_u * 2 == width && - dst_stride_v * 2 == width) { - width *= height; - height = 1; - src_stride_argb = dst_stride_y = dst_stride_u = dst_stride_v = 0; - } -#if defined(HAS_ARGBTOUV422ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUV422Row = ARGBToUV422Row_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUV422Row = ARGBToUV422Row_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUV422Row = ARGBToUV422Row_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOUV422ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUV422Row = ARGBToUV422Row_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUV422Row = ARGBToUV422Row_NEON; - } - } -#endif - -#if defined(HAS_ARGBTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToUV422Row(src_argb, dst_u, dst_v, width); - ARGBToYRow(src_argb, dst_y, width); - src_argb += src_stride_argb; - dst_y += dst_stride_y; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - return 0; -} - -// ARGB little endian (bgra in memory) to I411 -LIBYUV_API -int ARGBToI411(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*ARGBToUV411Row)(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) = ARGBToUV411Row_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - if (!src_argb || !dst_y || !dst_u || !dst_v || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_y == width && - dst_stride_u * 4 == width && - dst_stride_v * 4 == width) { - width *= height; - height = 1; - src_stride_argb = dst_stride_y = dst_stride_u = dst_stride_v = 0; - } -#if defined(HAS_ARGBTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif -#if defined(HAS_ARGBTOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - ARGBToYRow = ARGBToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - ARGBToYRow = ARGBToYRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif -#if defined(HAS_ARGBTOUV411ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 32) { - ARGBToUV411Row = ARGBToUV411Row_Any_NEON; - if (IS_ALIGNED(width, 32)) { - ARGBToUV411Row = ARGBToUV411Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToUV411Row(src_argb, dst_u, dst_v, width); - ARGBToYRow(src_argb, dst_y, width); - src_argb += src_stride_argb; - dst_y += dst_stride_y; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - return 0; -} - -LIBYUV_API -int ARGBToNV12(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_uv, int dst_stride_uv, - int width, int height) { - int y; - int halfwidth = (width + 1) >> 1; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - void (*MergeUVRow_)(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) = MergeUVRow_C; - if (!src_argb || - !dst_y || !dst_uv || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } -#if defined(HAS_ARGBTOYROW_SSSE3) && defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_Unaligned_SSSE3; - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } - } -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_NEON; - } - } -#endif -#if defined(HAS_MERGEUVROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && halfwidth >= 16) { - MergeUVRow_ = MergeUVRow_Any_SSE2; - if (IS_ALIGNED(halfwidth, 16)) { - MergeUVRow_ = MergeUVRow_Unaligned_SSE2; - if (IS_ALIGNED(dst_uv, 16) && IS_ALIGNED(dst_stride_uv, 16)) { - MergeUVRow_ = MergeUVRow_SSE2; - } - } - } -#endif -#if defined(HAS_MERGEUVROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && halfwidth >= 32) { - MergeUVRow_ = MergeUVRow_Any_AVX2; - if (IS_ALIGNED(halfwidth, 32)) { - MergeUVRow_ = MergeUVRow_AVX2; - } - } -#endif -#if defined(HAS_MERGEUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && halfwidth >= 16) { - MergeUVRow_ = MergeUVRow_Any_NEON; - if (IS_ALIGNED(halfwidth, 16)) { - MergeUVRow_ = MergeUVRow_NEON; - } - } -#endif - { - // Allocate a rows of uv. - align_buffer_64(row_u, ((halfwidth + 15) & ~15) * 2); - uint8* row_v = row_u + ((halfwidth + 15) & ~15); - - for (y = 0; y < height - 1; y += 2) { - ARGBToUVRow(src_argb, src_stride_argb, row_u, row_v, width); - MergeUVRow_(row_u, row_v, dst_uv, halfwidth); - ARGBToYRow(src_argb, dst_y, width); - ARGBToYRow(src_argb + src_stride_argb, dst_y + dst_stride_y, width); - src_argb += src_stride_argb * 2; - dst_y += dst_stride_y * 2; - dst_uv += dst_stride_uv; - } - if (height & 1) { - ARGBToUVRow(src_argb, 0, row_u, row_v, width); - MergeUVRow_(row_u, row_v, dst_uv, halfwidth); - ARGBToYRow(src_argb, dst_y, width); - } - free_aligned_buffer_64(row_u); - } - return 0; -} - -// Same as NV12 but U and V swapped. -LIBYUV_API -int ARGBToNV21(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - uint8* dst_uv, int dst_stride_uv, - int width, int height) { - int y; - int halfwidth = (width + 1) >> 1; - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - void (*MergeUVRow_)(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) = MergeUVRow_C; - if (!src_argb || - !dst_y || !dst_uv || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } -#if defined(HAS_ARGBTOYROW_SSSE3) && defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_Unaligned_SSSE3; - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUVRow = ARGBToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } - } -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_NEON; - } - } -#endif -#if defined(HAS_MERGEUVROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && halfwidth >= 16) { - MergeUVRow_ = MergeUVRow_Any_SSE2; - if (IS_ALIGNED(halfwidth, 16)) { - MergeUVRow_ = MergeUVRow_Unaligned_SSE2; - if (IS_ALIGNED(dst_uv, 16) && IS_ALIGNED(dst_stride_uv, 16)) { - MergeUVRow_ = MergeUVRow_SSE2; - } - } - } -#endif -#if defined(HAS_MERGEUVROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && halfwidth >= 32) { - MergeUVRow_ = MergeUVRow_Any_AVX2; - if (IS_ALIGNED(halfwidth, 32)) { - MergeUVRow_ = MergeUVRow_AVX2; - } - } -#endif -#if defined(HAS_MERGEUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && halfwidth >= 16) { - MergeUVRow_ = MergeUVRow_Any_NEON; - if (IS_ALIGNED(halfwidth, 16)) { - MergeUVRow_ = MergeUVRow_NEON; - } - } -#endif - { - // Allocate a rows of uv. - align_buffer_64(row_u, ((halfwidth + 15) & ~15) * 2); - uint8* row_v = row_u + ((halfwidth + 15) & ~15); - - for (y = 0; y < height - 1; y += 2) { - ARGBToUVRow(src_argb, src_stride_argb, row_u, row_v, width); - MergeUVRow_(row_v, row_u, dst_uv, halfwidth); - ARGBToYRow(src_argb, dst_y, width); - ARGBToYRow(src_argb + src_stride_argb, dst_y + dst_stride_y, width); - src_argb += src_stride_argb * 2; - dst_y += dst_stride_y * 2; - dst_uv += dst_stride_uv; - } - if (height & 1) { - ARGBToUVRow(src_argb, 0, row_u, row_v, width); - MergeUVRow_(row_v, row_u, dst_uv, halfwidth); - ARGBToYRow(src_argb, dst_y, width); - } - free_aligned_buffer_64(row_u); - } - return 0; -} - -// Convert ARGB to YUY2. -LIBYUV_API -int ARGBToYUY2(const uint8* src_argb, int src_stride_argb, - uint8* dst_yuy2, int dst_stride_yuy2, - int width, int height) { - int y; - void (*ARGBToUV422Row)(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) = ARGBToUV422Row_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - void (*I422ToYUY2Row)(const uint8* src_y, const uint8* src_u, - const uint8* src_v, uint8* dst_yuy2, int width) = I422ToYUY2Row_C; - - if (!src_argb || !dst_yuy2 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_yuy2 = dst_yuy2 + (height - 1) * dst_stride_yuy2; - dst_stride_yuy2 = -dst_stride_yuy2; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_yuy2 == width * 2) { - width *= height; - height = 1; - src_stride_argb = dst_stride_yuy2 = 0; - } -#if defined(HAS_ARGBTOUV422ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUV422Row = ARGBToUV422Row_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUV422Row = ARGBToUV422Row_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUV422Row = ARGBToUV422Row_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOUV422ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUV422Row = ARGBToUV422Row_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUV422Row = ARGBToUV422Row_NEON; - } - } -#endif -#if defined(HAS_ARGBTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif - -#if defined(HAS_I422TOYUY2ROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - I422ToYUY2Row = I422ToYUY2Row_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - I422ToYUY2Row = I422ToYUY2Row_SSE2; - } - } -#elif defined(HAS_I422TOYUY2ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - I422ToYUY2Row = I422ToYUY2Row_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToYUY2Row = I422ToYUY2Row_NEON; - } - } -#endif - - { - // Allocate a rows of yuv. - align_buffer_64(row_y, ((width + 63) & ~63) * 2); - uint8* row_u = row_y + ((width + 63) & ~63); - uint8* row_v = row_u + ((width + 63) & ~63) / 2; - - for (y = 0; y < height; ++y) { - ARGBToUV422Row(src_argb, row_u, row_v, width); - ARGBToYRow(src_argb, row_y, width); - I422ToYUY2Row(row_y, row_u, row_v, dst_yuy2, width); - src_argb += src_stride_argb; - dst_yuy2 += dst_stride_yuy2; - } - - free_aligned_buffer_64(row_y); - } - return 0; -} - -// Convert ARGB to UYVY. -LIBYUV_API -int ARGBToUYVY(const uint8* src_argb, int src_stride_argb, - uint8* dst_uyvy, int dst_stride_uyvy, - int width, int height) { - int y; - void (*ARGBToUV422Row)(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) = ARGBToUV422Row_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - void (*I422ToUYVYRow)(const uint8* src_y, const uint8* src_u, - const uint8* src_v, uint8* dst_uyvy, int width) = I422ToUYVYRow_C; - - if (!src_argb || !dst_uyvy || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_uyvy = dst_uyvy + (height - 1) * dst_stride_uyvy; - dst_stride_uyvy = -dst_stride_uyvy; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_uyvy == width * 2) { - width *= height; - height = 1; - src_stride_argb = dst_stride_uyvy = 0; - } -#if defined(HAS_ARGBTOUV422ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUV422Row = ARGBToUV422Row_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUV422Row = ARGBToUV422Row_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUV422Row = ARGBToUV422Row_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOUV422ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUV422Row = ARGBToUV422Row_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUV422Row = ARGBToUV422Row_NEON; - } - } -#endif -#if defined(HAS_ARGBTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif - -#if defined(HAS_I422TOUYVYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - I422ToUYVYRow = I422ToUYVYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - I422ToUYVYRow = I422ToUYVYRow_SSE2; - } - } -#elif defined(HAS_I422TOUYVYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - I422ToUYVYRow = I422ToUYVYRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToUYVYRow = I422ToUYVYRow_NEON; - } - } -#endif - - { - // Allocate a rows of yuv. - align_buffer_64(row_y, ((width + 63) & ~63) * 2); - uint8* row_u = row_y + ((width + 63) & ~63); - uint8* row_v = row_u + ((width + 63) & ~63) / 2; - - for (y = 0; y < height; ++y) { - ARGBToUV422Row(src_argb, row_u, row_v, width); - ARGBToYRow(src_argb, row_y, width); - I422ToUYVYRow(row_y, row_u, row_v, dst_uyvy, width); - src_argb += src_stride_argb; - dst_uyvy += dst_stride_uyvy; - } - - free_aligned_buffer_64(row_y); - } - return 0; -} - -// Convert ARGB to I400. -LIBYUV_API -int ARGBToI400(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - int width, int height) { - int y; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - if (!src_argb || !dst_y || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_y == width) { - width *= height; - height = 1; - src_stride_argb = dst_stride_y = 0; - } -#if defined(HAS_ARGBTOYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#endif -#if defined(HAS_ARGBTOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - ARGBToYRow = ARGBToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - ARGBToYRow = ARGBToYRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToYRow(src_argb, dst_y, width); - src_argb += src_stride_argb; - dst_y += dst_stride_y; - } - return 0; -} - -// Shuffle table for converting ARGB to RGBA. -static uvec8 kShuffleMaskARGBToRGBA = { - 3u, 0u, 1u, 2u, 7u, 4u, 5u, 6u, 11u, 8u, 9u, 10u, 15u, 12u, 13u, 14u -}; - -// Convert ARGB to RGBA. -LIBYUV_API -int ARGBToRGBA(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgba, int dst_stride_rgba, - int width, int height) { - return ARGBShuffle(src_argb, src_stride_argb, - dst_rgba, dst_stride_rgba, - (const uint8*)(&kShuffleMaskARGBToRGBA), - width, height); -} - -// Convert ARGB To RGB24. -LIBYUV_API -int ARGBToRGB24(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgb24, int dst_stride_rgb24, - int width, int height) { - int y; - void (*ARGBToRGB24Row)(const uint8* src_argb, uint8* dst_rgb, int pix) = - ARGBToRGB24Row_C; - if (!src_argb || !dst_rgb24 || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_rgb24 == width * 3) { - width *= height; - height = 1; - src_stride_argb = dst_stride_rgb24 = 0; - } -#if defined(HAS_ARGBTORGB24ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToRGB24Row = ARGBToRGB24Row_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToRGB24Row = ARGBToRGB24Row_SSSE3; - } - } -#elif defined(HAS_ARGBTORGB24ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToRGB24Row = ARGBToRGB24Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToRGB24Row = ARGBToRGB24Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToRGB24Row(src_argb, dst_rgb24, width); - src_argb += src_stride_argb; - dst_rgb24 += dst_stride_rgb24; - } - return 0; -} - -// Convert ARGB To RAW. -LIBYUV_API -int ARGBToRAW(const uint8* src_argb, int src_stride_argb, - uint8* dst_raw, int dst_stride_raw, - int width, int height) { - int y; - void (*ARGBToRAWRow)(const uint8* src_argb, uint8* dst_rgb, int pix) = - ARGBToRAWRow_C; - if (!src_argb || !dst_raw || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_raw == width * 3) { - width *= height; - height = 1; - src_stride_argb = dst_stride_raw = 0; - } -#if defined(HAS_ARGBTORAWROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToRAWRow = ARGBToRAWRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToRAWRow = ARGBToRAWRow_SSSE3; - } - } -#elif defined(HAS_ARGBTORAWROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToRAWRow = ARGBToRAWRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToRAWRow = ARGBToRAWRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToRAWRow(src_argb, dst_raw, width); - src_argb += src_stride_argb; - dst_raw += dst_stride_raw; - } - return 0; -} - -// Convert ARGB To RGB565. -LIBYUV_API -int ARGBToRGB565(const uint8* src_argb, int src_stride_argb, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height) { - int y; - void (*ARGBToRGB565Row)(const uint8* src_argb, uint8* dst_rgb, int pix) = - ARGBToRGB565Row_C; - if (!src_argb || !dst_rgb565 || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_rgb565 == width * 2) { - width *= height; - height = 1; - src_stride_argb = dst_stride_rgb565 = 0; - } -#if defined(HAS_ARGBTORGB565ROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToRGB565Row = ARGBToRGB565Row_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBToRGB565Row = ARGBToRGB565Row_SSE2; - } - } -#elif defined(HAS_ARGBTORGB565ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToRGB565Row = ARGBToRGB565Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToRGB565Row = ARGBToRGB565Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToRGB565Row(src_argb, dst_rgb565, width); - src_argb += src_stride_argb; - dst_rgb565 += dst_stride_rgb565; - } - return 0; -} - -// Convert ARGB To ARGB1555. -LIBYUV_API -int ARGBToARGB1555(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb1555, int dst_stride_argb1555, - int width, int height) { - int y; - void (*ARGBToARGB1555Row)(const uint8* src_argb, uint8* dst_rgb, int pix) = - ARGBToARGB1555Row_C; - if (!src_argb || !dst_argb1555 || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb1555 == width * 2) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb1555 = 0; - } -#if defined(HAS_ARGBTOARGB1555ROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToARGB1555Row = ARGBToARGB1555Row_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBToARGB1555Row = ARGBToARGB1555Row_SSE2; - } - } -#elif defined(HAS_ARGBTOARGB1555ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToARGB1555Row = ARGBToARGB1555Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToARGB1555Row = ARGBToARGB1555Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToARGB1555Row(src_argb, dst_argb1555, width); - src_argb += src_stride_argb; - dst_argb1555 += dst_stride_argb1555; - } - return 0; -} - -// Convert ARGB To ARGB4444. -LIBYUV_API -int ARGBToARGB4444(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb4444, int dst_stride_argb4444, - int width, int height) { - int y; - void (*ARGBToARGB4444Row)(const uint8* src_argb, uint8* dst_rgb, int pix) = - ARGBToARGB4444Row_C; - if (!src_argb || !dst_argb4444 || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb4444 == width * 2) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb4444 = 0; - } -#if defined(HAS_ARGBTOARGB4444ROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToARGB4444Row = ARGBToARGB4444Row_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBToARGB4444Row = ARGBToARGB4444Row_SSE2; - } - } -#elif defined(HAS_ARGBTOARGB4444ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToARGB4444Row = ARGBToARGB4444Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToARGB4444Row = ARGBToARGB4444Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToARGB4444Row(src_argb, dst_argb4444, width); - src_argb += src_stride_argb; - dst_argb4444 += dst_stride_argb4444; - } - return 0; -} - -// Convert ARGB to J420. (JPeg full range I420). -LIBYUV_API -int ARGBToJ420(const uint8* src_argb, int src_stride_argb, - uint8* dst_yj, int dst_stride_yj, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*ARGBToUVJRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVJRow_C; - void (*ARGBToYJRow)(const uint8* src_argb, uint8* dst_yj, int pix) = - ARGBToYJRow_C; - if (!src_argb || - !dst_yj || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } -#if defined(HAS_ARGBTOYJROW_SSSE3) && defined(HAS_ARGBTOUVJROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVJRow = ARGBToUVJRow_Any_SSSE3; - ARGBToYJRow = ARGBToYJRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToUVJRow = ARGBToUVJRow_Unaligned_SSSE3; - ARGBToYJRow = ARGBToYJRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToUVJRow = ARGBToUVJRow_SSSE3; - if (IS_ALIGNED(dst_yj, 16) && IS_ALIGNED(dst_stride_yj, 16)) { - ARGBToYJRow = ARGBToYJRow_SSSE3; - } - } - } - } -#endif -#if defined(HAS_ARGBTOYJROW_AVX2) && defined(HAS_ARGBTOUVJROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - ARGBToYJRow = ARGBToYJRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - ARGBToYJRow = ARGBToYJRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBTOYJROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYJRow = ARGBToYJRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYJRow = ARGBToYJRow_NEON; - } - } -#endif -#if defined(HAS_ARGBTOUVJROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUVJRow = ARGBToUVJRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUVJRow = ARGBToUVJRow_NEON; - } - } -#endif - - for (y = 0; y < height - 1; y += 2) { - ARGBToUVJRow(src_argb, src_stride_argb, dst_u, dst_v, width); - ARGBToYJRow(src_argb, dst_yj, width); - ARGBToYJRow(src_argb + src_stride_argb, dst_yj + dst_stride_yj, width); - src_argb += src_stride_argb * 2; - dst_yj += dst_stride_yj * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - ARGBToUVJRow(src_argb, 0, dst_u, dst_v, width); - ARGBToYJRow(src_argb, dst_yj, width); - } - return 0; -} - -// Convert ARGB to J400. -LIBYUV_API -int ARGBToJ400(const uint8* src_argb, int src_stride_argb, - uint8* dst_yj, int dst_stride_yj, - int width, int height) { - int y; - void (*ARGBToYJRow)(const uint8* src_argb, uint8* dst_yj, int pix) = - ARGBToYJRow_C; - if (!src_argb || !dst_yj || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_yj == width) { - width *= height; - height = 1; - src_stride_argb = dst_stride_yj = 0; - } -#if defined(HAS_ARGBTOYJROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToYJRow = ARGBToYJRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYJRow = ARGBToYJRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_yj, 16) && IS_ALIGNED(dst_stride_yj, 16)) { - ARGBToYJRow = ARGBToYJRow_SSSE3; - } - } - } -#endif -#if defined(HAS_ARGBTOYJROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - ARGBToYJRow = ARGBToYJRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - ARGBToYJRow = ARGBToYJRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBTOYJROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYJRow = ARGBToYJRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYJRow = ARGBToYJRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBToYJRow(src_argb, dst_yj, width); - src_argb += src_stride_argb; - dst_yj += dst_stride_yj; - } - return 0; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert_jpeg.cc b/src/main/jni/libyuv/source/convert_jpeg.cc deleted file mode 100644 index bcb980f7f..000000000 --- a/src/main/jni/libyuv/source/convert_jpeg.cc +++ /dev/null @@ -1,392 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/convert.h" - -#ifdef HAVE_JPEG -#include "libyuv/mjpeg_decoder.h" -#endif - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#ifdef HAVE_JPEG -struct I420Buffers { - uint8* y; - int y_stride; - uint8* u; - int u_stride; - uint8* v; - int v_stride; - int w; - int h; -}; - -static void JpegCopyI420(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - I420Buffers* dest = (I420Buffers*)(opaque); - I420Copy(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->y, dest->y_stride, - dest->u, dest->u_stride, - dest->v, dest->v_stride, - dest->w, rows); - dest->y += rows * dest->y_stride; - dest->u += ((rows + 1) >> 1) * dest->u_stride; - dest->v += ((rows + 1) >> 1) * dest->v_stride; - dest->h -= rows; -} - -static void JpegI422ToI420(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - I420Buffers* dest = (I420Buffers*)(opaque); - I422ToI420(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->y, dest->y_stride, - dest->u, dest->u_stride, - dest->v, dest->v_stride, - dest->w, rows); - dest->y += rows * dest->y_stride; - dest->u += ((rows + 1) >> 1) * dest->u_stride; - dest->v += ((rows + 1) >> 1) * dest->v_stride; - dest->h -= rows; -} - -static void JpegI444ToI420(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - I420Buffers* dest = (I420Buffers*)(opaque); - I444ToI420(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->y, dest->y_stride, - dest->u, dest->u_stride, - dest->v, dest->v_stride, - dest->w, rows); - dest->y += rows * dest->y_stride; - dest->u += ((rows + 1) >> 1) * dest->u_stride; - dest->v += ((rows + 1) >> 1) * dest->v_stride; - dest->h -= rows; -} - -static void JpegI411ToI420(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - I420Buffers* dest = (I420Buffers*)(opaque); - I411ToI420(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->y, dest->y_stride, - dest->u, dest->u_stride, - dest->v, dest->v_stride, - dest->w, rows); - dest->y += rows * dest->y_stride; - dest->u += ((rows + 1) >> 1) * dest->u_stride; - dest->v += ((rows + 1) >> 1) * dest->v_stride; - dest->h -= rows; -} - -static void JpegI400ToI420(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - I420Buffers* dest = (I420Buffers*)(opaque); - I400ToI420(data[0], strides[0], - dest->y, dest->y_stride, - dest->u, dest->u_stride, - dest->v, dest->v_stride, - dest->w, rows); - dest->y += rows * dest->y_stride; - dest->u += ((rows + 1) >> 1) * dest->u_stride; - dest->v += ((rows + 1) >> 1) * dest->v_stride; - dest->h -= rows; -} - -// Query size of MJPG in pixels. -LIBYUV_API -int MJPGSize(const uint8* sample, size_t sample_size, - int* width, int* height) { - MJpegDecoder mjpeg_decoder; - LIBYUV_BOOL ret = mjpeg_decoder.LoadFrame(sample, sample_size); - if (ret) { - *width = mjpeg_decoder.GetWidth(); - *height = mjpeg_decoder.GetHeight(); - } - mjpeg_decoder.UnloadFrame(); - return ret ? 0 : -1; // -1 for runtime failure. -} - -// MJPG (Motion JPeg) to I420 -// TODO(fbarchard): review w and h requirement. dw and dh may be enough. -LIBYUV_API -int MJPGToI420(const uint8* sample, - size_t sample_size, - uint8* y, int y_stride, - uint8* u, int u_stride, - uint8* v, int v_stride, - int w, int h, - int dw, int dh) { - if (sample_size == kUnknownDataSize) { - // ERROR: MJPEG frame size unknown - return -1; - } - - // TODO(fbarchard): Port MJpeg to C. - MJpegDecoder mjpeg_decoder; - LIBYUV_BOOL ret = mjpeg_decoder.LoadFrame(sample, sample_size); - if (ret && (mjpeg_decoder.GetWidth() != w || - mjpeg_decoder.GetHeight() != h)) { - // ERROR: MJPEG frame has unexpected dimensions - mjpeg_decoder.UnloadFrame(); - return 1; // runtime failure - } - if (ret) { - I420Buffers bufs = { y, y_stride, u, u_stride, v, v_stride, dw, dh }; - // YUV420 - if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 2 && - mjpeg_decoder.GetHorizSampFactor(0) == 2 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegCopyI420, &bufs, dw, dh); - // YUV422 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 2 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI422ToI420, &bufs, dw, dh); - // YUV444 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 1 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI444ToI420, &bufs, dw, dh); - // YUV411 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 4 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI411ToI420, &bufs, dw, dh); - // YUV400 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceGrayscale && - mjpeg_decoder.GetNumComponents() == 1 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI400ToI420, &bufs, dw, dh); - } else { - // TODO(fbarchard): Implement conversion for any other colorspace/sample - // factors that occur in practice. 411 is supported by libjpeg - // ERROR: Unable to convert MJPEG frame because format is not supported - mjpeg_decoder.UnloadFrame(); - return 1; - } - } - return ret ? 0 : 1; -} - -#ifdef HAVE_JPEG -struct ARGBBuffers { - uint8* argb; - int argb_stride; - int w; - int h; -}; - -static void JpegI420ToARGB(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - ARGBBuffers* dest = (ARGBBuffers*)(opaque); - I420ToARGB(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->argb, dest->argb_stride, - dest->w, rows); - dest->argb += rows * dest->argb_stride; - dest->h -= rows; -} - -static void JpegI422ToARGB(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - ARGBBuffers* dest = (ARGBBuffers*)(opaque); - I422ToARGB(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->argb, dest->argb_stride, - dest->w, rows); - dest->argb += rows * dest->argb_stride; - dest->h -= rows; -} - -static void JpegI444ToARGB(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - ARGBBuffers* dest = (ARGBBuffers*)(opaque); - I444ToARGB(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->argb, dest->argb_stride, - dest->w, rows); - dest->argb += rows * dest->argb_stride; - dest->h -= rows; -} - -static void JpegI411ToARGB(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - ARGBBuffers* dest = (ARGBBuffers*)(opaque); - I411ToARGB(data[0], strides[0], - data[1], strides[1], - data[2], strides[2], - dest->argb, dest->argb_stride, - dest->w, rows); - dest->argb += rows * dest->argb_stride; - dest->h -= rows; -} - -static void JpegI400ToARGB(void* opaque, - const uint8* const* data, - const int* strides, - int rows) { - ARGBBuffers* dest = (ARGBBuffers*)(opaque); - I400ToARGB(data[0], strides[0], - dest->argb, dest->argb_stride, - dest->w, rows); - dest->argb += rows * dest->argb_stride; - dest->h -= rows; -} - -// MJPG (Motion JPeg) to ARGB -// TODO(fbarchard): review w and h requirement. dw and dh may be enough. -LIBYUV_API -int MJPGToARGB(const uint8* sample, - size_t sample_size, - uint8* argb, int argb_stride, - int w, int h, - int dw, int dh) { - if (sample_size == kUnknownDataSize) { - // ERROR: MJPEG frame size unknown - return -1; - } - - // TODO(fbarchard): Port MJpeg to C. - MJpegDecoder mjpeg_decoder; - LIBYUV_BOOL ret = mjpeg_decoder.LoadFrame(sample, sample_size); - if (ret && (mjpeg_decoder.GetWidth() != w || - mjpeg_decoder.GetHeight() != h)) { - // ERROR: MJPEG frame has unexpected dimensions - mjpeg_decoder.UnloadFrame(); - return 1; // runtime failure - } - if (ret) { - ARGBBuffers bufs = { argb, argb_stride, dw, dh }; - // YUV420 - if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 2 && - mjpeg_decoder.GetHorizSampFactor(0) == 2 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI420ToARGB, &bufs, dw, dh); - // YUV422 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 2 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI422ToARGB, &bufs, dw, dh); - // YUV444 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 1 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI444ToARGB, &bufs, dw, dh); - // YUV411 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceYCbCr && - mjpeg_decoder.GetNumComponents() == 3 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 4 && - mjpeg_decoder.GetVertSampFactor(1) == 1 && - mjpeg_decoder.GetHorizSampFactor(1) == 1 && - mjpeg_decoder.GetVertSampFactor(2) == 1 && - mjpeg_decoder.GetHorizSampFactor(2) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI411ToARGB, &bufs, dw, dh); - // YUV400 - } else if (mjpeg_decoder.GetColorSpace() == - MJpegDecoder::kColorSpaceGrayscale && - mjpeg_decoder.GetNumComponents() == 1 && - mjpeg_decoder.GetVertSampFactor(0) == 1 && - mjpeg_decoder.GetHorizSampFactor(0) == 1) { - ret = mjpeg_decoder.DecodeToCallback(&JpegI400ToARGB, &bufs, dw, dh); - } else { - // TODO(fbarchard): Implement conversion for any other colorspace/sample - // factors that occur in practice. 411 is supported by libjpeg - // ERROR: Unable to convert MJPEG frame because format is not supported - mjpeg_decoder.UnloadFrame(); - return 1; - } - } - return ret ? 0 : 1; -} -#endif - -#endif - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert_to_argb.cc b/src/main/jni/libyuv/source/convert_to_argb.cc deleted file mode 100644 index 1b228a7b4..000000000 --- a/src/main/jni/libyuv/source/convert_to_argb.cc +++ /dev/null @@ -1,327 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/convert_argb.h" - -#include "libyuv/cpu_id.h" -#include "libyuv/format_conversion.h" -#ifdef HAVE_JPEG -#include "libyuv/mjpeg_decoder.h" -#endif -#include "libyuv/rotate_argb.h" -#include "libyuv/row.h" -#include "libyuv/video_common.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Convert camera sample to I420 with cropping, rotation and vertical flip. -// src_width is used for source stride computation -// src_height is used to compute location of planes, and indicate inversion -// sample_size is measured in bytes and is the size of the frame. -// With MJPEG it is the compressed size of the frame. -LIBYUV_API -int ConvertToARGB(const uint8* sample, size_t sample_size, - uint8* crop_argb, int argb_stride, - int crop_x, int crop_y, - int src_width, int src_height, - int crop_width, int crop_height, - enum RotationMode rotation, - uint32 fourcc) { - uint32 format = CanonicalFourCC(fourcc); - int aligned_src_width = (src_width + 1) & ~1; - const uint8* src; - const uint8* src_uv; - int abs_src_height = (src_height < 0) ? -src_height : src_height; - int inv_crop_height = (crop_height < 0) ? -crop_height : crop_height; - int r = 0; - - // One pass rotation is available for some formats. For the rest, convert - // to I420 (with optional vertical flipping) into a temporary I420 buffer, - // and then rotate the I420 to the final destination buffer. - // For in-place conversion, if destination crop_argb is same as source sample, - // also enable temporary buffer. - LIBYUV_BOOL need_buf = (rotation && format != FOURCC_ARGB) || - crop_argb == sample; - uint8* tmp_argb = crop_argb; - int tmp_argb_stride = argb_stride; - uint8* rotate_buffer = NULL; - int abs_crop_height = (crop_height < 0) ? -crop_height : crop_height; - - if (crop_argb == NULL || sample == NULL || - src_width <= 0 || crop_width <= 0 || - src_height == 0 || crop_height == 0) { - return -1; - } - if (src_height < 0) { - inv_crop_height = -inv_crop_height; - } - - if (need_buf) { - int argb_size = crop_width * abs_crop_height * 4; - rotate_buffer = (uint8*)malloc(argb_size); - if (!rotate_buffer) { - return 1; // Out of memory runtime error. - } - crop_argb = rotate_buffer; - argb_stride = crop_width; - } - - switch (format) { - // Single plane formats - case FOURCC_YUY2: - src = sample + (aligned_src_width * crop_y + crop_x) * 2; - r = YUY2ToARGB(src, aligned_src_width * 2, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_UYVY: - src = sample + (aligned_src_width * crop_y + crop_x) * 2; - r = UYVYToARGB(src, aligned_src_width * 2, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_24BG: - src = sample + (src_width * crop_y + crop_x) * 3; - r = RGB24ToARGB(src, src_width * 3, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RAW: - src = sample + (src_width * crop_y + crop_x) * 3; - r = RAWToARGB(src, src_width * 3, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_ARGB: - src = sample + (src_width * crop_y + crop_x) * 4; - r = ARGBToARGB(src, src_width * 4, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_BGRA: - src = sample + (src_width * crop_y + crop_x) * 4; - r = BGRAToARGB(src, src_width * 4, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_ABGR: - src = sample + (src_width * crop_y + crop_x) * 4; - r = ABGRToARGB(src, src_width * 4, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGBA: - src = sample + (src_width * crop_y + crop_x) * 4; - r = RGBAToARGB(src, src_width * 4, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGBP: - src = sample + (src_width * crop_y + crop_x) * 2; - r = RGB565ToARGB(src, src_width * 2, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGBO: - src = sample + (src_width * crop_y + crop_x) * 2; - r = ARGB1555ToARGB(src, src_width * 2, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_R444: - src = sample + (src_width * crop_y + crop_x) * 2; - r = ARGB4444ToARGB(src, src_width * 2, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - // TODO(fbarchard): Support cropping Bayer by odd numbers - // by adjusting fourcc. - case FOURCC_BGGR: - src = sample + (src_width * crop_y + crop_x); - r = BayerBGGRToARGB(src, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - - case FOURCC_GBRG: - src = sample + (src_width * crop_y + crop_x); - r = BayerGBRGToARGB(src, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - - case FOURCC_GRBG: - src = sample + (src_width * crop_y + crop_x); - r = BayerGRBGToARGB(src, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - - case FOURCC_RGGB: - src = sample + (src_width * crop_y + crop_x); - r = BayerRGGBToARGB(src, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - - case FOURCC_I400: - src = sample + src_width * crop_y + crop_x; - r = I400ToARGB(src, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - - // Biplanar formats - case FOURCC_NV12: - src = sample + (src_width * crop_y + crop_x); - src_uv = sample + aligned_src_width * (src_height + crop_y / 2) + crop_x; - r = NV12ToARGB(src, src_width, - src_uv, aligned_src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_NV21: - src = sample + (src_width * crop_y + crop_x); - src_uv = sample + aligned_src_width * (src_height + crop_y / 2) + crop_x; - // Call NV12 but with u and v parameters swapped. - r = NV21ToARGB(src, src_width, - src_uv, aligned_src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - case FOURCC_M420: - src = sample + (src_width * crop_y) * 12 / 8 + crop_x; - r = M420ToARGB(src, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; -// case FOURCC_Q420: -// src = sample + (src_width + aligned_src_width * 2) * crop_y + crop_x; -// src_uv = sample + (src_width + aligned_src_width * 2) * crop_y + -// src_width + crop_x * 2; -// r = Q420ToARGB(src, src_width * 3, -// src_uv, src_width * 3, -// crop_argb, argb_stride, -// crop_width, inv_crop_height); -// break; - // Triplanar formats - case FOURCC_I420: - case FOURCC_YU12: - case FOURCC_YV12: { - const uint8* src_y = sample + (src_width * crop_y + crop_x); - const uint8* src_u; - const uint8* src_v; - int halfwidth = (src_width + 1) / 2; - int halfheight = (abs_src_height + 1) / 2; - if (format == FOURCC_YV12) { - src_v = sample + src_width * abs_src_height + - (halfwidth * crop_y + crop_x) / 2; - src_u = sample + src_width * abs_src_height + - halfwidth * (halfheight + crop_y / 2) + crop_x / 2; - } else { - src_u = sample + src_width * abs_src_height + - (halfwidth * crop_y + crop_x) / 2; - src_v = sample + src_width * abs_src_height + - halfwidth * (halfheight + crop_y / 2) + crop_x / 2; - } - r = I420ToARGB(src_y, src_width, - src_u, halfwidth, - src_v, halfwidth, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - } - case FOURCC_I422: - case FOURCC_YV16: { - const uint8* src_y = sample + src_width * crop_y + crop_x; - const uint8* src_u; - const uint8* src_v; - int halfwidth = (src_width + 1) / 2; - if (format == FOURCC_YV16) { - src_v = sample + src_width * abs_src_height + - halfwidth * crop_y + crop_x / 2; - src_u = sample + src_width * abs_src_height + - halfwidth * (abs_src_height + crop_y) + crop_x / 2; - } else { - src_u = sample + src_width * abs_src_height + - halfwidth * crop_y + crop_x / 2; - src_v = sample + src_width * abs_src_height + - halfwidth * (abs_src_height + crop_y) + crop_x / 2; - } - r = I422ToARGB(src_y, src_width, - src_u, halfwidth, - src_v, halfwidth, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - } - case FOURCC_I444: - case FOURCC_YV24: { - const uint8* src_y = sample + src_width * crop_y + crop_x; - const uint8* src_u; - const uint8* src_v; - if (format == FOURCC_YV24) { - src_v = sample + src_width * (abs_src_height + crop_y) + crop_x; - src_u = sample + src_width * (abs_src_height * 2 + crop_y) + crop_x; - } else { - src_u = sample + src_width * (abs_src_height + crop_y) + crop_x; - src_v = sample + src_width * (abs_src_height * 2 + crop_y) + crop_x; - } - r = I444ToARGB(src_y, src_width, - src_u, src_width, - src_v, src_width, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - } - case FOURCC_I411: { - int quarterwidth = (src_width + 3) / 4; - const uint8* src_y = sample + src_width * crop_y + crop_x; - const uint8* src_u = sample + src_width * abs_src_height + - quarterwidth * crop_y + crop_x / 4; - const uint8* src_v = sample + src_width * abs_src_height + - quarterwidth * (abs_src_height + crop_y) + crop_x / 4; - r = I411ToARGB(src_y, src_width, - src_u, quarterwidth, - src_v, quarterwidth, - crop_argb, argb_stride, - crop_width, inv_crop_height); - break; - } -#ifdef HAVE_JPEG - case FOURCC_MJPG: - r = MJPGToARGB(sample, sample_size, - crop_argb, argb_stride, - src_width, abs_src_height, crop_width, inv_crop_height); - break; -#endif - default: - r = -1; // unknown fourcc - return failure code. - } - - if (need_buf) { - if (!r) { - r = ARGBRotate(crop_argb, argb_stride, - tmp_argb, tmp_argb_stride, - crop_width, abs_crop_height, rotation); - } - free(rotate_buffer); - } - - return r; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/convert_to_i420.cc b/src/main/jni/libyuv/source/convert_to_i420.cc deleted file mode 100644 index 7b194fff7..000000000 --- a/src/main/jni/libyuv/source/convert_to_i420.cc +++ /dev/null @@ -1,383 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include <stdlib.h> - -#include "libyuv/convert.h" - -#include "libyuv/format_conversion.h" -#include "libyuv/video_common.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Convert camera sample to I420 with cropping, rotation and vertical flip. -// src_width is used for source stride computation -// src_height is used to compute location of planes, and indicate inversion -// sample_size is measured in bytes and is the size of the frame. -// With MJPEG it is the compressed size of the frame. -LIBYUV_API -int ConvertToI420(const uint8* sample, - size_t sample_size, - uint8* y, int y_stride, - uint8* u, int u_stride, - uint8* v, int v_stride, - int crop_x, int crop_y, - int src_width, int src_height, - int crop_width, int crop_height, - enum RotationMode rotation, - uint32 fourcc) { - uint32 format = CanonicalFourCC(fourcc); - int aligned_src_width = (src_width + 1) & ~1; - const uint8* src; - const uint8* src_uv; - int abs_src_height = (src_height < 0) ? -src_height : src_height; - int inv_crop_height = (crop_height < 0) ? -crop_height : crop_height; - int r = 0; - LIBYUV_BOOL need_buf = (rotation && format != FOURCC_I420 && - format != FOURCC_NV12 && format != FOURCC_NV21 && - format != FOURCC_YU12 && format != FOURCC_YV12) || y == sample; - uint8* tmp_y = y; - uint8* tmp_u = u; - uint8* tmp_v = v; - int tmp_y_stride = y_stride; - int tmp_u_stride = u_stride; - int tmp_v_stride = v_stride; - uint8* rotate_buffer = NULL; - int abs_crop_height = (crop_height < 0) ? -crop_height : crop_height; - - if (!y || !u || !v || !sample || - src_width <= 0 || crop_width <= 0 || - src_height == 0 || crop_height == 0) { - return -1; - } - if (src_height < 0) { - inv_crop_height = -inv_crop_height; - } - - // One pass rotation is available for some formats. For the rest, convert - // to I420 (with optional vertical flipping) into a temporary I420 buffer, - // and then rotate the I420 to the final destination buffer. - // For in-place conversion, if destination y is same as source sample, - // also enable temporary buffer. - if (need_buf) { - int y_size = crop_width * abs_crop_height; - int uv_size = ((crop_width + 1) / 2) * ((abs_crop_height + 1) / 2); - rotate_buffer = (uint8*)malloc(y_size + uv_size * 2); - if (!rotate_buffer) { - return 1; // Out of memory runtime error. - } - y = rotate_buffer; - u = y + y_size; - v = u + uv_size; - y_stride = crop_width; - u_stride = v_stride = ((crop_width + 1) / 2); - } - - switch (format) { - // Single plane formats - case FOURCC_YUY2: - src = sample + (aligned_src_width * crop_y + crop_x) * 2; - r = YUY2ToI420(src, aligned_src_width * 2, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_UYVY: - src = sample + (aligned_src_width * crop_y + crop_x) * 2; - r = UYVYToI420(src, aligned_src_width * 2, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGBP: - src = sample + (src_width * crop_y + crop_x) * 2; - r = RGB565ToI420(src, src_width * 2, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGBO: - src = sample + (src_width * crop_y + crop_x) * 2; - r = ARGB1555ToI420(src, src_width * 2, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_R444: - src = sample + (src_width * crop_y + crop_x) * 2; - r = ARGB4444ToI420(src, src_width * 2, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_24BG: - src = sample + (src_width * crop_y + crop_x) * 3; - r = RGB24ToI420(src, src_width * 3, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RAW: - src = sample + (src_width * crop_y + crop_x) * 3; - r = RAWToI420(src, src_width * 3, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_ARGB: - src = sample + (src_width * crop_y + crop_x) * 4; - r = ARGBToI420(src, src_width * 4, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_BGRA: - src = sample + (src_width * crop_y + crop_x) * 4; - r = BGRAToI420(src, src_width * 4, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_ABGR: - src = sample + (src_width * crop_y + crop_x) * 4; - r = ABGRToI420(src, src_width * 4, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGBA: - src = sample + (src_width * crop_y + crop_x) * 4; - r = RGBAToI420(src, src_width * 4, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - // TODO(fbarchard): Support cropping Bayer by odd numbers - // by adjusting fourcc. - case FOURCC_BGGR: - src = sample + (src_width * crop_y + crop_x); - r = BayerBGGRToI420(src, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_GBRG: - src = sample + (src_width * crop_y + crop_x); - r = BayerGBRGToI420(src, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_GRBG: - src = sample + (src_width * crop_y + crop_x); - r = BayerGRBGToI420(src, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_RGGB: - src = sample + (src_width * crop_y + crop_x); - r = BayerRGGBToI420(src, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_I400: - src = sample + src_width * crop_y + crop_x; - r = I400ToI420(src, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - // Biplanar formats - case FOURCC_NV12: - src = sample + (src_width * crop_y + crop_x); - src_uv = sample + aligned_src_width * (src_height + crop_y / 2) + crop_x; - r = NV12ToI420Rotate(src, src_width, - src_uv, aligned_src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height, rotation); - break; - case FOURCC_NV21: - src = sample + (src_width * crop_y + crop_x); - src_uv = sample + aligned_src_width * (src_height + crop_y / 2) + crop_x; - // Call NV12 but with u and v parameters swapped. - r = NV12ToI420Rotate(src, src_width, - src_uv, aligned_src_width, - y, y_stride, - v, v_stride, - u, u_stride, - crop_width, inv_crop_height, rotation); - break; - case FOURCC_M420: - src = sample + (src_width * crop_y) * 12 / 8 + crop_x; - r = M420ToI420(src, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - case FOURCC_Q420: - src = sample + (src_width + aligned_src_width * 2) * crop_y + crop_x; - src_uv = sample + (src_width + aligned_src_width * 2) * crop_y + - src_width + crop_x * 2; - r = Q420ToI420(src, src_width * 3, - src_uv, src_width * 3, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - // Triplanar formats - case FOURCC_I420: - case FOURCC_YU12: - case FOURCC_YV12: { - const uint8* src_y = sample + (src_width * crop_y + crop_x); - const uint8* src_u; - const uint8* src_v; - int halfwidth = (src_width + 1) / 2; - int halfheight = (abs_src_height + 1) / 2; - if (format == FOURCC_YV12) { - src_v = sample + src_width * abs_src_height + - (halfwidth * crop_y + crop_x) / 2; - src_u = sample + src_width * abs_src_height + - halfwidth * (halfheight + crop_y / 2) + crop_x / 2; - } else { - src_u = sample + src_width * abs_src_height + - (halfwidth * crop_y + crop_x) / 2; - src_v = sample + src_width * abs_src_height + - halfwidth * (halfheight + crop_y / 2) + crop_x / 2; - } - r = I420Rotate(src_y, src_width, - src_u, halfwidth, - src_v, halfwidth, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height, rotation); - break; - } - case FOURCC_I422: - case FOURCC_YV16: { - const uint8* src_y = sample + src_width * crop_y + crop_x; - const uint8* src_u; - const uint8* src_v; - int halfwidth = (src_width + 1) / 2; - if (format == FOURCC_YV16) { - src_v = sample + src_width * abs_src_height + - halfwidth * crop_y + crop_x / 2; - src_u = sample + src_width * abs_src_height + - halfwidth * (abs_src_height + crop_y) + crop_x / 2; - } else { - src_u = sample + src_width * abs_src_height + - halfwidth * crop_y + crop_x / 2; - src_v = sample + src_width * abs_src_height + - halfwidth * (abs_src_height + crop_y) + crop_x / 2; - } - r = I422ToI420(src_y, src_width, - src_u, halfwidth, - src_v, halfwidth, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - } - case FOURCC_I444: - case FOURCC_YV24: { - const uint8* src_y = sample + src_width * crop_y + crop_x; - const uint8* src_u; - const uint8* src_v; - if (format == FOURCC_YV24) { - src_v = sample + src_width * (abs_src_height + crop_y) + crop_x; - src_u = sample + src_width * (abs_src_height * 2 + crop_y) + crop_x; - } else { - src_u = sample + src_width * (abs_src_height + crop_y) + crop_x; - src_v = sample + src_width * (abs_src_height * 2 + crop_y) + crop_x; - } - r = I444ToI420(src_y, src_width, - src_u, src_width, - src_v, src_width, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - } - case FOURCC_I411: { - int quarterwidth = (src_width + 3) / 4; - const uint8* src_y = sample + src_width * crop_y + crop_x; - const uint8* src_u = sample + src_width * abs_src_height + - quarterwidth * crop_y + crop_x / 4; - const uint8* src_v = sample + src_width * abs_src_height + - quarterwidth * (abs_src_height + crop_y) + crop_x / 4; - r = I411ToI420(src_y, src_width, - src_u, quarterwidth, - src_v, quarterwidth, - y, y_stride, - u, u_stride, - v, v_stride, - crop_width, inv_crop_height); - break; - } -#ifdef HAVE_JPEG - case FOURCC_MJPG: - r = MJPGToI420(sample, sample_size, - y, y_stride, - u, u_stride, - v, v_stride, - src_width, abs_src_height, crop_width, inv_crop_height); - break; -#endif - default: - r = -1; // unknown fourcc - return failure code. - } - - if (need_buf) { - if (!r) { - r = I420Rotate(y, y_stride, - u, u_stride, - v, v_stride, - tmp_y, tmp_y_stride, - tmp_u, tmp_u_stride, - tmp_v, tmp_v_stride, - crop_width, abs_crop_height, rotation); - } - free(rotate_buffer); - } - - return r; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/cpu_id.cc b/src/main/jni/libyuv/source/cpu_id.cc deleted file mode 100644 index deb4c4465..000000000 --- a/src/main/jni/libyuv/source/cpu_id.cc +++ /dev/null @@ -1,293 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/cpu_id.h" - -#if defined(_MSC_VER) && !defined(__clang__) -#include <intrin.h> // For __cpuidex() -#endif -#if !defined(__pnacl__) && !defined(__CLR_VER) && \ - !defined(__native_client__) && \ - defined(_MSC_VER) && (_MSC_FULL_VER >= 160040219) -#include <immintrin.h> // For _xgetbv() -#endif - -#if !defined(__native_client__) -#include <stdlib.h> // For getenv() -#endif - -// For ArmCpuCaps() but unittested on all platforms -#include <stdio.h> -#include <string.h> - -#include "libyuv/basic_types.h" // For CPU_X86 - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// For functions that use the stack and have runtime checks for overflow, -// use SAFEBUFFERS to avoid additional check. -#if defined(_MSC_VER) && (_MSC_FULL_VER >= 160040219) -#define SAFEBUFFERS __declspec(safebuffers) -#else -#define SAFEBUFFERS -#endif - -// Low level cpuid for X86. Returns zeros on other CPUs. -#if !defined(__pnacl__) && !defined(__CLR_VER) && \ - (defined(_M_IX86) || defined(_M_X64) || \ - defined(__i386__) || defined(__x86_64__)) -LIBYUV_API -void CpuId(uint32 info_eax, uint32 info_ecx, uint32* cpu_info) { -#if defined(_MSC_VER) && !defined(__clang__) -#if (_MSC_FULL_VER >= 160040219) - __cpuidex((int*)(cpu_info), info_eax, info_ecx); -#elif defined(_M_IX86) - __asm { - mov eax, info_eax - mov ecx, info_ecx - mov edi, cpu_info - cpuid - mov [edi], eax - mov [edi + 4], ebx - mov [edi + 8], ecx - mov [edi + 12], edx - } -#else - if (info_ecx == 0) { - __cpuid((int*)(cpu_info), info_eax); - } else { - cpu_info[3] = cpu_info[2] = cpu_info[1] = cpu_info[0] = 0; - } -#endif -#else // defined(_MSC_VER) - uint32 info_ebx, info_edx; - asm volatile ( // NOLINT -#if defined( __i386__) && defined(__PIC__) - // Preserve ebx for fpic 32 bit. - "mov %%ebx, %%edi \n" - "cpuid \n" - "xchg %%edi, %%ebx \n" - : "=D" (info_ebx), -#else - "cpuid \n" - : "=b" (info_ebx), -#endif // defined( __i386__) && defined(__PIC__) - "+a" (info_eax), "+c" (info_ecx), "=d" (info_edx)); - cpu_info[0] = info_eax; - cpu_info[1] = info_ebx; - cpu_info[2] = info_ecx; - cpu_info[3] = info_edx; -#endif // defined(_MSC_VER) -} - -#if !defined(__native_client__) -#define HAS_XGETBV -// X86 CPUs have xgetbv to detect OS saves high parts of ymm registers. -int TestOsSaveYmm() { - uint32 xcr0 = 0u; -#if defined(_MSC_VER) && (_MSC_FULL_VER >= 160040219) - xcr0 = (uint32)(_xgetbv(0)); // VS2010 SP1 required. -#elif defined(_M_IX86) && defined(_MSC_VER) - __asm { - xor ecx, ecx // xcr 0 - _asm _emit 0x0f _asm _emit 0x01 _asm _emit 0xd0 // For VS2010 and earlier. - mov xcr0, eax - } -#elif defined(__i386__) || defined(__x86_64__) - asm(".byte 0x0f, 0x01, 0xd0" : "=a" (xcr0) : "c" (0) : "%edx"); -#endif // defined(_MSC_VER) - return((xcr0 & 6) == 6); // Is ymm saved? -} -#endif // !defined(__native_client__) -#else -LIBYUV_API -void CpuId(uint32 eax, uint32 ecx, uint32* cpu_info) { - cpu_info[0] = cpu_info[1] = cpu_info[2] = cpu_info[3] = 0; -} -#endif - -// based on libvpx arm_cpudetect.c -// For Arm, but public to allow testing on any CPU -LIBYUV_API SAFEBUFFERS -int ArmCpuCaps(const char* cpuinfo_name) { - char cpuinfo_line[512]; - FILE* f = fopen(cpuinfo_name, "r"); - if (!f) { - // Assume Neon if /proc/cpuinfo is unavailable. - // This will occur for Chrome sandbox for Pepper or Render process. - return kCpuHasNEON; - } - while (fgets(cpuinfo_line, sizeof(cpuinfo_line) - 1, f)) { - if (memcmp(cpuinfo_line, "Features", 8) == 0) { - char* p = strstr(cpuinfo_line, " neon"); - if (p && (p[5] == ' ' || p[5] == '\n')) { - fclose(f); - return kCpuHasNEON; - } - } - } - fclose(f); - return 0; -} - -#if defined(__mips__) && defined(__linux__) -static int MipsCpuCaps(const char* search_string) { - char cpuinfo_line[512]; - const char* file_name = "/proc/cpuinfo"; - FILE* f = fopen(file_name, "r"); - if (!f) { - // Assume DSP if /proc/cpuinfo is unavailable. - // This will occur for Chrome sandbox for Pepper or Render process. - return kCpuHasMIPS_DSP; - } - while (fgets(cpuinfo_line, sizeof(cpuinfo_line) - 1, f) != NULL) { - if (strstr(cpuinfo_line, search_string) != NULL) { - fclose(f); - return kCpuHasMIPS_DSP; - } - } - fclose(f); - return 0; -} -#endif - -// CPU detect function for SIMD instruction sets. -LIBYUV_API -int cpu_info_ = kCpuInit; // cpu_info is not initialized yet. - -// Test environment variable for disabling CPU features. Any non-zero value -// to disable. Zero ignored to make it easy to set the variable on/off. -#if !defined(__native_client__) && !defined(_M_ARM) - -static LIBYUV_BOOL TestEnv(const char* name) { - const char* var = getenv(name); - if (var) { - if (var[0] != '0') { - return LIBYUV_TRUE; - } - } - return LIBYUV_FALSE; -} -#else // nacl does not support getenv(). -static LIBYUV_BOOL TestEnv(const char*) { - return LIBYUV_FALSE; -} -#endif - -LIBYUV_API SAFEBUFFERS -int InitCpuFlags(void) { -#if !defined(__pnacl__) && !defined(__CLR_VER) && defined(CPU_X86) - - uint32 cpu_info0[4] = { 0, 0, 0, 0 }; - uint32 cpu_info1[4] = { 0, 0, 0, 0 }; - uint32 cpu_info7[4] = { 0, 0, 0, 0 }; - CpuId(0, 0, cpu_info0); - CpuId(1, 0, cpu_info1); - if (cpu_info0[0] >= 7) { - CpuId(7, 0, cpu_info7); - } - cpu_info_ = ((cpu_info1[3] & 0x04000000) ? kCpuHasSSE2 : 0) | - ((cpu_info1[2] & 0x00000200) ? kCpuHasSSSE3 : 0) | - ((cpu_info1[2] & 0x00080000) ? kCpuHasSSE41 : 0) | - ((cpu_info1[2] & 0x00100000) ? kCpuHasSSE42 : 0) | - ((cpu_info7[1] & 0x00000200) ? kCpuHasERMS : 0) | - ((cpu_info1[2] & 0x00001000) ? kCpuHasFMA3 : 0) | - kCpuHasX86; - -#ifdef HAS_XGETBV - if ((cpu_info1[2] & 0x18000000) == 0x18000000 && // AVX and OSSave - TestOsSaveYmm()) { // Saves YMM. - cpu_info_ |= ((cpu_info7[1] & 0x00000020) ? kCpuHasAVX2 : 0) | - kCpuHasAVX; - } -#endif - // Environment variable overrides for testing. - if (TestEnv("LIBYUV_DISABLE_X86")) { - cpu_info_ &= ~kCpuHasX86; - } - if (TestEnv("LIBYUV_DISABLE_SSE2")) { - cpu_info_ &= ~kCpuHasSSE2; - } - if (TestEnv("LIBYUV_DISABLE_SSSE3")) { - cpu_info_ &= ~kCpuHasSSSE3; - } - if (TestEnv("LIBYUV_DISABLE_SSE41")) { - cpu_info_ &= ~kCpuHasSSE41; - } - if (TestEnv("LIBYUV_DISABLE_SSE42")) { - cpu_info_ &= ~kCpuHasSSE42; - } - if (TestEnv("LIBYUV_DISABLE_AVX")) { - cpu_info_ &= ~kCpuHasAVX; - } - if (TestEnv("LIBYUV_DISABLE_AVX2")) { - cpu_info_ &= ~kCpuHasAVX2; - } - if (TestEnv("LIBYUV_DISABLE_ERMS")) { - cpu_info_ &= ~kCpuHasERMS; - } - if (TestEnv("LIBYUV_DISABLE_FMA3")) { - cpu_info_ &= ~kCpuHasFMA3; - } -#elif defined(__mips__) && defined(__linux__) - // Linux mips parse text file for dsp detect. - cpu_info_ = MipsCpuCaps("dsp"); // set kCpuHasMIPS_DSP. -#if defined(__mips_dspr2) - cpu_info_ |= kCpuHasMIPS_DSPR2; -#endif - cpu_info_ |= kCpuHasMIPS; - - if (getenv("LIBYUV_DISABLE_MIPS")) { - cpu_info_ &= ~kCpuHasMIPS; - } - if (getenv("LIBYUV_DISABLE_MIPS_DSP")) { - cpu_info_ &= ~kCpuHasMIPS_DSP; - } - if (getenv("LIBYUV_DISABLE_MIPS_DSPR2")) { - cpu_info_ &= ~kCpuHasMIPS_DSPR2; - } -#elif defined(__arm__) || defined(__aarch64__) -// gcc -mfpu=neon defines __ARM_NEON__ -// __ARM_NEON__ generates code that requires Neon. NaCL also requires Neon. -// For Linux, /proc/cpuinfo can be tested but without that assume Neon. -#if defined(__ARM_NEON__) || defined(__native_client__) || !defined(__linux__) - cpu_info_ = kCpuHasNEON; -// For aarch64(arm64), /proc/cpuinfo's feature is not complete, e.g. no neon -// flag in it. -// So for aarch64, neon enabling is hard coded here. -#elif defined(__aarch64__) - cpu_info_ = kCpuHasNEON; -#else - // Linux arm parse text file for neon detect. - cpu_info_ = ArmCpuCaps("/proc/cpuinfo"); -#endif - cpu_info_ |= kCpuHasARM; - if (TestEnv("LIBYUV_DISABLE_NEON")) { - cpu_info_ &= ~kCpuHasNEON; - } -#endif // __arm__ - if (TestEnv("LIBYUV_DISABLE_ASM")) { - cpu_info_ = 0; - } - return cpu_info_; -} - -LIBYUV_API -void MaskCpuFlags(int enable_flags) { - cpu_info_ = InitCpuFlags() & enable_flags; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/format_conversion.cc b/src/main/jni/libyuv/source/format_conversion.cc deleted file mode 100644 index 3c1737153..000000000 --- a/src/main/jni/libyuv/source/format_conversion.cc +++ /dev/null @@ -1,554 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/format_conversion.h" - -#include "libyuv/basic_types.h" -#include "libyuv/cpu_id.h" -#include "libyuv/video_common.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// generate a selector mask useful for pshufb -static uint32 GenerateSelector(int select0, int select1) { - return (uint32)(select0) | - (uint32)((select1 + 4) << 8) | - (uint32)((select0 + 8) << 16) | - (uint32)((select1 + 12) << 24); -} - -static int MakeSelectors(const int blue_index, - const int green_index, - const int red_index, - uint32 dst_fourcc_bayer, - uint32* index_map) { - // Now build a lookup table containing the indices for the four pixels in each - // 2x2 Bayer grid. - switch (dst_fourcc_bayer) { - case FOURCC_BGGR: - index_map[0] = GenerateSelector(blue_index, green_index); - index_map[1] = GenerateSelector(green_index, red_index); - break; - case FOURCC_GBRG: - index_map[0] = GenerateSelector(green_index, blue_index); - index_map[1] = GenerateSelector(red_index, green_index); - break; - case FOURCC_RGGB: - index_map[0] = GenerateSelector(red_index, green_index); - index_map[1] = GenerateSelector(green_index, blue_index); - break; - case FOURCC_GRBG: - index_map[0] = GenerateSelector(green_index, red_index); - index_map[1] = GenerateSelector(blue_index, green_index); - break; - default: - return -1; // Bad FourCC - } - return 0; -} - -// Converts 32 bit ARGB to Bayer RGB formats. -LIBYUV_API -int ARGBToBayer(const uint8* src_argb, int src_stride_argb, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height, - uint32 dst_fourcc_bayer) { - int y; - const int blue_index = 0; // Offsets for ARGB format - const int green_index = 1; - const int red_index = 2; - uint32 index_map[2]; - void (*ARGBToBayerRow)(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) = ARGBToBayerRow_C; - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } -#if defined(HAS_ARGBTOBAYERROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToBayerRow = ARGBToBayerRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerRow_SSSE3; - } - } -#elif defined(HAS_ARGBTOBAYERROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToBayerRow = ARGBToBayerRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerRow_NEON; - } - } -#endif - if (MakeSelectors(blue_index, green_index, red_index, - dst_fourcc_bayer, index_map)) { - return -1; // Bad FourCC - } - - for (y = 0; y < height; ++y) { - ARGBToBayerRow(src_argb, dst_bayer, index_map[y & 1], width); - src_argb += src_stride_argb; - dst_bayer += dst_stride_bayer; - } - return 0; -} - -#define AVG(a, b) (((a) + (b)) >> 1) - -static void BayerRowBG(const uint8* src_bayer0, int src_stride_bayer, - uint8* dst_argb, int pix) { - const uint8* src_bayer1 = src_bayer0 + src_stride_bayer; - uint8 g = src_bayer0[1]; - uint8 r = src_bayer1[1]; - int x; - for (x = 0; x < pix - 2; x += 2) { - dst_argb[0] = src_bayer0[0]; - dst_argb[1] = AVG(g, src_bayer0[1]); - dst_argb[2] = AVG(r, src_bayer1[1]); - dst_argb[3] = 255U; - dst_argb[4] = AVG(src_bayer0[0], src_bayer0[2]); - dst_argb[5] = src_bayer0[1]; - dst_argb[6] = src_bayer1[1]; - dst_argb[7] = 255U; - g = src_bayer0[1]; - r = src_bayer1[1]; - src_bayer0 += 2; - src_bayer1 += 2; - dst_argb += 8; - } - dst_argb[0] = src_bayer0[0]; - dst_argb[1] = AVG(g, src_bayer0[1]); - dst_argb[2] = AVG(r, src_bayer1[1]); - dst_argb[3] = 255U; - if (!(pix & 1)) { - dst_argb[4] = src_bayer0[0]; - dst_argb[5] = src_bayer0[1]; - dst_argb[6] = src_bayer1[1]; - dst_argb[7] = 255U; - } -} - -static void BayerRowRG(const uint8* src_bayer0, int src_stride_bayer, - uint8* dst_argb, int pix) { - const uint8* src_bayer1 = src_bayer0 + src_stride_bayer; - uint8 g = src_bayer0[1]; - uint8 b = src_bayer1[1]; - int x; - for (x = 0; x < pix - 2; x += 2) { - dst_argb[0] = AVG(b, src_bayer1[1]); - dst_argb[1] = AVG(g, src_bayer0[1]); - dst_argb[2] = src_bayer0[0]; - dst_argb[3] = 255U; - dst_argb[4] = src_bayer1[1]; - dst_argb[5] = src_bayer0[1]; - dst_argb[6] = AVG(src_bayer0[0], src_bayer0[2]); - dst_argb[7] = 255U; - g = src_bayer0[1]; - b = src_bayer1[1]; - src_bayer0 += 2; - src_bayer1 += 2; - dst_argb += 8; - } - dst_argb[0] = AVG(b, src_bayer1[1]); - dst_argb[1] = AVG(g, src_bayer0[1]); - dst_argb[2] = src_bayer0[0]; - dst_argb[3] = 255U; - if (!(pix & 1)) { - dst_argb[4] = src_bayer1[1]; - dst_argb[5] = src_bayer0[1]; - dst_argb[6] = src_bayer0[0]; - dst_argb[7] = 255U; - } -} - -static void BayerRowGB(const uint8* src_bayer0, int src_stride_bayer, - uint8* dst_argb, int pix) { - const uint8* src_bayer1 = src_bayer0 + src_stride_bayer; - uint8 b = src_bayer0[1]; - int x; - for (x = 0; x < pix - 2; x += 2) { - dst_argb[0] = AVG(b, src_bayer0[1]); - dst_argb[1] = src_bayer0[0]; - dst_argb[2] = src_bayer1[0]; - dst_argb[3] = 255U; - dst_argb[4] = src_bayer0[1]; - dst_argb[5] = AVG(src_bayer0[0], src_bayer0[2]); - dst_argb[6] = AVG(src_bayer1[0], src_bayer1[2]); - dst_argb[7] = 255U; - b = src_bayer0[1]; - src_bayer0 += 2; - src_bayer1 += 2; - dst_argb += 8; - } - dst_argb[0] = AVG(b, src_bayer0[1]); - dst_argb[1] = src_bayer0[0]; - dst_argb[2] = src_bayer1[0]; - dst_argb[3] = 255U; - if (!(pix & 1)) { - dst_argb[4] = src_bayer0[1]; - dst_argb[5] = src_bayer0[0]; - dst_argb[6] = src_bayer1[0]; - dst_argb[7] = 255U; - } -} - -static void BayerRowGR(const uint8* src_bayer0, int src_stride_bayer, - uint8* dst_argb, int pix) { - const uint8* src_bayer1 = src_bayer0 + src_stride_bayer; - uint8 r = src_bayer0[1]; - int x; - for (x = 0; x < pix - 2; x += 2) { - dst_argb[0] = src_bayer1[0]; - dst_argb[1] = src_bayer0[0]; - dst_argb[2] = AVG(r, src_bayer0[1]); - dst_argb[3] = 255U; - dst_argb[4] = AVG(src_bayer1[0], src_bayer1[2]); - dst_argb[5] = AVG(src_bayer0[0], src_bayer0[2]); - dst_argb[6] = src_bayer0[1]; - dst_argb[7] = 255U; - r = src_bayer0[1]; - src_bayer0 += 2; - src_bayer1 += 2; - dst_argb += 8; - } - dst_argb[0] = src_bayer1[0]; - dst_argb[1] = src_bayer0[0]; - dst_argb[2] = AVG(r, src_bayer0[1]); - dst_argb[3] = 255U; - if (!(pix & 1)) { - dst_argb[4] = src_bayer1[0]; - dst_argb[5] = src_bayer0[0]; - dst_argb[6] = src_bayer0[1]; - dst_argb[7] = 255U; - } -} - -// Converts any Bayer RGB format to ARGB. -LIBYUV_API -int BayerToARGB(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int dst_stride_argb, - int width, int height, - uint32 src_fourcc_bayer) { - int y; - void (*BayerRow0)(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int pix); - void (*BayerRow1)(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int pix); - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - switch (src_fourcc_bayer) { - case FOURCC_BGGR: - BayerRow0 = BayerRowBG; - BayerRow1 = BayerRowGR; - break; - case FOURCC_GBRG: - BayerRow0 = BayerRowGB; - BayerRow1 = BayerRowRG; - break; - case FOURCC_GRBG: - BayerRow0 = BayerRowGR; - BayerRow1 = BayerRowBG; - break; - case FOURCC_RGGB: - BayerRow0 = BayerRowRG; - BayerRow1 = BayerRowGB; - break; - default: - return -1; // Bad FourCC - } - - for (y = 0; y < height - 1; y += 2) { - BayerRow0(src_bayer, src_stride_bayer, dst_argb, width); - BayerRow1(src_bayer + src_stride_bayer, -src_stride_bayer, - dst_argb + dst_stride_argb, width); - src_bayer += src_stride_bayer * 2; - dst_argb += dst_stride_argb * 2; - } - if (height & 1) { - BayerRow0(src_bayer, src_stride_bayer, dst_argb, width); - } - return 0; -} - -// Converts any Bayer RGB format to ARGB. -LIBYUV_API -int BayerToI420(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height, - uint32 src_fourcc_bayer) { - void (*BayerRow0)(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int pix); - void (*BayerRow1)(const uint8* src_bayer, int src_stride_bayer, - uint8* dst_argb, int pix); - - void (*ARGBToUVRow)(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) = ARGBToUVRow_C; - void (*ARGBToYRow)(const uint8* src_argb, uint8* dst_y, int pix) = - ARGBToYRow_C; - // Negative height means invert the image. - if (height < 0) { - int halfheight; - height = -height; - halfheight = (height + 1) >> 1; - dst_y = dst_y + (height - 1) * dst_stride_y; - dst_u = dst_u + (halfheight - 1) * dst_stride_u; - dst_v = dst_v + (halfheight - 1) * dst_stride_v; - dst_stride_y = -dst_stride_y; - dst_stride_u = -dst_stride_u; - dst_stride_v = -dst_stride_v; - } -#if defined(HAS_ARGBTOYROW_SSSE3) && defined(HAS_ARGBTOUVROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_SSSE3; - ARGBToYRow = ARGBToYRow_Any_SSSE3; - if (IS_ALIGNED(width, 16)) { - ARGBToYRow = ARGBToYRow_Unaligned_SSSE3; - ARGBToUVRow = ARGBToUVRow_SSSE3; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - ARGBToYRow = ARGBToYRow_SSSE3; - } - } - } -#elif defined(HAS_ARGBTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToYRow = ARGBToYRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToYRow = ARGBToYRow_NEON; - } - } -#endif -#if defined(HAS_ARGBTOUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 16) { - ARGBToUVRow = ARGBToUVRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - ARGBToUVRow = ARGBToUVRow_NEON; - } - } -#endif - - switch (src_fourcc_bayer) { - case FOURCC_BGGR: - BayerRow0 = BayerRowBG; - BayerRow1 = BayerRowGR; - break; - case FOURCC_GBRG: - BayerRow0 = BayerRowGB; - BayerRow1 = BayerRowRG; - break; - case FOURCC_GRBG: - BayerRow0 = BayerRowGR; - BayerRow1 = BayerRowBG; - break; - case FOURCC_RGGB: - BayerRow0 = BayerRowRG; - BayerRow1 = BayerRowGB; - break; - default: - return -1; // Bad FourCC - } - - { - // Allocate 2 rows of ARGB. - const int kRowSize = (width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); - int y; - for (y = 0; y < height - 1; y += 2) { - BayerRow0(src_bayer, src_stride_bayer, row, width); - BayerRow1(src_bayer + src_stride_bayer, -src_stride_bayer, - row + kRowSize, width); - ARGBToUVRow(row, kRowSize, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - ARGBToYRow(row + kRowSize, dst_y + dst_stride_y, width); - src_bayer += src_stride_bayer * 2; - dst_y += dst_stride_y * 2; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - if (height & 1) { - BayerRow0(src_bayer, src_stride_bayer, row, width); - ARGBToUVRow(row, 0, dst_u, dst_v, width); - ARGBToYRow(row, dst_y, width); - } - free_aligned_buffer_64(row); - } - return 0; -} - -// Convert I420 to Bayer. -LIBYUV_API -int I420ToBayer(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_bayer, int dst_stride_bayer, - int width, int height, - uint32 dst_fourcc_bayer) { - void (*I422ToARGBRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToARGBRow_C; - void (*ARGBToBayerRow)(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) = ARGBToBayerRow_C; - const int blue_index = 0; // Offsets for ARGB format - const int green_index = 1; - const int red_index = 2; - uint32 index_map[2]; - // Negative height means invert the image. - if (height < 0) { - int halfheight; - height = -height; - halfheight = (height + 1) >> 1; - src_y = src_y + (height - 1) * src_stride_y; - src_u = src_u + (halfheight - 1) * src_stride_u; - src_v = src_v + (halfheight - 1) * src_stride_v; - src_stride_y = -src_stride_y; - src_stride_u = -src_stride_u; - src_stride_v = -src_stride_v; - } -#if defined(HAS_I422TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToARGBRow = I422ToARGBRow_SSSE3; - } - } -#endif -#if defined(HAS_I422TOARGBROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 16) { - I422ToARGBRow = I422ToARGBRow_Any_AVX2; - if (IS_ALIGNED(width, 16)) { - I422ToARGBRow = I422ToARGBRow_AVX2; - } - } -#endif -#if defined(HAS_I422TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - I422ToARGBRow = I422ToARGBRow_NEON; - } - } -#endif -#if defined(HAS_I422TOARGBROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_y, 4) && IS_ALIGNED(src_stride_y, 4) && - IS_ALIGNED(src_u, 2) && IS_ALIGNED(src_stride_u, 2) && - IS_ALIGNED(src_v, 2) && IS_ALIGNED(src_stride_v, 2)) { - I422ToARGBRow = I422ToARGBRow_MIPS_DSPR2; - } -#endif - -#if defined(HAS_ARGBTOBAYERROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - ARGBToBayerRow = ARGBToBayerRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerRow_SSSE3; - } - } -#elif defined(HAS_ARGBTOBAYERROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToBayerRow = ARGBToBayerRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerRow_NEON; - } - } -#endif - - if (MakeSelectors(blue_index, green_index, red_index, - dst_fourcc_bayer, index_map)) { - return -1; // Bad FourCC - } - { - // Allocate a row of ARGB. - align_buffer_64(row, width * 4); - int y; - for (y = 0; y < height; ++y) { - I422ToARGBRow(src_y, src_u, src_v, row, width); - ARGBToBayerRow(row, dst_bayer, index_map[y & 1], width); - dst_bayer += dst_stride_bayer; - src_y += src_stride_y; - if (y & 1) { - src_u += src_stride_u; - src_v += src_stride_v; - } - } - free_aligned_buffer_64(row); - } - return 0; -} - -#define MAKEBAYERFOURCC(BAYER) \ -LIBYUV_API \ -int Bayer##BAYER##ToI420(const uint8* src_bayer, int src_stride_bayer, \ - uint8* dst_y, int dst_stride_y, \ - uint8* dst_u, int dst_stride_u, \ - uint8* dst_v, int dst_stride_v, \ - int width, int height) { \ - return BayerToI420(src_bayer, src_stride_bayer, \ - dst_y, dst_stride_y, \ - dst_u, dst_stride_u, \ - dst_v, dst_stride_v, \ - width, height, \ - FOURCC_##BAYER); \ -} \ - \ -LIBYUV_API \ -int I420ToBayer##BAYER(const uint8* src_y, int src_stride_y, \ - const uint8* src_u, int src_stride_u, \ - const uint8* src_v, int src_stride_v, \ - uint8* dst_bayer, int dst_stride_bayer, \ - int width, int height) { \ - return I420ToBayer(src_y, src_stride_y, \ - src_u, src_stride_u, \ - src_v, src_stride_v, \ - dst_bayer, dst_stride_bayer, \ - width, height, \ - FOURCC_##BAYER); \ -} \ - \ -LIBYUV_API \ -int ARGBToBayer##BAYER(const uint8* src_argb, int src_stride_argb, \ - uint8* dst_bayer, int dst_stride_bayer, \ - int width, int height) { \ - return ARGBToBayer(src_argb, src_stride_argb, \ - dst_bayer, dst_stride_bayer, \ - width, height, \ - FOURCC_##BAYER); \ -} \ - \ -LIBYUV_API \ -int Bayer##BAYER##ToARGB(const uint8* src_bayer, int src_stride_bayer, \ - uint8* dst_argb, int dst_stride_argb, \ - int width, int height) { \ - return BayerToARGB(src_bayer, src_stride_bayer, \ - dst_argb, dst_stride_argb, \ - width, height, \ - FOURCC_##BAYER); \ -} - -MAKEBAYERFOURCC(BGGR) -MAKEBAYERFOURCC(GBRG) -MAKEBAYERFOURCC(GRBG) -MAKEBAYERFOURCC(RGGB) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/mjpeg_decoder.cc b/src/main/jni/libyuv/source/mjpeg_decoder.cc deleted file mode 100644 index 36028c3cc..000000000 --- a/src/main/jni/libyuv/source/mjpeg_decoder.cc +++ /dev/null @@ -1,566 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/mjpeg_decoder.h" - -#ifdef HAVE_JPEG -#include <assert.h> - -#if !defined(__pnacl__) && !defined(__CLR_VER) && \ - !defined(COVERAGE_ENABLED) && !defined(TARGET_IPHONE_SIMULATOR) -// Must be included before jpeglib. -#include <setjmp.h> -#define HAVE_SETJMP -#endif -struct FILE; // For jpeglib.h. - -// C++ build requires extern C for jpeg internals. -#ifdef __cplusplus -extern "C" { -#endif - -#include <jpeglib.h> - -#ifdef __cplusplus -} // extern "C" -#endif - -#include "libyuv/planar_functions.h" // For CopyPlane(). - -namespace libyuv { - -#ifdef HAVE_SETJMP -struct SetJmpErrorMgr { - jpeg_error_mgr base; // Must be at the top - jmp_buf setjmp_buffer; -}; -#endif - -const int MJpegDecoder::kColorSpaceUnknown = JCS_UNKNOWN; -const int MJpegDecoder::kColorSpaceGrayscale = JCS_GRAYSCALE; -const int MJpegDecoder::kColorSpaceRgb = JCS_RGB; -const int MJpegDecoder::kColorSpaceYCbCr = JCS_YCbCr; -const int MJpegDecoder::kColorSpaceCMYK = JCS_CMYK; -const int MJpegDecoder::kColorSpaceYCCK = JCS_YCCK; - -// Methods that are passed to jpeglib. -boolean fill_input_buffer(jpeg_decompress_struct* cinfo); -void init_source(jpeg_decompress_struct* cinfo); -void skip_input_data(jpeg_decompress_struct* cinfo, - long num_bytes); // NOLINT -void term_source(jpeg_decompress_struct* cinfo); -void ErrorHandler(jpeg_common_struct* cinfo); - -MJpegDecoder::MJpegDecoder() - : has_scanline_padding_(LIBYUV_FALSE), - num_outbufs_(0), - scanlines_(NULL), - scanlines_sizes_(NULL), - databuf_(NULL), - databuf_strides_(NULL) { - decompress_struct_ = new jpeg_decompress_struct; - source_mgr_ = new jpeg_source_mgr; -#ifdef HAVE_SETJMP - error_mgr_ = new SetJmpErrorMgr; - decompress_struct_->err = jpeg_std_error(&error_mgr_->base); - // Override standard exit()-based error handler. - error_mgr_->base.error_exit = &ErrorHandler; -#endif - decompress_struct_->client_data = NULL; - source_mgr_->init_source = &init_source; - source_mgr_->fill_input_buffer = &fill_input_buffer; - source_mgr_->skip_input_data = &skip_input_data; - source_mgr_->resync_to_restart = &jpeg_resync_to_restart; - source_mgr_->term_source = &term_source; - jpeg_create_decompress(decompress_struct_); - decompress_struct_->src = source_mgr_; - buf_vec_.buffers = &buf_; - buf_vec_.len = 1; -} - -MJpegDecoder::~MJpegDecoder() { - jpeg_destroy_decompress(decompress_struct_); - delete decompress_struct_; - delete source_mgr_; -#ifdef HAVE_SETJMP - delete error_mgr_; -#endif - DestroyOutputBuffers(); -} - -LIBYUV_BOOL MJpegDecoder::LoadFrame(const uint8* src, size_t src_len) { - if (!ValidateJpeg(src, src_len)) { - return LIBYUV_FALSE; - } - - buf_.data = src; - buf_.len = static_cast<int>(src_len); - buf_vec_.pos = 0; - decompress_struct_->client_data = &buf_vec_; -#ifdef HAVE_SETJMP - if (setjmp(error_mgr_->setjmp_buffer)) { - // We called jpeg_read_header, it experienced an error, and we called - // longjmp() and rewound the stack to here. Return error. - return LIBYUV_FALSE; - } -#endif - if (jpeg_read_header(decompress_struct_, TRUE) != JPEG_HEADER_OK) { - // ERROR: Bad MJPEG header - return LIBYUV_FALSE; - } - AllocOutputBuffers(GetNumComponents()); - for (int i = 0; i < num_outbufs_; ++i) { - int scanlines_size = GetComponentScanlinesPerImcuRow(i); - if (scanlines_sizes_[i] != scanlines_size) { - if (scanlines_[i]) { - delete scanlines_[i]; - } - scanlines_[i] = new uint8* [scanlines_size]; - scanlines_sizes_[i] = scanlines_size; - } - - // We allocate padding for the final scanline to pad it up to DCTSIZE bytes - // to avoid memory errors, since jpeglib only reads full MCUs blocks. For - // the preceding scanlines, the padding is not needed/wanted because the - // following addresses will already be valid (they are the initial bytes of - // the next scanline) and will be overwritten when jpeglib writes out that - // next scanline. - int databuf_stride = GetComponentStride(i); - int databuf_size = scanlines_size * databuf_stride; - if (databuf_strides_[i] != databuf_stride) { - if (databuf_[i]) { - delete databuf_[i]; - } - databuf_[i] = new uint8[databuf_size]; - databuf_strides_[i] = databuf_stride; - } - - if (GetComponentStride(i) != GetComponentWidth(i)) { - has_scanline_padding_ = LIBYUV_TRUE; - } - } - return LIBYUV_TRUE; -} - -static int DivideAndRoundUp(int numerator, int denominator) { - return (numerator + denominator - 1) / denominator; -} - -static int DivideAndRoundDown(int numerator, int denominator) { - return numerator / denominator; -} - -// Returns width of the last loaded frame. -int MJpegDecoder::GetWidth() { - return decompress_struct_->image_width; -} - -// Returns height of the last loaded frame. -int MJpegDecoder::GetHeight() { - return decompress_struct_->image_height; -} - -// Returns format of the last loaded frame. The return value is one of the -// kColorSpace* constants. -int MJpegDecoder::GetColorSpace() { - return decompress_struct_->jpeg_color_space; -} - -// Number of color components in the color space. -int MJpegDecoder::GetNumComponents() { - return decompress_struct_->num_components; -} - -// Sample factors of the n-th component. -int MJpegDecoder::GetHorizSampFactor(int component) { - return decompress_struct_->comp_info[component].h_samp_factor; -} - -int MJpegDecoder::GetVertSampFactor(int component) { - return decompress_struct_->comp_info[component].v_samp_factor; -} - -int MJpegDecoder::GetHorizSubSampFactor(int component) { - return decompress_struct_->max_h_samp_factor / - GetHorizSampFactor(component); -} - -int MJpegDecoder::GetVertSubSampFactor(int component) { - return decompress_struct_->max_v_samp_factor / - GetVertSampFactor(component); -} - -int MJpegDecoder::GetImageScanlinesPerImcuRow() { - return decompress_struct_->max_v_samp_factor * DCTSIZE; -} - -int MJpegDecoder::GetComponentScanlinesPerImcuRow(int component) { - int vs = GetVertSubSampFactor(component); - return DivideAndRoundUp(GetImageScanlinesPerImcuRow(), vs); -} - -int MJpegDecoder::GetComponentWidth(int component) { - int hs = GetHorizSubSampFactor(component); - return DivideAndRoundUp(GetWidth(), hs); -} - -int MJpegDecoder::GetComponentHeight(int component) { - int vs = GetVertSubSampFactor(component); - return DivideAndRoundUp(GetHeight(), vs); -} - -// Get width in bytes padded out to a multiple of DCTSIZE -int MJpegDecoder::GetComponentStride(int component) { - return (GetComponentWidth(component) + DCTSIZE - 1) & ~(DCTSIZE - 1); -} - -int MJpegDecoder::GetComponentSize(int component) { - return GetComponentWidth(component) * GetComponentHeight(component); -} - -LIBYUV_BOOL MJpegDecoder::UnloadFrame() { -#ifdef HAVE_SETJMP - if (setjmp(error_mgr_->setjmp_buffer)) { - // We called jpeg_abort_decompress, it experienced an error, and we called - // longjmp() and rewound the stack to here. Return error. - return LIBYUV_FALSE; - } -#endif - jpeg_abort_decompress(decompress_struct_); - return LIBYUV_TRUE; -} - -// TODO(fbarchard): Allow rectangle to be specified: x, y, width, height. -LIBYUV_BOOL MJpegDecoder::DecodeToBuffers( - uint8** planes, int dst_width, int dst_height) { - if (dst_width != GetWidth() || - dst_height > GetHeight()) { - // ERROR: Bad dimensions - return LIBYUV_FALSE; - } -#ifdef HAVE_SETJMP - if (setjmp(error_mgr_->setjmp_buffer)) { - // We called into jpeglib, it experienced an error sometime during this - // function call, and we called longjmp() and rewound the stack to here. - // Return error. - return LIBYUV_FALSE; - } -#endif - if (!StartDecode()) { - return LIBYUV_FALSE; - } - SetScanlinePointers(databuf_); - int lines_left = dst_height; - // Compute amount of lines to skip to implement vertical crop. - // TODO(fbarchard): Ensure skip is a multiple of maximum component - // subsample. ie 2 - int skip = (GetHeight() - dst_height) / 2; - if (skip > 0) { - // There is no API to skip lines in the output data, so we read them - // into the temp buffer. - while (skip >= GetImageScanlinesPerImcuRow()) { - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - skip -= GetImageScanlinesPerImcuRow(); - } - if (skip > 0) { - // Have a partial iMCU row left over to skip. Must read it and then - // copy the parts we want into the destination. - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - for (int i = 0; i < num_outbufs_; ++i) { - // TODO(fbarchard): Compute skip to avoid this - assert(skip % GetVertSubSampFactor(i) == 0); - int rows_to_skip = - DivideAndRoundDown(skip, GetVertSubSampFactor(i)); - int scanlines_to_copy = GetComponentScanlinesPerImcuRow(i) - - rows_to_skip; - int data_to_skip = rows_to_skip * GetComponentStride(i); - CopyPlane(databuf_[i] + data_to_skip, GetComponentStride(i), - planes[i], GetComponentWidth(i), - GetComponentWidth(i), scanlines_to_copy); - planes[i] += scanlines_to_copy * GetComponentWidth(i); - } - lines_left -= (GetImageScanlinesPerImcuRow() - skip); - } - } - - // Read full MCUs but cropped horizontally - for (; lines_left > GetImageScanlinesPerImcuRow(); - lines_left -= GetImageScanlinesPerImcuRow()) { - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - for (int i = 0; i < num_outbufs_; ++i) { - int scanlines_to_copy = GetComponentScanlinesPerImcuRow(i); - CopyPlane(databuf_[i], GetComponentStride(i), - planes[i], GetComponentWidth(i), - GetComponentWidth(i), scanlines_to_copy); - planes[i] += scanlines_to_copy * GetComponentWidth(i); - } - } - - if (lines_left > 0) { - // Have a partial iMCU row left over to decode. - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - for (int i = 0; i < num_outbufs_; ++i) { - int scanlines_to_copy = - DivideAndRoundUp(lines_left, GetVertSubSampFactor(i)); - CopyPlane(databuf_[i], GetComponentStride(i), - planes[i], GetComponentWidth(i), - GetComponentWidth(i), scanlines_to_copy); - planes[i] += scanlines_to_copy * GetComponentWidth(i); - } - } - return FinishDecode(); -} - -LIBYUV_BOOL MJpegDecoder::DecodeToCallback(CallbackFunction fn, void* opaque, - int dst_width, int dst_height) { - if (dst_width != GetWidth() || - dst_height > GetHeight()) { - // ERROR: Bad dimensions - return LIBYUV_FALSE; - } -#ifdef HAVE_SETJMP - if (setjmp(error_mgr_->setjmp_buffer)) { - // We called into jpeglib, it experienced an error sometime during this - // function call, and we called longjmp() and rewound the stack to here. - // Return error. - return LIBYUV_FALSE; - } -#endif - if (!StartDecode()) { - return LIBYUV_FALSE; - } - SetScanlinePointers(databuf_); - int lines_left = dst_height; - // TODO(fbarchard): Compute amount of lines to skip to implement vertical crop - int skip = (GetHeight() - dst_height) / 2; - if (skip > 0) { - while (skip >= GetImageScanlinesPerImcuRow()) { - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - skip -= GetImageScanlinesPerImcuRow(); - } - if (skip > 0) { - // Have a partial iMCU row left over to skip. - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - for (int i = 0; i < num_outbufs_; ++i) { - // TODO(fbarchard): Compute skip to avoid this - assert(skip % GetVertSubSampFactor(i) == 0); - int rows_to_skip = DivideAndRoundDown(skip, GetVertSubSampFactor(i)); - int data_to_skip = rows_to_skip * GetComponentStride(i); - // Change our own data buffer pointers so we can pass them to the - // callback. - databuf_[i] += data_to_skip; - } - int scanlines_to_copy = GetImageScanlinesPerImcuRow() - skip; - (*fn)(opaque, databuf_, databuf_strides_, scanlines_to_copy); - // Now change them back. - for (int i = 0; i < num_outbufs_; ++i) { - int rows_to_skip = DivideAndRoundDown(skip, GetVertSubSampFactor(i)); - int data_to_skip = rows_to_skip * GetComponentStride(i); - databuf_[i] -= data_to_skip; - } - lines_left -= scanlines_to_copy; - } - } - // Read full MCUs until we get to the crop point. - for (; lines_left >= GetImageScanlinesPerImcuRow(); - lines_left -= GetImageScanlinesPerImcuRow()) { - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - (*fn)(opaque, databuf_, databuf_strides_, GetImageScanlinesPerImcuRow()); - } - if (lines_left > 0) { - // Have a partial iMCU row left over to decode. - if (!DecodeImcuRow()) { - FinishDecode(); - return LIBYUV_FALSE; - } - (*fn)(opaque, databuf_, databuf_strides_, lines_left); - } - return FinishDecode(); -} - -void init_source(j_decompress_ptr cinfo) { - fill_input_buffer(cinfo); -} - -boolean fill_input_buffer(j_decompress_ptr cinfo) { - BufferVector* buf_vec = reinterpret_cast<BufferVector*>(cinfo->client_data); - if (buf_vec->pos >= buf_vec->len) { - assert(0 && "No more data"); - // ERROR: No more data - return FALSE; - } - cinfo->src->next_input_byte = buf_vec->buffers[buf_vec->pos].data; - cinfo->src->bytes_in_buffer = buf_vec->buffers[buf_vec->pos].len; - ++buf_vec->pos; - return TRUE; -} - -void skip_input_data(j_decompress_ptr cinfo, - long num_bytes) { // NOLINT - cinfo->src->next_input_byte += num_bytes; -} - -void term_source(j_decompress_ptr cinfo) { - // Nothing to do. -} - -#ifdef HAVE_SETJMP -void ErrorHandler(j_common_ptr cinfo) { - // This is called when a jpeglib command experiences an error. Unfortunately - // jpeglib's error handling model is not very flexible, because it expects the - // error handler to not return--i.e., it wants the program to terminate. To - // recover from errors we use setjmp() as shown in their example. setjmp() is - // C's implementation for the "call with current continuation" functionality - // seen in some functional programming languages. - // A formatted message can be output, but is unsafe for release. -#ifdef DEBUG - char buf[JMSG_LENGTH_MAX]; - (*cinfo->err->format_message)(cinfo, buf); - // ERROR: Error in jpeglib: buf -#endif - - SetJmpErrorMgr* mgr = reinterpret_cast<SetJmpErrorMgr*>(cinfo->err); - // This rewinds the call stack to the point of the corresponding setjmp() - // and causes it to return (for a second time) with value 1. - longjmp(mgr->setjmp_buffer, 1); -} -#endif - -void MJpegDecoder::AllocOutputBuffers(int num_outbufs) { - if (num_outbufs != num_outbufs_) { - // We could perhaps optimize this case to resize the output buffers without - // necessarily having to delete and recreate each one, but it's not worth - // it. - DestroyOutputBuffers(); - - scanlines_ = new uint8** [num_outbufs]; - scanlines_sizes_ = new int[num_outbufs]; - databuf_ = new uint8* [num_outbufs]; - databuf_strides_ = new int[num_outbufs]; - - for (int i = 0; i < num_outbufs; ++i) { - scanlines_[i] = NULL; - scanlines_sizes_[i] = 0; - databuf_[i] = NULL; - databuf_strides_[i] = 0; - } - - num_outbufs_ = num_outbufs; - } -} - -void MJpegDecoder::DestroyOutputBuffers() { - for (int i = 0; i < num_outbufs_; ++i) { - delete [] scanlines_[i]; - delete [] databuf_[i]; - } - delete [] scanlines_; - delete [] databuf_; - delete [] scanlines_sizes_; - delete [] databuf_strides_; - scanlines_ = NULL; - databuf_ = NULL; - scanlines_sizes_ = NULL; - databuf_strides_ = NULL; - num_outbufs_ = 0; -} - -// JDCT_IFAST and do_block_smoothing improve performance substantially. -LIBYUV_BOOL MJpegDecoder::StartDecode() { - decompress_struct_->raw_data_out = TRUE; - decompress_struct_->dct_method = JDCT_IFAST; // JDCT_ISLOW is default - decompress_struct_->dither_mode = JDITHER_NONE; - // Not applicable to 'raw': - decompress_struct_->do_fancy_upsampling = (boolean)(LIBYUV_FALSE); - // Only for buffered mode: - decompress_struct_->enable_2pass_quant = (boolean)(LIBYUV_FALSE); - // Blocky but fast: - decompress_struct_->do_block_smoothing = (boolean)(LIBYUV_FALSE); - - if (!jpeg_start_decompress(decompress_struct_)) { - // ERROR: Couldn't start JPEG decompressor"; - return LIBYUV_FALSE; - } - return LIBYUV_TRUE; -} - -LIBYUV_BOOL MJpegDecoder::FinishDecode() { - // jpeglib considers it an error if we finish without decoding the whole - // image, so we call "abort" rather than "finish". - jpeg_abort_decompress(decompress_struct_); - return LIBYUV_TRUE; -} - -void MJpegDecoder::SetScanlinePointers(uint8** data) { - for (int i = 0; i < num_outbufs_; ++i) { - uint8* data_i = data[i]; - for (int j = 0; j < scanlines_sizes_[i]; ++j) { - scanlines_[i][j] = data_i; - data_i += GetComponentStride(i); - } - } -} - -inline LIBYUV_BOOL MJpegDecoder::DecodeImcuRow() { - return (unsigned int)(GetImageScanlinesPerImcuRow()) == - jpeg_read_raw_data(decompress_struct_, - scanlines_, - GetImageScanlinesPerImcuRow()); -} - -// The helper function which recognizes the jpeg sub-sampling type. -JpegSubsamplingType MJpegDecoder::JpegSubsamplingTypeHelper( - int* subsample_x, int* subsample_y, int number_of_components) { - if (number_of_components == 3) { // Color images. - if (subsample_x[0] == 1 && subsample_y[0] == 1 && - subsample_x[1] == 2 && subsample_y[1] == 2 && - subsample_x[2] == 2 && subsample_y[2] == 2) { - return kJpegYuv420; - } else if (subsample_x[0] == 1 && subsample_y[0] == 1 && - subsample_x[1] == 2 && subsample_y[1] == 1 && - subsample_x[2] == 2 && subsample_y[2] == 1) { - return kJpegYuv422; - } else if (subsample_x[0] == 1 && subsample_y[0] == 1 && - subsample_x[1] == 1 && subsample_y[1] == 1 && - subsample_x[2] == 1 && subsample_y[2] == 1) { - return kJpegYuv444; - } - } else if (number_of_components == 1) { // Grey-scale images. - if (subsample_x[0] == 1 && subsample_y[0] == 1) { - return kJpegYuv400; - } - } - return kJpegUnknown; -} - -} // namespace libyuv -#endif // HAVE_JPEG - diff --git a/src/main/jni/libyuv/source/mjpeg_validate.cc b/src/main/jni/libyuv/source/mjpeg_validate.cc deleted file mode 100644 index 23d22d099..000000000 --- a/src/main/jni/libyuv/source/mjpeg_validate.cc +++ /dev/null @@ -1,47 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/mjpeg_decoder.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Helper function to validate the jpeg appears intact. -// TODO(fbarchard): Optimize case where SOI is found but EOI is not. -LIBYUV_BOOL ValidateJpeg(const uint8* sample, size_t sample_size) { - size_t i; - if (sample_size < 64) { - // ERROR: Invalid jpeg size: sample_size - return LIBYUV_FALSE; - } - if (sample[0] != 0xff || sample[1] != 0xd8) { // Start Of Image - // ERROR: Invalid jpeg initial start code - return LIBYUV_FALSE; - } - for (i = sample_size - 2; i > 1;) { - if (sample[i] != 0xd9) { - if (sample[i] == 0xff && sample[i + 1] == 0xd9) { // End Of Image - return LIBYUV_TRUE; // Success: Valid jpeg. - } - --i; - } - --i; - } - // ERROR: Invalid jpeg end code not found. Size sample_size - return LIBYUV_FALSE; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - diff --git a/src/main/jni/libyuv/source/planar_functions.cc b/src/main/jni/libyuv/source/planar_functions.cc deleted file mode 100644 index 3857008ca..000000000 --- a/src/main/jni/libyuv/source/planar_functions.cc +++ /dev/null @@ -1,2291 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/planar_functions.h" - -#include <string.h> // for memset() - -#include "libyuv/cpu_id.h" -#ifdef HAVE_JPEG -#include "libyuv/mjpeg_decoder.h" -#endif -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Copy a plane of data -LIBYUV_API -void CopyPlane(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height) { - int y; - void (*CopyRow)(const uint8* src, uint8* dst, int width) = CopyRow_C; - // Coalesce rows. - if (src_stride_y == width && - dst_stride_y == width) { - width *= height; - height = 1; - src_stride_y = dst_stride_y = 0; - } - // Nothing to do. - if (src_y == dst_y && src_stride_y == dst_stride_y) { - return; - } -#if defined(HAS_COPYROW_X86) - if (TestCpuFlag(kCpuHasX86) && IS_ALIGNED(width, 4)) { - CopyRow = CopyRow_X86; - } -#endif -#if defined(HAS_COPYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 32) && - IS_ALIGNED(src_y, 16) && IS_ALIGNED(src_stride_y, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - CopyRow = CopyRow_SSE2; - } -#endif -#if defined(HAS_COPYROW_ERMS) - if (TestCpuFlag(kCpuHasERMS)) { - CopyRow = CopyRow_ERMS; - } -#endif -#if defined(HAS_COPYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 32)) { - CopyRow = CopyRow_NEON; - } -#endif -#if defined(HAS_COPYROW_MIPS) - if (TestCpuFlag(kCpuHasMIPS)) { - CopyRow = CopyRow_MIPS; - } -#endif - - // Copy plane - for (y = 0; y < height; ++y) { - CopyRow(src_y, dst_y, width); - src_y += src_stride_y; - dst_y += dst_stride_y; - } -} - -LIBYUV_API -void CopyPlane_16(const uint16* src_y, int src_stride_y, - uint16* dst_y, int dst_stride_y, - int width, int height) { - int y; - void (*CopyRow)(const uint16* src, uint16* dst, int width) = CopyRow_16_C; - // Coalesce rows. - if (src_stride_y == width && - dst_stride_y == width) { - width *= height; - height = 1; - src_stride_y = dst_stride_y = 0; - } -#if defined(HAS_COPYROW_16_X86) - if (TestCpuFlag(kCpuHasX86) && IS_ALIGNED(width, 4)) { - CopyRow = CopyRow_16_X86; - } -#endif -#if defined(HAS_COPYROW_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 32) && - IS_ALIGNED(src_y, 16) && IS_ALIGNED(src_stride_y, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - CopyRow = CopyRow_16_SSE2; - } -#endif -#if defined(HAS_COPYROW_16_ERMS) - if (TestCpuFlag(kCpuHasERMS)) { - CopyRow = CopyRow_16_ERMS; - } -#endif -#if defined(HAS_COPYROW_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 32)) { - CopyRow = CopyRow_16_NEON; - } -#endif -#if defined(HAS_COPYROW_16_MIPS) - if (TestCpuFlag(kCpuHasMIPS)) { - CopyRow = CopyRow_16_MIPS; - } -#endif - - // Copy plane - for (y = 0; y < height; ++y) { - CopyRow(src_y, dst_y, width); - src_y += src_stride_y; - dst_y += dst_stride_y; - } -} - -// Copy I422. -LIBYUV_API -int I422Copy(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int halfwidth = (width + 1) >> 1; - if (!src_y || !src_u || !src_v || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_u = src_u + (height - 1) * src_stride_u; - src_v = src_v + (height - 1) * src_stride_v; - src_stride_y = -src_stride_y; - src_stride_u = -src_stride_u; - src_stride_v = -src_stride_v; - } - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - CopyPlane(src_u, src_stride_u, dst_u, dst_stride_u, halfwidth, height); - CopyPlane(src_v, src_stride_v, dst_v, dst_stride_v, halfwidth, height); - return 0; -} - -// Copy I444. -LIBYUV_API -int I444Copy(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - if (!src_y || !src_u || !src_v || - !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_u = src_u + (height - 1) * src_stride_u; - src_v = src_v + (height - 1) * src_stride_v; - src_stride_y = -src_stride_y; - src_stride_u = -src_stride_u; - src_stride_v = -src_stride_v; - } - - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - CopyPlane(src_u, src_stride_u, dst_u, dst_stride_u, width, height); - CopyPlane(src_v, src_stride_v, dst_v, dst_stride_v, width, height); - return 0; -} - -// Copy I400. -LIBYUV_API -int I400ToI400(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height) { - if (!src_y || !dst_y || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - return 0; -} - -// Convert I420 to I400. -LIBYUV_API -int I420ToI400(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - int width, int height) { - if (!src_y || !dst_y || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - CopyPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - return 0; -} - -// Mirror a plane of data. -void MirrorPlane(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height) { - int y; - void (*MirrorRow)(const uint8* src, uint8* dst, int width) = MirrorRow_C; - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } -#if defined(HAS_MIRRORROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 16)) { - MirrorRow = MirrorRow_NEON; - } -#endif -#if defined(HAS_MIRRORROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 16)) { - MirrorRow = MirrorRow_SSE2; - } -#endif -#if defined(HAS_MIRRORROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 16) && - IS_ALIGNED(src_y, 16) && IS_ALIGNED(src_stride_y, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - MirrorRow = MirrorRow_SSSE3; - } -#endif -#if defined(HAS_MIRRORROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && IS_ALIGNED(width, 32)) { - MirrorRow = MirrorRow_AVX2; - } -#endif - - // Mirror plane - for (y = 0; y < height; ++y) { - MirrorRow(src_y, dst_y, width); - src_y += src_stride_y; - dst_y += dst_stride_y; - } -} - -// Convert YUY2 to I422. -LIBYUV_API -int YUY2ToI422(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*YUY2ToUV422Row)(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix) = - YUY2ToUV422Row_C; - void (*YUY2ToYRow)(const uint8* src_yuy2, uint8* dst_y, int pix) = - YUY2ToYRow_C; - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_yuy2 = src_yuy2 + (height - 1) * src_stride_yuy2; - src_stride_yuy2 = -src_stride_yuy2; - } - // Coalesce rows. - if (src_stride_yuy2 == width * 2 && - dst_stride_y == width && - dst_stride_u * 2 == width && - dst_stride_v * 2 == width) { - width *= height; - height = 1; - src_stride_yuy2 = dst_stride_y = dst_stride_u = dst_stride_v = 0; - } -#if defined(HAS_YUY2TOYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - YUY2ToUV422Row = YUY2ToUV422Row_Any_SSE2; - YUY2ToYRow = YUY2ToYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - YUY2ToUV422Row = YUY2ToUV422Row_Unaligned_SSE2; - YUY2ToYRow = YUY2ToYRow_Unaligned_SSE2; - if (IS_ALIGNED(src_yuy2, 16) && IS_ALIGNED(src_stride_yuy2, 16)) { - YUY2ToUV422Row = YUY2ToUV422Row_SSE2; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - YUY2ToYRow = YUY2ToYRow_SSE2; - } - } - } - } -#endif -#if defined(HAS_YUY2TOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - YUY2ToUV422Row = YUY2ToUV422Row_Any_AVX2; - YUY2ToYRow = YUY2ToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - YUY2ToUV422Row = YUY2ToUV422Row_AVX2; - YUY2ToYRow = YUY2ToYRow_AVX2; - } - } -#endif -#if defined(HAS_YUY2TOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - YUY2ToYRow = YUY2ToYRow_Any_NEON; - if (width >= 16) { - YUY2ToUV422Row = YUY2ToUV422Row_Any_NEON; - } - if (IS_ALIGNED(width, 16)) { - YUY2ToYRow = YUY2ToYRow_NEON; - YUY2ToUV422Row = YUY2ToUV422Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - YUY2ToUV422Row(src_yuy2, dst_u, dst_v, width); - YUY2ToYRow(src_yuy2, dst_y, width); - src_yuy2 += src_stride_yuy2; - dst_y += dst_stride_y; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - return 0; -} - -// Convert UYVY to I422. -LIBYUV_API -int UYVYToI422(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int y; - void (*UYVYToUV422Row)(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix) = - UYVYToUV422Row_C; - void (*UYVYToYRow)(const uint8* src_uyvy, - uint8* dst_y, int pix) = UYVYToYRow_C; - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_uyvy = src_uyvy + (height - 1) * src_stride_uyvy; - src_stride_uyvy = -src_stride_uyvy; - } - // Coalesce rows. - if (src_stride_uyvy == width * 2 && - dst_stride_y == width && - dst_stride_u * 2 == width && - dst_stride_v * 2 == width) { - width *= height; - height = 1; - src_stride_uyvy = dst_stride_y = dst_stride_u = dst_stride_v = 0; - } -#if defined(HAS_UYVYTOYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 16) { - UYVYToUV422Row = UYVYToUV422Row_Any_SSE2; - UYVYToYRow = UYVYToYRow_Any_SSE2; - if (IS_ALIGNED(width, 16)) { - UYVYToUV422Row = UYVYToUV422Row_Unaligned_SSE2; - UYVYToYRow = UYVYToYRow_Unaligned_SSE2; - if (IS_ALIGNED(src_uyvy, 16) && IS_ALIGNED(src_stride_uyvy, 16)) { - UYVYToUV422Row = UYVYToUV422Row_SSE2; - if (IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - UYVYToYRow = UYVYToYRow_SSE2; - } - } - } - } -#endif -#if defined(HAS_UYVYTOYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 32) { - UYVYToUV422Row = UYVYToUV422Row_Any_AVX2; - UYVYToYRow = UYVYToYRow_Any_AVX2; - if (IS_ALIGNED(width, 32)) { - UYVYToUV422Row = UYVYToUV422Row_AVX2; - UYVYToYRow = UYVYToYRow_AVX2; - } - } -#endif -#if defined(HAS_UYVYTOYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - UYVYToYRow = UYVYToYRow_Any_NEON; - if (width >= 16) { - UYVYToUV422Row = UYVYToUV422Row_Any_NEON; - } - if (IS_ALIGNED(width, 16)) { - UYVYToYRow = UYVYToYRow_NEON; - UYVYToUV422Row = UYVYToUV422Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - UYVYToUV422Row(src_uyvy, dst_u, dst_v, width); - UYVYToYRow(src_uyvy, dst_y, width); - src_uyvy += src_stride_uyvy; - dst_y += dst_stride_y; - dst_u += dst_stride_u; - dst_v += dst_stride_v; - } - return 0; -} - -// Mirror I400 with optional flipping -LIBYUV_API -int I400Mirror(const uint8* src_y, int src_stride_y, - uint8* dst_y, int dst_stride_y, - int width, int height) { - if (!src_y || !dst_y || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - - MirrorPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - return 0; -} - -// Mirror I420 with optional flipping -LIBYUV_API -int I420Mirror(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height) { - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - if (!src_y || !src_u || !src_v || !dst_y || !dst_u || !dst_v || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - src_y = src_y + (height - 1) * src_stride_y; - src_u = src_u + (halfheight - 1) * src_stride_u; - src_v = src_v + (halfheight - 1) * src_stride_v; - src_stride_y = -src_stride_y; - src_stride_u = -src_stride_u; - src_stride_v = -src_stride_v; - } - - if (dst_y) { - MirrorPlane(src_y, src_stride_y, dst_y, dst_stride_y, width, height); - } - MirrorPlane(src_u, src_stride_u, dst_u, dst_stride_u, halfwidth, halfheight); - MirrorPlane(src_v, src_stride_v, dst_v, dst_stride_v, halfwidth, halfheight); - return 0; -} - -// ARGB mirror. -LIBYUV_API -int ARGBMirror(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBMirrorRow)(const uint8* src, uint8* dst, int width) = - ARGBMirrorRow_C; - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - -#if defined(HAS_ARGBMIRRORROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBMirrorRow = ARGBMirrorRow_SSSE3; - } -#endif -#if defined(HAS_ARGBMIRRORROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && IS_ALIGNED(width, 8)) { - ARGBMirrorRow = ARGBMirrorRow_AVX2; - } -#endif -#if defined(HAS_ARGBMIRRORROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 4)) { - ARGBMirrorRow = ARGBMirrorRow_NEON; - } -#endif - - // Mirror plane - for (y = 0; y < height; ++y) { - ARGBMirrorRow(src_argb, dst_argb, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Get a blender that optimized for the CPU, alignment and pixel count. -// As there are 6 blenders to choose from, the caller should try to use -// the same blend function for all pixels if possible. -LIBYUV_API -ARGBBlendRow GetARGBBlend() { - void (*ARGBBlendRow)(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width) = ARGBBlendRow_C; -#if defined(HAS_ARGBBLENDROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3)) { - ARGBBlendRow = ARGBBlendRow_SSSE3; - return ARGBBlendRow; - } -#endif -#if defined(HAS_ARGBBLENDROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2)) { - ARGBBlendRow = ARGBBlendRow_SSE2; - } -#endif -#if defined(HAS_ARGBBLENDROW_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - ARGBBlendRow = ARGBBlendRow_NEON; - } -#endif - return ARGBBlendRow; -} - -// Alpha Blend 2 ARGB images and store to destination. -LIBYUV_API -int ARGBBlend(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBBlendRow)(const uint8* src_argb, const uint8* src_argb1, - uint8* dst_argb, int width) = GetARGBBlend(); - if (!src_argb0 || !src_argb1 || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_argb0 == width * 4 && - src_stride_argb1 == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb0 = src_stride_argb1 = dst_stride_argb = 0; - } - - for (y = 0; y < height; ++y) { - ARGBBlendRow(src_argb0, src_argb1, dst_argb, width); - src_argb0 += src_stride_argb0; - src_argb1 += src_stride_argb1; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Multiply 2 ARGB images and store to destination. -LIBYUV_API -int ARGBMultiply(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBMultiplyRow)(const uint8* src0, const uint8* src1, uint8* dst, - int width) = ARGBMultiplyRow_C; - if (!src_argb0 || !src_argb1 || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_argb0 == width * 4 && - src_stride_argb1 == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb0 = src_stride_argb1 = dst_stride_argb = 0; - } -#if defined(HAS_ARGBMULTIPLYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4) { - ARGBMultiplyRow = ARGBMultiplyRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBMultiplyRow = ARGBMultiplyRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBMULTIPLYROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 8) { - ARGBMultiplyRow = ARGBMultiplyRow_Any_AVX2; - if (IS_ALIGNED(width, 8)) { - ARGBMultiplyRow = ARGBMultiplyRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBMULTIPLYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBMultiplyRow = ARGBMultiplyRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBMultiplyRow = ARGBMultiplyRow_NEON; - } - } -#endif - - // Multiply plane - for (y = 0; y < height; ++y) { - ARGBMultiplyRow(src_argb0, src_argb1, dst_argb, width); - src_argb0 += src_stride_argb0; - src_argb1 += src_stride_argb1; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Add 2 ARGB images and store to destination. -LIBYUV_API -int ARGBAdd(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBAddRow)(const uint8* src0, const uint8* src1, uint8* dst, - int width) = ARGBAddRow_C; - if (!src_argb0 || !src_argb1 || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_argb0 == width * 4 && - src_stride_argb1 == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb0 = src_stride_argb1 = dst_stride_argb = 0; - } -#if defined(HAS_ARGBADDROW_SSE2) && defined(_MSC_VER) - if (TestCpuFlag(kCpuHasSSE2)) { - ARGBAddRow = ARGBAddRow_SSE2; - } -#endif -#if defined(HAS_ARGBADDROW_SSE2) && !defined(_MSC_VER) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4) { - ARGBAddRow = ARGBAddRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBAddRow = ARGBAddRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBADDROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 8) { - ARGBAddRow = ARGBAddRow_Any_AVX2; - if (IS_ALIGNED(width, 8)) { - ARGBAddRow = ARGBAddRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBADDROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBAddRow = ARGBAddRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBAddRow = ARGBAddRow_NEON; - } - } -#endif - - // Add plane - for (y = 0; y < height; ++y) { - ARGBAddRow(src_argb0, src_argb1, dst_argb, width); - src_argb0 += src_stride_argb0; - src_argb1 += src_stride_argb1; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Subtract 2 ARGB images and store to destination. -LIBYUV_API -int ARGBSubtract(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBSubtractRow)(const uint8* src0, const uint8* src1, uint8* dst, - int width) = ARGBSubtractRow_C; - if (!src_argb0 || !src_argb1 || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_argb0 == width * 4 && - src_stride_argb1 == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb0 = src_stride_argb1 = dst_stride_argb = 0; - } -#if defined(HAS_ARGBSUBTRACTROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4) { - ARGBSubtractRow = ARGBSubtractRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBSubtractRow = ARGBSubtractRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBSUBTRACTROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 8) { - ARGBSubtractRow = ARGBSubtractRow_Any_AVX2; - if (IS_ALIGNED(width, 8)) { - ARGBSubtractRow = ARGBSubtractRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBSUBTRACTROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBSubtractRow = ARGBSubtractRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBSubtractRow = ARGBSubtractRow_NEON; - } - } -#endif - - // Subtract plane - for (y = 0; y < height; ++y) { - ARGBSubtractRow(src_argb0, src_argb1, dst_argb, width); - src_argb0 += src_stride_argb0; - src_argb1 += src_stride_argb1; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert I422 to BGRA. -LIBYUV_API -int I422ToBGRA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_bgra, int dst_stride_bgra, - int width, int height) { - int y; - void (*I422ToBGRARow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToBGRARow_C; - if (!src_y || !src_u || !src_v || - !dst_bgra || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_bgra = dst_bgra + (height - 1) * dst_stride_bgra; - dst_stride_bgra = -dst_stride_bgra; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 2 == width && - src_stride_v * 2 == width && - dst_stride_bgra == width * 4) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_bgra = 0; - } -#if defined(HAS_I422TOBGRAROW_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - I422ToBGRARow = I422ToBGRARow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToBGRARow = I422ToBGRARow_NEON; - } - } -#elif defined(HAS_I422TOBGRAROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToBGRARow = I422ToBGRARow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToBGRARow = I422ToBGRARow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_bgra, 16) && IS_ALIGNED(dst_stride_bgra, 16)) { - I422ToBGRARow = I422ToBGRARow_SSSE3; - } - } - } -#elif defined(HAS_I422TOBGRAROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_y, 4) && IS_ALIGNED(src_stride_y, 4) && - IS_ALIGNED(src_u, 2) && IS_ALIGNED(src_stride_u, 2) && - IS_ALIGNED(src_v, 2) && IS_ALIGNED(src_stride_v, 2) && - IS_ALIGNED(dst_bgra, 4) && IS_ALIGNED(dst_stride_bgra, 4)) { - I422ToBGRARow = I422ToBGRARow_MIPS_DSPR2; - } -#endif - - for (y = 0; y < height; ++y) { - I422ToBGRARow(src_y, src_u, src_v, dst_bgra, width); - dst_bgra += dst_stride_bgra; - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - } - return 0; -} - -// Convert I422 to ABGR. -LIBYUV_API -int I422ToABGR(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_abgr, int dst_stride_abgr, - int width, int height) { - int y; - void (*I422ToABGRRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToABGRRow_C; - if (!src_y || !src_u || !src_v || - !dst_abgr || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_abgr = dst_abgr + (height - 1) * dst_stride_abgr; - dst_stride_abgr = -dst_stride_abgr; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 2 == width && - src_stride_v * 2 == width && - dst_stride_abgr == width * 4) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_abgr = 0; - } -#if defined(HAS_I422TOABGRROW_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - I422ToABGRRow = I422ToABGRRow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToABGRRow = I422ToABGRRow_NEON; - } - } -#elif defined(HAS_I422TOABGRROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToABGRRow = I422ToABGRRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToABGRRow = I422ToABGRRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_abgr, 16) && IS_ALIGNED(dst_stride_abgr, 16)) { - I422ToABGRRow = I422ToABGRRow_SSSE3; - } - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToABGRRow(src_y, src_u, src_v, dst_abgr, width); - dst_abgr += dst_stride_abgr; - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - } - return 0; -} - -// Convert I422 to RGBA. -LIBYUV_API -int I422ToRGBA(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_rgba, int dst_stride_rgba, - int width, int height) { - int y; - void (*I422ToRGBARow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToRGBARow_C; - if (!src_y || !src_u || !src_v || - !dst_rgba || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_rgba = dst_rgba + (height - 1) * dst_stride_rgba; - dst_stride_rgba = -dst_stride_rgba; - } - // Coalesce rows. - if (src_stride_y == width && - src_stride_u * 2 == width && - src_stride_v * 2 == width && - dst_stride_rgba == width * 4) { - width *= height; - height = 1; - src_stride_y = src_stride_u = src_stride_v = dst_stride_rgba = 0; - } -#if defined(HAS_I422TORGBAROW_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - I422ToRGBARow = I422ToRGBARow_Any_NEON; - if (IS_ALIGNED(width, 16)) { - I422ToRGBARow = I422ToRGBARow_NEON; - } - } -#elif defined(HAS_I422TORGBAROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - I422ToRGBARow = I422ToRGBARow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - I422ToRGBARow = I422ToRGBARow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_rgba, 16) && IS_ALIGNED(dst_stride_rgba, 16)) { - I422ToRGBARow = I422ToRGBARow_SSSE3; - } - } - } -#endif - - for (y = 0; y < height; ++y) { - I422ToRGBARow(src_y, src_u, src_v, dst_rgba, width); - dst_rgba += dst_stride_rgba; - src_y += src_stride_y; - src_u += src_stride_u; - src_v += src_stride_v; - } - return 0; -} - -// Convert NV12 to RGB565. -LIBYUV_API -int NV12ToRGB565(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height) { - int y; - void (*NV12ToRGB565Row)(const uint8* y_buf, - const uint8* uv_buf, - uint8* rgb_buf, - int width) = NV12ToRGB565Row_C; - if (!src_y || !src_uv || !dst_rgb565 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_rgb565 = dst_rgb565 + (height - 1) * dst_stride_rgb565; - dst_stride_rgb565 = -dst_stride_rgb565; - } -#if defined(HAS_NV12TORGB565ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - NV12ToRGB565Row = NV12ToRGB565Row_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - NV12ToRGB565Row = NV12ToRGB565Row_SSSE3; - } - } -#elif defined(HAS_NV12TORGB565ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - NV12ToRGB565Row = NV12ToRGB565Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - NV12ToRGB565Row = NV12ToRGB565Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - NV12ToRGB565Row(src_y, src_uv, dst_rgb565, width); - dst_rgb565 += dst_stride_rgb565; - src_y += src_stride_y; - if (y & 1) { - src_uv += src_stride_uv; - } - } - return 0; -} - -// Convert NV21 to RGB565. -LIBYUV_API -int NV21ToRGB565(const uint8* src_y, int src_stride_y, - const uint8* src_vu, int src_stride_vu, - uint8* dst_rgb565, int dst_stride_rgb565, - int width, int height) { - int y; - void (*NV21ToRGB565Row)(const uint8* y_buf, - const uint8* src_vu, - uint8* rgb_buf, - int width) = NV21ToRGB565Row_C; - if (!src_y || !src_vu || !dst_rgb565 || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_rgb565 = dst_rgb565 + (height - 1) * dst_stride_rgb565; - dst_stride_rgb565 = -dst_stride_rgb565; - } -#if defined(HAS_NV21TORGB565ROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - NV21ToRGB565Row = NV21ToRGB565Row_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - NV21ToRGB565Row = NV21ToRGB565Row_SSSE3; - } - } -#elif defined(HAS_NV21TORGB565ROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - NV21ToRGB565Row = NV21ToRGB565Row_Any_NEON; - if (IS_ALIGNED(width, 8)) { - NV21ToRGB565Row = NV21ToRGB565Row_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - NV21ToRGB565Row(src_y, src_vu, dst_rgb565, width); - dst_rgb565 += dst_stride_rgb565; - src_y += src_stride_y; - if (y & 1) { - src_vu += src_stride_vu; - } - } - return 0; -} - -LIBYUV_API -void SetPlane(uint8* dst_y, int dst_stride_y, - int width, int height, - uint32 value) { - int y; - uint32 v32 = value | (value << 8) | (value << 16) | (value << 24); - void (*SetRow)(uint8* dst, uint32 value, int pix) = SetRow_C; - // Coalesce rows. - if (dst_stride_y == width) { - width *= height; - height = 1; - dst_stride_y = 0; - } -#if defined(HAS_SETROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && - IS_ALIGNED(width, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - SetRow = SetRow_NEON; - } -#endif -#if defined(HAS_SETROW_X86) - if (TestCpuFlag(kCpuHasX86) && IS_ALIGNED(width, 4)) { - SetRow = SetRow_X86; - } -#endif - - // Set plane - for (y = 0; y < height; ++y) { - SetRow(dst_y, v32, width); - dst_y += dst_stride_y; - } -} - -// Draw a rectangle into I420 -LIBYUV_API -int I420Rect(uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int x, int y, - int width, int height, - int value_y, int value_u, int value_v) { - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - uint8* start_y = dst_y + y * dst_stride_y + x; - uint8* start_u = dst_u + (y / 2) * dst_stride_u + (x / 2); - uint8* start_v = dst_v + (y / 2) * dst_stride_v + (x / 2); - if (!dst_y || !dst_u || !dst_v || - width <= 0 || height <= 0 || - x < 0 || y < 0 || - value_y < 0 || value_y > 255 || - value_u < 0 || value_u > 255 || - value_v < 0 || value_v > 255) { - return -1; - } - - SetPlane(start_y, dst_stride_y, width, height, value_y); - SetPlane(start_u, dst_stride_u, halfwidth, halfheight, value_u); - SetPlane(start_v, dst_stride_v, halfwidth, halfheight, value_v); - return 0; -} - -// Draw a rectangle into ARGB -LIBYUV_API -int ARGBRect(uint8* dst_argb, int dst_stride_argb, - int dst_x, int dst_y, - int width, int height, - uint32 value) { - if (!dst_argb || - width <= 0 || height <= 0 || - dst_x < 0 || dst_y < 0) { - return -1; - } - dst_argb += dst_y * dst_stride_argb + dst_x * 4; - // Coalesce rows. - if (dst_stride_argb == width * 4) { - width *= height; - height = 1; - dst_stride_argb = 0; - } -#if defined(HAS_SETROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBSetRows_NEON(dst_argb, value, width, dst_stride_argb, height); - return 0; - } -#endif -#if defined(HAS_SETROW_X86) - if (TestCpuFlag(kCpuHasX86)) { - ARGBSetRows_X86(dst_argb, value, width, dst_stride_argb, height); - return 0; - } -#endif - ARGBSetRows_C(dst_argb, value, width, dst_stride_argb, height); - return 0; -} - -// Convert unattentuated ARGB to preattenuated ARGB. -// An unattenutated ARGB alpha blend uses the formula -// p = a * f + (1 - a) * b -// where -// p is output pixel -// f is foreground pixel -// b is background pixel -// a is alpha value from foreground pixel -// An preattenutated ARGB alpha blend uses the formula -// p = f + (1 - a) * b -// where -// f is foreground pixel premultiplied by alpha - -LIBYUV_API -int ARGBAttenuate(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBAttenuateRow)(const uint8* src_argb, uint8* dst_argb, - int width) = ARGBAttenuateRow_C; - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBATTENUATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBAttenuateRow = ARGBAttenuateRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBAttenuateRow = ARGBAttenuateRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBATTENUATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 4) { - ARGBAttenuateRow = ARGBAttenuateRow_Any_SSSE3; - if (IS_ALIGNED(width, 4)) { - ARGBAttenuateRow = ARGBAttenuateRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBATTENUATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 8) { - ARGBAttenuateRow = ARGBAttenuateRow_Any_AVX2; - if (IS_ALIGNED(width, 8)) { - ARGBAttenuateRow = ARGBAttenuateRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBATTENUATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBAttenuateRow = ARGBAttenuateRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBAttenuateRow = ARGBAttenuateRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBAttenuateRow(src_argb, dst_argb, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert preattentuated ARGB to unattenuated ARGB. -LIBYUV_API -int ARGBUnattenuate(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBUnattenuateRow)(const uint8* src_argb, uint8* dst_argb, - int width) = ARGBUnattenuateRow_C; - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBUNATTENUATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4) { - ARGBUnattenuateRow = ARGBUnattenuateRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBUnattenuateRow = ARGBUnattenuateRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBUNATTENUATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 8) { - ARGBUnattenuateRow = ARGBUnattenuateRow_Any_AVX2; - if (IS_ALIGNED(width, 8)) { - ARGBUnattenuateRow = ARGBUnattenuateRow_AVX2; - } - } -#endif -// TODO(fbarchard): Neon version. - - for (y = 0; y < height; ++y) { - ARGBUnattenuateRow(src_argb, dst_argb, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Convert ARGB to Grayed ARGB. -LIBYUV_API -int ARGBGrayTo(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBGrayRow)(const uint8* src_argb, uint8* dst_argb, - int width) = ARGBGrayRow_C; - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBGRAYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 8) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBGrayRow = ARGBGrayRow_SSSE3; - } -#elif defined(HAS_ARGBGRAYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - ARGBGrayRow = ARGBGrayRow_NEON; - } -#endif - - for (y = 0; y < height; ++y) { - ARGBGrayRow(src_argb, dst_argb, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Make a rectangle of ARGB gray scale. -LIBYUV_API -int ARGBGray(uint8* dst_argb, int dst_stride_argb, - int dst_x, int dst_y, - int width, int height) { - int y; - void (*ARGBGrayRow)(const uint8* src_argb, uint8* dst_argb, - int width) = ARGBGrayRow_C; - uint8* dst = dst_argb + dst_y * dst_stride_argb + dst_x * 4; - if (!dst_argb || width <= 0 || height <= 0 || dst_x < 0 || dst_y < 0) { - return -1; - } - // Coalesce rows. - if (dst_stride_argb == width * 4) { - width *= height; - height = 1; - dst_stride_argb = 0; - } -#if defined(HAS_ARGBGRAYROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 8) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBGrayRow = ARGBGrayRow_SSSE3; - } -#elif defined(HAS_ARGBGRAYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - ARGBGrayRow = ARGBGrayRow_NEON; - } -#endif - for (y = 0; y < height; ++y) { - ARGBGrayRow(dst, dst, width); - dst += dst_stride_argb; - } - return 0; -} - -// Make a rectangle of ARGB Sepia tone. -LIBYUV_API -int ARGBSepia(uint8* dst_argb, int dst_stride_argb, - int dst_x, int dst_y, int width, int height) { - int y; - void (*ARGBSepiaRow)(uint8* dst_argb, int width) = ARGBSepiaRow_C; - uint8* dst = dst_argb + dst_y * dst_stride_argb + dst_x * 4; - if (!dst_argb || width <= 0 || height <= 0 || dst_x < 0 || dst_y < 0) { - return -1; - } - // Coalesce rows. - if (dst_stride_argb == width * 4) { - width *= height; - height = 1; - dst_stride_argb = 0; - } -#if defined(HAS_ARGBSEPIAROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 8) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBSepiaRow = ARGBSepiaRow_SSSE3; - } -#elif defined(HAS_ARGBSEPIAROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - ARGBSepiaRow = ARGBSepiaRow_NEON; - } -#endif - for (y = 0; y < height; ++y) { - ARGBSepiaRow(dst, width); - dst += dst_stride_argb; - } - return 0; -} - -// Apply a 4x4 matrix to each ARGB pixel. -// Note: Normally for shading, but can be used to swizzle or invert. -LIBYUV_API -int ARGBColorMatrix(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - const int8* matrix_argb, - int width, int height) { - int y; - void (*ARGBColorMatrixRow)(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width) = ARGBColorMatrixRow_C; - if (!src_argb || !dst_argb || !matrix_argb || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBCOLORMATRIXROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 8) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBColorMatrixRow = ARGBColorMatrixRow_SSSE3; - } -#elif defined(HAS_ARGBCOLORMATRIXROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - ARGBColorMatrixRow = ARGBColorMatrixRow_NEON; - } -#endif - for (y = 0; y < height; ++y) { - ARGBColorMatrixRow(src_argb, dst_argb, matrix_argb, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Apply a 4x3 matrix to each ARGB pixel. -// Deprecated. -LIBYUV_API -int RGBColorMatrix(uint8* dst_argb, int dst_stride_argb, - const int8* matrix_rgb, - int dst_x, int dst_y, int width, int height) { - SIMD_ALIGNED(int8 matrix_argb[16]); - uint8* dst = dst_argb + dst_y * dst_stride_argb + dst_x * 4; - if (!dst_argb || !matrix_rgb || width <= 0 || height <= 0 || - dst_x < 0 || dst_y < 0) { - return -1; - } - - // Convert 4x3 7 bit matrix to 4x4 6 bit matrix. - matrix_argb[0] = matrix_rgb[0] / 2; - matrix_argb[1] = matrix_rgb[1] / 2; - matrix_argb[2] = matrix_rgb[2] / 2; - matrix_argb[3] = matrix_rgb[3] / 2; - matrix_argb[4] = matrix_rgb[4] / 2; - matrix_argb[5] = matrix_rgb[5] / 2; - matrix_argb[6] = matrix_rgb[6] / 2; - matrix_argb[7] = matrix_rgb[7] / 2; - matrix_argb[8] = matrix_rgb[8] / 2; - matrix_argb[9] = matrix_rgb[9] / 2; - matrix_argb[10] = matrix_rgb[10] / 2; - matrix_argb[11] = matrix_rgb[11] / 2; - matrix_argb[14] = matrix_argb[13] = matrix_argb[12] = 0; - matrix_argb[15] = 64; // 1.0 - - return ARGBColorMatrix((const uint8*)(dst), dst_stride_argb, - dst, dst_stride_argb, - &matrix_argb[0], width, height); -} - -// Apply a color table each ARGB pixel. -// Table contains 256 ARGB values. -LIBYUV_API -int ARGBColorTable(uint8* dst_argb, int dst_stride_argb, - const uint8* table_argb, - int dst_x, int dst_y, int width, int height) { - int y; - void (*ARGBColorTableRow)(uint8* dst_argb, const uint8* table_argb, - int width) = ARGBColorTableRow_C; - uint8* dst = dst_argb + dst_y * dst_stride_argb + dst_x * 4; - if (!dst_argb || !table_argb || width <= 0 || height <= 0 || - dst_x < 0 || dst_y < 0) { - return -1; - } - // Coalesce rows. - if (dst_stride_argb == width * 4) { - width *= height; - height = 1; - dst_stride_argb = 0; - } -#if defined(HAS_ARGBCOLORTABLEROW_X86) - if (TestCpuFlag(kCpuHasX86)) { - ARGBColorTableRow = ARGBColorTableRow_X86; - } -#endif - for (y = 0; y < height; ++y) { - ARGBColorTableRow(dst, table_argb, width); - dst += dst_stride_argb; - } - return 0; -} - -// Apply a color table each ARGB pixel but preserve destination alpha. -// Table contains 256 ARGB values. -LIBYUV_API -int RGBColorTable(uint8* dst_argb, int dst_stride_argb, - const uint8* table_argb, - int dst_x, int dst_y, int width, int height) { - int y; - void (*RGBColorTableRow)(uint8* dst_argb, const uint8* table_argb, - int width) = RGBColorTableRow_C; - uint8* dst = dst_argb + dst_y * dst_stride_argb + dst_x * 4; - if (!dst_argb || !table_argb || width <= 0 || height <= 0 || - dst_x < 0 || dst_y < 0) { - return -1; - } - // Coalesce rows. - if (dst_stride_argb == width * 4) { - width *= height; - height = 1; - dst_stride_argb = 0; - } -#if defined(HAS_RGBCOLORTABLEROW_X86) - if (TestCpuFlag(kCpuHasX86)) { - RGBColorTableRow = RGBColorTableRow_X86; - } -#endif - for (y = 0; y < height; ++y) { - RGBColorTableRow(dst, table_argb, width); - dst += dst_stride_argb; - } - return 0; -} - -// ARGBQuantize is used to posterize art. -// e.g. rgb / qvalue * qvalue + qvalue / 2 -// But the low levels implement efficiently with 3 parameters, and could be -// used for other high level operations. -// dst_argb[0] = (b * scale >> 16) * interval_size + interval_offset; -// where scale is 1 / interval_size as a fixed point value. -// The divide is replaces with a multiply by reciprocal fixed point multiply. -// Caveat - although SSE2 saturates, the C function does not and should be used -// with care if doing anything but quantization. -LIBYUV_API -int ARGBQuantize(uint8* dst_argb, int dst_stride_argb, - int scale, int interval_size, int interval_offset, - int dst_x, int dst_y, int width, int height) { - int y; - void (*ARGBQuantizeRow)(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width) = ARGBQuantizeRow_C; - uint8* dst = dst_argb + dst_y * dst_stride_argb + dst_x * 4; - if (!dst_argb || width <= 0 || height <= 0 || dst_x < 0 || dst_y < 0 || - interval_size < 1 || interval_size > 255) { - return -1; - } - // Coalesce rows. - if (dst_stride_argb == width * 4) { - width *= height; - height = 1; - dst_stride_argb = 0; - } -#if defined(HAS_ARGBQUANTIZEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBQuantizeRow = ARGBQuantizeRow_SSE2; - } -#elif defined(HAS_ARGBQUANTIZEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - ARGBQuantizeRow = ARGBQuantizeRow_NEON; - } -#endif - for (y = 0; y < height; ++y) { - ARGBQuantizeRow(dst, scale, interval_size, interval_offset, width); - dst += dst_stride_argb; - } - return 0; -} - -// Computes table of cumulative sum for image where the value is the sum -// of all values above and to the left of the entry. Used by ARGBBlur. -LIBYUV_API -int ARGBComputeCumulativeSum(const uint8* src_argb, int src_stride_argb, - int32* dst_cumsum, int dst_stride32_cumsum, - int width, int height) { - int y; - void (*ComputeCumulativeSumRow)(const uint8* row, int32* cumsum, - const int32* previous_cumsum, int width) = ComputeCumulativeSumRow_C; - int32* previous_cumsum = dst_cumsum; - if (!dst_cumsum || !src_argb || width <= 0 || height <= 0) { - return -1; - } -#if defined(HAS_CUMULATIVESUMTOAVERAGEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2)) { - ComputeCumulativeSumRow = ComputeCumulativeSumRow_SSE2; - } -#endif - memset(dst_cumsum, 0, width * sizeof(dst_cumsum[0]) * 4); // 4 int per pixel. - for (y = 0; y < height; ++y) { - ComputeCumulativeSumRow(src_argb, dst_cumsum, previous_cumsum, width); - previous_cumsum = dst_cumsum; - dst_cumsum += dst_stride32_cumsum; - src_argb += src_stride_argb; - } - return 0; -} - -// Blur ARGB image. -// Caller should allocate CumulativeSum table of width * height * 16 bytes -// aligned to 16 byte boundary. height can be radius * 2 + 2 to save memory -// as the buffer is treated as circular. -LIBYUV_API -int ARGBBlur(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int32* dst_cumsum, int dst_stride32_cumsum, - int width, int height, int radius) { - int y; - void (*ComputeCumulativeSumRow)(const uint8 *row, int32 *cumsum, - const int32* previous_cumsum, int width) = ComputeCumulativeSumRow_C; - void (*CumulativeSumToAverageRow)(const int32* topleft, const int32* botleft, - int width, int area, uint8* dst, int count) = CumulativeSumToAverageRow_C; - int32* cumsum_bot_row; - int32* max_cumsum_bot_row; - int32* cumsum_top_row; - - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - if (radius > height) { - radius = height; - } - if (radius > (width / 2 - 1)) { - radius = width / 2 - 1; - } - if (radius <= 0) { - return -1; - } -#if defined(HAS_CUMULATIVESUMTOAVERAGEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2)) { - ComputeCumulativeSumRow = ComputeCumulativeSumRow_SSE2; - CumulativeSumToAverageRow = CumulativeSumToAverageRow_SSE2; - } -#endif - // Compute enough CumulativeSum for first row to be blurred. After this - // one row of CumulativeSum is updated at a time. - ARGBComputeCumulativeSum(src_argb, src_stride_argb, - dst_cumsum, dst_stride32_cumsum, - width, radius); - - src_argb = src_argb + radius * src_stride_argb; - cumsum_bot_row = &dst_cumsum[(radius - 1) * dst_stride32_cumsum]; - - max_cumsum_bot_row = &dst_cumsum[(radius * 2 + 2) * dst_stride32_cumsum]; - cumsum_top_row = &dst_cumsum[0]; - - for (y = 0; y < height; ++y) { - int top_y = ((y - radius - 1) >= 0) ? (y - radius - 1) : 0; - int bot_y = ((y + radius) < height) ? (y + radius) : (height - 1); - int area = radius * (bot_y - top_y); - int boxwidth = radius * 4; - int x; - int n; - - // Increment cumsum_top_row pointer with circular buffer wrap around. - if (top_y) { - cumsum_top_row += dst_stride32_cumsum; - if (cumsum_top_row >= max_cumsum_bot_row) { - cumsum_top_row = dst_cumsum; - } - } - // Increment cumsum_bot_row pointer with circular buffer wrap around and - // then fill in a row of CumulativeSum. - if ((y + radius) < height) { - const int32* prev_cumsum_bot_row = cumsum_bot_row; - cumsum_bot_row += dst_stride32_cumsum; - if (cumsum_bot_row >= max_cumsum_bot_row) { - cumsum_bot_row = dst_cumsum; - } - ComputeCumulativeSumRow(src_argb, cumsum_bot_row, prev_cumsum_bot_row, - width); - src_argb += src_stride_argb; - } - - // Left clipped. - for (x = 0; x < radius + 1; ++x) { - CumulativeSumToAverageRow(cumsum_top_row, cumsum_bot_row, - boxwidth, area, &dst_argb[x * 4], 1); - area += (bot_y - top_y); - boxwidth += 4; - } - - // Middle unclipped. - n = (width - 1) - radius - x + 1; - CumulativeSumToAverageRow(cumsum_top_row, cumsum_bot_row, - boxwidth, area, &dst_argb[x * 4], n); - - // Right clipped. - for (x += n; x <= width - 1; ++x) { - area -= (bot_y - top_y); - boxwidth -= 4; - CumulativeSumToAverageRow(cumsum_top_row + (x - radius - 1) * 4, - cumsum_bot_row + (x - radius - 1) * 4, - boxwidth, area, &dst_argb[x * 4], 1); - } - dst_argb += dst_stride_argb; - } - return 0; -} - -// Multiply ARGB image by a specified ARGB value. -LIBYUV_API -int ARGBShade(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height, uint32 value) { - int y; - void (*ARGBShadeRow)(const uint8* src_argb, uint8* dst_argb, - int width, uint32 value) = ARGBShadeRow_C; - if (!src_argb || !dst_argb || width <= 0 || height == 0 || value == 0u) { - return -1; - } - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBSHADEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBShadeRow = ARGBShadeRow_SSE2; - } -#elif defined(HAS_ARGBSHADEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - ARGBShadeRow = ARGBShadeRow_NEON; - } -#endif - - for (y = 0; y < height; ++y) { - ARGBShadeRow(src_argb, dst_argb, width, value); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Interpolate 2 ARGB images by specified amount (0 to 255). -LIBYUV_API -int ARGBInterpolate(const uint8* src_argb0, int src_stride_argb0, - const uint8* src_argb1, int src_stride_argb1, - uint8* dst_argb, int dst_stride_argb, - int width, int height, int interpolation) { - int y; - void (*InterpolateRow)(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) = InterpolateRow_C; - if (!src_argb0 || !src_argb1 || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - dst_argb = dst_argb + (height - 1) * dst_stride_argb; - dst_stride_argb = -dst_stride_argb; - } - // Coalesce rows. - if (src_stride_argb0 == width * 4 && - src_stride_argb1 == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb0 = src_stride_argb1 = dst_stride_argb = 0; - } -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(src_argb0, 16) && IS_ALIGNED(src_stride_argb0, 16) && - IS_ALIGNED(src_argb1, 16) && IS_ALIGNED(src_stride_argb1, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 4) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(width, 4)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb0, 16) && IS_ALIGNED(src_stride_argb0, 16) && - IS_ALIGNED(src_argb1, 16) && IS_ALIGNED(src_stride_argb1, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 8) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(width, 8)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 4) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(width, 4)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROWS_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && width >= 1 && - IS_ALIGNED(src_argb0, 4) && IS_ALIGNED(src_stride_argb0, 4) && - IS_ALIGNED(src_argb1, 4) && IS_ALIGNED(src_stride_argb1, 4) && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride_argb, 4)) { - ScaleARGBFilterRows = InterpolateRow_MIPS_DSPR2; - } -#endif - - for (y = 0; y < height; ++y) { - InterpolateRow(dst_argb, src_argb0, src_argb1 - src_argb0, - width * 4, interpolation); - src_argb0 += src_stride_argb0; - src_argb1 += src_stride_argb1; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Shuffle ARGB channel order. e.g. BGRA to ARGB. -LIBYUV_API -int ARGBShuffle(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_argb, int dst_stride_argb, - const uint8* shuffler, int width, int height) { - int y; - void (*ARGBShuffleRow)(const uint8* src_bgra, uint8* dst_argb, - const uint8* shuffler, int pix) = ARGBShuffleRow_C; - if (!src_bgra || !dst_argb || - width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_bgra = src_bgra + (height - 1) * src_stride_bgra; - src_stride_bgra = -src_stride_bgra; - } - // Coalesce rows. - if (src_stride_bgra == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_bgra = dst_stride_argb = 0; - } -#if defined(HAS_ARGBSHUFFLEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 4) { - ARGBShuffleRow = ARGBShuffleRow_Any_SSE2; - if (IS_ALIGNED(width, 4)) { - ARGBShuffleRow = ARGBShuffleRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBSHUFFLEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8) { - ARGBShuffleRow = ARGBShuffleRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - ARGBShuffleRow = ARGBShuffleRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_bgra, 16) && IS_ALIGNED(src_stride_bgra, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - ARGBShuffleRow = ARGBShuffleRow_SSSE3; - } - } - } -#endif -#if defined(HAS_ARGBSHUFFLEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && width >= 16) { - ARGBShuffleRow = ARGBShuffleRow_Any_AVX2; - if (IS_ALIGNED(width, 16)) { - ARGBShuffleRow = ARGBShuffleRow_AVX2; - } - } -#endif -#if defined(HAS_ARGBSHUFFLEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 4) { - ARGBShuffleRow = ARGBShuffleRow_Any_NEON; - if (IS_ALIGNED(width, 4)) { - ARGBShuffleRow = ARGBShuffleRow_NEON; - } - } -#endif - - for (y = 0; y < height; ++y) { - ARGBShuffleRow(src_bgra, dst_argb, shuffler, width); - src_bgra += src_stride_bgra; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Sobel ARGB effect. -static int ARGBSobelize(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height, - void (*SobelRow)(const uint8* src_sobelx, - const uint8* src_sobely, - uint8* dst, int width)) { - int y; - void (*ARGBToBayerRow)(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) = ARGBToBayerGGRow_C; - void (*SobelYRow)(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width) = SobelYRow_C; - void (*SobelXRow)(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobely, int width) = - SobelXRow_C; - const int kEdge = 16; // Extra pixels at start of row for extrude/align. - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // ARGBToBayer used to select G channel from ARGB. -#if defined(HAS_ARGBTOBAYERGGROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && width >= 8 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToBayerRow = ARGBToBayerGGRow_Any_SSE2; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerGGRow_SSE2; - } - } -#endif -#if defined(HAS_ARGBTOBAYERROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && width >= 8 && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16)) { - ARGBToBayerRow = ARGBToBayerRow_Any_SSSE3; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerRow_SSSE3; - } - } -#endif -#if defined(HAS_ARGBTOBAYERGGROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && width >= 8) { - ARGBToBayerRow = ARGBToBayerGGRow_Any_NEON; - if (IS_ALIGNED(width, 8)) { - ARGBToBayerRow = ARGBToBayerGGRow_NEON; - } - } -#endif -#if defined(HAS_SOBELYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2)) { - SobelYRow = SobelYRow_SSE2; - } -#endif -#if defined(HAS_SOBELYROW_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - SobelYRow = SobelYRow_NEON; - } -#endif -#if defined(HAS_SOBELXROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2)) { - SobelXRow = SobelXRow_SSE2; - } -#endif -#if defined(HAS_SOBELXROW_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - SobelXRow = SobelXRow_NEON; - } -#endif - { - // 3 rows with edges before/after. - const int kRowSize = (width + kEdge + 15) & ~15; - align_buffer_64(rows, kRowSize * 2 + (kEdge + kRowSize * 3 + kEdge)); - uint8* row_sobelx = rows; - uint8* row_sobely = rows + kRowSize; - uint8* row_y = rows + kRowSize * 2; - - // Convert first row. - uint8* row_y0 = row_y + kEdge; - uint8* row_y1 = row_y0 + kRowSize; - uint8* row_y2 = row_y1 + kRowSize; - ARGBToBayerRow(src_argb, row_y0, 0x0d090501, width); - row_y0[-1] = row_y0[0]; - memset(row_y0 + width, row_y0[width - 1], 16); // Extrude 16 for valgrind. - ARGBToBayerRow(src_argb, row_y1, 0x0d090501, width); - row_y1[-1] = row_y1[0]; - memset(row_y1 + width, row_y1[width - 1], 16); - memset(row_y2 + width, 0, 16); - - for (y = 0; y < height; ++y) { - // Convert next row of ARGB to Y. - if (y < (height - 1)) { - src_argb += src_stride_argb; - } - ARGBToBayerRow(src_argb, row_y2, 0x0d090501, width); - row_y2[-1] = row_y2[0]; - row_y2[width] = row_y2[width - 1]; - - SobelXRow(row_y0 - 1, row_y1 - 1, row_y2 - 1, row_sobelx, width); - SobelYRow(row_y0 - 1, row_y2 - 1, row_sobely, width); - SobelRow(row_sobelx, row_sobely, dst_argb, width); - - // Cycle thru circular queue of 3 row_y buffers. - { - uint8* row_yt = row_y0; - row_y0 = row_y1; - row_y1 = row_y2; - row_y2 = row_yt; - } - - dst_argb += dst_stride_argb; - } - free_aligned_buffer_64(rows); - } - return 0; -} - -// Sobel ARGB effect. -LIBYUV_API -int ARGBSobel(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - void (*SobelRow)(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) = SobelRow_C; -#if defined(HAS_SOBELROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - SobelRow = SobelRow_SSE2; - } -#endif -#if defined(HAS_SOBELROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - SobelRow = SobelRow_NEON; - } -#endif - return ARGBSobelize(src_argb, src_stride_argb, dst_argb, dst_stride_argb, - width, height, SobelRow); -} - -// Sobel ARGB effect with planar output. -LIBYUV_API -int ARGBSobelToPlane(const uint8* src_argb, int src_stride_argb, - uint8* dst_y, int dst_stride_y, - int width, int height) { - void (*SobelToPlaneRow)(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_, int width) = SobelToPlaneRow_C; -#if defined(HAS_SOBELTOPLANEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 16) && - IS_ALIGNED(dst_y, 16) && IS_ALIGNED(dst_stride_y, 16)) { - SobelToPlaneRow = SobelToPlaneRow_SSE2; - } -#endif -#if defined(HAS_SOBELTOPLANEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 16)) { - SobelToPlaneRow = SobelToPlaneRow_NEON; - } -#endif - return ARGBSobelize(src_argb, src_stride_argb, dst_y, dst_stride_y, - width, height, SobelToPlaneRow); -} - -// SobelXY ARGB effect. -// Similar to Sobel, but also stores Sobel X in R and Sobel Y in B. G = Sobel. -LIBYUV_API -int ARGBSobelXY(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - void (*SobelXYRow)(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) = SobelXYRow_C; -#if defined(HAS_SOBELXYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - SobelXYRow = SobelXYRow_SSE2; - } -#endif -#if defined(HAS_SOBELXYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - SobelXYRow = SobelXYRow_NEON; - } -#endif - return ARGBSobelize(src_argb, src_stride_argb, dst_argb, dst_stride_argb, - width, height, SobelXYRow); -} - -// Apply a 4x4 polynomial to each ARGB pixel. -LIBYUV_API -int ARGBPolynomial(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - const float* poly, - int width, int height) { - int y; - void (*ARGBPolynomialRow)(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width) = ARGBPolynomialRow_C; - if (!src_argb || !dst_argb || !poly || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBPOLYNOMIALROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 2)) { - ARGBPolynomialRow = ARGBPolynomialRow_SSE2; - } -#endif -#if defined(HAS_ARGBPOLYNOMIALROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && TestCpuFlag(kCpuHasFMA3) && - IS_ALIGNED(width, 2)) { - ARGBPolynomialRow = ARGBPolynomialRow_AVX2; - } -#endif - - for (y = 0; y < height; ++y) { - ARGBPolynomialRow(src_argb, dst_argb, poly, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Apply a lumacolortable to each ARGB pixel. -LIBYUV_API -int ARGBLumaColorTable(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - const uint8* luma, - int width, int height) { - int y; - void (*ARGBLumaColorTableRow)(const uint8* src_argb, uint8* dst_argb, - int width, const uint8* luma, const uint32 lumacoeff) = - ARGBLumaColorTableRow_C; - if (!src_argb || !dst_argb || !luma || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBLUMACOLORTABLEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 4)) { - ARGBLumaColorTableRow = ARGBLumaColorTableRow_SSSE3; - } -#endif - - for (y = 0; y < height; ++y) { - ARGBLumaColorTableRow(src_argb, dst_argb, width, luma, 0x00264b0f); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Copy Alpha from one ARGB image to another. -LIBYUV_API -int ARGBCopyAlpha(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBCopyAlphaRow)(const uint8* src_argb, uint8* dst_argb, int width) = - ARGBCopyAlphaRow_C; - if (!src_argb || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - // Coalesce rows. - if (src_stride_argb == width * 4 && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_argb = dst_stride_argb = 0; - } -#if defined(HAS_ARGBCOPYALPHAROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride_argb, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16) && - IS_ALIGNED(width, 8)) { - ARGBCopyAlphaRow = ARGBCopyAlphaRow_SSE2; - } -#endif -#if defined(HAS_ARGBCOPYALPHAROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && IS_ALIGNED(width, 16)) { - ARGBCopyAlphaRow = ARGBCopyAlphaRow_AVX2; - } -#endif - - for (y = 0; y < height; ++y) { - ARGBCopyAlphaRow(src_argb, dst_argb, width); - src_argb += src_stride_argb; - dst_argb += dst_stride_argb; - } - return 0; -} - -// Copy a planar Y channel to the alpha channel of a destination ARGB image. -LIBYUV_API -int ARGBCopyYToAlpha(const uint8* src_y, int src_stride_y, - uint8* dst_argb, int dst_stride_argb, - int width, int height) { - int y; - void (*ARGBCopyYToAlphaRow)(const uint8* src_y, uint8* dst_argb, int width) = - ARGBCopyYToAlphaRow_C; - if (!src_y || !dst_argb || width <= 0 || height == 0) { - return -1; - } - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_y = src_y + (height - 1) * src_stride_y; - src_stride_y = -src_stride_y; - } - // Coalesce rows. - if (src_stride_y == width && - dst_stride_argb == width * 4) { - width *= height; - height = 1; - src_stride_y = dst_stride_argb = 0; - } -#if defined(HAS_ARGBCOPYYTOALPHAROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && - IS_ALIGNED(src_y, 16) && IS_ALIGNED(src_stride_y, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16) && - IS_ALIGNED(width, 8)) { - ARGBCopyYToAlphaRow = ARGBCopyYToAlphaRow_SSE2; - } -#endif -#if defined(HAS_ARGBCOPYYTOALPHAROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && IS_ALIGNED(width, 16)) { - ARGBCopyYToAlphaRow = ARGBCopyYToAlphaRow_AVX2; - } -#endif - - for (y = 0; y < height; ++y) { - ARGBCopyYToAlphaRow(src_y, dst_argb, width); - src_y += src_stride_y; - dst_argb += dst_stride_argb; - } - return 0; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/rotate.cc b/src/main/jni/libyuv/source/rotate.cc deleted file mode 100644 index fe0e72b13..000000000 --- a/src/main/jni/libyuv/source/rotate.cc +++ /dev/null @@ -1,1315 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/rotate.h" - -#include "libyuv/cpu_id.h" -#include "libyuv/convert.h" -#include "libyuv/planar_functions.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) -#if defined(__APPLE__) && defined(__i386__) -#define DECLARE_FUNCTION(name) \ - ".text \n" \ - ".private_extern _" #name " \n" \ - ".align 4,0x90 \n" \ -"_" #name ": \n" -#elif defined(__MINGW32__) || defined(__CYGWIN__) && defined(__i386__) -#define DECLARE_FUNCTION(name) \ - ".text \n" \ - ".align 4,0x90 \n" \ -"_" #name ": \n" -#else -#define DECLARE_FUNCTION(name) \ - ".text \n" \ - ".align 4,0x90 \n" \ -#name ": \n" -#endif -#endif - -#if !defined(LIBYUV_DISABLE_NEON) && !defined(__native_client__) && \ - (defined(__ARM_NEON__) || defined(LIBYUV_NEON)) -#define HAS_MIRRORROW_NEON -void MirrorRow_NEON(const uint8* src, uint8* dst, int width); -#define HAS_MIRRORROW_UV_NEON -void MirrorUVRow_NEON(const uint8* src, uint8* dst_a, uint8* dst_b, int width); -#define HAS_TRANSPOSE_WX8_NEON -void TransposeWx8_NEON(const uint8* src, int src_stride, - uint8* dst, int dst_stride, int width); -#define HAS_TRANSPOSE_UVWX8_NEON -void TransposeUVWx8_NEON(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width); -#elif !defined(LIBYUV_DISABLE_NEON) && !defined(__native_client__) && \ - (defined(__aarch64__) || defined(LIBYUV_NEON)) -// #define HAS_MIRRORROW_NEON -// void MirrorRow_NEON(const uint8* src, uint8* dst, int width); -// #define HAS_MIRRORROW_UV_NEON -// void MirrorUVRow_NEON(const uint8* src, uint8* dst_a, uint8* dst_b, int width); -// #define HAS_TRANSPOSE_WX8_NEON -// void TransposeWx8_NEON(const uint8* src, int src_stride, -// uint8* dst, int dst_stride, int width); -// #define HAS_TRANSPOSE_UVWX8_NEON -// void TransposeUVWx8_NEON(const uint8* src, int src_stride, -// uint8* dst_a, int dst_stride_a, -// uint8* dst_b, int dst_stride_b, -// int width); -#endif // defined(__ARM_NEON__) - -#if !defined(LIBYUV_DISABLE_MIPS) && !defined(__native_client__) && \ - defined(__mips__) && \ - defined(__mips_dsp) && (__mips_dsp_rev >= 2) -#define HAS_TRANSPOSE_WX8_MIPS_DSPR2 -void TransposeWx8_MIPS_DSPR2(const uint8* src, int src_stride, - uint8* dst, int dst_stride, int width); - -void TransposeWx8_FAST_MIPS_DSPR2(const uint8* src, int src_stride, - uint8* dst, int dst_stride, int width); -#define HAS_TRANSPOSE_UVWx8_MIPS_DSPR2 -void TransposeUVWx8_MIPS_DSPR2(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width); -#endif // defined(__mips__) - -#if !defined(LIBYUV_DISABLE_X86) && \ - defined(_M_IX86) && defined(_MSC_VER) -#define HAS_TRANSPOSE_WX8_SSSE3 -__declspec(naked) __declspec(align(16)) -static void TransposeWx8_SSSE3(const uint8* src, int src_stride, - uint8* dst, int dst_stride, int width) { - __asm { - push edi - push esi - push ebp - mov eax, [esp + 12 + 4] // src - mov edi, [esp + 12 + 8] // src_stride - mov edx, [esp + 12 + 12] // dst - mov esi, [esp + 12 + 16] // dst_stride - mov ecx, [esp + 12 + 20] // width - - // Read in the data from the source pointer. - // First round of bit swap. - align 4 - convertloop: - movq xmm0, qword ptr [eax] - lea ebp, [eax + 8] - movq xmm1, qword ptr [eax + edi] - lea eax, [eax + 2 * edi] - punpcklbw xmm0, xmm1 - movq xmm2, qword ptr [eax] - movdqa xmm1, xmm0 - palignr xmm1, xmm1, 8 - movq xmm3, qword ptr [eax + edi] - lea eax, [eax + 2 * edi] - punpcklbw xmm2, xmm3 - movdqa xmm3, xmm2 - movq xmm4, qword ptr [eax] - palignr xmm3, xmm3, 8 - movq xmm5, qword ptr [eax + edi] - punpcklbw xmm4, xmm5 - lea eax, [eax + 2 * edi] - movdqa xmm5, xmm4 - movq xmm6, qword ptr [eax] - palignr xmm5, xmm5, 8 - movq xmm7, qword ptr [eax + edi] - punpcklbw xmm6, xmm7 - mov eax, ebp - movdqa xmm7, xmm6 - palignr xmm7, xmm7, 8 - // Second round of bit swap. - punpcklwd xmm0, xmm2 - punpcklwd xmm1, xmm3 - movdqa xmm2, xmm0 - movdqa xmm3, xmm1 - palignr xmm2, xmm2, 8 - palignr xmm3, xmm3, 8 - punpcklwd xmm4, xmm6 - punpcklwd xmm5, xmm7 - movdqa xmm6, xmm4 - movdqa xmm7, xmm5 - palignr xmm6, xmm6, 8 - palignr xmm7, xmm7, 8 - // Third round of bit swap. - // Write to the destination pointer. - punpckldq xmm0, xmm4 - movq qword ptr [edx], xmm0 - movdqa xmm4, xmm0 - palignr xmm4, xmm4, 8 - movq qword ptr [edx + esi], xmm4 - lea edx, [edx + 2 * esi] - punpckldq xmm2, xmm6 - movdqa xmm6, xmm2 - palignr xmm6, xmm6, 8 - movq qword ptr [edx], xmm2 - punpckldq xmm1, xmm5 - movq qword ptr [edx + esi], xmm6 - lea edx, [edx + 2 * esi] - movdqa xmm5, xmm1 - movq qword ptr [edx], xmm1 - palignr xmm5, xmm5, 8 - punpckldq xmm3, xmm7 - movq qword ptr [edx + esi], xmm5 - lea edx, [edx + 2 * esi] - movq qword ptr [edx], xmm3 - movdqa xmm7, xmm3 - palignr xmm7, xmm7, 8 - sub ecx, 8 - movq qword ptr [edx + esi], xmm7 - lea edx, [edx + 2 * esi] - jg convertloop - - pop ebp - pop esi - pop edi - ret - } -} - -#define HAS_TRANSPOSE_UVWX8_SSE2 -__declspec(naked) __declspec(align(16)) -static void TransposeUVWx8_SSE2(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int w) { - __asm { - push ebx - push esi - push edi - push ebp - mov eax, [esp + 16 + 4] // src - mov edi, [esp + 16 + 8] // src_stride - mov edx, [esp + 16 + 12] // dst_a - mov esi, [esp + 16 + 16] // dst_stride_a - mov ebx, [esp + 16 + 20] // dst_b - mov ebp, [esp + 16 + 24] // dst_stride_b - mov ecx, esp - sub esp, 4 + 16 - and esp, ~15 - mov [esp + 16], ecx - mov ecx, [ecx + 16 + 28] // w - - align 4 - convertloop: - // Read in the data from the source pointer. - // First round of bit swap. - movdqa xmm0, [eax] - movdqa xmm1, [eax + edi] - lea eax, [eax + 2 * edi] - movdqa xmm7, xmm0 // use xmm7 as temp register. - punpcklbw xmm0, xmm1 - punpckhbw xmm7, xmm1 - movdqa xmm1, xmm7 - movdqa xmm2, [eax] - movdqa xmm3, [eax + edi] - lea eax, [eax + 2 * edi] - movdqa xmm7, xmm2 - punpcklbw xmm2, xmm3 - punpckhbw xmm7, xmm3 - movdqa xmm3, xmm7 - movdqa xmm4, [eax] - movdqa xmm5, [eax + edi] - lea eax, [eax + 2 * edi] - movdqa xmm7, xmm4 - punpcklbw xmm4, xmm5 - punpckhbw xmm7, xmm5 - movdqa xmm5, xmm7 - movdqa xmm6, [eax] - movdqa xmm7, [eax + edi] - lea eax, [eax + 2 * edi] - movdqa [esp], xmm5 // backup xmm5 - neg edi - movdqa xmm5, xmm6 // use xmm5 as temp register. - punpcklbw xmm6, xmm7 - punpckhbw xmm5, xmm7 - movdqa xmm7, xmm5 - lea eax, [eax + 8 * edi + 16] - neg edi - // Second round of bit swap. - movdqa xmm5, xmm0 - punpcklwd xmm0, xmm2 - punpckhwd xmm5, xmm2 - movdqa xmm2, xmm5 - movdqa xmm5, xmm1 - punpcklwd xmm1, xmm3 - punpckhwd xmm5, xmm3 - movdqa xmm3, xmm5 - movdqa xmm5, xmm4 - punpcklwd xmm4, xmm6 - punpckhwd xmm5, xmm6 - movdqa xmm6, xmm5 - movdqa xmm5, [esp] // restore xmm5 - movdqa [esp], xmm6 // backup xmm6 - movdqa xmm6, xmm5 // use xmm6 as temp register. - punpcklwd xmm5, xmm7 - punpckhwd xmm6, xmm7 - movdqa xmm7, xmm6 - // Third round of bit swap. - // Write to the destination pointer. - movdqa xmm6, xmm0 - punpckldq xmm0, xmm4 - punpckhdq xmm6, xmm4 - movdqa xmm4, xmm6 - movdqa xmm6, [esp] // restore xmm6 - movlpd qword ptr [edx], xmm0 - movhpd qword ptr [ebx], xmm0 - movlpd qword ptr [edx + esi], xmm4 - lea edx, [edx + 2 * esi] - movhpd qword ptr [ebx + ebp], xmm4 - lea ebx, [ebx + 2 * ebp] - movdqa xmm0, xmm2 // use xmm0 as the temp register. - punpckldq xmm2, xmm6 - movlpd qword ptr [edx], xmm2 - movhpd qword ptr [ebx], xmm2 - punpckhdq xmm0, xmm6 - movlpd qword ptr [edx + esi], xmm0 - lea edx, [edx + 2 * esi] - movhpd qword ptr [ebx + ebp], xmm0 - lea ebx, [ebx + 2 * ebp] - movdqa xmm0, xmm1 // use xmm0 as the temp register. - punpckldq xmm1, xmm5 - movlpd qword ptr [edx], xmm1 - movhpd qword ptr [ebx], xmm1 - punpckhdq xmm0, xmm5 - movlpd qword ptr [edx + esi], xmm0 - lea edx, [edx + 2 * esi] - movhpd qword ptr [ebx + ebp], xmm0 - lea ebx, [ebx + 2 * ebp] - movdqa xmm0, xmm3 // use xmm0 as the temp register. - punpckldq xmm3, xmm7 - movlpd qword ptr [edx], xmm3 - movhpd qword ptr [ebx], xmm3 - punpckhdq xmm0, xmm7 - sub ecx, 8 - movlpd qword ptr [edx + esi], xmm0 - lea edx, [edx + 2 * esi] - movhpd qword ptr [ebx + ebp], xmm0 - lea ebx, [ebx + 2 * ebp] - jg convertloop - - mov esp, [esp + 16] - pop ebp - pop edi - pop esi - pop ebx - ret - } -} -#elif !defined(LIBYUV_DISABLE_X86) && \ - (defined(__i386__) || (defined(__x86_64__) && !defined(__native_client__))) -#define HAS_TRANSPOSE_WX8_SSSE3 -static void TransposeWx8_SSSE3(const uint8* src, int src_stride, - uint8* dst, int dst_stride, int width) { - asm volatile ( - // Read in the data from the source pointer. - // First round of bit swap. - ".p2align 2 \n" - "1: \n" - "movq (%0),%%xmm0 \n" - "movq (%0,%3),%%xmm1 \n" - "lea (%0,%3,2),%0 \n" - "punpcklbw %%xmm1,%%xmm0 \n" - "movq (%0),%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "palignr $0x8,%%xmm1,%%xmm1 \n" - "movq (%0,%3),%%xmm3 \n" - "lea (%0,%3,2),%0 \n" - "punpcklbw %%xmm3,%%xmm2 \n" - "movdqa %%xmm2,%%xmm3 \n" - "movq (%0),%%xmm4 \n" - "palignr $0x8,%%xmm3,%%xmm3 \n" - "movq (%0,%3),%%xmm5 \n" - "lea (%0,%3,2),%0 \n" - "punpcklbw %%xmm5,%%xmm4 \n" - "movdqa %%xmm4,%%xmm5 \n" - "movq (%0),%%xmm6 \n" - "palignr $0x8,%%xmm5,%%xmm5 \n" - "movq (%0,%3),%%xmm7 \n" - "lea (%0,%3,2),%0 \n" - "punpcklbw %%xmm7,%%xmm6 \n" - "neg %3 \n" - "movdqa %%xmm6,%%xmm7 \n" - "lea 0x8(%0,%3,8),%0 \n" - "palignr $0x8,%%xmm7,%%xmm7 \n" - "neg %3 \n" - // Second round of bit swap. - "punpcklwd %%xmm2,%%xmm0 \n" - "punpcklwd %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "movdqa %%xmm1,%%xmm3 \n" - "palignr $0x8,%%xmm2,%%xmm2 \n" - "palignr $0x8,%%xmm3,%%xmm3 \n" - "punpcklwd %%xmm6,%%xmm4 \n" - "punpcklwd %%xmm7,%%xmm5 \n" - "movdqa %%xmm4,%%xmm6 \n" - "movdqa %%xmm5,%%xmm7 \n" - "palignr $0x8,%%xmm6,%%xmm6 \n" - "palignr $0x8,%%xmm7,%%xmm7 \n" - // Third round of bit swap. - // Write to the destination pointer. - "punpckldq %%xmm4,%%xmm0 \n" - "movq %%xmm0,(%1) \n" - "movdqa %%xmm0,%%xmm4 \n" - "palignr $0x8,%%xmm4,%%xmm4 \n" - "movq %%xmm4,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm6,%%xmm2 \n" - "movdqa %%xmm2,%%xmm6 \n" - "movq %%xmm2,(%1) \n" - "palignr $0x8,%%xmm6,%%xmm6 \n" - "punpckldq %%xmm5,%%xmm1 \n" - "movq %%xmm6,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "movdqa %%xmm1,%%xmm5 \n" - "movq %%xmm1,(%1) \n" - "palignr $0x8,%%xmm5,%%xmm5 \n" - "movq %%xmm5,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm7,%%xmm3 \n" - "movq %%xmm3,(%1) \n" - "movdqa %%xmm3,%%xmm7 \n" - "palignr $0x8,%%xmm7,%%xmm7 \n" - "sub $0x8,%2 \n" - "movq %%xmm7,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : "r"((intptr_t)(src_stride)), // %3 - "r"((intptr_t)(dst_stride)) // %4 - : "memory", "cc" - #if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" - #endif - ); -} - -#if !defined(LIBYUV_DISABLE_X86) && defined(__i386__) -#define HAS_TRANSPOSE_UVWX8_SSE2 -void TransposeUVWx8_SSE2(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int w); - asm ( - DECLARE_FUNCTION(TransposeUVWx8_SSE2) - "push %ebx \n" - "push %esi \n" - "push %edi \n" - "push %ebp \n" - "mov 0x14(%esp),%eax \n" - "mov 0x18(%esp),%edi \n" - "mov 0x1c(%esp),%edx \n" - "mov 0x20(%esp),%esi \n" - "mov 0x24(%esp),%ebx \n" - "mov 0x28(%esp),%ebp \n" - "mov %esp,%ecx \n" - "sub $0x14,%esp \n" - "and $0xfffffff0,%esp \n" - "mov %ecx,0x10(%esp) \n" - "mov 0x2c(%ecx),%ecx \n" - -"1: \n" - "movdqa (%eax),%xmm0 \n" - "movdqa (%eax,%edi,1),%xmm1 \n" - "lea (%eax,%edi,2),%eax \n" - "movdqa %xmm0,%xmm7 \n" - "punpcklbw %xmm1,%xmm0 \n" - "punpckhbw %xmm1,%xmm7 \n" - "movdqa %xmm7,%xmm1 \n" - "movdqa (%eax),%xmm2 \n" - "movdqa (%eax,%edi,1),%xmm3 \n" - "lea (%eax,%edi,2),%eax \n" - "movdqa %xmm2,%xmm7 \n" - "punpcklbw %xmm3,%xmm2 \n" - "punpckhbw %xmm3,%xmm7 \n" - "movdqa %xmm7,%xmm3 \n" - "movdqa (%eax),%xmm4 \n" - "movdqa (%eax,%edi,1),%xmm5 \n" - "lea (%eax,%edi,2),%eax \n" - "movdqa %xmm4,%xmm7 \n" - "punpcklbw %xmm5,%xmm4 \n" - "punpckhbw %xmm5,%xmm7 \n" - "movdqa %xmm7,%xmm5 \n" - "movdqa (%eax),%xmm6 \n" - "movdqa (%eax,%edi,1),%xmm7 \n" - "lea (%eax,%edi,2),%eax \n" - "movdqa %xmm5,(%esp) \n" - "neg %edi \n" - "movdqa %xmm6,%xmm5 \n" - "punpcklbw %xmm7,%xmm6 \n" - "punpckhbw %xmm7,%xmm5 \n" - "movdqa %xmm5,%xmm7 \n" - "lea 0x10(%eax,%edi,8),%eax \n" - "neg %edi \n" - "movdqa %xmm0,%xmm5 \n" - "punpcklwd %xmm2,%xmm0 \n" - "punpckhwd %xmm2,%xmm5 \n" - "movdqa %xmm5,%xmm2 \n" - "movdqa %xmm1,%xmm5 \n" - "punpcklwd %xmm3,%xmm1 \n" - "punpckhwd %xmm3,%xmm5 \n" - "movdqa %xmm5,%xmm3 \n" - "movdqa %xmm4,%xmm5 \n" - "punpcklwd %xmm6,%xmm4 \n" - "punpckhwd %xmm6,%xmm5 \n" - "movdqa %xmm5,%xmm6 \n" - "movdqa (%esp),%xmm5 \n" - "movdqa %xmm6,(%esp) \n" - "movdqa %xmm5,%xmm6 \n" - "punpcklwd %xmm7,%xmm5 \n" - "punpckhwd %xmm7,%xmm6 \n" - "movdqa %xmm6,%xmm7 \n" - "movdqa %xmm0,%xmm6 \n" - "punpckldq %xmm4,%xmm0 \n" - "punpckhdq %xmm4,%xmm6 \n" - "movdqa %xmm6,%xmm4 \n" - "movdqa (%esp),%xmm6 \n" - "movlpd %xmm0,(%edx) \n" - "movhpd %xmm0,(%ebx) \n" - "movlpd %xmm4,(%edx,%esi,1) \n" - "lea (%edx,%esi,2),%edx \n" - "movhpd %xmm4,(%ebx,%ebp,1) \n" - "lea (%ebx,%ebp,2),%ebx \n" - "movdqa %xmm2,%xmm0 \n" - "punpckldq %xmm6,%xmm2 \n" - "movlpd %xmm2,(%edx) \n" - "movhpd %xmm2,(%ebx) \n" - "punpckhdq %xmm6,%xmm0 \n" - "movlpd %xmm0,(%edx,%esi,1) \n" - "lea (%edx,%esi,2),%edx \n" - "movhpd %xmm0,(%ebx,%ebp,1) \n" - "lea (%ebx,%ebp,2),%ebx \n" - "movdqa %xmm1,%xmm0 \n" - "punpckldq %xmm5,%xmm1 \n" - "movlpd %xmm1,(%edx) \n" - "movhpd %xmm1,(%ebx) \n" - "punpckhdq %xmm5,%xmm0 \n" - "movlpd %xmm0,(%edx,%esi,1) \n" - "lea (%edx,%esi,2),%edx \n" - "movhpd %xmm0,(%ebx,%ebp,1) \n" - "lea (%ebx,%ebp,2),%ebx \n" - "movdqa %xmm3,%xmm0 \n" - "punpckldq %xmm7,%xmm3 \n" - "movlpd %xmm3,(%edx) \n" - "movhpd %xmm3,(%ebx) \n" - "punpckhdq %xmm7,%xmm0 \n" - "sub $0x8,%ecx \n" - "movlpd %xmm0,(%edx,%esi,1) \n" - "lea (%edx,%esi,2),%edx \n" - "movhpd %xmm0,(%ebx,%ebp,1) \n" - "lea (%ebx,%ebp,2),%ebx \n" - "jg 1b \n" - "mov 0x10(%esp),%esp \n" - "pop %ebp \n" - "pop %edi \n" - "pop %esi \n" - "pop %ebx \n" -#if defined(__native_client__) - "pop %ecx \n" - "and $0xffffffe0,%ecx \n" - "jmp *%ecx \n" -#else - "ret \n" -#endif -); -#elif !defined(LIBYUV_DISABLE_X86) && !defined(__native_client__) && \ - defined(__x86_64__) -// 64 bit version has enough registers to do 16x8 to 8x16 at a time. -#define HAS_TRANSPOSE_WX8_FAST_SSSE3 -static void TransposeWx8_FAST_SSSE3(const uint8* src, int src_stride, - uint8* dst, int dst_stride, int width) { - asm volatile ( - // Read in the data from the source pointer. - // First round of bit swap. - ".p2align 2 \n" -"1: \n" - "movdqa (%0),%%xmm0 \n" - "movdqa (%0,%3),%%xmm1 \n" - "lea (%0,%3,2),%0 \n" - "movdqa %%xmm0,%%xmm8 \n" - "punpcklbw %%xmm1,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm8 \n" - "movdqa (%0),%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm8,%%xmm9 \n" - "palignr $0x8,%%xmm1,%%xmm1 \n" - "palignr $0x8,%%xmm9,%%xmm9 \n" - "movdqa (%0,%3),%%xmm3 \n" - "lea (%0,%3,2),%0 \n" - "movdqa %%xmm2,%%xmm10 \n" - "punpcklbw %%xmm3,%%xmm2 \n" - "punpckhbw %%xmm3,%%xmm10 \n" - "movdqa %%xmm2,%%xmm3 \n" - "movdqa %%xmm10,%%xmm11 \n" - "movdqa (%0),%%xmm4 \n" - "palignr $0x8,%%xmm3,%%xmm3 \n" - "palignr $0x8,%%xmm11,%%xmm11 \n" - "movdqa (%0,%3),%%xmm5 \n" - "lea (%0,%3,2),%0 \n" - "movdqa %%xmm4,%%xmm12 \n" - "punpcklbw %%xmm5,%%xmm4 \n" - "punpckhbw %%xmm5,%%xmm12 \n" - "movdqa %%xmm4,%%xmm5 \n" - "movdqa %%xmm12,%%xmm13 \n" - "movdqa (%0),%%xmm6 \n" - "palignr $0x8,%%xmm5,%%xmm5 \n" - "palignr $0x8,%%xmm13,%%xmm13 \n" - "movdqa (%0,%3),%%xmm7 \n" - "lea (%0,%3,2),%0 \n" - "movdqa %%xmm6,%%xmm14 \n" - "punpcklbw %%xmm7,%%xmm6 \n" - "punpckhbw %%xmm7,%%xmm14 \n" - "neg %3 \n" - "movdqa %%xmm6,%%xmm7 \n" - "movdqa %%xmm14,%%xmm15 \n" - "lea 0x10(%0,%3,8),%0 \n" - "palignr $0x8,%%xmm7,%%xmm7 \n" - "palignr $0x8,%%xmm15,%%xmm15 \n" - "neg %3 \n" - // Second round of bit swap. - "punpcklwd %%xmm2,%%xmm0 \n" - "punpcklwd %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "movdqa %%xmm1,%%xmm3 \n" - "palignr $0x8,%%xmm2,%%xmm2 \n" - "palignr $0x8,%%xmm3,%%xmm3 \n" - "punpcklwd %%xmm6,%%xmm4 \n" - "punpcklwd %%xmm7,%%xmm5 \n" - "movdqa %%xmm4,%%xmm6 \n" - "movdqa %%xmm5,%%xmm7 \n" - "palignr $0x8,%%xmm6,%%xmm6 \n" - "palignr $0x8,%%xmm7,%%xmm7 \n" - "punpcklwd %%xmm10,%%xmm8 \n" - "punpcklwd %%xmm11,%%xmm9 \n" - "movdqa %%xmm8,%%xmm10 \n" - "movdqa %%xmm9,%%xmm11 \n" - "palignr $0x8,%%xmm10,%%xmm10 \n" - "palignr $0x8,%%xmm11,%%xmm11 \n" - "punpcklwd %%xmm14,%%xmm12 \n" - "punpcklwd %%xmm15,%%xmm13 \n" - "movdqa %%xmm12,%%xmm14 \n" - "movdqa %%xmm13,%%xmm15 \n" - "palignr $0x8,%%xmm14,%%xmm14 \n" - "palignr $0x8,%%xmm15,%%xmm15 \n" - // Third round of bit swap. - // Write to the destination pointer. - "punpckldq %%xmm4,%%xmm0 \n" - "movq %%xmm0,(%1) \n" - "movdqa %%xmm0,%%xmm4 \n" - "palignr $0x8,%%xmm4,%%xmm4 \n" - "movq %%xmm4,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm6,%%xmm2 \n" - "movdqa %%xmm2,%%xmm6 \n" - "movq %%xmm2,(%1) \n" - "palignr $0x8,%%xmm6,%%xmm6 \n" - "punpckldq %%xmm5,%%xmm1 \n" - "movq %%xmm6,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "movdqa %%xmm1,%%xmm5 \n" - "movq %%xmm1,(%1) \n" - "palignr $0x8,%%xmm5,%%xmm5 \n" - "movq %%xmm5,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm7,%%xmm3 \n" - "movq %%xmm3,(%1) \n" - "movdqa %%xmm3,%%xmm7 \n" - "palignr $0x8,%%xmm7,%%xmm7 \n" - "movq %%xmm7,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm12,%%xmm8 \n" - "movq %%xmm8,(%1) \n" - "movdqa %%xmm8,%%xmm12 \n" - "palignr $0x8,%%xmm12,%%xmm12 \n" - "movq %%xmm12,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm14,%%xmm10 \n" - "movdqa %%xmm10,%%xmm14 \n" - "movq %%xmm10,(%1) \n" - "palignr $0x8,%%xmm14,%%xmm14 \n" - "punpckldq %%xmm13,%%xmm9 \n" - "movq %%xmm14,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "movdqa %%xmm9,%%xmm13 \n" - "movq %%xmm9,(%1) \n" - "palignr $0x8,%%xmm13,%%xmm13 \n" - "movq %%xmm13,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "punpckldq %%xmm15,%%xmm11 \n" - "movq %%xmm11,(%1) \n" - "movdqa %%xmm11,%%xmm15 \n" - "palignr $0x8,%%xmm15,%%xmm15 \n" - "sub $0x10,%2 \n" - "movq %%xmm15,(%1,%4) \n" - "lea (%1,%4,2),%1 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : "r"((intptr_t)(src_stride)), // %3 - "r"((intptr_t)(dst_stride)) // %4 - : "memory", "cc", - "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7", - "xmm8", "xmm9", "xmm10", "xmm11", "xmm12", "xmm13", "xmm14", "xmm15" -); -} - -#define HAS_TRANSPOSE_UVWX8_SSE2 -static void TransposeUVWx8_SSE2(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int w) { - asm volatile ( - // Read in the data from the source pointer. - // First round of bit swap. - ".p2align 2 \n" -"1: \n" - "movdqa (%0),%%xmm0 \n" - "movdqa (%0,%4),%%xmm1 \n" - "lea (%0,%4,2),%0 \n" - "movdqa %%xmm0,%%xmm8 \n" - "punpcklbw %%xmm1,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm8 \n" - "movdqa %%xmm8,%%xmm1 \n" - "movdqa (%0),%%xmm2 \n" - "movdqa (%0,%4),%%xmm3 \n" - "lea (%0,%4,2),%0 \n" - "movdqa %%xmm2,%%xmm8 \n" - "punpcklbw %%xmm3,%%xmm2 \n" - "punpckhbw %%xmm3,%%xmm8 \n" - "movdqa %%xmm8,%%xmm3 \n" - "movdqa (%0),%%xmm4 \n" - "movdqa (%0,%4),%%xmm5 \n" - "lea (%0,%4,2),%0 \n" - "movdqa %%xmm4,%%xmm8 \n" - "punpcklbw %%xmm5,%%xmm4 \n" - "punpckhbw %%xmm5,%%xmm8 \n" - "movdqa %%xmm8,%%xmm5 \n" - "movdqa (%0),%%xmm6 \n" - "movdqa (%0,%4),%%xmm7 \n" - "lea (%0,%4,2),%0 \n" - "movdqa %%xmm6,%%xmm8 \n" - "punpcklbw %%xmm7,%%xmm6 \n" - "neg %4 \n" - "lea 0x10(%0,%4,8),%0 \n" - "punpckhbw %%xmm7,%%xmm8 \n" - "movdqa %%xmm8,%%xmm7 \n" - "neg %4 \n" - // Second round of bit swap. - "movdqa %%xmm0,%%xmm8 \n" - "movdqa %%xmm1,%%xmm9 \n" - "punpckhwd %%xmm2,%%xmm8 \n" - "punpckhwd %%xmm3,%%xmm9 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpcklwd %%xmm3,%%xmm1 \n" - "movdqa %%xmm8,%%xmm2 \n" - "movdqa %%xmm9,%%xmm3 \n" - "movdqa %%xmm4,%%xmm8 \n" - "movdqa %%xmm5,%%xmm9 \n" - "punpckhwd %%xmm6,%%xmm8 \n" - "punpckhwd %%xmm7,%%xmm9 \n" - "punpcklwd %%xmm6,%%xmm4 \n" - "punpcklwd %%xmm7,%%xmm5 \n" - "movdqa %%xmm8,%%xmm6 \n" - "movdqa %%xmm9,%%xmm7 \n" - // Third round of bit swap. - // Write to the destination pointer. - "movdqa %%xmm0,%%xmm8 \n" - "punpckldq %%xmm4,%%xmm0 \n" - "movlpd %%xmm0,(%1) \n" // Write back U channel - "movhpd %%xmm0,(%2) \n" // Write back V channel - "punpckhdq %%xmm4,%%xmm8 \n" - "movlpd %%xmm8,(%1,%5) \n" - "lea (%1,%5,2),%1 \n" - "movhpd %%xmm8,(%2,%6) \n" - "lea (%2,%6,2),%2 \n" - "movdqa %%xmm2,%%xmm8 \n" - "punpckldq %%xmm6,%%xmm2 \n" - "movlpd %%xmm2,(%1) \n" - "movhpd %%xmm2,(%2) \n" - "punpckhdq %%xmm6,%%xmm8 \n" - "movlpd %%xmm8,(%1,%5) \n" - "lea (%1,%5,2),%1 \n" - "movhpd %%xmm8,(%2,%6) \n" - "lea (%2,%6,2),%2 \n" - "movdqa %%xmm1,%%xmm8 \n" - "punpckldq %%xmm5,%%xmm1 \n" - "movlpd %%xmm1,(%1) \n" - "movhpd %%xmm1,(%2) \n" - "punpckhdq %%xmm5,%%xmm8 \n" - "movlpd %%xmm8,(%1,%5) \n" - "lea (%1,%5,2),%1 \n" - "movhpd %%xmm8,(%2,%6) \n" - "lea (%2,%6,2),%2 \n" - "movdqa %%xmm3,%%xmm8 \n" - "punpckldq %%xmm7,%%xmm3 \n" - "movlpd %%xmm3,(%1) \n" - "movhpd %%xmm3,(%2) \n" - "punpckhdq %%xmm7,%%xmm8 \n" - "sub $0x8,%3 \n" - "movlpd %%xmm8,(%1,%5) \n" - "lea (%1,%5,2),%1 \n" - "movhpd %%xmm8,(%2,%6) \n" - "lea (%2,%6,2),%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst_a), // %1 - "+r"(dst_b), // %2 - "+r"(w) // %3 - : "r"((intptr_t)(src_stride)), // %4 - "r"((intptr_t)(dst_stride_a)), // %5 - "r"((intptr_t)(dst_stride_b)) // %6 - : "memory", "cc", - "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7", - "xmm8", "xmm9" -); -} -#endif -#endif - -static void TransposeWx8_C(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width) { - int i; - for (i = 0; i < width; ++i) { - dst[0] = src[0 * src_stride]; - dst[1] = src[1 * src_stride]; - dst[2] = src[2 * src_stride]; - dst[3] = src[3 * src_stride]; - dst[4] = src[4 * src_stride]; - dst[5] = src[5 * src_stride]; - dst[6] = src[6 * src_stride]; - dst[7] = src[7 * src_stride]; - ++src; - dst += dst_stride; - } -} - -static void TransposeWxH_C(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - int i; - for (i = 0; i < width; ++i) { - int j; - for (j = 0; j < height; ++j) { - dst[i * dst_stride + j] = src[j * src_stride + i]; - } - } -} - -LIBYUV_API -void TransposePlane(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - int i = height; - void (*TransposeWx8)(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width) = TransposeWx8_C; -#if defined(HAS_TRANSPOSE_WX8_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - TransposeWx8 = TransposeWx8_NEON; - } -#endif -#if defined(HAS_TRANSPOSE_WX8_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 8)) { - TransposeWx8 = TransposeWx8_SSSE3; - } -#endif -#if defined(HAS_TRANSPOSE_WX8_FAST_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && - IS_ALIGNED(width, 16) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16)) { - TransposeWx8 = TransposeWx8_FAST_SSSE3; - } -#endif -#if defined(HAS_TRANSPOSE_WX8_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2)) { - if (IS_ALIGNED(width, 4) && - IS_ALIGNED(src, 4) && IS_ALIGNED(src_stride, 4)) { - TransposeWx8 = TransposeWx8_FAST_MIPS_DSPR2; - } else { - TransposeWx8 = TransposeWx8_MIPS_DSPR2; - } - } -#endif - - // Work across the source in 8x8 tiles - while (i >= 8) { - TransposeWx8(src, src_stride, dst, dst_stride, width); - src += 8 * src_stride; // Go down 8 rows. - dst += 8; // Move over 8 columns. - i -= 8; - } - - TransposeWxH_C(src, src_stride, dst, dst_stride, width, i); -} - -LIBYUV_API -void RotatePlane90(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - // Rotate by 90 is a transpose with the source read - // from bottom to top. So set the source pointer to the end - // of the buffer and flip the sign of the source stride. - src += src_stride * (height - 1); - src_stride = -src_stride; - TransposePlane(src, src_stride, dst, dst_stride, width, height); -} - -LIBYUV_API -void RotatePlane270(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - // Rotate by 270 is a transpose with the destination written - // from bottom to top. So set the destination pointer to the end - // of the buffer and flip the sign of the destination stride. - dst += dst_stride * (width - 1); - dst_stride = -dst_stride; - TransposePlane(src, src_stride, dst, dst_stride, width, height); -} - -LIBYUV_API -void RotatePlane180(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - // Swap first and last row and mirror the content. Uses a temporary row. - align_buffer_64(row, width); - const uint8* src_bot = src + src_stride * (height - 1); - uint8* dst_bot = dst + dst_stride * (height - 1); - int half_height = (height + 1) >> 1; - int y; - void (*MirrorRow)(const uint8* src, uint8* dst, int width) = MirrorRow_C; - void (*CopyRow)(const uint8* src, uint8* dst, int width) = CopyRow_C; -#if defined(HAS_MIRRORROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 16)) { - MirrorRow = MirrorRow_NEON; - } -#endif -#if defined(HAS_MIRRORROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 16) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - MirrorRow = MirrorRow_SSE2; - } -#endif -#if defined(HAS_MIRRORROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 16) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - MirrorRow = MirrorRow_SSSE3; - } -#endif -#if defined(HAS_MIRRORROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && IS_ALIGNED(width, 32)) { - MirrorRow = MirrorRow_AVX2; - } -#endif -#if defined(HAS_MIRRORROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && - IS_ALIGNED(src, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst, 4) && IS_ALIGNED(dst_stride, 4)) { - MirrorRow = MirrorRow_MIPS_DSPR2; - } -#endif -#if defined(HAS_COPYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 32)) { - CopyRow = CopyRow_NEON; - } -#endif -#if defined(HAS_COPYROW_X86) - if (TestCpuFlag(kCpuHasX86) && IS_ALIGNED(width, 4)) { - CopyRow = CopyRow_X86; - } -#endif -#if defined(HAS_COPYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width, 32) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - CopyRow = CopyRow_SSE2; - } -#endif -#if defined(HAS_COPYROW_ERMS) - if (TestCpuFlag(kCpuHasERMS)) { - CopyRow = CopyRow_ERMS; - } -#endif -#if defined(HAS_COPYROW_MIPS) - if (TestCpuFlag(kCpuHasMIPS)) { - CopyRow = CopyRow_MIPS; - } -#endif - - // Odd height will harmlessly mirror the middle row twice. - for (y = 0; y < half_height; ++y) { - MirrorRow(src, row, width); // Mirror first row into a buffer - src += src_stride; - MirrorRow(src_bot, dst, width); // Mirror last row into first row - dst += dst_stride; - CopyRow(row, dst_bot, width); // Copy first mirrored row into last - src_bot -= src_stride; - dst_bot -= dst_stride; - } - free_aligned_buffer_64(row); -} - -static void TransposeUVWx8_C(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width) { - int i; - for (i = 0; i < width; ++i) { - dst_a[0] = src[0 * src_stride + 0]; - dst_b[0] = src[0 * src_stride + 1]; - dst_a[1] = src[1 * src_stride + 0]; - dst_b[1] = src[1 * src_stride + 1]; - dst_a[2] = src[2 * src_stride + 0]; - dst_b[2] = src[2 * src_stride + 1]; - dst_a[3] = src[3 * src_stride + 0]; - dst_b[3] = src[3 * src_stride + 1]; - dst_a[4] = src[4 * src_stride + 0]; - dst_b[4] = src[4 * src_stride + 1]; - dst_a[5] = src[5 * src_stride + 0]; - dst_b[5] = src[5 * src_stride + 1]; - dst_a[6] = src[6 * src_stride + 0]; - dst_b[6] = src[6 * src_stride + 1]; - dst_a[7] = src[7 * src_stride + 0]; - dst_b[7] = src[7 * src_stride + 1]; - src += 2; - dst_a += dst_stride_a; - dst_b += dst_stride_b; - } -} - -static void TransposeUVWxH_C(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height) { - int i; - for (i = 0; i < width * 2; i += 2) { - int j; - for (j = 0; j < height; ++j) { - dst_a[j + ((i >> 1) * dst_stride_a)] = src[i + (j * src_stride)]; - dst_b[j + ((i >> 1) * dst_stride_b)] = src[i + (j * src_stride) + 1]; - } - } -} - -LIBYUV_API -void TransposeUV(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height) { - int i = height; - void (*TransposeUVWx8)(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width) = TransposeUVWx8_C; -#if defined(HAS_TRANSPOSE_UVWX8_NEON) - if (TestCpuFlag(kCpuHasNEON)) { - TransposeUVWx8 = TransposeUVWx8_NEON; - } -#elif defined(HAS_TRANSPOSE_UVWX8_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && - IS_ALIGNED(width, 8) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16)) { - TransposeUVWx8 = TransposeUVWx8_SSE2; - } -#elif defined(HAS_TRANSPOSE_UVWx8_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(width, 2) && - IS_ALIGNED(src, 4) && IS_ALIGNED(src_stride, 4)) { - TransposeUVWx8 = TransposeUVWx8_MIPS_DSPR2; - } -#endif - - // Work through the source in 8x8 tiles. - while (i >= 8) { - TransposeUVWx8(src, src_stride, - dst_a, dst_stride_a, - dst_b, dst_stride_b, - width); - src += 8 * src_stride; // Go down 8 rows. - dst_a += 8; // Move over 8 columns. - dst_b += 8; // Move over 8 columns. - i -= 8; - } - - TransposeUVWxH_C(src, src_stride, - dst_a, dst_stride_a, - dst_b, dst_stride_b, - width, i); -} - -LIBYUV_API -void RotateUV90(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height) { - src += src_stride * (height - 1); - src_stride = -src_stride; - - TransposeUV(src, src_stride, - dst_a, dst_stride_a, - dst_b, dst_stride_b, - width, height); -} - -LIBYUV_API -void RotateUV270(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height) { - dst_a += dst_stride_a * (width - 1); - dst_b += dst_stride_b * (width - 1); - dst_stride_a = -dst_stride_a; - dst_stride_b = -dst_stride_b; - - TransposeUV(src, src_stride, - dst_a, dst_stride_a, - dst_b, dst_stride_b, - width, height); -} - -// Rotate 180 is a horizontal and vertical flip. -LIBYUV_API -void RotateUV180(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width, int height) { - int i; - void (*MirrorRowUV)(const uint8* src, uint8* dst_u, uint8* dst_v, int width) = - MirrorUVRow_C; -#if defined(HAS_MIRRORUVROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 8)) { - MirrorRowUV = MirrorUVRow_NEON; - } -#elif defined(HAS_MIRRORROW_UV_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 16) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16)) { - MirrorRowUV = MirrorUVRow_SSSE3; - } -#elif defined(HAS_MIRRORUVROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && - IS_ALIGNED(src, 4) && IS_ALIGNED(src_stride, 4)) { - MirrorRowUV = MirrorUVRow_MIPS_DSPR2; - } -#endif - - dst_a += dst_stride_a * (height - 1); - dst_b += dst_stride_b * (height - 1); - - for (i = 0; i < height; ++i) { - MirrorRowUV(src, dst_a, dst_b, width); - src += src_stride; - dst_a -= dst_stride_a; - dst_b -= dst_stride_b; - } -} - -LIBYUV_API -int RotatePlane(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height, - enum RotationMode mode) { - if (!src || width <= 0 || height == 0 || !dst) { - return -1; - } - - // Negative height means invert the image. - if (height < 0) { - height = -height; - src = src + (height - 1) * src_stride; - src_stride = -src_stride; - } - - switch (mode) { - case kRotate0: - // copy frame - CopyPlane(src, src_stride, - dst, dst_stride, - width, height); - return 0; - case kRotate90: - RotatePlane90(src, src_stride, - dst, dst_stride, - width, height); - return 0; - case kRotate270: - RotatePlane270(src, src_stride, - dst, dst_stride, - width, height); - return 0; - case kRotate180: - RotatePlane180(src, src_stride, - dst, dst_stride, - width, height); - return 0; - default: - break; - } - return -1; -} - -LIBYUV_API -int I420Rotate(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height, - enum RotationMode mode) { - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - if (!src_y || !src_u || !src_v || width <= 0 || height == 0 || - !dst_y || !dst_u || !dst_v) { - return -1; - } - - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - src_y = src_y + (height - 1) * src_stride_y; - src_u = src_u + (halfheight - 1) * src_stride_u; - src_v = src_v + (halfheight - 1) * src_stride_v; - src_stride_y = -src_stride_y; - src_stride_u = -src_stride_u; - src_stride_v = -src_stride_v; - } - - switch (mode) { - case kRotate0: - // copy frame - return I420Copy(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height); - case kRotate90: - RotatePlane90(src_y, src_stride_y, - dst_y, dst_stride_y, - width, height); - RotatePlane90(src_u, src_stride_u, - dst_u, dst_stride_u, - halfwidth, halfheight); - RotatePlane90(src_v, src_stride_v, - dst_v, dst_stride_v, - halfwidth, halfheight); - return 0; - case kRotate270: - RotatePlane270(src_y, src_stride_y, - dst_y, dst_stride_y, - width, height); - RotatePlane270(src_u, src_stride_u, - dst_u, dst_stride_u, - halfwidth, halfheight); - RotatePlane270(src_v, src_stride_v, - dst_v, dst_stride_v, - halfwidth, halfheight); - return 0; - case kRotate180: - RotatePlane180(src_y, src_stride_y, - dst_y, dst_stride_y, - width, height); - RotatePlane180(src_u, src_stride_u, - dst_u, dst_stride_u, - halfwidth, halfheight); - RotatePlane180(src_v, src_stride_v, - dst_v, dst_stride_v, - halfwidth, halfheight); - return 0; - default: - break; - } - return -1; -} - -LIBYUV_API -int NV12ToI420Rotate(const uint8* src_y, int src_stride_y, - const uint8* src_uv, int src_stride_uv, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int width, int height, - enum RotationMode mode) { - int halfwidth = (width + 1) >> 1; - int halfheight = (height + 1) >> 1; - if (!src_y || !src_uv || width <= 0 || height == 0 || - !dst_y || !dst_u || !dst_v) { - return -1; - } - - // Negative height means invert the image. - if (height < 0) { - height = -height; - halfheight = (height + 1) >> 1; - src_y = src_y + (height - 1) * src_stride_y; - src_uv = src_uv + (halfheight - 1) * src_stride_uv; - src_stride_y = -src_stride_y; - src_stride_uv = -src_stride_uv; - } - - switch (mode) { - case kRotate0: - // copy frame - return NV12ToI420(src_y, src_stride_y, - src_uv, src_stride_uv, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - width, height); - case kRotate90: - RotatePlane90(src_y, src_stride_y, - dst_y, dst_stride_y, - width, height); - RotateUV90(src_uv, src_stride_uv, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - halfwidth, halfheight); - return 0; - case kRotate270: - RotatePlane270(src_y, src_stride_y, - dst_y, dst_stride_y, - width, height); - RotateUV270(src_uv, src_stride_uv, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - halfwidth, halfheight); - return 0; - case kRotate180: - RotatePlane180(src_y, src_stride_y, - dst_y, dst_stride_y, - width, height); - RotateUV180(src_uv, src_stride_uv, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - halfwidth, halfheight); - return 0; - default: - break; - } - return -1; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/rotate_argb.cc b/src/main/jni/libyuv/source/rotate_argb.cc deleted file mode 100644 index ab0f9ce07..000000000 --- a/src/main/jni/libyuv/source/rotate_argb.cc +++ /dev/null @@ -1,209 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/rotate.h" - -#include "libyuv/cpu_id.h" -#include "libyuv/convert.h" -#include "libyuv/planar_functions.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// ARGBScale has a function to copy pixels to a row, striding each source -// pixel by a constant. -#if !defined(LIBYUV_DISABLE_X86) && \ - (defined(_M_IX86) || \ - (defined(__x86_64__) && !defined(__native_client__)) || defined(__i386__)) -#define HAS_SCALEARGBROWDOWNEVEN_SSE2 -void ScaleARGBRowDownEven_SSE2(const uint8* src_ptr, int src_stride, - int src_stepx, - uint8* dst_ptr, int dst_width); -#endif -#if !defined(LIBYUV_DISABLE_NEON) && !defined(__native_client__) && \ - (defined(__ARM_NEON__) || defined(LIBYUV_NEON)) -#define HAS_SCALEARGBROWDOWNEVEN_NEON -void ScaleARGBRowDownEven_NEON(const uint8* src_ptr, int src_stride, - int src_stepx, - uint8* dst_ptr, int dst_width); -#endif - -void ScaleARGBRowDownEven_C(const uint8* src_ptr, int, - int src_stepx, - uint8* dst_ptr, int dst_width); - -static void ARGBTranspose(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - int i; - int src_pixel_step = src_stride >> 2; - void (*ScaleARGBRowDownEven)(const uint8* src_ptr, int src_stride, - int src_step, uint8* dst_ptr, int dst_width) = ScaleARGBRowDownEven_C; -#if defined(HAS_SCALEARGBROWDOWNEVEN_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(height, 4) && // Width of dest. - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBRowDownEven = ScaleARGBRowDownEven_SSE2; - } -#elif defined(HAS_SCALEARGBROWDOWNEVEN_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(height, 4) && // Width of dest. - IS_ALIGNED(src, 4)) { - ScaleARGBRowDownEven = ScaleARGBRowDownEven_NEON; - } -#endif - - for (i = 0; i < width; ++i) { // column of source to row of dest. - ScaleARGBRowDownEven(src, 0, src_pixel_step, dst, height); - dst += dst_stride; - src += 4; - } -} - -void ARGBRotate90(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - // Rotate by 90 is a ARGBTranspose with the source read - // from bottom to top. So set the source pointer to the end - // of the buffer and flip the sign of the source stride. - src += src_stride * (height - 1); - src_stride = -src_stride; - ARGBTranspose(src, src_stride, dst, dst_stride, width, height); -} - -void ARGBRotate270(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - // Rotate by 270 is a ARGBTranspose with the destination written - // from bottom to top. So set the destination pointer to the end - // of the buffer and flip the sign of the destination stride. - dst += dst_stride * (width - 1); - dst_stride = -dst_stride; - ARGBTranspose(src, src_stride, dst, dst_stride, width, height); -} - -void ARGBRotate180(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width, int height) { - // Swap first and last row and mirror the content. Uses a temporary row. - align_buffer_64(row, width * 4); - const uint8* src_bot = src + src_stride * (height - 1); - uint8* dst_bot = dst + dst_stride * (height - 1); - int half_height = (height + 1) >> 1; - int y; - void (*ARGBMirrorRow)(const uint8* src, uint8* dst, int width) = - ARGBMirrorRow_C; - void (*CopyRow)(const uint8* src, uint8* dst, int width) = CopyRow_C; -#if defined(HAS_ARGBMIRRORROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && IS_ALIGNED(width, 4) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - ARGBMirrorRow = ARGBMirrorRow_SSSE3; - } -#endif -#if defined(HAS_ARGBMIRRORROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && IS_ALIGNED(width, 8)) { - ARGBMirrorRow = ARGBMirrorRow_AVX2; - } -#endif -#if defined(HAS_ARGBMIRRORROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width, 4)) { - ARGBMirrorRow = ARGBMirrorRow_NEON; - } -#endif -#if defined(HAS_COPYROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(width * 4, 32)) { - CopyRow = CopyRow_NEON; - } -#endif -#if defined(HAS_COPYROW_X86) - if (TestCpuFlag(kCpuHasX86)) { - CopyRow = CopyRow_X86; - } -#endif -#if defined(HAS_COPYROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(width * 4, 32) && - IS_ALIGNED(src, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst, 16) && IS_ALIGNED(dst_stride, 16)) { - CopyRow = CopyRow_SSE2; - } -#endif -#if defined(HAS_COPYROW_ERMS) - if (TestCpuFlag(kCpuHasERMS)) { - CopyRow = CopyRow_ERMS; - } -#endif -#if defined(HAS_COPYROW_MIPS) - if (TestCpuFlag(kCpuHasMIPS)) { - CopyRow = CopyRow_MIPS; - } -#endif - - // Odd height will harmlessly mirror the middle row twice. - for (y = 0; y < half_height; ++y) { - ARGBMirrorRow(src, row, width); // Mirror first row into a buffer - ARGBMirrorRow(src_bot, dst, width); // Mirror last row into first row - CopyRow(row, dst_bot, width * 4); // Copy first mirrored row into last - src += src_stride; - dst += dst_stride; - src_bot -= src_stride; - dst_bot -= dst_stride; - } - free_aligned_buffer_64(row); -} - -LIBYUV_API -int ARGBRotate(const uint8* src_argb, int src_stride_argb, - uint8* dst_argb, int dst_stride_argb, - int width, int height, - enum RotationMode mode) { - if (!src_argb || width <= 0 || height == 0 || !dst_argb) { - return -1; - } - - // Negative height means invert the image. - if (height < 0) { - height = -height; - src_argb = src_argb + (height - 1) * src_stride_argb; - src_stride_argb = -src_stride_argb; - } - - switch (mode) { - case kRotate0: - // copy frame - return ARGBCopy(src_argb, src_stride_argb, - dst_argb, dst_stride_argb, - width, height); - case kRotate90: - ARGBRotate90(src_argb, src_stride_argb, - dst_argb, dst_stride_argb, - width, height); - return 0; - case kRotate270: - ARGBRotate270(src_argb, src_stride_argb, - dst_argb, dst_stride_argb, - width, height); - return 0; - case kRotate180: - ARGBRotate180(src_argb, src_stride_argb, - dst_argb, dst_stride_argb, - width, height); - return 0; - default: - break; - } - return -1; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/rotate_mips.cc b/src/main/jni/libyuv/source/rotate_mips.cc deleted file mode 100644 index 70770fd06..000000000 --- a/src/main/jni/libyuv/source/rotate_mips.cc +++ /dev/null @@ -1,485 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_MIPS) && \ - defined(__mips_dsp) && (__mips_dsp_rev >= 2) && \ - (_MIPS_SIM == _MIPS_SIM_ABI32) - -void TransposeWx8_MIPS_DSPR2(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "sll $t2, %[src_stride], 0x1 \n" // src_stride x 2 - "sll $t4, %[src_stride], 0x2 \n" // src_stride x 4 - "sll $t9, %[src_stride], 0x3 \n" // src_stride x 8 - "addu $t3, $t2, %[src_stride] \n" - "addu $t5, $t4, %[src_stride] \n" - "addu $t6, $t2, $t4 \n" - "andi $t0, %[dst], 0x3 \n" - "andi $t1, %[dst_stride], 0x3 \n" - "or $t0, $t0, $t1 \n" - "bnez $t0, 11f \n" - " subu $t7, $t9, %[src_stride] \n" -//dst + dst_stride word aligned - "1: \n" - "lbu $t0, 0(%[src]) \n" - "lbux $t1, %[src_stride](%[src]) \n" - "lbux $t8, $t2(%[src]) \n" - "lbux $t9, $t3(%[src]) \n" - "sll $t1, $t1, 16 \n" - "sll $t9, $t9, 16 \n" - "or $t0, $t0, $t1 \n" - "or $t8, $t8, $t9 \n" - "precr.qb.ph $s0, $t8, $t0 \n" - "lbux $t0, $t4(%[src]) \n" - "lbux $t1, $t5(%[src]) \n" - "lbux $t8, $t6(%[src]) \n" - "lbux $t9, $t7(%[src]) \n" - "sll $t1, $t1, 16 \n" - "sll $t9, $t9, 16 \n" - "or $t0, $t0, $t1 \n" - "or $t8, $t8, $t9 \n" - "precr.qb.ph $s1, $t8, $t0 \n" - "sw $s0, 0(%[dst]) \n" - "addiu %[width], -1 \n" - "addiu %[src], 1 \n" - "sw $s1, 4(%[dst]) \n" - "bnez %[width], 1b \n" - " addu %[dst], %[dst], %[dst_stride] \n" - "b 2f \n" -//dst + dst_stride unaligned - "11: \n" - "lbu $t0, 0(%[src]) \n" - "lbux $t1, %[src_stride](%[src]) \n" - "lbux $t8, $t2(%[src]) \n" - "lbux $t9, $t3(%[src]) \n" - "sll $t1, $t1, 16 \n" - "sll $t9, $t9, 16 \n" - "or $t0, $t0, $t1 \n" - "or $t8, $t8, $t9 \n" - "precr.qb.ph $s0, $t8, $t0 \n" - "lbux $t0, $t4(%[src]) \n" - "lbux $t1, $t5(%[src]) \n" - "lbux $t8, $t6(%[src]) \n" - "lbux $t9, $t7(%[src]) \n" - "sll $t1, $t1, 16 \n" - "sll $t9, $t9, 16 \n" - "or $t0, $t0, $t1 \n" - "or $t8, $t8, $t9 \n" - "precr.qb.ph $s1, $t8, $t0 \n" - "swr $s0, 0(%[dst]) \n" - "swl $s0, 3(%[dst]) \n" - "addiu %[width], -1 \n" - "addiu %[src], 1 \n" - "swr $s1, 4(%[dst]) \n" - "swl $s1, 7(%[dst]) \n" - "bnez %[width], 11b \n" - "addu %[dst], %[dst], %[dst_stride] \n" - "2: \n" - ".set pop \n" - :[src] "+r" (src), - [dst] "+r" (dst), - [width] "+r" (width) - :[src_stride] "r" (src_stride), - [dst_stride] "r" (dst_stride) - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9", - "s0", "s1" - ); -} - -void TransposeWx8_FAST_MIPS_DSPR2(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width) { - __asm__ __volatile__ ( - ".set noat \n" - ".set push \n" - ".set noreorder \n" - "beqz %[width], 2f \n" - " sll $t2, %[src_stride], 0x1 \n" // src_stride x 2 - "sll $t4, %[src_stride], 0x2 \n" // src_stride x 4 - "sll $t9, %[src_stride], 0x3 \n" // src_stride x 8 - "addu $t3, $t2, %[src_stride] \n" - "addu $t5, $t4, %[src_stride] \n" - "addu $t6, $t2, $t4 \n" - - "srl $AT, %[width], 0x2 \n" - "andi $t0, %[dst], 0x3 \n" - "andi $t1, %[dst_stride], 0x3 \n" - "or $t0, $t0, $t1 \n" - "bnez $t0, 11f \n" - " subu $t7, $t9, %[src_stride] \n" -//dst + dst_stride word aligned - "1: \n" - "lw $t0, 0(%[src]) \n" - "lwx $t1, %[src_stride](%[src]) \n" - "lwx $t8, $t2(%[src]) \n" - "lwx $t9, $t3(%[src]) \n" - -// t0 = | 30 | 20 | 10 | 00 | -// t1 = | 31 | 21 | 11 | 01 | -// t8 = | 32 | 22 | 12 | 02 | -// t9 = | 33 | 23 | 13 | 03 | - - "precr.qb.ph $s0, $t1, $t0 \n" - "precr.qb.ph $s1, $t9, $t8 \n" - "precrq.qb.ph $s2, $t1, $t0 \n" - "precrq.qb.ph $s3, $t9, $t8 \n" - - // s0 = | 21 | 01 | 20 | 00 | - // s1 = | 23 | 03 | 22 | 02 | - // s2 = | 31 | 11 | 30 | 10 | - // s3 = | 33 | 13 | 32 | 12 | - - "precr.qb.ph $s4, $s1, $s0 \n" - "precrq.qb.ph $s5, $s1, $s0 \n" - "precr.qb.ph $s6, $s3, $s2 \n" - "precrq.qb.ph $s7, $s3, $s2 \n" - - // s4 = | 03 | 02 | 01 | 00 | - // s5 = | 23 | 22 | 21 | 20 | - // s6 = | 13 | 12 | 11 | 10 | - // s7 = | 33 | 32 | 31 | 30 | - - "lwx $t0, $t4(%[src]) \n" - "lwx $t1, $t5(%[src]) \n" - "lwx $t8, $t6(%[src]) \n" - "lwx $t9, $t7(%[src]) \n" - -// t0 = | 34 | 24 | 14 | 04 | -// t1 = | 35 | 25 | 15 | 05 | -// t8 = | 36 | 26 | 16 | 06 | -// t9 = | 37 | 27 | 17 | 07 | - - "precr.qb.ph $s0, $t1, $t0 \n" - "precr.qb.ph $s1, $t9, $t8 \n" - "precrq.qb.ph $s2, $t1, $t0 \n" - "precrq.qb.ph $s3, $t9, $t8 \n" - - // s0 = | 25 | 05 | 24 | 04 | - // s1 = | 27 | 07 | 26 | 06 | - // s2 = | 35 | 15 | 34 | 14 | - // s3 = | 37 | 17 | 36 | 16 | - - "precr.qb.ph $t0, $s1, $s0 \n" - "precrq.qb.ph $t1, $s1, $s0 \n" - "precr.qb.ph $t8, $s3, $s2 \n" - "precrq.qb.ph $t9, $s3, $s2 \n" - - // t0 = | 07 | 06 | 05 | 04 | - // t1 = | 27 | 26 | 25 | 24 | - // t8 = | 17 | 16 | 15 | 14 | - // t9 = | 37 | 36 | 35 | 34 | - - "addu $s0, %[dst], %[dst_stride] \n" - "addu $s1, $s0, %[dst_stride] \n" - "addu $s2, $s1, %[dst_stride] \n" - - "sw $s4, 0(%[dst]) \n" - "sw $t0, 4(%[dst]) \n" - "sw $s6, 0($s0) \n" - "sw $t8, 4($s0) \n" - "sw $s5, 0($s1) \n" - "sw $t1, 4($s1) \n" - "sw $s7, 0($s2) \n" - "sw $t9, 4($s2) \n" - - "addiu $AT, -1 \n" - "addiu %[src], 4 \n" - - "bnez $AT, 1b \n" - " addu %[dst], $s2, %[dst_stride] \n" - "b 2f \n" -//dst + dst_stride unaligned - "11: \n" - "lw $t0, 0(%[src]) \n" - "lwx $t1, %[src_stride](%[src]) \n" - "lwx $t8, $t2(%[src]) \n" - "lwx $t9, $t3(%[src]) \n" - -// t0 = | 30 | 20 | 10 | 00 | -// t1 = | 31 | 21 | 11 | 01 | -// t8 = | 32 | 22 | 12 | 02 | -// t9 = | 33 | 23 | 13 | 03 | - - "precr.qb.ph $s0, $t1, $t0 \n" - "precr.qb.ph $s1, $t9, $t8 \n" - "precrq.qb.ph $s2, $t1, $t0 \n" - "precrq.qb.ph $s3, $t9, $t8 \n" - - // s0 = | 21 | 01 | 20 | 00 | - // s1 = | 23 | 03 | 22 | 02 | - // s2 = | 31 | 11 | 30 | 10 | - // s3 = | 33 | 13 | 32 | 12 | - - "precr.qb.ph $s4, $s1, $s0 \n" - "precrq.qb.ph $s5, $s1, $s0 \n" - "precr.qb.ph $s6, $s3, $s2 \n" - "precrq.qb.ph $s7, $s3, $s2 \n" - - // s4 = | 03 | 02 | 01 | 00 | - // s5 = | 23 | 22 | 21 | 20 | - // s6 = | 13 | 12 | 11 | 10 | - // s7 = | 33 | 32 | 31 | 30 | - - "lwx $t0, $t4(%[src]) \n" - "lwx $t1, $t5(%[src]) \n" - "lwx $t8, $t6(%[src]) \n" - "lwx $t9, $t7(%[src]) \n" - -// t0 = | 34 | 24 | 14 | 04 | -// t1 = | 35 | 25 | 15 | 05 | -// t8 = | 36 | 26 | 16 | 06 | -// t9 = | 37 | 27 | 17 | 07 | - - "precr.qb.ph $s0, $t1, $t0 \n" - "precr.qb.ph $s1, $t9, $t8 \n" - "precrq.qb.ph $s2, $t1, $t0 \n" - "precrq.qb.ph $s3, $t9, $t8 \n" - - // s0 = | 25 | 05 | 24 | 04 | - // s1 = | 27 | 07 | 26 | 06 | - // s2 = | 35 | 15 | 34 | 14 | - // s3 = | 37 | 17 | 36 | 16 | - - "precr.qb.ph $t0, $s1, $s0 \n" - "precrq.qb.ph $t1, $s1, $s0 \n" - "precr.qb.ph $t8, $s3, $s2 \n" - "precrq.qb.ph $t9, $s3, $s2 \n" - - // t0 = | 07 | 06 | 05 | 04 | - // t1 = | 27 | 26 | 25 | 24 | - // t8 = | 17 | 16 | 15 | 14 | - // t9 = | 37 | 36 | 35 | 34 | - - "addu $s0, %[dst], %[dst_stride] \n" - "addu $s1, $s0, %[dst_stride] \n" - "addu $s2, $s1, %[dst_stride] \n" - - "swr $s4, 0(%[dst]) \n" - "swl $s4, 3(%[dst]) \n" - "swr $t0, 4(%[dst]) \n" - "swl $t0, 7(%[dst]) \n" - "swr $s6, 0($s0) \n" - "swl $s6, 3($s0) \n" - "swr $t8, 4($s0) \n" - "swl $t8, 7($s0) \n" - "swr $s5, 0($s1) \n" - "swl $s5, 3($s1) \n" - "swr $t1, 4($s1) \n" - "swl $t1, 7($s1) \n" - "swr $s7, 0($s2) \n" - "swl $s7, 3($s2) \n" - "swr $t9, 4($s2) \n" - "swl $t9, 7($s2) \n" - - "addiu $AT, -1 \n" - "addiu %[src], 4 \n" - - "bnez $AT, 11b \n" - " addu %[dst], $s2, %[dst_stride] \n" - "2: \n" - ".set pop \n" - ".set at \n" - :[src] "+r" (src), - [dst] "+r" (dst), - [width] "+r" (width) - :[src_stride] "r" (src_stride), - [dst_stride] "r" (dst_stride) - : "t0", "t1", "t2", "t3", "t4", "t5", "t6", "t7", "t8", "t9", - "s0", "s1", "s2", "s3", "s4", "s5", "s6", "s7" - ); -} - -void TransposeUVWx8_MIPS_DSPR2(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "beqz %[width], 2f \n" - " sll $t2, %[src_stride], 0x1 \n" // src_stride x 2 - "sll $t4, %[src_stride], 0x2 \n" // src_stride x 4 - "sll $t9, %[src_stride], 0x3 \n" // src_stride x 8 - "addu $t3, $t2, %[src_stride] \n" - "addu $t5, $t4, %[src_stride] \n" - "addu $t6, $t2, $t4 \n" - "subu $t7, $t9, %[src_stride] \n" - "srl $t1, %[width], 1 \n" - -// check word aligment for dst_a, dst_b, dst_stride_a and dst_stride_b - "andi $t0, %[dst_a], 0x3 \n" - "andi $t8, %[dst_b], 0x3 \n" - "or $t0, $t0, $t8 \n" - "andi $t8, %[dst_stride_a], 0x3 \n" - "andi $s5, %[dst_stride_b], 0x3 \n" - "or $t8, $t8, $s5 \n" - "or $t0, $t0, $t8 \n" - "bnez $t0, 11f \n" - " nop \n" -// dst + dst_stride word aligned (both, a & b dst addresses) - "1: \n" - "lw $t0, 0(%[src]) \n" // |B0|A0|b0|a0| - "lwx $t8, %[src_stride](%[src]) \n" // |B1|A1|b1|a1| - "addu $s5, %[dst_a], %[dst_stride_a] \n" - "lwx $t9, $t2(%[src]) \n" // |B2|A2|b2|a2| - "lwx $s0, $t3(%[src]) \n" // |B3|A3|b3|a3| - "addu $s6, %[dst_b], %[dst_stride_b] \n" - - "precrq.ph.w $s1, $t8, $t0 \n" // |B1|A1|B0|A0| - "precrq.ph.w $s2, $s0, $t9 \n" // |B3|A3|B2|A2| - "precr.qb.ph $s3, $s2, $s1 \n" // |A3|A2|A1|A0| - "precrq.qb.ph $s4, $s2, $s1 \n" // |B3|B2|B1|B0| - - "sll $t0, $t0, 16 \n" - "packrl.ph $s1, $t8, $t0 \n" // |b1|a1|b0|a0| - "sll $t9, $t9, 16 \n" - "packrl.ph $s2, $s0, $t9 \n" // |b3|a3|b2|a2| - - "sw $s3, 0($s5) \n" - "sw $s4, 0($s6) \n" - - "precr.qb.ph $s3, $s2, $s1 \n" // |a3|a2|a1|a0| - "precrq.qb.ph $s4, $s2, $s1 \n" // |b3|b2|b1|b0| - - "lwx $t0, $t4(%[src]) \n" // |B4|A4|b4|a4| - "lwx $t8, $t5(%[src]) \n" // |B5|A5|b5|a5| - "lwx $t9, $t6(%[src]) \n" // |B6|A6|b6|a6| - "lwx $s0, $t7(%[src]) \n" // |B7|A7|b7|a7| - "sw $s3, 0(%[dst_a]) \n" - "sw $s4, 0(%[dst_b]) \n" - - "precrq.ph.w $s1, $t8, $t0 \n" // |B5|A5|B4|A4| - "precrq.ph.w $s2, $s0, $t9 \n" // |B6|A6|B7|A7| - "precr.qb.ph $s3, $s2, $s1 \n" // |A7|A6|A5|A4| - "precrq.qb.ph $s4, $s2, $s1 \n" // |B7|B6|B5|B4| - - "sll $t0, $t0, 16 \n" - "packrl.ph $s1, $t8, $t0 \n" // |b5|a5|b4|a4| - "sll $t9, $t9, 16 \n" - "packrl.ph $s2, $s0, $t9 \n" // |b7|a7|b6|a6| - "sw $s3, 4($s5) \n" - "sw $s4, 4($s6) \n" - - "precr.qb.ph $s3, $s2, $s1 \n" // |a7|a6|a5|a4| - "precrq.qb.ph $s4, $s2, $s1 \n" // |b7|b6|b5|b4| - - "addiu %[src], 4 \n" - "addiu $t1, -1 \n" - "sll $t0, %[dst_stride_a], 1 \n" - "sll $t8, %[dst_stride_b], 1 \n" - "sw $s3, 4(%[dst_a]) \n" - "sw $s4, 4(%[dst_b]) \n" - "addu %[dst_a], %[dst_a], $t0 \n" - "bnez $t1, 1b \n" - " addu %[dst_b], %[dst_b], $t8 \n" - "b 2f \n" - " nop \n" - -// dst_a or dst_b or dst_stride_a or dst_stride_b not word aligned - "11: \n" - "lw $t0, 0(%[src]) \n" // |B0|A0|b0|a0| - "lwx $t8, %[src_stride](%[src]) \n" // |B1|A1|b1|a1| - "addu $s5, %[dst_a], %[dst_stride_a] \n" - "lwx $t9, $t2(%[src]) \n" // |B2|A2|b2|a2| - "lwx $s0, $t3(%[src]) \n" // |B3|A3|b3|a3| - "addu $s6, %[dst_b], %[dst_stride_b] \n" - - "precrq.ph.w $s1, $t8, $t0 \n" // |B1|A1|B0|A0| - "precrq.ph.w $s2, $s0, $t9 \n" // |B3|A3|B2|A2| - "precr.qb.ph $s3, $s2, $s1 \n" // |A3|A2|A1|A0| - "precrq.qb.ph $s4, $s2, $s1 \n" // |B3|B2|B1|B0| - - "sll $t0, $t0, 16 \n" - "packrl.ph $s1, $t8, $t0 \n" // |b1|a1|b0|a0| - "sll $t9, $t9, 16 \n" - "packrl.ph $s2, $s0, $t9 \n" // |b3|a3|b2|a2| - - "swr $s3, 0($s5) \n" - "swl $s3, 3($s5) \n" - "swr $s4, 0($s6) \n" - "swl $s4, 3($s6) \n" - - "precr.qb.ph $s3, $s2, $s1 \n" // |a3|a2|a1|a0| - "precrq.qb.ph $s4, $s2, $s1 \n" // |b3|b2|b1|b0| - - "lwx $t0, $t4(%[src]) \n" // |B4|A4|b4|a4| - "lwx $t8, $t5(%[src]) \n" // |B5|A5|b5|a5| - "lwx $t9, $t6(%[src]) \n" // |B6|A6|b6|a6| - "lwx $s0, $t7(%[src]) \n" // |B7|A7|b7|a7| - "swr $s3, 0(%[dst_a]) \n" - "swl $s3, 3(%[dst_a]) \n" - "swr $s4, 0(%[dst_b]) \n" - "swl $s4, 3(%[dst_b]) \n" - - "precrq.ph.w $s1, $t8, $t0 \n" // |B5|A5|B4|A4| - "precrq.ph.w $s2, $s0, $t9 \n" // |B6|A6|B7|A7| - "precr.qb.ph $s3, $s2, $s1 \n" // |A7|A6|A5|A4| - "precrq.qb.ph $s4, $s2, $s1 \n" // |B7|B6|B5|B4| - - "sll $t0, $t0, 16 \n" - "packrl.ph $s1, $t8, $t0 \n" // |b5|a5|b4|a4| - "sll $t9, $t9, 16 \n" - "packrl.ph $s2, $s0, $t9 \n" // |b7|a7|b6|a6| - - "swr $s3, 4($s5) \n" - "swl $s3, 7($s5) \n" - "swr $s4, 4($s6) \n" - "swl $s4, 7($s6) \n" - - "precr.qb.ph $s3, $s2, $s1 \n" // |a7|a6|a5|a4| - "precrq.qb.ph $s4, $s2, $s1 \n" // |b7|b6|b5|b4| - - "addiu %[src], 4 \n" - "addiu $t1, -1 \n" - "sll $t0, %[dst_stride_a], 1 \n" - "sll $t8, %[dst_stride_b], 1 \n" - "swr $s3, 4(%[dst_a]) \n" - "swl $s3, 7(%[dst_a]) \n" - "swr $s4, 4(%[dst_b]) \n" - "swl $s4, 7(%[dst_b]) \n" - "addu %[dst_a], %[dst_a], $t0 \n" - "bnez $t1, 11b \n" - " addu %[dst_b], %[dst_b], $t8 \n" - - "2: \n" - ".set pop \n" - : [src] "+r" (src), - [dst_a] "+r" (dst_a), - [dst_b] "+r" (dst_b), - [width] "+r" (width), - [src_stride] "+r" (src_stride) - : [dst_stride_a] "r" (dst_stride_a), - [dst_stride_b] "r" (dst_stride_b) - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9", - "s0", "s1", "s2", "s3", - "s4", "s5", "s6" - ); -} - -#endif // defined(__mips_dsp) && (__mips_dsp_rev >= 2) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/rotate_neon.cc b/src/main/jni/libyuv/source/rotate_neon.cc deleted file mode 100644 index d354e11fa..000000000 --- a/src/main/jni/libyuv/source/rotate_neon.cc +++ /dev/null @@ -1,533 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_NEON) && defined(__ARM_NEON__) - -static uvec8 kVTbl4x4Transpose = - { 0, 4, 8, 12, 1, 5, 9, 13, 2, 6, 10, 14, 3, 7, 11, 15 }; - -void TransposeWx8_NEON(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width) { - const uint8* src_temp = NULL; - asm volatile ( - // loops are on blocks of 8. loop will stop when - // counter gets to or below 0. starting the counter - // at w-8 allow for this - "sub %5, #8 \n" - - // handle 8x8 blocks. this should be the majority of the plane - ".p2align 2 \n" - "1: \n" - "mov %0, %1 \n" - - MEMACCESS(0) - "vld1.8 {d0}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d1}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d2}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d3}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d4}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d5}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d6}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d7}, [%0] \n" - - "vtrn.8 d1, d0 \n" - "vtrn.8 d3, d2 \n" - "vtrn.8 d5, d4 \n" - "vtrn.8 d7, d6 \n" - - "vtrn.16 d1, d3 \n" - "vtrn.16 d0, d2 \n" - "vtrn.16 d5, d7 \n" - "vtrn.16 d4, d6 \n" - - "vtrn.32 d1, d5 \n" - "vtrn.32 d0, d4 \n" - "vtrn.32 d3, d7 \n" - "vtrn.32 d2, d6 \n" - - "vrev16.8 q0, q0 \n" - "vrev16.8 q1, q1 \n" - "vrev16.8 q2, q2 \n" - "vrev16.8 q3, q3 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.8 {d1}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d3}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d2}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d5}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d4}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d7}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d6}, [%0] \n" - - "add %1, #8 \n" // src += 8 - "add %3, %3, %4, lsl #3 \n" // dst += 8 * dst_stride - "subs %5, #8 \n" // w -= 8 - "bge 1b \n" - - // add 8 back to counter. if the result is 0 there are - // no residuals. - "adds %5, #8 \n" - "beq 4f \n" - - // some residual, so between 1 and 7 lines left to transpose - "cmp %5, #2 \n" - "blt 3f \n" - - "cmp %5, #4 \n" - "blt 2f \n" - - // 4x8 block - "mov %0, %1 \n" - MEMACCESS(0) - "vld1.32 {d0[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d0[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d1[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d1[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d2[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d2[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d3[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d3[1]}, [%0] \n" - - "mov %0, %3 \n" - - MEMACCESS(6) - "vld1.8 {q3}, [%6] \n" - - "vtbl.8 d4, {d0, d1}, d6 \n" - "vtbl.8 d5, {d0, d1}, d7 \n" - "vtbl.8 d0, {d2, d3}, d6 \n" - "vtbl.8 d1, {d2, d3}, d7 \n" - - // TODO(frkoenig): Rework shuffle above to - // write out with 4 instead of 8 writes. - MEMACCESS(0) - "vst1.32 {d4[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d4[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d5[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d5[1]}, [%0] \n" - - "add %0, %3, #4 \n" - MEMACCESS(0) - "vst1.32 {d0[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d0[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d1[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d1[1]}, [%0] \n" - - "add %1, #4 \n" // src += 4 - "add %3, %3, %4, lsl #2 \n" // dst += 4 * dst_stride - "subs %5, #4 \n" // w -= 4 - "beq 4f \n" - - // some residual, check to see if it includes a 2x8 block, - // or less - "cmp %5, #2 \n" - "blt 3f \n" - - // 2x8 block - "2: \n" - "mov %0, %1 \n" - MEMACCESS(0) - "vld1.16 {d0[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d0[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d0[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d0[3]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[3]}, [%0] \n" - - "vtrn.8 d0, d1 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.64 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.64 {d1}, [%0] \n" - - "add %1, #2 \n" // src += 2 - "add %3, %3, %4, lsl #1 \n" // dst += 2 * dst_stride - "subs %5, #2 \n" // w -= 2 - "beq 4f \n" - - // 1x8 block - "3: \n" - MEMACCESS(1) - "vld1.8 {d0[0]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[1]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[2]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[3]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[4]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[5]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[6]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[7]}, [%1] \n" - - MEMACCESS(3) - "vst1.64 {d0}, [%3] \n" - - "4: \n" - - : "+r"(src_temp), // %0 - "+r"(src), // %1 - "+r"(src_stride), // %2 - "+r"(dst), // %3 - "+r"(dst_stride), // %4 - "+r"(width) // %5 - : "r"(&kVTbl4x4Transpose) // %6 - : "memory", "cc", "q0", "q1", "q2", "q3" - ); -} - -static uvec8 kVTbl4x4TransposeDi = - { 0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15 }; - -void TransposeUVWx8_NEON(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width) { - const uint8* src_temp = NULL; - asm volatile ( - // loops are on blocks of 8. loop will stop when - // counter gets to or below 0. starting the counter - // at w-8 allow for this - "sub %7, #8 \n" - - // handle 8x8 blocks. this should be the majority of the plane - ".p2align 2 \n" - "1: \n" - "mov %0, %1 \n" - - MEMACCESS(0) - "vld2.8 {d0, d1}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d2, d3}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d4, d5}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d6, d7}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d16, d17}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d18, d19}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d20, d21}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d22, d23}, [%0] \n" - - "vtrn.8 q1, q0 \n" - "vtrn.8 q3, q2 \n" - "vtrn.8 q9, q8 \n" - "vtrn.8 q11, q10 \n" - - "vtrn.16 q1, q3 \n" - "vtrn.16 q0, q2 \n" - "vtrn.16 q9, q11 \n" - "vtrn.16 q8, q10 \n" - - "vtrn.32 q1, q9 \n" - "vtrn.32 q0, q8 \n" - "vtrn.32 q3, q11 \n" - "vtrn.32 q2, q10 \n" - - "vrev16.8 q0, q0 \n" - "vrev16.8 q1, q1 \n" - "vrev16.8 q2, q2 \n" - "vrev16.8 q3, q3 \n" - "vrev16.8 q8, q8 \n" - "vrev16.8 q9, q9 \n" - "vrev16.8 q10, q10 \n" - "vrev16.8 q11, q11 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.8 {d2}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d6}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d4}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d18}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d16}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d22}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d20}, [%0] \n" - - "mov %0, %5 \n" - - MEMACCESS(0) - "vst1.8 {d3}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d1}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d7}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d5}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d19}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d17}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d23}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d21}, [%0] \n" - - "add %1, #8*2 \n" // src += 8*2 - "add %3, %3, %4, lsl #3 \n" // dst_a += 8 * dst_stride_a - "add %5, %5, %6, lsl #3 \n" // dst_b += 8 * dst_stride_b - "subs %7, #8 \n" // w -= 8 - "bge 1b \n" - - // add 8 back to counter. if the result is 0 there are - // no residuals. - "adds %7, #8 \n" - "beq 4f \n" - - // some residual, so between 1 and 7 lines left to transpose - "cmp %7, #2 \n" - "blt 3f \n" - - "cmp %7, #4 \n" - "blt 2f \n" - - // TODO(frkoenig): Clean this up - // 4x8 block - "mov %0, %1 \n" - MEMACCESS(0) - "vld1.64 {d0}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d1}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d2}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d3}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d4}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d5}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d6}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d7}, [%0] \n" - - MEMACCESS(8) - "vld1.8 {q15}, [%8] \n" - - "vtrn.8 q0, q1 \n" - "vtrn.8 q2, q3 \n" - - "vtbl.8 d16, {d0, d1}, d30 \n" - "vtbl.8 d17, {d0, d1}, d31 \n" - "vtbl.8 d18, {d2, d3}, d30 \n" - "vtbl.8 d19, {d2, d3}, d31 \n" - "vtbl.8 d20, {d4, d5}, d30 \n" - "vtbl.8 d21, {d4, d5}, d31 \n" - "vtbl.8 d22, {d6, d7}, d30 \n" - "vtbl.8 d23, {d6, d7}, d31 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.32 {d16[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d16[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d17[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d17[1]}, [%0], %4 \n" - - "add %0, %3, #4 \n" - MEMACCESS(0) - "vst1.32 {d20[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d20[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d21[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d21[1]}, [%0] \n" - - "mov %0, %5 \n" - - MEMACCESS(0) - "vst1.32 {d18[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d18[1]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d19[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d19[1]}, [%0], %6 \n" - - "add %0, %5, #4 \n" - MEMACCESS(0) - "vst1.32 {d22[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d22[1]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d23[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d23[1]}, [%0] \n" - - "add %1, #4*2 \n" // src += 4 * 2 - "add %3, %3, %4, lsl #2 \n" // dst_a += 4 * dst_stride_a - "add %5, %5, %6, lsl #2 \n" // dst_b += 4 * dst_stride_b - "subs %7, #4 \n" // w -= 4 - "beq 4f \n" - - // some residual, check to see if it includes a 2x8 block, - // or less - "cmp %7, #2 \n" - "blt 3f \n" - - // 2x8 block - "2: \n" - "mov %0, %1 \n" - MEMACCESS(0) - "vld2.16 {d0[0], d2[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[0], d3[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d0[1], d2[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[1], d3[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d0[2], d2[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[2], d3[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d0[3], d2[3]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[3], d3[3]}, [%0] \n" - - "vtrn.8 d0, d1 \n" - "vtrn.8 d2, d3 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.64 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.64 {d2}, [%0] \n" - - "mov %0, %5 \n" - - MEMACCESS(0) - "vst1.64 {d1}, [%0], %6 \n" - MEMACCESS(0) - "vst1.64 {d3}, [%0] \n" - - "add %1, #2*2 \n" // src += 2 * 2 - "add %3, %3, %4, lsl #1 \n" // dst_a += 2 * dst_stride_a - "add %5, %5, %6, lsl #1 \n" // dst_b += 2 * dst_stride_b - "subs %7, #2 \n" // w -= 2 - "beq 4f \n" - - // 1x8 block - "3: \n" - MEMACCESS(1) - "vld2.8 {d0[0], d1[0]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[1], d1[1]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[2], d1[2]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[3], d1[3]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[4], d1[4]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[5], d1[5]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[6], d1[6]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[7], d1[7]}, [%1] \n" - - MEMACCESS(3) - "vst1.64 {d0}, [%3] \n" - MEMACCESS(5) - "vst1.64 {d1}, [%5] \n" - - "4: \n" - - : "+r"(src_temp), // %0 - "+r"(src), // %1 - "+r"(src_stride), // %2 - "+r"(dst_a), // %3 - "+r"(dst_stride_a), // %4 - "+r"(dst_b), // %5 - "+r"(dst_stride_b), // %6 - "+r"(width) // %7 - : "r"(&kVTbl4x4TransposeDi) // %8 - : "memory", "cc", - "q0", "q1", "q2", "q3", "q8", "q9", "q10", "q11" - ); -} -#endif - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/rotate_neon64.cc b/src/main/jni/libyuv/source/rotate_neon64.cc deleted file mode 100644 index b080a2c6a..000000000 --- a/src/main/jni/libyuv/source/rotate_neon64.cc +++ /dev/null @@ -1,540 +0,0 @@ -/* - * Copyright 2014 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#if !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__) -//this ifdef should be removed if TransposeWx8_NEON's aarch64 has -//been done -#ifdef HAS_TRANSPOSE_WX8_NEON -static uvec8 kVTbl4x4Transpose = - { 0, 4, 8, 12, 1, 5, 9, 13, 2, 6, 10, 14, 3, 7, 11, 15 }; - -void TransposeWx8_NEON(const uint8* src, int src_stride, - uint8* dst, int dst_stride, - int width) { - const uint8* src_temp = NULL; - asm volatile ( - // loops are on blocks of 8. loop will stop when - // counter gets to or below 0. starting the counter - // at w-8 allow for this - "sub %5, #8 \n" - - // handle 8x8 blocks. this should be the majority of the plane - ".p2align 2 \n" - "1: \n" - "mov %0, %1 \n" - - MEMACCESS(0) - "vld1.8 {d0}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d1}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d2}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d3}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d4}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d5}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d6}, [%0], %2 \n" - MEMACCESS(0) - "vld1.8 {d7}, [%0] \n" - - "vtrn.8 d1, d0 \n" - "vtrn.8 d3, d2 \n" - "vtrn.8 d5, d4 \n" - "vtrn.8 d7, d6 \n" - - "vtrn.16 d1, d3 \n" - "vtrn.16 d0, d2 \n" - "vtrn.16 d5, d7 \n" - "vtrn.16 d4, d6 \n" - - "vtrn.32 d1, d5 \n" - "vtrn.32 d0, d4 \n" - "vtrn.32 d3, d7 \n" - "vtrn.32 d2, d6 \n" - - "vrev16.8 q0, q0 \n" - "vrev16.8 q1, q1 \n" - "vrev16.8 q2, q2 \n" - "vrev16.8 q3, q3 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.8 {d1}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d3}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d2}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d5}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d4}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d7}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d6}, [%0] \n" - - "add %1, #8 \n" // src += 8 - "add %3, %3, %4, lsl #3 \n" // dst += 8 * dst_stride - "subs %5, #8 \n" // w -= 8 - "bge 1b \n" - - // add 8 back to counter. if the result is 0 there are - // no residuals. - "adds %5, #8 \n" - "beq 4f \n" - - // some residual, so between 1 and 7 lines left to transpose - "cmp %5, #2 \n" - "blt 3f \n" - - "cmp %5, #4 \n" - "blt 2f \n" - - // 4x8 block - "mov %0, %1 \n" - MEMACCESS(0) - "vld1.32 {d0[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d0[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d1[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d1[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d2[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d2[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d3[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.32 {d3[1]}, [%0] \n" - - "mov %0, %3 \n" - - MEMACCESS(6) - "vld1.8 {q3}, [%6] \n" - - "vtbl.8 d4, {d0, d1}, d6 \n" - "vtbl.8 d5, {d0, d1}, d7 \n" - "vtbl.8 d0, {d2, d3}, d6 \n" - "vtbl.8 d1, {d2, d3}, d7 \n" - - // TODO(frkoenig): Rework shuffle above to - // write out with 4 instead of 8 writes. - MEMACCESS(0) - "vst1.32 {d4[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d4[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d5[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d5[1]}, [%0] \n" - - "add %0, %3, #4 \n" - MEMACCESS(0) - "vst1.32 {d0[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d0[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d1[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d1[1]}, [%0] \n" - - "add %1, #4 \n" // src += 4 - "add %3, %3, %4, lsl #2 \n" // dst += 4 * dst_stride - "subs %5, #4 \n" // w -= 4 - "beq 4f \n" - - // some residual, check to see if it includes a 2x8 block, - // or less - "cmp %5, #2 \n" - "blt 3f \n" - - // 2x8 block - "2: \n" - "mov %0, %1 \n" - MEMACCESS(0) - "vld1.16 {d0[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d0[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d0[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d0[3]}, [%0], %2 \n" - MEMACCESS(0) - "vld1.16 {d1[3]}, [%0] \n" - - "vtrn.8 d0, d1 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.64 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.64 {d1}, [%0] \n" - - "add %1, #2 \n" // src += 2 - "add %3, %3, %4, lsl #1 \n" // dst += 2 * dst_stride - "subs %5, #2 \n" // w -= 2 - "beq 4f \n" - - // 1x8 block - "3: \n" - MEMACCESS(1) - "vld1.8 {d0[0]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[1]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[2]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[3]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[4]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[5]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[6]}, [%1], %2 \n" - MEMACCESS(1) - "vld1.8 {d0[7]}, [%1] \n" - - MEMACCESS(3) - "vst1.64 {d0}, [%3] \n" - - "4: \n" - - : "+r"(src_temp), // %0 - "+r"(src), // %1 - "+r"(src_stride), // %2 - "+r"(dst), // %3 - "+r"(dst_stride), // %4 - "+r"(width) // %5 - : "r"(&kVTbl4x4Transpose) // %6 - : "memory", "cc", "q0", "q1", "q2", "q3" - ); -} -#endif //HAS_TRANSPOSE_WX8_NEON - -//this ifdef should be removed if TransposeUVWx8_NEON's aarch64 has -//been done -#ifdef HAS_TRANSPOSE_UVWX8_NEON -static uvec8 kVTbl4x4TransposeDi = - { 0, 8, 1, 9, 2, 10, 3, 11, 4, 12, 5, 13, 6, 14, 7, 15 }; - -void TransposeUVWx8_NEON(const uint8* src, int src_stride, - uint8* dst_a, int dst_stride_a, - uint8* dst_b, int dst_stride_b, - int width) { - const uint8* src_temp = NULL; - asm volatile ( - // loops are on blocks of 8. loop will stop when - // counter gets to or below 0. starting the counter - // at w-8 allow for this - "sub %7, #8 \n" - - // handle 8x8 blocks. this should be the majority of the plane - ".p2align 2 \n" - "1: \n" - "mov %0, %1 \n" - - MEMACCESS(0) - "vld2.8 {d0, d1}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d2, d3}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d4, d5}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d6, d7}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d16, d17}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d18, d19}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d20, d21}, [%0], %2 \n" - MEMACCESS(0) - "vld2.8 {d22, d23}, [%0] \n" - - "vtrn.8 q1, q0 \n" - "vtrn.8 q3, q2 \n" - "vtrn.8 q9, q8 \n" - "vtrn.8 q11, q10 \n" - - "vtrn.16 q1, q3 \n" - "vtrn.16 q0, q2 \n" - "vtrn.16 q9, q11 \n" - "vtrn.16 q8, q10 \n" - - "vtrn.32 q1, q9 \n" - "vtrn.32 q0, q8 \n" - "vtrn.32 q3, q11 \n" - "vtrn.32 q2, q10 \n" - - "vrev16.8 q0, q0 \n" - "vrev16.8 q1, q1 \n" - "vrev16.8 q2, q2 \n" - "vrev16.8 q3, q3 \n" - "vrev16.8 q8, q8 \n" - "vrev16.8 q9, q9 \n" - "vrev16.8 q10, q10 \n" - "vrev16.8 q11, q11 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.8 {d2}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d6}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d4}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d18}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d16}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d22}, [%0], %4 \n" - MEMACCESS(0) - "vst1.8 {d20}, [%0] \n" - - "mov %0, %5 \n" - - MEMACCESS(0) - "vst1.8 {d3}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d1}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d7}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d5}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d19}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d17}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d23}, [%0], %6 \n" - MEMACCESS(0) - "vst1.8 {d21}, [%0] \n" - - "add %1, #8*2 \n" // src += 8*2 - "add %3, %3, %4, lsl #3 \n" // dst_a += 8 * dst_stride_a - "add %5, %5, %6, lsl #3 \n" // dst_b += 8 * dst_stride_b - "subs %7, #8 \n" // w -= 8 - "bge 1b \n" - - // add 8 back to counter. if the result is 0 there are - // no residuals. - "adds %7, #8 \n" - "beq 4f \n" - - // some residual, so between 1 and 7 lines left to transpose - "cmp %7, #2 \n" - "blt 3f \n" - - "cmp %7, #4 \n" - "blt 2f \n" - - // TODO(frkoenig): Clean this up - // 4x8 block - "mov %0, %1 \n" - MEMACCESS(0) - "vld1.64 {d0}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d1}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d2}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d3}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d4}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d5}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d6}, [%0], %2 \n" - MEMACCESS(0) - "vld1.64 {d7}, [%0] \n" - - MEMACCESS(8) - "vld1.8 {q15}, [%8] \n" - - "vtrn.8 q0, q1 \n" - "vtrn.8 q2, q3 \n" - - "vtbl.8 d16, {d0, d1}, d30 \n" - "vtbl.8 d17, {d0, d1}, d31 \n" - "vtbl.8 d18, {d2, d3}, d30 \n" - "vtbl.8 d19, {d2, d3}, d31 \n" - "vtbl.8 d20, {d4, d5}, d30 \n" - "vtbl.8 d21, {d4, d5}, d31 \n" - "vtbl.8 d22, {d6, d7}, d30 \n" - "vtbl.8 d23, {d6, d7}, d31 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.32 {d16[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d16[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d17[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d17[1]}, [%0], %4 \n" - - "add %0, %3, #4 \n" - MEMACCESS(0) - "vst1.32 {d20[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d20[1]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d21[0]}, [%0], %4 \n" - MEMACCESS(0) - "vst1.32 {d21[1]}, [%0] \n" - - "mov %0, %5 \n" - - MEMACCESS(0) - "vst1.32 {d18[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d18[1]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d19[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d19[1]}, [%0], %6 \n" - - "add %0, %5, #4 \n" - MEMACCESS(0) - "vst1.32 {d22[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d22[1]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d23[0]}, [%0], %6 \n" - MEMACCESS(0) - "vst1.32 {d23[1]}, [%0] \n" - - "add %1, #4*2 \n" // src += 4 * 2 - "add %3, %3, %4, lsl #2 \n" // dst_a += 4 * dst_stride_a - "add %5, %5, %6, lsl #2 \n" // dst_b += 4 * dst_stride_b - "subs %7, #4 \n" // w -= 4 - "beq 4f \n" - - // some residual, check to see if it includes a 2x8 block, - // or less - "cmp %7, #2 \n" - "blt 3f \n" - - // 2x8 block - "2: \n" - "mov %0, %1 \n" - MEMACCESS(0) - "vld2.16 {d0[0], d2[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[0], d3[0]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d0[1], d2[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[1], d3[1]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d0[2], d2[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[2], d3[2]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d0[3], d2[3]}, [%0], %2 \n" - MEMACCESS(0) - "vld2.16 {d1[3], d3[3]}, [%0] \n" - - "vtrn.8 d0, d1 \n" - "vtrn.8 d2, d3 \n" - - "mov %0, %3 \n" - - MEMACCESS(0) - "vst1.64 {d0}, [%0], %4 \n" - MEMACCESS(0) - "vst1.64 {d2}, [%0] \n" - - "mov %0, %5 \n" - - MEMACCESS(0) - "vst1.64 {d1}, [%0], %6 \n" - MEMACCESS(0) - "vst1.64 {d3}, [%0] \n" - - "add %1, #2*2 \n" // src += 2 * 2 - "add %3, %3, %4, lsl #1 \n" // dst_a += 2 * dst_stride_a - "add %5, %5, %6, lsl #1 \n" // dst_b += 2 * dst_stride_b - "subs %7, #2 \n" // w -= 2 - "beq 4f \n" - - // 1x8 block - "3: \n" - MEMACCESS(1) - "vld2.8 {d0[0], d1[0]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[1], d1[1]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[2], d1[2]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[3], d1[3]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[4], d1[4]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[5], d1[5]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[6], d1[6]}, [%1], %2 \n" - MEMACCESS(1) - "vld2.8 {d0[7], d1[7]}, [%1] \n" - - MEMACCESS(3) - "vst1.64 {d0}, [%3] \n" - MEMACCESS(5) - "vst1.64 {d1}, [%5] \n" - - "4: \n" - - : "+r"(src_temp), // %0 - "+r"(src), // %1 - "+r"(src_stride), // %2 - "+r"(dst_a), // %3 - "+r"(dst_stride_a), // %4 - "+r"(dst_b), // %5 - "+r"(dst_stride_b), // %6 - "+r"(width) // %7 - : "r"(&kVTbl4x4TransposeDi) // %8 - : "memory", "cc", - "q0", "q1", "q2", "q3", "q8", "q9", "q10", "q11" - ); -} -#endif // HAS_TRANSPOSE_UVWX8_NEON -#endif // __aarch64__ - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_any.cc b/src/main/jni/libyuv/source/row_any.cc deleted file mode 100644 index aaa0378d7..000000000 --- a/src/main/jni/libyuv/source/row_any.cc +++ /dev/null @@ -1,602 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// TODO(fbarchard): Consider 'any' functions handling any quantity of pixels. -// TODO(fbarchard): Consider 'any' functions handling odd alignment. -// YUV to RGB does multiple of 8 with SIMD and remainder with C. -#define YANY(NAMEANY, I420TORGB_SIMD, I420TORGB_C, UV_SHIFT, BPP, MASK) \ - void NAMEANY(const uint8* y_buf, \ - const uint8* u_buf, \ - const uint8* v_buf, \ - uint8* rgb_buf, \ - int width) { \ - int n = width & ~MASK; \ - I420TORGB_SIMD(y_buf, u_buf, v_buf, rgb_buf, n); \ - I420TORGB_C(y_buf + n, \ - u_buf + (n >> UV_SHIFT), \ - v_buf + (n >> UV_SHIFT), \ - rgb_buf + n * BPP, width & MASK); \ - } - -#ifdef HAS_I422TOARGBROW_SSSE3 -YANY(I422ToARGBRow_Any_SSSE3, I422ToARGBRow_Unaligned_SSSE3, I422ToARGBRow_C, - 1, 4, 7) -#endif // HAS_I422TOARGBROW_SSSE3 -#ifdef HAS_I444TOARGBROW_SSSE3 -YANY(I444ToARGBRow_Any_SSSE3, I444ToARGBRow_Unaligned_SSSE3, I444ToARGBRow_C, - 0, 4, 7) -YANY(I411ToARGBRow_Any_SSSE3, I411ToARGBRow_Unaligned_SSSE3, I411ToARGBRow_C, - 2, 4, 7) -YANY(I422ToBGRARow_Any_SSSE3, I422ToBGRARow_Unaligned_SSSE3, I422ToBGRARow_C, - 1, 4, 7) -YANY(I422ToABGRRow_Any_SSSE3, I422ToABGRRow_Unaligned_SSSE3, I422ToABGRRow_C, - 1, 4, 7) -YANY(I422ToRGBARow_Any_SSSE3, I422ToRGBARow_Unaligned_SSSE3, I422ToRGBARow_C, - 1, 4, 7) -// I422ToRGB565Row_SSSE3 is unaligned. -YANY(I422ToARGB4444Row_Any_SSSE3, I422ToARGB4444Row_SSSE3, I422ToARGB4444Row_C, - 1, 2, 7) -YANY(I422ToARGB1555Row_Any_SSSE3, I422ToARGB1555Row_SSSE3, I422ToARGB1555Row_C, - 1, 2, 7) -YANY(I422ToRGB565Row_Any_SSSE3, I422ToRGB565Row_SSSE3, I422ToRGB565Row_C, - 1, 2, 7) -// I422ToRGB24Row_SSSE3 is unaligned. -YANY(I422ToRGB24Row_Any_SSSE3, I422ToRGB24Row_SSSE3, I422ToRGB24Row_C, 1, 3, 7) -YANY(I422ToRAWRow_Any_SSSE3, I422ToRAWRow_SSSE3, I422ToRAWRow_C, 1, 3, 7) -YANY(I422ToYUY2Row_Any_SSE2, I422ToYUY2Row_SSE2, I422ToYUY2Row_C, 1, 2, 15) -YANY(I422ToUYVYRow_Any_SSE2, I422ToUYVYRow_SSE2, I422ToUYVYRow_C, 1, 2, 15) -#endif // HAS_I444TOARGBROW_SSSE3 -#ifdef HAS_I422TOARGBROW_AVX2 -YANY(I422ToARGBRow_Any_AVX2, I422ToARGBRow_AVX2, I422ToARGBRow_C, 1, 4, 15) -#endif // HAS_I422TOARGBROW_AVX2 -#ifdef HAS_I422TOARGBROW_NEON -YANY(I444ToARGBRow_Any_NEON, I444ToARGBRow_NEON, I444ToARGBRow_C, 0, 4, 7) -YANY(I422ToARGBRow_Any_NEON, I422ToARGBRow_NEON, I422ToARGBRow_C, 1, 4, 7) -YANY(I411ToARGBRow_Any_NEON, I411ToARGBRow_NEON, I411ToARGBRow_C, 2, 4, 7) -YANY(I422ToBGRARow_Any_NEON, I422ToBGRARow_NEON, I422ToBGRARow_C, 1, 4, 7) -YANY(I422ToABGRRow_Any_NEON, I422ToABGRRow_NEON, I422ToABGRRow_C, 1, 4, 7) -YANY(I422ToRGBARow_Any_NEON, I422ToRGBARow_NEON, I422ToRGBARow_C, 1, 4, 7) -YANY(I422ToRGB24Row_Any_NEON, I422ToRGB24Row_NEON, I422ToRGB24Row_C, 1, 3, 7) -YANY(I422ToRAWRow_Any_NEON, I422ToRAWRow_NEON, I422ToRAWRow_C, 1, 3, 7) -YANY(I422ToARGB4444Row_Any_NEON, I422ToARGB4444Row_NEON, I422ToARGB4444Row_C, - 1, 2, 7) -YANY(I422ToARGB1555Row_Any_NEON, I422ToARGB1555Row_NEON, I422ToARGB1555Row_C, - 1, 2, 7) -YANY(I422ToRGB565Row_Any_NEON, I422ToRGB565Row_NEON, I422ToRGB565Row_C, 1, 2, 7) -#endif // HAS_I422TOARGBROW_NEON -#ifdef HAS_I422TOYUY2ROW_NEON -YANY(I422ToYUY2Row_Any_NEON, I422ToYUY2Row_NEON, I422ToYUY2Row_C, 1, 2, 15) -#endif // HAS_I422TOYUY2ROW_NEON -#ifdef HAS_I422TOUYVYROW_NEON -YANY(I422ToUYVYRow_Any_NEON, I422ToUYVYRow_NEON, I422ToUYVYRow_C, 1, 2, 15) -#endif // HAS_I422TOUYVYROW_NEON -#undef YANY - -// Wrappers to handle odd width -#define NV2NY(NAMEANY, NV12TORGB_SIMD, NV12TORGB_C, UV_SHIFT, BPP) \ - void NAMEANY(const uint8* y_buf, \ - const uint8* uv_buf, \ - uint8* rgb_buf, \ - int width) { \ - int n = width & ~7; \ - NV12TORGB_SIMD(y_buf, uv_buf, rgb_buf, n); \ - NV12TORGB_C(y_buf + n, \ - uv_buf + (n >> UV_SHIFT), \ - rgb_buf + n * BPP, width & 7); \ - } - -#ifdef HAS_NV12TOARGBROW_SSSE3 -NV2NY(NV12ToARGBRow_Any_SSSE3, NV12ToARGBRow_Unaligned_SSSE3, NV12ToARGBRow_C, - 0, 4) -NV2NY(NV21ToARGBRow_Any_SSSE3, NV21ToARGBRow_Unaligned_SSSE3, NV21ToARGBRow_C, - 0, 4) -#endif // HAS_NV12TOARGBROW_SSSE3 -#ifdef HAS_NV12TOARGBROW_NEON -NV2NY(NV12ToARGBRow_Any_NEON, NV12ToARGBRow_NEON, NV12ToARGBRow_C, 0, 4) -NV2NY(NV21ToARGBRow_Any_NEON, NV21ToARGBRow_NEON, NV21ToARGBRow_C, 0, 4) -#endif // HAS_NV12TOARGBROW_NEON -#ifdef HAS_NV12TORGB565ROW_SSSE3 -NV2NY(NV12ToRGB565Row_Any_SSSE3, NV12ToRGB565Row_SSSE3, NV12ToRGB565Row_C, - 0, 2) -NV2NY(NV21ToRGB565Row_Any_SSSE3, NV21ToRGB565Row_SSSE3, NV21ToRGB565Row_C, - 0, 2) -#endif // HAS_NV12TORGB565ROW_SSSE3 -#ifdef HAS_NV12TORGB565ROW_NEON -NV2NY(NV12ToRGB565Row_Any_NEON, NV12ToRGB565Row_NEON, NV12ToRGB565Row_C, 0, 2) -NV2NY(NV21ToRGB565Row_Any_NEON, NV21ToRGB565Row_NEON, NV21ToRGB565Row_C, 0, 2) -#endif // HAS_NV12TORGB565ROW_NEON -#undef NVANY - -#define RGBANY(NAMEANY, ARGBTORGB_SIMD, ARGBTORGB_C, MASK, SBPP, BPP) \ - void NAMEANY(const uint8* src, \ - uint8* dst, \ - int width) { \ - int n = width & ~MASK; \ - ARGBTORGB_SIMD(src, dst, n); \ - ARGBTORGB_C(src + n * SBPP, dst + n * BPP, width & MASK); \ - } - -#if defined(HAS_ARGBTORGB24ROW_SSSE3) -RGBANY(ARGBToRGB24Row_Any_SSSE3, ARGBToRGB24Row_SSSE3, ARGBToRGB24Row_C, - 15, 4, 3) -RGBANY(ARGBToRAWRow_Any_SSSE3, ARGBToRAWRow_SSSE3, ARGBToRAWRow_C, - 15, 4, 3) -RGBANY(ARGBToRGB565Row_Any_SSE2, ARGBToRGB565Row_SSE2, ARGBToRGB565Row_C, - 3, 4, 2) -RGBANY(ARGBToARGB1555Row_Any_SSE2, ARGBToARGB1555Row_SSE2, ARGBToARGB1555Row_C, - 3, 4, 2) -RGBANY(ARGBToARGB4444Row_Any_SSE2, ARGBToARGB4444Row_SSE2, ARGBToARGB4444Row_C, - 3, 4, 2) -#endif -#if defined(HAS_I400TOARGBROW_SSE2) -RGBANY(I400ToARGBRow_Any_SSE2, I400ToARGBRow_Unaligned_SSE2, I400ToARGBRow_C, - 7, 1, 4) -#endif -#if defined(HAS_YTOARGBROW_SSE2) -RGBANY(YToARGBRow_Any_SSE2, YToARGBRow_SSE2, YToARGBRow_C, - 7, 1, 4) -RGBANY(YUY2ToARGBRow_Any_SSSE3, YUY2ToARGBRow_Unaligned_SSSE3, YUY2ToARGBRow_C, - 15, 2, 4) -RGBANY(UYVYToARGBRow_Any_SSSE3, UYVYToARGBRow_Unaligned_SSSE3, UYVYToARGBRow_C, - 15, 2, 4) -// These require alignment on ARGB, so C is used for remainder. -RGBANY(RGB24ToARGBRow_Any_SSSE3, RGB24ToARGBRow_SSSE3, RGB24ToARGBRow_C, - 15, 3, 4) -RGBANY(RAWToARGBRow_Any_SSSE3, RAWToARGBRow_SSSE3, RAWToARGBRow_C, - 15, 3, 4) -RGBANY(RGB565ToARGBRow_Any_SSE2, RGB565ToARGBRow_SSE2, RGB565ToARGBRow_C, - 7, 2, 4) -RGBANY(ARGB1555ToARGBRow_Any_SSE2, ARGB1555ToARGBRow_SSE2, ARGB1555ToARGBRow_C, - 7, 2, 4) -RGBANY(ARGB4444ToARGBRow_Any_SSE2, ARGB4444ToARGBRow_SSE2, ARGB4444ToARGBRow_C, - 7, 2, 4) -#endif -#if defined(HAS_ARGBTORGB24ROW_NEON) -RGBANY(ARGBToRGB24Row_Any_NEON, ARGBToRGB24Row_NEON, ARGBToRGB24Row_C, 7, 4, 3) -RGBANY(ARGBToRAWRow_Any_NEON, ARGBToRAWRow_NEON, ARGBToRAWRow_C, 7, 4, 3) -RGBANY(ARGBToRGB565Row_Any_NEON, ARGBToRGB565Row_NEON, ARGBToRGB565Row_C, - 7, 4, 2) -RGBANY(ARGBToARGB1555Row_Any_NEON, ARGBToARGB1555Row_NEON, ARGBToARGB1555Row_C, - 7, 4, 2) -RGBANY(ARGBToARGB4444Row_Any_NEON, ARGBToARGB4444Row_NEON, ARGBToARGB4444Row_C, - 7, 4, 2) -RGBANY(I400ToARGBRow_Any_NEON, I400ToARGBRow_NEON, I400ToARGBRow_C, - 7, 1, 4) -RGBANY(YToARGBRow_Any_NEON, YToARGBRow_NEON, YToARGBRow_C, - 7, 1, 4) -RGBANY(YUY2ToARGBRow_Any_NEON, YUY2ToARGBRow_NEON, YUY2ToARGBRow_C, - 7, 2, 4) -RGBANY(UYVYToARGBRow_Any_NEON, UYVYToARGBRow_NEON, UYVYToARGBRow_C, - 7, 2, 4) -#endif -#undef RGBANY - -// ARGB to Bayer does multiple of 4 pixels, SSSE3 aligned src, unaligned dst. -#define BAYERANY(NAMEANY, ARGBTORGB_SIMD, ARGBTORGB_C, MASK, SBPP, BPP) \ - void NAMEANY(const uint8* src, \ - uint8* dst, uint32 selector, \ - int width) { \ - int n = width & ~MASK; \ - ARGBTORGB_SIMD(src, dst, selector, n); \ - ARGBTORGB_C(src + n * SBPP, dst + n * BPP, selector, width & MASK); \ - } - -#if defined(HAS_ARGBTOBAYERROW_SSSE3) -BAYERANY(ARGBToBayerRow_Any_SSSE3, ARGBToBayerRow_SSSE3, ARGBToBayerRow_C, - 7, 4, 1) -#endif -#if defined(HAS_ARGBTOBAYERROW_NEON) -BAYERANY(ARGBToBayerRow_Any_NEON, ARGBToBayerRow_NEON, ARGBToBayerRow_C, - 7, 4, 1) -#endif -#if defined(HAS_ARGBTOBAYERGGROW_SSE2) -BAYERANY(ARGBToBayerGGRow_Any_SSE2, ARGBToBayerGGRow_SSE2, ARGBToBayerGGRow_C, - 7, 4, 1) -#endif -#if defined(HAS_ARGBTOBAYERGGROW_NEON) -BAYERANY(ARGBToBayerGGRow_Any_NEON, ARGBToBayerGGRow_NEON, ARGBToBayerGGRow_C, - 7, 4, 1) -#endif - -#undef BAYERANY - -// RGB/YUV to Y does multiple of 16 with SIMD and last 16 with SIMD. -#define YANY(NAMEANY, ARGBTOY_SIMD, SBPP, BPP, NUM) \ - void NAMEANY(const uint8* src_argb, uint8* dst_y, int width) { \ - ARGBTOY_SIMD(src_argb, dst_y, width - NUM); \ - ARGBTOY_SIMD(src_argb + (width - NUM) * SBPP, \ - dst_y + (width - NUM) * BPP, NUM); \ - } - -#ifdef HAS_ARGBTOYROW_AVX2 -YANY(ARGBToYRow_Any_AVX2, ARGBToYRow_AVX2, 4, 1, 32) -YANY(ARGBToYJRow_Any_AVX2, ARGBToYJRow_AVX2, 4, 1, 32) -YANY(YUY2ToYRow_Any_AVX2, YUY2ToYRow_AVX2, 2, 1, 32) -YANY(UYVYToYRow_Any_AVX2, UYVYToYRow_AVX2, 2, 1, 32) -#endif -#ifdef HAS_ARGBTOYROW_SSSE3 -YANY(ARGBToYRow_Any_SSSE3, ARGBToYRow_Unaligned_SSSE3, 4, 1, 16) -#endif -#ifdef HAS_BGRATOYROW_SSSE3 -YANY(BGRAToYRow_Any_SSSE3, BGRAToYRow_Unaligned_SSSE3, 4, 1, 16) -YANY(ABGRToYRow_Any_SSSE3, ABGRToYRow_Unaligned_SSSE3, 4, 1, 16) -YANY(RGBAToYRow_Any_SSSE3, RGBAToYRow_Unaligned_SSSE3, 4, 1, 16) -YANY(YUY2ToYRow_Any_SSE2, YUY2ToYRow_Unaligned_SSE2, 2, 1, 16) -YANY(UYVYToYRow_Any_SSE2, UYVYToYRow_Unaligned_SSE2, 2, 1, 16) -#endif -#ifdef HAS_ARGBTOYJROW_SSSE3 -YANY(ARGBToYJRow_Any_SSSE3, ARGBToYJRow_Unaligned_SSSE3, 4, 1, 16) -#endif -#ifdef HAS_ARGBTOYROW_NEON -YANY(ARGBToYRow_Any_NEON, ARGBToYRow_NEON, 4, 1, 8) -#endif -#ifdef HAS_ARGBTOYJROW_NEON -YANY(ARGBToYJRow_Any_NEON, ARGBToYJRow_NEON, 4, 1, 8) -#endif -#ifdef HAS_BGRATOYROW_NEON -YANY(BGRAToYRow_Any_NEON, BGRAToYRow_NEON, 4, 1, 8) -#endif -#ifdef HAS_ABGRTOYROW_NEON -YANY(ABGRToYRow_Any_NEON, ABGRToYRow_NEON, 4, 1, 8) -#endif -#ifdef HAS_RGBATOYROW_NEON -YANY(RGBAToYRow_Any_NEON, RGBAToYRow_NEON, 4, 1, 8) -#endif -#ifdef HAS_RGB24TOYROW_NEON -YANY(RGB24ToYRow_Any_NEON, RGB24ToYRow_NEON, 3, 1, 8) -#endif -#ifdef HAS_RAWTOYROW_NEON -YANY(RAWToYRow_Any_NEON, RAWToYRow_NEON, 3, 1, 8) -#endif -#ifdef HAS_RGB565TOYROW_NEON -YANY(RGB565ToYRow_Any_NEON, RGB565ToYRow_NEON, 2, 1, 8) -#endif -#ifdef HAS_ARGB1555TOYROW_NEON -YANY(ARGB1555ToYRow_Any_NEON, ARGB1555ToYRow_NEON, 2, 1, 8) -#endif -#ifdef HAS_ARGB4444TOYROW_NEON -YANY(ARGB4444ToYRow_Any_NEON, ARGB4444ToYRow_NEON, 2, 1, 8) -#endif -#ifdef HAS_YUY2TOYROW_NEON -YANY(YUY2ToYRow_Any_NEON, YUY2ToYRow_NEON, 2, 1, 16) -#endif -#ifdef HAS_UYVYTOYROW_NEON -YANY(UYVYToYRow_Any_NEON, UYVYToYRow_NEON, 2, 1, 16) -#endif -#ifdef HAS_RGB24TOARGBROW_NEON -YANY(RGB24ToARGBRow_Any_NEON, RGB24ToARGBRow_NEON, 3, 4, 8) -#endif -#ifdef HAS_RAWTOARGBROW_NEON -YANY(RAWToARGBRow_Any_NEON, RAWToARGBRow_NEON, 3, 4, 8) -#endif -#ifdef HAS_RGB565TOARGBROW_NEON -YANY(RGB565ToARGBRow_Any_NEON, RGB565ToARGBRow_NEON, 2, 4, 8) -#endif -#ifdef HAS_ARGB1555TOARGBROW_NEON -YANY(ARGB1555ToARGBRow_Any_NEON, ARGB1555ToARGBRow_NEON, 2, 4, 8) -#endif -#ifdef HAS_ARGB4444TOARGBROW_NEON -YANY(ARGB4444ToARGBRow_Any_NEON, ARGB4444ToARGBRow_NEON, 2, 4, 8) -#endif -#undef YANY - -#define YANY(NAMEANY, ARGBTOY_SIMD, ARGBTOY_C, SBPP, BPP, MASK) \ - void NAMEANY(const uint8* src_argb, uint8* dst_y, int width) { \ - int n = width & ~MASK; \ - ARGBTOY_SIMD(src_argb, dst_y, n); \ - ARGBTOY_C(src_argb + n * SBPP, \ - dst_y + n * BPP, width & MASK); \ - } - -// Attenuate is destructive so last16 method can not be used due to overlap. -#ifdef HAS_ARGBATTENUATEROW_SSSE3 -YANY(ARGBAttenuateRow_Any_SSSE3, ARGBAttenuateRow_SSSE3, ARGBAttenuateRow_C, - 4, 4, 3) -#endif -#ifdef HAS_ARGBATTENUATEROW_SSE2 -YANY(ARGBAttenuateRow_Any_SSE2, ARGBAttenuateRow_SSE2, ARGBAttenuateRow_C, - 4, 4, 3) -#endif -#ifdef HAS_ARGBUNATTENUATEROW_SSE2 -YANY(ARGBUnattenuateRow_Any_SSE2, ARGBUnattenuateRow_SSE2, ARGBUnattenuateRow_C, - 4, 4, 3) -#endif -#ifdef HAS_ARGBATTENUATEROW_AVX2 -YANY(ARGBAttenuateRow_Any_AVX2, ARGBAttenuateRow_AVX2, ARGBAttenuateRow_C, - 4, 4, 7) -#endif -#ifdef HAS_ARGBUNATTENUATEROW_AVX2 -YANY(ARGBUnattenuateRow_Any_AVX2, ARGBUnattenuateRow_AVX2, ARGBUnattenuateRow_C, - 4, 4, 7) -#endif -#ifdef HAS_ARGBATTENUATEROW_NEON -YANY(ARGBAttenuateRow_Any_NEON, ARGBAttenuateRow_NEON, ARGBAttenuateRow_C, - 4, 4, 7) -#endif -#undef YANY - -// RGB/YUV to UV does multiple of 16 with SIMD and remainder with C. -#define UVANY(NAMEANY, ANYTOUV_SIMD, ANYTOUV_C, BPP, MASK) \ - void NAMEANY(const uint8* src_argb, int src_stride_argb, \ - uint8* dst_u, uint8* dst_v, int width) { \ - int n = width & ~MASK; \ - ANYTOUV_SIMD(src_argb, src_stride_argb, dst_u, dst_v, n); \ - ANYTOUV_C(src_argb + n * BPP, src_stride_argb, \ - dst_u + (n >> 1), \ - dst_v + (n >> 1), \ - width & MASK); \ - } - -#ifdef HAS_ARGBTOUVROW_AVX2 -UVANY(ARGBToUVRow_Any_AVX2, ARGBToUVRow_AVX2, ARGBToUVRow_C, 4, 31) -UVANY(YUY2ToUVRow_Any_AVX2, YUY2ToUVRow_AVX2, YUY2ToUVRow_C, 2, 31) -UVANY(UYVYToUVRow_Any_AVX2, UYVYToUVRow_AVX2, UYVYToUVRow_C, 2, 31) -#endif -#ifdef HAS_ARGBTOUVROW_SSSE3 -UVANY(ARGBToUVRow_Any_SSSE3, ARGBToUVRow_Unaligned_SSSE3, ARGBToUVRow_C, 4, 15) -UVANY(ARGBToUVJRow_Any_SSSE3, ARGBToUVJRow_Unaligned_SSSE3, ARGBToUVJRow_C, - 4, 15) -UVANY(BGRAToUVRow_Any_SSSE3, BGRAToUVRow_Unaligned_SSSE3, BGRAToUVRow_C, 4, 15) -UVANY(ABGRToUVRow_Any_SSSE3, ABGRToUVRow_Unaligned_SSSE3, ABGRToUVRow_C, 4, 15) -UVANY(RGBAToUVRow_Any_SSSE3, RGBAToUVRow_Unaligned_SSSE3, RGBAToUVRow_C, 4, 15) -UVANY(YUY2ToUVRow_Any_SSE2, YUY2ToUVRow_Unaligned_SSE2, YUY2ToUVRow_C, 2, 15) -UVANY(UYVYToUVRow_Any_SSE2, UYVYToUVRow_Unaligned_SSE2, UYVYToUVRow_C, 2, 15) -#endif -#ifdef HAS_ARGBTOUVROW_NEON -UVANY(ARGBToUVRow_Any_NEON, ARGBToUVRow_NEON, ARGBToUVRow_C, 4, 15) -#endif -#ifdef HAS_ARGBTOUVJROW_NEON -UVANY(ARGBToUVJRow_Any_NEON, ARGBToUVJRow_NEON, ARGBToUVJRow_C, 4, 15) -#endif -#ifdef HAS_BGRATOUVROW_NEON -UVANY(BGRAToUVRow_Any_NEON, BGRAToUVRow_NEON, BGRAToUVRow_C, 4, 15) -#endif -#ifdef HAS_ABGRTOUVROW_NEON -UVANY(ABGRToUVRow_Any_NEON, ABGRToUVRow_NEON, ABGRToUVRow_C, 4, 15) -#endif -#ifdef HAS_RGBATOUVROW_NEON -UVANY(RGBAToUVRow_Any_NEON, RGBAToUVRow_NEON, RGBAToUVRow_C, 4, 15) -#endif -#ifdef HAS_RGB24TOUVROW_NEON -UVANY(RGB24ToUVRow_Any_NEON, RGB24ToUVRow_NEON, RGB24ToUVRow_C, 3, 15) -#endif -#ifdef HAS_RAWTOUVROW_NEON -UVANY(RAWToUVRow_Any_NEON, RAWToUVRow_NEON, RAWToUVRow_C, 3, 15) -#endif -#ifdef HAS_RGB565TOUVROW_NEON -UVANY(RGB565ToUVRow_Any_NEON, RGB565ToUVRow_NEON, RGB565ToUVRow_C, 2, 15) -#endif -#ifdef HAS_ARGB1555TOUVROW_NEON -UVANY(ARGB1555ToUVRow_Any_NEON, ARGB1555ToUVRow_NEON, ARGB1555ToUVRow_C, 2, 15) -#endif -#ifdef HAS_ARGB4444TOUVROW_NEON -UVANY(ARGB4444ToUVRow_Any_NEON, ARGB4444ToUVRow_NEON, ARGB4444ToUVRow_C, 2, 15) -#endif -#ifdef HAS_YUY2TOUVROW_NEON -UVANY(YUY2ToUVRow_Any_NEON, YUY2ToUVRow_NEON, YUY2ToUVRow_C, 2, 15) -#endif -#ifdef HAS_UYVYTOUVROW_NEON -UVANY(UYVYToUVRow_Any_NEON, UYVYToUVRow_NEON, UYVYToUVRow_C, 2, 15) -#endif -#undef UVANY - -#define UV422ANY(NAMEANY, ANYTOUV_SIMD, ANYTOUV_C, BPP, MASK, SHIFT) \ - void NAMEANY(const uint8* src_uv, \ - uint8* dst_u, uint8* dst_v, int width) { \ - int n = width & ~MASK; \ - ANYTOUV_SIMD(src_uv, dst_u, dst_v, n); \ - ANYTOUV_C(src_uv + n * BPP, \ - dst_u + (n >> SHIFT), \ - dst_v + (n >> SHIFT), \ - width & MASK); \ - } - -#ifdef HAS_ARGBTOUV444ROW_SSSE3 -UV422ANY(ARGBToUV444Row_Any_SSSE3, ARGBToUV444Row_Unaligned_SSSE3, - ARGBToUV444Row_C, 4, 15, 0) -#endif -#ifdef HAS_YUY2TOUV422ROW_AVX2 -UV422ANY(YUY2ToUV422Row_Any_AVX2, YUY2ToUV422Row_AVX2, - YUY2ToUV422Row_C, 2, 31, 1) -UV422ANY(UYVYToUV422Row_Any_AVX2, UYVYToUV422Row_AVX2, - UYVYToUV422Row_C, 2, 31, 1) -#endif -#ifdef HAS_ARGBTOUVROW_SSSE3 -UV422ANY(ARGBToUV422Row_Any_SSSE3, ARGBToUV422Row_Unaligned_SSSE3, - ARGBToUV422Row_C, 4, 15, 1) -UV422ANY(YUY2ToUV422Row_Any_SSE2, YUY2ToUV422Row_Unaligned_SSE2, - YUY2ToUV422Row_C, 2, 15, 1) -UV422ANY(UYVYToUV422Row_Any_SSE2, UYVYToUV422Row_Unaligned_SSE2, - UYVYToUV422Row_C, 2, 15, 1) -#endif -#ifdef HAS_YUY2TOUV422ROW_NEON -UV422ANY(ARGBToUV444Row_Any_NEON, ARGBToUV444Row_NEON, - ARGBToUV444Row_C, 4, 7, 0) -UV422ANY(ARGBToUV422Row_Any_NEON, ARGBToUV422Row_NEON, - ARGBToUV422Row_C, 4, 15, 1) -UV422ANY(ARGBToUV411Row_Any_NEON, ARGBToUV411Row_NEON, - ARGBToUV411Row_C, 4, 31, 2) -UV422ANY(YUY2ToUV422Row_Any_NEON, YUY2ToUV422Row_NEON, - YUY2ToUV422Row_C, 2, 15, 1) -UV422ANY(UYVYToUV422Row_Any_NEON, UYVYToUV422Row_NEON, - UYVYToUV422Row_C, 2, 15, 1) -#endif -#undef UV422ANY - -#define SPLITUVROWANY(NAMEANY, ANYTOUV_SIMD, ANYTOUV_C, MASK) \ - void NAMEANY(const uint8* src_uv, \ - uint8* dst_u, uint8* dst_v, int width) { \ - int n = width & ~MASK; \ - ANYTOUV_SIMD(src_uv, dst_u, dst_v, n); \ - ANYTOUV_C(src_uv + n * 2, \ - dst_u + n, \ - dst_v + n, \ - width & MASK); \ - } - -#ifdef HAS_SPLITUVROW_SSE2 -SPLITUVROWANY(SplitUVRow_Any_SSE2, SplitUVRow_Unaligned_SSE2, SplitUVRow_C, 15) -#endif -#ifdef HAS_SPLITUVROW_AVX2 -SPLITUVROWANY(SplitUVRow_Any_AVX2, SplitUVRow_AVX2, SplitUVRow_C, 31) -#endif -#ifdef HAS_SPLITUVROW_NEON -SPLITUVROWANY(SplitUVRow_Any_NEON, SplitUVRow_NEON, SplitUVRow_C, 15) -#endif -#ifdef HAS_SPLITUVROW_MIPS_DSPR2 -SPLITUVROWANY(SplitUVRow_Any_MIPS_DSPR2, SplitUVRow_Unaligned_MIPS_DSPR2, - SplitUVRow_C, 15) -#endif -#undef SPLITUVROWANY - -#define MERGEUVROW_ANY(NAMEANY, ANYTOUV_SIMD, ANYTOUV_C, MASK) \ - void NAMEANY(const uint8* src_u, const uint8* src_v, \ - uint8* dst_uv, int width) { \ - int n = width & ~MASK; \ - ANYTOUV_SIMD(src_u, src_v, dst_uv, n); \ - ANYTOUV_C(src_u + n, \ - src_v + n, \ - dst_uv + n * 2, \ - width & MASK); \ - } - -#ifdef HAS_MERGEUVROW_SSE2 -MERGEUVROW_ANY(MergeUVRow_Any_SSE2, MergeUVRow_Unaligned_SSE2, MergeUVRow_C, 15) -#endif -#ifdef HAS_MERGEUVROW_AVX2 -MERGEUVROW_ANY(MergeUVRow_Any_AVX2, MergeUVRow_AVX2, MergeUVRow_C, 31) -#endif -#ifdef HAS_MERGEUVROW_NEON -MERGEUVROW_ANY(MergeUVRow_Any_NEON, MergeUVRow_NEON, MergeUVRow_C, 15) -#endif -#undef MERGEUVROW_ANY - -#define MATHROW_ANY(NAMEANY, ARGBMATH_SIMD, ARGBMATH_C, MASK) \ - void NAMEANY(const uint8* src_argb0, const uint8* src_argb1, \ - uint8* dst_argb, int width) { \ - int n = width & ~MASK; \ - ARGBMATH_SIMD(src_argb0, src_argb1, dst_argb, n); \ - ARGBMATH_C(src_argb0 + n * 4, \ - src_argb1 + n * 4, \ - dst_argb + n * 4, \ - width & MASK); \ - } - -#ifdef HAS_ARGBMULTIPLYROW_SSE2 -MATHROW_ANY(ARGBMultiplyRow_Any_SSE2, ARGBMultiplyRow_SSE2, ARGBMultiplyRow_C, - 3) -#endif -#ifdef HAS_ARGBADDROW_SSE2 -MATHROW_ANY(ARGBAddRow_Any_SSE2, ARGBAddRow_SSE2, ARGBAddRow_C, 3) -#endif -#ifdef HAS_ARGBSUBTRACTROW_SSE2 -MATHROW_ANY(ARGBSubtractRow_Any_SSE2, ARGBSubtractRow_SSE2, ARGBSubtractRow_C, - 3) -#endif -#ifdef HAS_ARGBMULTIPLYROW_AVX2 -MATHROW_ANY(ARGBMultiplyRow_Any_AVX2, ARGBMultiplyRow_AVX2, ARGBMultiplyRow_C, - 7) -#endif -#ifdef HAS_ARGBADDROW_AVX2 -MATHROW_ANY(ARGBAddRow_Any_AVX2, ARGBAddRow_AVX2, ARGBAddRow_C, 7) -#endif -#ifdef HAS_ARGBSUBTRACTROW_AVX2 -MATHROW_ANY(ARGBSubtractRow_Any_AVX2, ARGBSubtractRow_AVX2, ARGBSubtractRow_C, - 7) -#endif -#ifdef HAS_ARGBMULTIPLYROW_NEON -MATHROW_ANY(ARGBMultiplyRow_Any_NEON, ARGBMultiplyRow_NEON, ARGBMultiplyRow_C, - 7) -#endif -#ifdef HAS_ARGBADDROW_NEON -MATHROW_ANY(ARGBAddRow_Any_NEON, ARGBAddRow_NEON, ARGBAddRow_C, 7) -#endif -#ifdef HAS_ARGBSUBTRACTROW_NEON -MATHROW_ANY(ARGBSubtractRow_Any_NEON, ARGBSubtractRow_NEON, ARGBSubtractRow_C, - 7) -#endif -#undef MATHROW_ANY - -// Shuffle may want to work in place, so last16 method can not be used. -#define YANY(NAMEANY, ARGBTOY_SIMD, ARGBTOY_C, SBPP, BPP, MASK) \ - void NAMEANY(const uint8* src_argb, uint8* dst_argb, \ - const uint8* shuffler, int width) { \ - int n = width & ~MASK; \ - ARGBTOY_SIMD(src_argb, dst_argb, shuffler, n); \ - ARGBTOY_C(src_argb + n * SBPP, \ - dst_argb + n * BPP, shuffler, width & MASK); \ - } - -#ifdef HAS_ARGBSHUFFLEROW_SSE2 -YANY(ARGBShuffleRow_Any_SSE2, ARGBShuffleRow_SSE2, - ARGBShuffleRow_C, 4, 4, 3) -#endif -#ifdef HAS_ARGBSHUFFLEROW_SSSE3 -YANY(ARGBShuffleRow_Any_SSSE3, ARGBShuffleRow_Unaligned_SSSE3, - ARGBShuffleRow_C, 4, 4, 7) -#endif -#ifdef HAS_ARGBSHUFFLEROW_AVX2 -YANY(ARGBShuffleRow_Any_AVX2, ARGBShuffleRow_AVX2, - ARGBShuffleRow_C, 4, 4, 15) -#endif -#ifdef HAS_ARGBSHUFFLEROW_NEON -YANY(ARGBShuffleRow_Any_NEON, ARGBShuffleRow_NEON, - ARGBShuffleRow_C, 4, 4, 3) -#endif -#undef YANY - -// Interpolate may want to work in place, so last16 method can not be used. -#define NANY(NAMEANY, TERP_SIMD, TERP_C, SBPP, BPP, MASK) \ - void NAMEANY(uint8* dst_ptr, const uint8* src_ptr, \ - ptrdiff_t src_stride_ptr, int width, \ - int source_y_fraction) { \ - int n = width & ~MASK; \ - TERP_SIMD(dst_ptr, src_ptr, src_stride_ptr, \ - n, source_y_fraction); \ - TERP_C(dst_ptr + n * BPP, \ - src_ptr + n * SBPP, src_stride_ptr, \ - width & MASK, source_y_fraction); \ - } - -#ifdef HAS_INTERPOLATEROW_AVX2 -NANY(InterpolateRow_Any_AVX2, InterpolateRow_AVX2, - InterpolateRow_C, 1, 1, 32) -#endif -#ifdef HAS_INTERPOLATEROW_SSSE3 -NANY(InterpolateRow_Any_SSSE3, InterpolateRow_Unaligned_SSSE3, - InterpolateRow_C, 1, 1, 15) -#endif -#ifdef HAS_INTERPOLATEROW_SSE2 -NANY(InterpolateRow_Any_SSE2, InterpolateRow_Unaligned_SSE2, - InterpolateRow_C, 1, 1, 15) -#endif -#ifdef HAS_INTERPOLATEROW_NEON -NANY(InterpolateRow_Any_NEON, InterpolateRow_NEON, - InterpolateRow_C, 1, 1, 15) -#endif -#ifdef HAS_INTERPOLATEROW_MIPS_DSPR2 -NANY(InterpolateRow_Any_MIPS_DSPR2, InterpolateRow_MIPS_DSPR2, - InterpolateRow_C, 1, 1, 3) -#endif -#undef NANY - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_common.cc b/src/main/jni/libyuv/source/row_common.cc deleted file mode 100644 index fa2b752a2..000000000 --- a/src/main/jni/libyuv/source/row_common.cc +++ /dev/null @@ -1,2286 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#include <string.h> // For memcpy and memset. - -#include "libyuv/basic_types.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// llvm x86 is poor at ternary operator, so use branchless min/max. - -#define USE_BRANCHLESS 1 -#if USE_BRANCHLESS -static __inline int32 clamp0(int32 v) { - return ((-(v) >> 31) & (v)); -} - -static __inline int32 clamp255(int32 v) { - return (((255 - (v)) >> 31) | (v)) & 255; -} - -static __inline uint32 Clamp(int32 val) { - int v = clamp0(val); - return (uint32)(clamp255(v)); -} - -static __inline uint32 Abs(int32 v) { - int m = v >> 31; - return (v + m) ^ m; -} -#else // USE_BRANCHLESS -static __inline int32 clamp0(int32 v) { - return (v < 0) ? 0 : v; -} - -static __inline int32 clamp255(int32 v) { - return (v > 255) ? 255 : v; -} - -static __inline uint32 Clamp(int32 val) { - int v = clamp0(val); - return (uint32)(clamp255(v)); -} - -static __inline uint32 Abs(int32 v) { - return (v < 0) ? -v : v; -} -#endif // USE_BRANCHLESS - -#ifdef LIBYUV_LITTLE_ENDIAN -#define WRITEWORD(p, v) *(uint32*)(p) = v -#else -static inline void WRITEWORD(uint8* p, uint32 v) { - p[0] = (uint8)(v & 255); - p[1] = (uint8)((v >> 8) & 255); - p[2] = (uint8)((v >> 16) & 255); - p[3] = (uint8)((v >> 24) & 255); -} -#endif - -void RGB24ToARGBRow_C(const uint8* src_rgb24, uint8* dst_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_rgb24[0]; - uint8 g = src_rgb24[1]; - uint8 r = src_rgb24[2]; - dst_argb[0] = b; - dst_argb[1] = g; - dst_argb[2] = r; - dst_argb[3] = 255u; - dst_argb += 4; - src_rgb24 += 3; - } -} - -void RAWToARGBRow_C(const uint8* src_raw, uint8* dst_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 r = src_raw[0]; - uint8 g = src_raw[1]; - uint8 b = src_raw[2]; - dst_argb[0] = b; - dst_argb[1] = g; - dst_argb[2] = r; - dst_argb[3] = 255u; - dst_argb += 4; - src_raw += 3; - } -} - -void RGB565ToARGBRow_C(const uint8* src_rgb565, uint8* dst_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_rgb565[0] & 0x1f; - uint8 g = (src_rgb565[0] >> 5) | ((src_rgb565[1] & 0x07) << 3); - uint8 r = src_rgb565[1] >> 3; - dst_argb[0] = (b << 3) | (b >> 2); - dst_argb[1] = (g << 2) | (g >> 4); - dst_argb[2] = (r << 3) | (r >> 2); - dst_argb[3] = 255u; - dst_argb += 4; - src_rgb565 += 2; - } -} - -void ARGB1555ToARGBRow_C(const uint8* src_argb1555, uint8* dst_argb, - int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_argb1555[0] & 0x1f; - uint8 g = (src_argb1555[0] >> 5) | ((src_argb1555[1] & 0x03) << 3); - uint8 r = (src_argb1555[1] & 0x7c) >> 2; - uint8 a = src_argb1555[1] >> 7; - dst_argb[0] = (b << 3) | (b >> 2); - dst_argb[1] = (g << 3) | (g >> 2); - dst_argb[2] = (r << 3) | (r >> 2); - dst_argb[3] = -a; - dst_argb += 4; - src_argb1555 += 2; - } -} - -void ARGB4444ToARGBRow_C(const uint8* src_argb4444, uint8* dst_argb, - int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_argb4444[0] & 0x0f; - uint8 g = src_argb4444[0] >> 4; - uint8 r = src_argb4444[1] & 0x0f; - uint8 a = src_argb4444[1] >> 4; - dst_argb[0] = (b << 4) | b; - dst_argb[1] = (g << 4) | g; - dst_argb[2] = (r << 4) | r; - dst_argb[3] = (a << 4) | a; - dst_argb += 4; - src_argb4444 += 2; - } -} - -void ARGBToRGB24Row_C(const uint8* src_argb, uint8* dst_rgb, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_argb[0]; - uint8 g = src_argb[1]; - uint8 r = src_argb[2]; - dst_rgb[0] = b; - dst_rgb[1] = g; - dst_rgb[2] = r; - dst_rgb += 3; - src_argb += 4; - } -} - -void ARGBToRAWRow_C(const uint8* src_argb, uint8* dst_rgb, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_argb[0]; - uint8 g = src_argb[1]; - uint8 r = src_argb[2]; - dst_rgb[0] = r; - dst_rgb[1] = g; - dst_rgb[2] = b; - dst_rgb += 3; - src_argb += 4; - } -} - -void ARGBToRGB565Row_C(const uint8* src_argb, uint8* dst_rgb, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 b0 = src_argb[0] >> 3; - uint8 g0 = src_argb[1] >> 2; - uint8 r0 = src_argb[2] >> 3; - uint8 b1 = src_argb[4] >> 3; - uint8 g1 = src_argb[5] >> 2; - uint8 r1 = src_argb[6] >> 3; - WRITEWORD(dst_rgb, b0 | (g0 << 5) | (r0 << 11) | - (b1 << 16) | (g1 << 21) | (r1 << 27)); - dst_rgb += 4; - src_argb += 8; - } - if (width & 1) { - uint8 b0 = src_argb[0] >> 3; - uint8 g0 = src_argb[1] >> 2; - uint8 r0 = src_argb[2] >> 3; - *(uint16*)(dst_rgb) = b0 | (g0 << 5) | (r0 << 11); - } -} - -void ARGBToARGB1555Row_C(const uint8* src_argb, uint8* dst_rgb, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 b0 = src_argb[0] >> 3; - uint8 g0 = src_argb[1] >> 3; - uint8 r0 = src_argb[2] >> 3; - uint8 a0 = src_argb[3] >> 7; - uint8 b1 = src_argb[4] >> 3; - uint8 g1 = src_argb[5] >> 3; - uint8 r1 = src_argb[6] >> 3; - uint8 a1 = src_argb[7] >> 7; - *(uint32*)(dst_rgb) = - b0 | (g0 << 5) | (r0 << 10) | (a0 << 15) | - (b1 << 16) | (g1 << 21) | (r1 << 26) | (a1 << 31); - dst_rgb += 4; - src_argb += 8; - } - if (width & 1) { - uint8 b0 = src_argb[0] >> 3; - uint8 g0 = src_argb[1] >> 3; - uint8 r0 = src_argb[2] >> 3; - uint8 a0 = src_argb[3] >> 7; - *(uint16*)(dst_rgb) = - b0 | (g0 << 5) | (r0 << 10) | (a0 << 15); - } -} - -void ARGBToARGB4444Row_C(const uint8* src_argb, uint8* dst_rgb, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 b0 = src_argb[0] >> 4; - uint8 g0 = src_argb[1] >> 4; - uint8 r0 = src_argb[2] >> 4; - uint8 a0 = src_argb[3] >> 4; - uint8 b1 = src_argb[4] >> 4; - uint8 g1 = src_argb[5] >> 4; - uint8 r1 = src_argb[6] >> 4; - uint8 a1 = src_argb[7] >> 4; - *(uint32*)(dst_rgb) = - b0 | (g0 << 4) | (r0 << 8) | (a0 << 12) | - (b1 << 16) | (g1 << 20) | (r1 << 24) | (a1 << 28); - dst_rgb += 4; - src_argb += 8; - } - if (width & 1) { - uint8 b0 = src_argb[0] >> 4; - uint8 g0 = src_argb[1] >> 4; - uint8 r0 = src_argb[2] >> 4; - uint8 a0 = src_argb[3] >> 4; - *(uint16*)(dst_rgb) = - b0 | (g0 << 4) | (r0 << 8) | (a0 << 12); - } -} - -static __inline int RGBToY(uint8 r, uint8 g, uint8 b) { - return (66 * r + 129 * g + 25 * b + 0x1080) >> 8; -} - -static __inline int RGBToU(uint8 r, uint8 g, uint8 b) { - return (112 * b - 74 * g - 38 * r + 0x8080) >> 8; -} -static __inline int RGBToV(uint8 r, uint8 g, uint8 b) { - return (112 * r - 94 * g - 18 * b + 0x8080) >> 8; -} - -#define MAKEROWY(NAME, R, G, B, BPP) \ -void NAME ## ToYRow_C(const uint8* src_argb0, uint8* dst_y, int width) { \ - int x; \ - for (x = 0; x < width; ++x) { \ - dst_y[0] = RGBToY(src_argb0[R], src_argb0[G], src_argb0[B]); \ - src_argb0 += BPP; \ - dst_y += 1; \ - } \ -} \ -void NAME ## ToUVRow_C(const uint8* src_rgb0, int src_stride_rgb, \ - uint8* dst_u, uint8* dst_v, int width) { \ - const uint8* src_rgb1 = src_rgb0 + src_stride_rgb; \ - int x; \ - for (x = 0; x < width - 1; x += 2) { \ - uint8 ab = (src_rgb0[B] + src_rgb0[B + BPP] + \ - src_rgb1[B] + src_rgb1[B + BPP]) >> 2; \ - uint8 ag = (src_rgb0[G] + src_rgb0[G + BPP] + \ - src_rgb1[G] + src_rgb1[G + BPP]) >> 2; \ - uint8 ar = (src_rgb0[R] + src_rgb0[R + BPP] + \ - src_rgb1[R] + src_rgb1[R + BPP]) >> 2; \ - dst_u[0] = RGBToU(ar, ag, ab); \ - dst_v[0] = RGBToV(ar, ag, ab); \ - src_rgb0 += BPP * 2; \ - src_rgb1 += BPP * 2; \ - dst_u += 1; \ - dst_v += 1; \ - } \ - if (width & 1) { \ - uint8 ab = (src_rgb0[B] + src_rgb1[B]) >> 1; \ - uint8 ag = (src_rgb0[G] + src_rgb1[G]) >> 1; \ - uint8 ar = (src_rgb0[R] + src_rgb1[R]) >> 1; \ - dst_u[0] = RGBToU(ar, ag, ab); \ - dst_v[0] = RGBToV(ar, ag, ab); \ - } \ -} - -MAKEROWY(ARGB, 2, 1, 0, 4) -MAKEROWY(BGRA, 1, 2, 3, 4) -MAKEROWY(ABGR, 0, 1, 2, 4) -MAKEROWY(RGBA, 3, 2, 1, 4) -MAKEROWY(RGB24, 2, 1, 0, 3) -MAKEROWY(RAW, 0, 1, 2, 3) -#undef MAKEROWY - -// JPeg uses a variation on BT.601-1 full range -// y = 0.29900 * r + 0.58700 * g + 0.11400 * b -// u = -0.16874 * r - 0.33126 * g + 0.50000 * b + center -// v = 0.50000 * r - 0.41869 * g - 0.08131 * b + center -// BT.601 Mpeg range uses: -// b 0.1016 * 255 = 25.908 = 25 -// g 0.5078 * 255 = 129.489 = 129 -// r 0.2578 * 255 = 65.739 = 66 -// JPeg 8 bit Y (not used): -// b 0.11400 * 256 = 29.184 = 29 -// g 0.58700 * 256 = 150.272 = 150 -// r 0.29900 * 256 = 76.544 = 77 -// JPeg 7 bit Y: -// b 0.11400 * 128 = 14.592 = 15 -// g 0.58700 * 128 = 75.136 = 75 -// r 0.29900 * 128 = 38.272 = 38 -// JPeg 8 bit U: -// b 0.50000 * 255 = 127.5 = 127 -// g -0.33126 * 255 = -84.4713 = -84 -// r -0.16874 * 255 = -43.0287 = -43 -// JPeg 8 bit V: -// b -0.08131 * 255 = -20.73405 = -20 -// g -0.41869 * 255 = -106.76595 = -107 -// r 0.50000 * 255 = 127.5 = 127 - -static __inline int RGBToYJ(uint8 r, uint8 g, uint8 b) { - return (38 * r + 75 * g + 15 * b + 64) >> 7; -} - -static __inline int RGBToUJ(uint8 r, uint8 g, uint8 b) { - return (127 * b - 84 * g - 43 * r + 0x8080) >> 8; -} -static __inline int RGBToVJ(uint8 r, uint8 g, uint8 b) { - return (127 * r - 107 * g - 20 * b + 0x8080) >> 8; -} - -#define AVGB(a, b) (((a) + (b) + 1) >> 1) - -#define MAKEROWYJ(NAME, R, G, B, BPP) \ -void NAME ## ToYJRow_C(const uint8* src_argb0, uint8* dst_y, int width) { \ - int x; \ - for (x = 0; x < width; ++x) { \ - dst_y[0] = RGBToYJ(src_argb0[R], src_argb0[G], src_argb0[B]); \ - src_argb0 += BPP; \ - dst_y += 1; \ - } \ -} \ -void NAME ## ToUVJRow_C(const uint8* src_rgb0, int src_stride_rgb, \ - uint8* dst_u, uint8* dst_v, int width) { \ - const uint8* src_rgb1 = src_rgb0 + src_stride_rgb; \ - int x; \ - for (x = 0; x < width - 1; x += 2) { \ - uint8 ab = AVGB(AVGB(src_rgb0[B], src_rgb1[B]), \ - AVGB(src_rgb0[B + BPP], src_rgb1[B + BPP])); \ - uint8 ag = AVGB(AVGB(src_rgb0[G], src_rgb1[G]), \ - AVGB(src_rgb0[G + BPP], src_rgb1[G + BPP])); \ - uint8 ar = AVGB(AVGB(src_rgb0[R], src_rgb1[R]), \ - AVGB(src_rgb0[R + BPP], src_rgb1[R + BPP])); \ - dst_u[0] = RGBToUJ(ar, ag, ab); \ - dst_v[0] = RGBToVJ(ar, ag, ab); \ - src_rgb0 += BPP * 2; \ - src_rgb1 += BPP * 2; \ - dst_u += 1; \ - dst_v += 1; \ - } \ - if (width & 1) { \ - uint8 ab = AVGB(src_rgb0[B], src_rgb1[B]); \ - uint8 ag = AVGB(src_rgb0[G], src_rgb1[G]); \ - uint8 ar = AVGB(src_rgb0[R], src_rgb1[R]); \ - dst_u[0] = RGBToUJ(ar, ag, ab); \ - dst_v[0] = RGBToVJ(ar, ag, ab); \ - } \ -} - -MAKEROWYJ(ARGB, 2, 1, 0, 4) -#undef MAKEROWYJ - -void RGB565ToYRow_C(const uint8* src_rgb565, uint8* dst_y, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_rgb565[0] & 0x1f; - uint8 g = (src_rgb565[0] >> 5) | ((src_rgb565[1] & 0x07) << 3); - uint8 r = src_rgb565[1] >> 3; - b = (b << 3) | (b >> 2); - g = (g << 2) | (g >> 4); - r = (r << 3) | (r >> 2); - dst_y[0] = RGBToY(r, g, b); - src_rgb565 += 2; - dst_y += 1; - } -} - -void ARGB1555ToYRow_C(const uint8* src_argb1555, uint8* dst_y, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_argb1555[0] & 0x1f; - uint8 g = (src_argb1555[0] >> 5) | ((src_argb1555[1] & 0x03) << 3); - uint8 r = (src_argb1555[1] & 0x7c) >> 2; - b = (b << 3) | (b >> 2); - g = (g << 3) | (g >> 2); - r = (r << 3) | (r >> 2); - dst_y[0] = RGBToY(r, g, b); - src_argb1555 += 2; - dst_y += 1; - } -} - -void ARGB4444ToYRow_C(const uint8* src_argb4444, uint8* dst_y, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 b = src_argb4444[0] & 0x0f; - uint8 g = src_argb4444[0] >> 4; - uint8 r = src_argb4444[1] & 0x0f; - b = (b << 4) | b; - g = (g << 4) | g; - r = (r << 4) | r; - dst_y[0] = RGBToY(r, g, b); - src_argb4444 += 2; - dst_y += 1; - } -} - -void RGB565ToUVRow_C(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int width) { - const uint8* next_rgb565 = src_rgb565 + src_stride_rgb565; - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 b0 = src_rgb565[0] & 0x1f; - uint8 g0 = (src_rgb565[0] >> 5) | ((src_rgb565[1] & 0x07) << 3); - uint8 r0 = src_rgb565[1] >> 3; - uint8 b1 = src_rgb565[2] & 0x1f; - uint8 g1 = (src_rgb565[2] >> 5) | ((src_rgb565[3] & 0x07) << 3); - uint8 r1 = src_rgb565[3] >> 3; - uint8 b2 = next_rgb565[0] & 0x1f; - uint8 g2 = (next_rgb565[0] >> 5) | ((next_rgb565[1] & 0x07) << 3); - uint8 r2 = next_rgb565[1] >> 3; - uint8 b3 = next_rgb565[2] & 0x1f; - uint8 g3 = (next_rgb565[2] >> 5) | ((next_rgb565[3] & 0x07) << 3); - uint8 r3 = next_rgb565[3] >> 3; - uint8 b = (b0 + b1 + b2 + b3); // 565 * 4 = 787. - uint8 g = (g0 + g1 + g2 + g3); - uint8 r = (r0 + r1 + r2 + r3); - b = (b << 1) | (b >> 6); // 787 -> 888. - r = (r << 1) | (r >> 6); - dst_u[0] = RGBToU(r, g, b); - dst_v[0] = RGBToV(r, g, b); - src_rgb565 += 4; - next_rgb565 += 4; - dst_u += 1; - dst_v += 1; - } - if (width & 1) { - uint8 b0 = src_rgb565[0] & 0x1f; - uint8 g0 = (src_rgb565[0] >> 5) | ((src_rgb565[1] & 0x07) << 3); - uint8 r0 = src_rgb565[1] >> 3; - uint8 b2 = next_rgb565[0] & 0x1f; - uint8 g2 = (next_rgb565[0] >> 5) | ((next_rgb565[1] & 0x07) << 3); - uint8 r2 = next_rgb565[1] >> 3; - uint8 b = (b0 + b2); // 565 * 2 = 676. - uint8 g = (g0 + g2); - uint8 r = (r0 + r2); - b = (b << 2) | (b >> 4); // 676 -> 888 - g = (g << 1) | (g >> 6); - r = (r << 2) | (r >> 4); - dst_u[0] = RGBToU(r, g, b); - dst_v[0] = RGBToV(r, g, b); - } -} - -void ARGB1555ToUVRow_C(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int width) { - const uint8* next_argb1555 = src_argb1555 + src_stride_argb1555; - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 b0 = src_argb1555[0] & 0x1f; - uint8 g0 = (src_argb1555[0] >> 5) | ((src_argb1555[1] & 0x03) << 3); - uint8 r0 = (src_argb1555[1] & 0x7c) >> 2; - uint8 b1 = src_argb1555[2] & 0x1f; - uint8 g1 = (src_argb1555[2] >> 5) | ((src_argb1555[3] & 0x03) << 3); - uint8 r1 = (src_argb1555[3] & 0x7c) >> 2; - uint8 b2 = next_argb1555[0] & 0x1f; - uint8 g2 = (next_argb1555[0] >> 5) | ((next_argb1555[1] & 0x03) << 3); - uint8 r2 = (next_argb1555[1] & 0x7c) >> 2; - uint8 b3 = next_argb1555[2] & 0x1f; - uint8 g3 = (next_argb1555[2] >> 5) | ((next_argb1555[3] & 0x03) << 3); - uint8 r3 = (next_argb1555[3] & 0x7c) >> 2; - uint8 b = (b0 + b1 + b2 + b3); // 555 * 4 = 777. - uint8 g = (g0 + g1 + g2 + g3); - uint8 r = (r0 + r1 + r2 + r3); - b = (b << 1) | (b >> 6); // 777 -> 888. - g = (g << 1) | (g >> 6); - r = (r << 1) | (r >> 6); - dst_u[0] = RGBToU(r, g, b); - dst_v[0] = RGBToV(r, g, b); - src_argb1555 += 4; - next_argb1555 += 4; - dst_u += 1; - dst_v += 1; - } - if (width & 1) { - uint8 b0 = src_argb1555[0] & 0x1f; - uint8 g0 = (src_argb1555[0] >> 5) | ((src_argb1555[1] & 0x03) << 3); - uint8 r0 = (src_argb1555[1] & 0x7c) >> 2; - uint8 b2 = next_argb1555[0] & 0x1f; - uint8 g2 = (next_argb1555[0] >> 5) | ((next_argb1555[1] & 0x03) << 3); - uint8 r2 = next_argb1555[1] >> 3; - uint8 b = (b0 + b2); // 555 * 2 = 666. - uint8 g = (g0 + g2); - uint8 r = (r0 + r2); - b = (b << 2) | (b >> 4); // 666 -> 888. - g = (g << 2) | (g >> 4); - r = (r << 2) | (r >> 4); - dst_u[0] = RGBToU(r, g, b); - dst_v[0] = RGBToV(r, g, b); - } -} - -void ARGB4444ToUVRow_C(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int width) { - const uint8* next_argb4444 = src_argb4444 + src_stride_argb4444; - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 b0 = src_argb4444[0] & 0x0f; - uint8 g0 = src_argb4444[0] >> 4; - uint8 r0 = src_argb4444[1] & 0x0f; - uint8 b1 = src_argb4444[2] & 0x0f; - uint8 g1 = src_argb4444[2] >> 4; - uint8 r1 = src_argb4444[3] & 0x0f; - uint8 b2 = next_argb4444[0] & 0x0f; - uint8 g2 = next_argb4444[0] >> 4; - uint8 r2 = next_argb4444[1] & 0x0f; - uint8 b3 = next_argb4444[2] & 0x0f; - uint8 g3 = next_argb4444[2] >> 4; - uint8 r3 = next_argb4444[3] & 0x0f; - uint8 b = (b0 + b1 + b2 + b3); // 444 * 4 = 666. - uint8 g = (g0 + g1 + g2 + g3); - uint8 r = (r0 + r1 + r2 + r3); - b = (b << 2) | (b >> 4); // 666 -> 888. - g = (g << 2) | (g >> 4); - r = (r << 2) | (r >> 4); - dst_u[0] = RGBToU(r, g, b); - dst_v[0] = RGBToV(r, g, b); - src_argb4444 += 4; - next_argb4444 += 4; - dst_u += 1; - dst_v += 1; - } - if (width & 1) { - uint8 b0 = src_argb4444[0] & 0x0f; - uint8 g0 = src_argb4444[0] >> 4; - uint8 r0 = src_argb4444[1] & 0x0f; - uint8 b2 = next_argb4444[0] & 0x0f; - uint8 g2 = next_argb4444[0] >> 4; - uint8 r2 = next_argb4444[1] & 0x0f; - uint8 b = (b0 + b2); // 444 * 2 = 555. - uint8 g = (g0 + g2); - uint8 r = (r0 + r2); - b = (b << 3) | (b >> 2); // 555 -> 888. - g = (g << 3) | (g >> 2); - r = (r << 3) | (r >> 2); - dst_u[0] = RGBToU(r, g, b); - dst_v[0] = RGBToV(r, g, b); - } -} - -void ARGBToUV444Row_C(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 ab = src_argb[0]; - uint8 ag = src_argb[1]; - uint8 ar = src_argb[2]; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - src_argb += 4; - dst_u += 1; - dst_v += 1; - } -} - -void ARGBToUV422Row_C(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 ab = (src_argb[0] + src_argb[4]) >> 1; - uint8 ag = (src_argb[1] + src_argb[5]) >> 1; - uint8 ar = (src_argb[2] + src_argb[6]) >> 1; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - src_argb += 8; - dst_u += 1; - dst_v += 1; - } - if (width & 1) { - uint8 ab = src_argb[0]; - uint8 ag = src_argb[1]; - uint8 ar = src_argb[2]; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - } -} - -void ARGBToUV411Row_C(const uint8* src_argb, - uint8* dst_u, uint8* dst_v, int width) { - int x; - for (x = 0; x < width - 3; x += 4) { - uint8 ab = (src_argb[0] + src_argb[4] + src_argb[8] + src_argb[12]) >> 2; - uint8 ag = (src_argb[1] + src_argb[5] + src_argb[9] + src_argb[13]) >> 2; - uint8 ar = (src_argb[2] + src_argb[6] + src_argb[10] + src_argb[14]) >> 2; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - src_argb += 16; - dst_u += 1; - dst_v += 1; - } - if ((width & 3) == 3) { - uint8 ab = (src_argb[0] + src_argb[4] + src_argb[8]) / 3; - uint8 ag = (src_argb[1] + src_argb[5] + src_argb[9]) / 3; - uint8 ar = (src_argb[2] + src_argb[6] + src_argb[10]) / 3; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - } else if ((width & 3) == 2) { - uint8 ab = (src_argb[0] + src_argb[4]) >> 1; - uint8 ag = (src_argb[1] + src_argb[5]) >> 1; - uint8 ar = (src_argb[2] + src_argb[6]) >> 1; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - } else if ((width & 3) == 1) { - uint8 ab = src_argb[0]; - uint8 ag = src_argb[1]; - uint8 ar = src_argb[2]; - dst_u[0] = RGBToU(ar, ag, ab); - dst_v[0] = RGBToV(ar, ag, ab); - } -} - -void ARGBGrayRow_C(const uint8* src_argb, uint8* dst_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - uint8 y = RGBToYJ(src_argb[2], src_argb[1], src_argb[0]); - dst_argb[2] = dst_argb[1] = dst_argb[0] = y; - dst_argb[3] = src_argb[3]; - dst_argb += 4; - src_argb += 4; - } -} - -// Convert a row of image to Sepia tone. -void ARGBSepiaRow_C(uint8* dst_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - int b = dst_argb[0]; - int g = dst_argb[1]; - int r = dst_argb[2]; - int sb = (b * 17 + g * 68 + r * 35) >> 7; - int sg = (b * 22 + g * 88 + r * 45) >> 7; - int sr = (b * 24 + g * 98 + r * 50) >> 7; - // b does not over flow. a is preserved from original. - dst_argb[0] = sb; - dst_argb[1] = clamp255(sg); - dst_argb[2] = clamp255(sr); - dst_argb += 4; - } -} - -// Apply color matrix to a row of image. Matrix is signed. -// TODO(fbarchard): Consider adding rounding (+32). -void ARGBColorMatrixRow_C(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - int b = src_argb[0]; - int g = src_argb[1]; - int r = src_argb[2]; - int a = src_argb[3]; - int sb = (b * matrix_argb[0] + g * matrix_argb[1] + - r * matrix_argb[2] + a * matrix_argb[3]) >> 6; - int sg = (b * matrix_argb[4] + g * matrix_argb[5] + - r * matrix_argb[6] + a * matrix_argb[7]) >> 6; - int sr = (b * matrix_argb[8] + g * matrix_argb[9] + - r * matrix_argb[10] + a * matrix_argb[11]) >> 6; - int sa = (b * matrix_argb[12] + g * matrix_argb[13] + - r * matrix_argb[14] + a * matrix_argb[15]) >> 6; - dst_argb[0] = Clamp(sb); - dst_argb[1] = Clamp(sg); - dst_argb[2] = Clamp(sr); - dst_argb[3] = Clamp(sa); - src_argb += 4; - dst_argb += 4; - } -} - -// Apply color table to a row of image. -void ARGBColorTableRow_C(uint8* dst_argb, const uint8* table_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - int b = dst_argb[0]; - int g = dst_argb[1]; - int r = dst_argb[2]; - int a = dst_argb[3]; - dst_argb[0] = table_argb[b * 4 + 0]; - dst_argb[1] = table_argb[g * 4 + 1]; - dst_argb[2] = table_argb[r * 4 + 2]; - dst_argb[3] = table_argb[a * 4 + 3]; - dst_argb += 4; - } -} - -// Apply color table to a row of image. -void RGBColorTableRow_C(uint8* dst_argb, const uint8* table_argb, int width) { - int x; - for (x = 0; x < width; ++x) { - int b = dst_argb[0]; - int g = dst_argb[1]; - int r = dst_argb[2]; - dst_argb[0] = table_argb[b * 4 + 0]; - dst_argb[1] = table_argb[g * 4 + 1]; - dst_argb[2] = table_argb[r * 4 + 2]; - dst_argb += 4; - } -} - -void ARGBQuantizeRow_C(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width) { - int x; - for (x = 0; x < width; ++x) { - int b = dst_argb[0]; - int g = dst_argb[1]; - int r = dst_argb[2]; - dst_argb[0] = (b * scale >> 16) * interval_size + interval_offset; - dst_argb[1] = (g * scale >> 16) * interval_size + interval_offset; - dst_argb[2] = (r * scale >> 16) * interval_size + interval_offset; - dst_argb += 4; - } -} - -#define REPEAT8(v) (v) | ((v) << 8) -#define SHADE(f, v) v * f >> 24 - -void ARGBShadeRow_C(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value) { - const uint32 b_scale = REPEAT8(value & 0xff); - const uint32 g_scale = REPEAT8((value >> 8) & 0xff); - const uint32 r_scale = REPEAT8((value >> 16) & 0xff); - const uint32 a_scale = REPEAT8(value >> 24); - - int i; - for (i = 0; i < width; ++i) { - const uint32 b = REPEAT8(src_argb[0]); - const uint32 g = REPEAT8(src_argb[1]); - const uint32 r = REPEAT8(src_argb[2]); - const uint32 a = REPEAT8(src_argb[3]); - dst_argb[0] = SHADE(b, b_scale); - dst_argb[1] = SHADE(g, g_scale); - dst_argb[2] = SHADE(r, r_scale); - dst_argb[3] = SHADE(a, a_scale); - src_argb += 4; - dst_argb += 4; - } -} -#undef REPEAT8 -#undef SHADE - -#define REPEAT8(v) (v) | ((v) << 8) -#define SHADE(f, v) v * f >> 16 - -void ARGBMultiplyRow_C(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - int i; - for (i = 0; i < width; ++i) { - const uint32 b = REPEAT8(src_argb0[0]); - const uint32 g = REPEAT8(src_argb0[1]); - const uint32 r = REPEAT8(src_argb0[2]); - const uint32 a = REPEAT8(src_argb0[3]); - const uint32 b_scale = src_argb1[0]; - const uint32 g_scale = src_argb1[1]; - const uint32 r_scale = src_argb1[2]; - const uint32 a_scale = src_argb1[3]; - dst_argb[0] = SHADE(b, b_scale); - dst_argb[1] = SHADE(g, g_scale); - dst_argb[2] = SHADE(r, r_scale); - dst_argb[3] = SHADE(a, a_scale); - src_argb0 += 4; - src_argb1 += 4; - dst_argb += 4; - } -} -#undef REPEAT8 -#undef SHADE - -#define SHADE(f, v) clamp255(v + f) - -void ARGBAddRow_C(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - int i; - for (i = 0; i < width; ++i) { - const int b = src_argb0[0]; - const int g = src_argb0[1]; - const int r = src_argb0[2]; - const int a = src_argb0[3]; - const int b_add = src_argb1[0]; - const int g_add = src_argb1[1]; - const int r_add = src_argb1[2]; - const int a_add = src_argb1[3]; - dst_argb[0] = SHADE(b, b_add); - dst_argb[1] = SHADE(g, g_add); - dst_argb[2] = SHADE(r, r_add); - dst_argb[3] = SHADE(a, a_add); - src_argb0 += 4; - src_argb1 += 4; - dst_argb += 4; - } -} -#undef SHADE - -#define SHADE(f, v) clamp0(f - v) - -void ARGBSubtractRow_C(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - int i; - for (i = 0; i < width; ++i) { - const int b = src_argb0[0]; - const int g = src_argb0[1]; - const int r = src_argb0[2]; - const int a = src_argb0[3]; - const int b_sub = src_argb1[0]; - const int g_sub = src_argb1[1]; - const int r_sub = src_argb1[2]; - const int a_sub = src_argb1[3]; - dst_argb[0] = SHADE(b, b_sub); - dst_argb[1] = SHADE(g, g_sub); - dst_argb[2] = SHADE(r, r_sub); - dst_argb[3] = SHADE(a, a_sub); - src_argb0 += 4; - src_argb1 += 4; - dst_argb += 4; - } -} -#undef SHADE - -// Sobel functions which mimics SSSE3. -void SobelXRow_C(const uint8* src_y0, const uint8* src_y1, const uint8* src_y2, - uint8* dst_sobelx, int width) { - int i; - for (i = 0; i < width; ++i) { - int a = src_y0[i]; - int b = src_y1[i]; - int c = src_y2[i]; - int a_sub = src_y0[i + 2]; - int b_sub = src_y1[i + 2]; - int c_sub = src_y2[i + 2]; - int a_diff = a - a_sub; - int b_diff = b - b_sub; - int c_diff = c - c_sub; - int sobel = Abs(a_diff + b_diff * 2 + c_diff); - dst_sobelx[i] = (uint8)(clamp255(sobel)); - } -} - -void SobelYRow_C(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width) { - int i; - for (i = 0; i < width; ++i) { - int a = src_y0[i + 0]; - int b = src_y0[i + 1]; - int c = src_y0[i + 2]; - int a_sub = src_y1[i + 0]; - int b_sub = src_y1[i + 1]; - int c_sub = src_y1[i + 2]; - int a_diff = a - a_sub; - int b_diff = b - b_sub; - int c_diff = c - c_sub; - int sobel = Abs(a_diff + b_diff * 2 + c_diff); - dst_sobely[i] = (uint8)(clamp255(sobel)); - } -} - -void SobelRow_C(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - int i; - for (i = 0; i < width; ++i) { - int r = src_sobelx[i]; - int b = src_sobely[i]; - int s = clamp255(r + b); - dst_argb[0] = (uint8)(s); - dst_argb[1] = (uint8)(s); - dst_argb[2] = (uint8)(s); - dst_argb[3] = (uint8)(255u); - dst_argb += 4; - } -} - -void SobelToPlaneRow_C(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width) { - int i; - for (i = 0; i < width; ++i) { - int r = src_sobelx[i]; - int b = src_sobely[i]; - int s = clamp255(r + b); - dst_y[i] = (uint8)(s); - } -} - -void SobelXYRow_C(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - int i; - for (i = 0; i < width; ++i) { - int r = src_sobelx[i]; - int b = src_sobely[i]; - int g = clamp255(r + b); - dst_argb[0] = (uint8)(b); - dst_argb[1] = (uint8)(g); - dst_argb[2] = (uint8)(r); - dst_argb[3] = (uint8)(255u); - dst_argb += 4; - } -} - -void I400ToARGBRow_C(const uint8* src_y, uint8* dst_argb, int width) { - // Copy a Y to RGB. - int x; - for (x = 0; x < width; ++x) { - uint8 y = src_y[0]; - dst_argb[2] = dst_argb[1] = dst_argb[0] = y; - dst_argb[3] = 255u; - dst_argb += 4; - ++src_y; - } -} - -// C reference code that mimics the YUV assembly. - -#define YG 74 /* (int8)(1.164 * 64 + 0.5) */ - -#define UB 127 /* min(63,(int8)(2.018 * 64)) */ -#define UG -25 /* (int8)(-0.391 * 64 - 0.5) */ -#define UR 0 - -#define VB 0 -#define VG -52 /* (int8)(-0.813 * 64 - 0.5) */ -#define VR 102 /* (int8)(1.596 * 64 + 0.5) */ - -// Bias -#define BB UB * 128 + VB * 128 -#define BG UG * 128 + VG * 128 -#define BR UR * 128 + VR * 128 - -static __inline void YuvPixel(uint8 y, uint8 u, uint8 v, - uint8* b, uint8* g, uint8* r) { - int32 y1 = ((int32)(y) - 16) * YG; - *b = Clamp((int32)((u * UB + v * VB) - (BB) + y1) >> 6); - *g = Clamp((int32)((u * UG + v * VG) - (BG) + y1) >> 6); - *r = Clamp((int32)((u * UR + v * VR) - (BR) + y1) >> 6); -} - -#if !defined(LIBYUV_DISABLE_NEON) && \ - (defined(__ARM_NEON__) || defined(LIBYUV_NEON)) -// C mimic assembly. -// TODO(fbarchard): Remove subsampling from Neon. -void I444ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - uint8 u = (src_u[0] + src_u[1] + 1) >> 1; - uint8 v = (src_v[0] + src_v[1] + 1) >> 1; - YuvPixel(src_y[0], u, v, rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_y[1], u, v, rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_y += 2; - src_u += 2; - src_v += 2; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - } -} -#else -void I444ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width; ++x) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - src_y += 1; - src_u += 1; - src_v += 1; - rgb_buf += 4; // Advance 1 pixel. - } -} -#endif -// Also used for 420 -void I422ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_y += 2; - src_u += 1; - src_v += 1; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void I422ToRGB24Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 3, rgb_buf + 4, rgb_buf + 5); - src_y += 2; - src_u += 1; - src_v += 1; - rgb_buf += 6; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - } -} - -void I422ToRAWRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 2, rgb_buf + 1, rgb_buf + 0); - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 5, rgb_buf + 4, rgb_buf + 3); - src_y += 2; - src_u += 1; - src_v += 1; - rgb_buf += 6; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 2, rgb_buf + 1, rgb_buf + 0); - } -} - -void I422ToARGB4444Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb4444, - int width) { - uint8 b0; - uint8 g0; - uint8 r0; - uint8 b1; - uint8 g1; - uint8 r1; - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], &b0, &g0, &r0); - YuvPixel(src_y[1], src_u[0], src_v[0], &b1, &g1, &r1); - b0 = b0 >> 4; - g0 = g0 >> 4; - r0 = r0 >> 4; - b1 = b1 >> 4; - g1 = g1 >> 4; - r1 = r1 >> 4; - *(uint32*)(dst_argb4444) = b0 | (g0 << 4) | (r0 << 8) | - (b1 << 16) | (g1 << 20) | (r1 << 24) | 0xf000f000; - src_y += 2; - src_u += 1; - src_v += 1; - dst_argb4444 += 4; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], &b0, &g0, &r0); - b0 = b0 >> 4; - g0 = g0 >> 4; - r0 = r0 >> 4; - *(uint16*)(dst_argb4444) = b0 | (g0 << 4) | (r0 << 8) | - 0xf000; - } -} - -void I422ToARGB1555Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb1555, - int width) { - uint8 b0; - uint8 g0; - uint8 r0; - uint8 b1; - uint8 g1; - uint8 r1; - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], &b0, &g0, &r0); - YuvPixel(src_y[1], src_u[0], src_v[0], &b1, &g1, &r1); - b0 = b0 >> 3; - g0 = g0 >> 3; - r0 = r0 >> 3; - b1 = b1 >> 3; - g1 = g1 >> 3; - r1 = r1 >> 3; - *(uint32*)(dst_argb1555) = b0 | (g0 << 5) | (r0 << 10) | - (b1 << 16) | (g1 << 21) | (r1 << 26) | 0x80008000; - src_y += 2; - src_u += 1; - src_v += 1; - dst_argb1555 += 4; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], &b0, &g0, &r0); - b0 = b0 >> 3; - g0 = g0 >> 3; - r0 = r0 >> 3; - *(uint16*)(dst_argb1555) = b0 | (g0 << 5) | (r0 << 10) | - 0x8000; - } -} - -void I422ToRGB565Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb565, - int width) { - uint8 b0; - uint8 g0; - uint8 r0; - uint8 b1; - uint8 g1; - uint8 r1; - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], &b0, &g0, &r0); - YuvPixel(src_y[1], src_u[0], src_v[0], &b1, &g1, &r1); - b0 = b0 >> 3; - g0 = g0 >> 2; - r0 = r0 >> 3; - b1 = b1 >> 3; - g1 = g1 >> 2; - r1 = r1 >> 3; - *(uint32*)(dst_rgb565) = b0 | (g0 << 5) | (r0 << 11) | - (b1 << 16) | (g1 << 21) | (r1 << 27); - src_y += 2; - src_u += 1; - src_v += 1; - dst_rgb565 += 4; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], &b0, &g0, &r0); - b0 = b0 >> 3; - g0 = g0 >> 2; - r0 = r0 >> 3; - *(uint16*)(dst_rgb565) = b0 | (g0 << 5) | (r0 << 11); - } -} - -void I411ToARGBRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 3; x += 4) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - YuvPixel(src_y[2], src_u[0], src_v[0], - rgb_buf + 8, rgb_buf + 9, rgb_buf + 10); - rgb_buf[11] = 255; - YuvPixel(src_y[3], src_u[0], src_v[0], - rgb_buf + 12, rgb_buf + 13, rgb_buf + 14); - rgb_buf[15] = 255; - src_y += 4; - src_u += 1; - src_v += 1; - rgb_buf += 16; // Advance 4 pixels. - } - if (width & 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_y += 2; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void NV12ToARGBRow_C(const uint8* src_y, - const uint8* usrc_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], usrc_v[0], usrc_v[1], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_y[1], usrc_v[0], usrc_v[1], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_y += 2; - usrc_v += 2; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], usrc_v[0], usrc_v[1], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void NV21ToARGBRow_C(const uint8* src_y, - const uint8* src_vu, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_vu[1], src_vu[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - - YuvPixel(src_y[1], src_vu[1], src_vu[0], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - - src_y += 2; - src_vu += 2; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_vu[1], src_vu[0], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void NV12ToRGB565Row_C(const uint8* src_y, - const uint8* usrc_v, - uint8* dst_rgb565, - int width) { - uint8 b0; - uint8 g0; - uint8 r0; - uint8 b1; - uint8 g1; - uint8 r1; - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], usrc_v[0], usrc_v[1], &b0, &g0, &r0); - YuvPixel(src_y[1], usrc_v[0], usrc_v[1], &b1, &g1, &r1); - b0 = b0 >> 3; - g0 = g0 >> 2; - r0 = r0 >> 3; - b1 = b1 >> 3; - g1 = g1 >> 2; - r1 = r1 >> 3; - *(uint32*)(dst_rgb565) = b0 | (g0 << 5) | (r0 << 11) | - (b1 << 16) | (g1 << 21) | (r1 << 27); - src_y += 2; - usrc_v += 2; - dst_rgb565 += 4; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], usrc_v[0], usrc_v[1], &b0, &g0, &r0); - b0 = b0 >> 3; - g0 = g0 >> 2; - r0 = r0 >> 3; - *(uint16*)(dst_rgb565) = b0 | (g0 << 5) | (r0 << 11); - } -} - -void NV21ToRGB565Row_C(const uint8* src_y, - const uint8* vsrc_u, - uint8* dst_rgb565, - int width) { - uint8 b0; - uint8 g0; - uint8 r0; - uint8 b1; - uint8 g1; - uint8 r1; - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], vsrc_u[1], vsrc_u[0], &b0, &g0, &r0); - YuvPixel(src_y[1], vsrc_u[1], vsrc_u[0], &b1, &g1, &r1); - b0 = b0 >> 3; - g0 = g0 >> 2; - r0 = r0 >> 3; - b1 = b1 >> 3; - g1 = g1 >> 2; - r1 = r1 >> 3; - *(uint32*)(dst_rgb565) = b0 | (g0 << 5) | (r0 << 11) | - (b1 << 16) | (g1 << 21) | (r1 << 27); - src_y += 2; - vsrc_u += 2; - dst_rgb565 += 4; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], vsrc_u[1], vsrc_u[0], &b0, &g0, &r0); - b0 = b0 >> 3; - g0 = g0 >> 2; - r0 = r0 >> 3; - *(uint16*)(dst_rgb565) = b0 | (g0 << 5) | (r0 << 11); - } -} - -void YUY2ToARGBRow_C(const uint8* src_yuy2, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_yuy2[0], src_yuy2[1], src_yuy2[3], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_yuy2[2], src_yuy2[1], src_yuy2[3], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_yuy2 += 4; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_yuy2[0], src_yuy2[1], src_yuy2[3], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void UYVYToARGBRow_C(const uint8* src_uyvy, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_uyvy[1], src_uyvy[0], src_uyvy[2], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_uyvy[3], src_uyvy[0], src_uyvy[2], - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_uyvy += 4; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_uyvy[1], src_uyvy[0], src_uyvy[2], - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void I422ToBGRARow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 3, rgb_buf + 2, rgb_buf + 1); - rgb_buf[0] = 255; - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 7, rgb_buf + 6, rgb_buf + 5); - rgb_buf[4] = 255; - src_y += 2; - src_u += 1; - src_v += 1; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 3, rgb_buf + 2, rgb_buf + 1); - rgb_buf[0] = 255; - } -} - -void I422ToABGRRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 2, rgb_buf + 1, rgb_buf + 0); - rgb_buf[3] = 255; - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 6, rgb_buf + 5, rgb_buf + 4); - rgb_buf[7] = 255; - src_y += 2; - src_u += 1; - src_v += 1; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 2, rgb_buf + 1, rgb_buf + 0); - rgb_buf[3] = 255; - } -} - -void I422ToRGBARow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 1, rgb_buf + 2, rgb_buf + 3); - rgb_buf[0] = 255; - YuvPixel(src_y[1], src_u[0], src_v[0], - rgb_buf + 5, rgb_buf + 6, rgb_buf + 7); - rgb_buf[4] = 255; - src_y += 2; - src_u += 1; - src_v += 1; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], src_u[0], src_v[0], - rgb_buf + 1, rgb_buf + 2, rgb_buf + 3); - rgb_buf[0] = 255; - } -} - -void YToARGBRow_C(const uint8* src_y, uint8* rgb_buf, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - YuvPixel(src_y[0], 128, 128, - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - YuvPixel(src_y[1], 128, 128, - rgb_buf + 4, rgb_buf + 5, rgb_buf + 6); - rgb_buf[7] = 255; - src_y += 2; - rgb_buf += 8; // Advance 2 pixels. - } - if (width & 1) { - YuvPixel(src_y[0], 128, 128, - rgb_buf + 0, rgb_buf + 1, rgb_buf + 2); - rgb_buf[3] = 255; - } -} - -void MirrorRow_C(const uint8* src, uint8* dst, int width) { - int x; - src += width - 1; - for (x = 0; x < width - 1; x += 2) { - dst[x] = src[0]; - dst[x + 1] = src[-1]; - src -= 2; - } - if (width & 1) { - dst[width - 1] = src[0]; - } -} - -void MirrorUVRow_C(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int width) { - int x; - src_uv += (width - 1) << 1; - for (x = 0; x < width - 1; x += 2) { - dst_u[x] = src_uv[0]; - dst_u[x + 1] = src_uv[-2]; - dst_v[x] = src_uv[1]; - dst_v[x + 1] = src_uv[-2 + 1]; - src_uv -= 4; - } - if (width & 1) { - dst_u[width - 1] = src_uv[0]; - dst_v[width - 1] = src_uv[1]; - } -} - -void ARGBMirrorRow_C(const uint8* src, uint8* dst, int width) { - int x; - const uint32* src32 = (const uint32*)(src); - uint32* dst32 = (uint32*)(dst); - src32 += width - 1; - for (x = 0; x < width - 1; x += 2) { - dst32[x] = src32[0]; - dst32[x + 1] = src32[-1]; - src32 -= 2; - } - if (width & 1) { - dst32[width - 1] = src32[0]; - } -} - -void SplitUVRow_C(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - dst_u[x] = src_uv[0]; - dst_u[x + 1] = src_uv[2]; - dst_v[x] = src_uv[1]; - dst_v[x + 1] = src_uv[3]; - src_uv += 4; - } - if (width & 1) { - dst_u[width - 1] = src_uv[0]; - dst_v[width - 1] = src_uv[1]; - } -} - -void MergeUVRow_C(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - dst_uv[0] = src_u[x]; - dst_uv[1] = src_v[x]; - dst_uv[2] = src_u[x + 1]; - dst_uv[3] = src_v[x + 1]; - dst_uv += 4; - } - if (width & 1) { - dst_uv[0] = src_u[width - 1]; - dst_uv[1] = src_v[width - 1]; - } -} - -void CopyRow_C(const uint8* src, uint8* dst, int count) { - memcpy(dst, src, count); -} - -void CopyRow_16_C(const uint16* src, uint16* dst, int count) { - memcpy(dst, src, count * 2); -} - -void SetRow_C(uint8* dst, uint32 v8, int count) { -#ifdef _MSC_VER - // VC will generate rep stosb. - int x; - for (x = 0; x < count; ++x) { - dst[x] = v8; - } -#else - memset(dst, v8, count); -#endif -} - -void ARGBSetRows_C(uint8* dst, uint32 v32, int width, - int dst_stride, int height) { - int y; - for (y = 0; y < height; ++y) { - uint32* d = (uint32*)(dst); - int x; - for (x = 0; x < width; ++x) { - d[x] = v32; - } - dst += dst_stride; - } -} - -// Filter 2 rows of YUY2 UV's (422) into U and V (420). -void YUY2ToUVRow_C(const uint8* src_yuy2, int src_stride_yuy2, - uint8* dst_u, uint8* dst_v, int width) { - // Output a row of UV values, filtering 2 rows of YUY2. - int x; - for (x = 0; x < width; x += 2) { - dst_u[0] = (src_yuy2[1] + src_yuy2[src_stride_yuy2 + 1] + 1) >> 1; - dst_v[0] = (src_yuy2[3] + src_yuy2[src_stride_yuy2 + 3] + 1) >> 1; - src_yuy2 += 4; - dst_u += 1; - dst_v += 1; - } -} - -// Copy row of YUY2 UV's (422) into U and V (422). -void YUY2ToUV422Row_C(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int width) { - // Output a row of UV values. - int x; - for (x = 0; x < width; x += 2) { - dst_u[0] = src_yuy2[1]; - dst_v[0] = src_yuy2[3]; - src_yuy2 += 4; - dst_u += 1; - dst_v += 1; - } -} - -// Copy row of YUY2 Y's (422) into Y (420/422). -void YUY2ToYRow_C(const uint8* src_yuy2, uint8* dst_y, int width) { - // Output a row of Y values. - int x; - for (x = 0; x < width - 1; x += 2) { - dst_y[x] = src_yuy2[0]; - dst_y[x + 1] = src_yuy2[2]; - src_yuy2 += 4; - } - if (width & 1) { - dst_y[width - 1] = src_yuy2[0]; - } -} - -// Filter 2 rows of UYVY UV's (422) into U and V (420). -void UYVYToUVRow_C(const uint8* src_uyvy, int src_stride_uyvy, - uint8* dst_u, uint8* dst_v, int width) { - // Output a row of UV values. - int x; - for (x = 0; x < width; x += 2) { - dst_u[0] = (src_uyvy[0] + src_uyvy[src_stride_uyvy + 0] + 1) >> 1; - dst_v[0] = (src_uyvy[2] + src_uyvy[src_stride_uyvy + 2] + 1) >> 1; - src_uyvy += 4; - dst_u += 1; - dst_v += 1; - } -} - -// Copy row of UYVY UV's (422) into U and V (422). -void UYVYToUV422Row_C(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int width) { - // Output a row of UV values. - int x; - for (x = 0; x < width; x += 2) { - dst_u[0] = src_uyvy[0]; - dst_v[0] = src_uyvy[2]; - src_uyvy += 4; - dst_u += 1; - dst_v += 1; - } -} - -// Copy row of UYVY Y's (422) into Y (420/422). -void UYVYToYRow_C(const uint8* src_uyvy, uint8* dst_y, int width) { - // Output a row of Y values. - int x; - for (x = 0; x < width - 1; x += 2) { - dst_y[x] = src_uyvy[1]; - dst_y[x + 1] = src_uyvy[3]; - src_uyvy += 4; - } - if (width & 1) { - dst_y[width - 1] = src_uyvy[1]; - } -} - -#define BLEND(f, b, a) (((256 - a) * b) >> 8) + f - -// Blend src_argb0 over src_argb1 and store to dst_argb. -// dst_argb may be src_argb0 or src_argb1. -// This code mimics the SSSE3 version for better testability. -void ARGBBlendRow_C(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - uint32 fb = src_argb0[0]; - uint32 fg = src_argb0[1]; - uint32 fr = src_argb0[2]; - uint32 a = src_argb0[3]; - uint32 bb = src_argb1[0]; - uint32 bg = src_argb1[1]; - uint32 br = src_argb1[2]; - dst_argb[0] = BLEND(fb, bb, a); - dst_argb[1] = BLEND(fg, bg, a); - dst_argb[2] = BLEND(fr, br, a); - dst_argb[3] = 255u; - - fb = src_argb0[4 + 0]; - fg = src_argb0[4 + 1]; - fr = src_argb0[4 + 2]; - a = src_argb0[4 + 3]; - bb = src_argb1[4 + 0]; - bg = src_argb1[4 + 1]; - br = src_argb1[4 + 2]; - dst_argb[4 + 0] = BLEND(fb, bb, a); - dst_argb[4 + 1] = BLEND(fg, bg, a); - dst_argb[4 + 2] = BLEND(fr, br, a); - dst_argb[4 + 3] = 255u; - src_argb0 += 8; - src_argb1 += 8; - dst_argb += 8; - } - - if (width & 1) { - uint32 fb = src_argb0[0]; - uint32 fg = src_argb0[1]; - uint32 fr = src_argb0[2]; - uint32 a = src_argb0[3]; - uint32 bb = src_argb1[0]; - uint32 bg = src_argb1[1]; - uint32 br = src_argb1[2]; - dst_argb[0] = BLEND(fb, bb, a); - dst_argb[1] = BLEND(fg, bg, a); - dst_argb[2] = BLEND(fr, br, a); - dst_argb[3] = 255u; - } -} -#undef BLEND -#define ATTENUATE(f, a) (a | (a << 8)) * (f | (f << 8)) >> 24 - -// Multiply source RGB by alpha and store to destination. -// This code mimics the SSSE3 version for better testability. -void ARGBAttenuateRow_C(const uint8* src_argb, uint8* dst_argb, int width) { - int i; - for (i = 0; i < width - 1; i += 2) { - uint32 b = src_argb[0]; - uint32 g = src_argb[1]; - uint32 r = src_argb[2]; - uint32 a = src_argb[3]; - dst_argb[0] = ATTENUATE(b, a); - dst_argb[1] = ATTENUATE(g, a); - dst_argb[2] = ATTENUATE(r, a); - dst_argb[3] = a; - b = src_argb[4]; - g = src_argb[5]; - r = src_argb[6]; - a = src_argb[7]; - dst_argb[4] = ATTENUATE(b, a); - dst_argb[5] = ATTENUATE(g, a); - dst_argb[6] = ATTENUATE(r, a); - dst_argb[7] = a; - src_argb += 8; - dst_argb += 8; - } - - if (width & 1) { - const uint32 b = src_argb[0]; - const uint32 g = src_argb[1]; - const uint32 r = src_argb[2]; - const uint32 a = src_argb[3]; - dst_argb[0] = ATTENUATE(b, a); - dst_argb[1] = ATTENUATE(g, a); - dst_argb[2] = ATTENUATE(r, a); - dst_argb[3] = a; - } -} -#undef ATTENUATE - -// Divide source RGB by alpha and store to destination. -// b = (b * 255 + (a / 2)) / a; -// g = (g * 255 + (a / 2)) / a; -// r = (r * 255 + (a / 2)) / a; -// Reciprocal method is off by 1 on some values. ie 125 -// 8.8 fixed point inverse table with 1.0 in upper short and 1 / a in lower. -#define T(a) 0x01000000 + (0x10000 / a) -const uint32 fixed_invtbl8[256] = { - 0x01000000, 0x0100ffff, T(0x02), T(0x03), T(0x04), T(0x05), T(0x06), T(0x07), - T(0x08), T(0x09), T(0x0a), T(0x0b), T(0x0c), T(0x0d), T(0x0e), T(0x0f), - T(0x10), T(0x11), T(0x12), T(0x13), T(0x14), T(0x15), T(0x16), T(0x17), - T(0x18), T(0x19), T(0x1a), T(0x1b), T(0x1c), T(0x1d), T(0x1e), T(0x1f), - T(0x20), T(0x21), T(0x22), T(0x23), T(0x24), T(0x25), T(0x26), T(0x27), - T(0x28), T(0x29), T(0x2a), T(0x2b), T(0x2c), T(0x2d), T(0x2e), T(0x2f), - T(0x30), T(0x31), T(0x32), T(0x33), T(0x34), T(0x35), T(0x36), T(0x37), - T(0x38), T(0x39), T(0x3a), T(0x3b), T(0x3c), T(0x3d), T(0x3e), T(0x3f), - T(0x40), T(0x41), T(0x42), T(0x43), T(0x44), T(0x45), T(0x46), T(0x47), - T(0x48), T(0x49), T(0x4a), T(0x4b), T(0x4c), T(0x4d), T(0x4e), T(0x4f), - T(0x50), T(0x51), T(0x52), T(0x53), T(0x54), T(0x55), T(0x56), T(0x57), - T(0x58), T(0x59), T(0x5a), T(0x5b), T(0x5c), T(0x5d), T(0x5e), T(0x5f), - T(0x60), T(0x61), T(0x62), T(0x63), T(0x64), T(0x65), T(0x66), T(0x67), - T(0x68), T(0x69), T(0x6a), T(0x6b), T(0x6c), T(0x6d), T(0x6e), T(0x6f), - T(0x70), T(0x71), T(0x72), T(0x73), T(0x74), T(0x75), T(0x76), T(0x77), - T(0x78), T(0x79), T(0x7a), T(0x7b), T(0x7c), T(0x7d), T(0x7e), T(0x7f), - T(0x80), T(0x81), T(0x82), T(0x83), T(0x84), T(0x85), T(0x86), T(0x87), - T(0x88), T(0x89), T(0x8a), T(0x8b), T(0x8c), T(0x8d), T(0x8e), T(0x8f), - T(0x90), T(0x91), T(0x92), T(0x93), T(0x94), T(0x95), T(0x96), T(0x97), - T(0x98), T(0x99), T(0x9a), T(0x9b), T(0x9c), T(0x9d), T(0x9e), T(0x9f), - T(0xa0), T(0xa1), T(0xa2), T(0xa3), T(0xa4), T(0xa5), T(0xa6), T(0xa7), - T(0xa8), T(0xa9), T(0xaa), T(0xab), T(0xac), T(0xad), T(0xae), T(0xaf), - T(0xb0), T(0xb1), T(0xb2), T(0xb3), T(0xb4), T(0xb5), T(0xb6), T(0xb7), - T(0xb8), T(0xb9), T(0xba), T(0xbb), T(0xbc), T(0xbd), T(0xbe), T(0xbf), - T(0xc0), T(0xc1), T(0xc2), T(0xc3), T(0xc4), T(0xc5), T(0xc6), T(0xc7), - T(0xc8), T(0xc9), T(0xca), T(0xcb), T(0xcc), T(0xcd), T(0xce), T(0xcf), - T(0xd0), T(0xd1), T(0xd2), T(0xd3), T(0xd4), T(0xd5), T(0xd6), T(0xd7), - T(0xd8), T(0xd9), T(0xda), T(0xdb), T(0xdc), T(0xdd), T(0xde), T(0xdf), - T(0xe0), T(0xe1), T(0xe2), T(0xe3), T(0xe4), T(0xe5), T(0xe6), T(0xe7), - T(0xe8), T(0xe9), T(0xea), T(0xeb), T(0xec), T(0xed), T(0xee), T(0xef), - T(0xf0), T(0xf1), T(0xf2), T(0xf3), T(0xf4), T(0xf5), T(0xf6), T(0xf7), - T(0xf8), T(0xf9), T(0xfa), T(0xfb), T(0xfc), T(0xfd), T(0xfe), 0x01000100 }; -#undef T - -void ARGBUnattenuateRow_C(const uint8* src_argb, uint8* dst_argb, int width) { - int i; - for (i = 0; i < width; ++i) { - uint32 b = src_argb[0]; - uint32 g = src_argb[1]; - uint32 r = src_argb[2]; - const uint32 a = src_argb[3]; - const uint32 ia = fixed_invtbl8[a] & 0xffff; // 8.8 fixed point - b = (b * ia) >> 8; - g = (g * ia) >> 8; - r = (r * ia) >> 8; - // Clamping should not be necessary but is free in assembly. - dst_argb[0] = clamp255(b); - dst_argb[1] = clamp255(g); - dst_argb[2] = clamp255(r); - dst_argb[3] = a; - src_argb += 4; - dst_argb += 4; - } -} - -void ComputeCumulativeSumRow_C(const uint8* row, int32* cumsum, - const int32* previous_cumsum, int width) { - int32 row_sum[4] = {0, 0, 0, 0}; - int x; - for (x = 0; x < width; ++x) { - row_sum[0] += row[x * 4 + 0]; - row_sum[1] += row[x * 4 + 1]; - row_sum[2] += row[x * 4 + 2]; - row_sum[3] += row[x * 4 + 3]; - cumsum[x * 4 + 0] = row_sum[0] + previous_cumsum[x * 4 + 0]; - cumsum[x * 4 + 1] = row_sum[1] + previous_cumsum[x * 4 + 1]; - cumsum[x * 4 + 2] = row_sum[2] + previous_cumsum[x * 4 + 2]; - cumsum[x * 4 + 3] = row_sum[3] + previous_cumsum[x * 4 + 3]; - } -} - -void CumulativeSumToAverageRow_C(const int32* tl, const int32* bl, - int w, int area, uint8* dst, int count) { - float ooa = 1.0f / area; - int i; - for (i = 0; i < count; ++i) { - dst[0] = (uint8)((bl[w + 0] + tl[0] - bl[0] - tl[w + 0]) * ooa); - dst[1] = (uint8)((bl[w + 1] + tl[1] - bl[1] - tl[w + 1]) * ooa); - dst[2] = (uint8)((bl[w + 2] + tl[2] - bl[2] - tl[w + 2]) * ooa); - dst[3] = (uint8)((bl[w + 3] + tl[3] - bl[3] - tl[w + 3]) * ooa); - dst += 4; - tl += 4; - bl += 4; - } -} - -// Copy pixels from rotated source to destination row with a slope. -LIBYUV_API -void ARGBAffineRow_C(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* uv_dudv, int width) { - int i; - // Render a row of pixels from source into a buffer. - float uv[2]; - uv[0] = uv_dudv[0]; - uv[1] = uv_dudv[1]; - for (i = 0; i < width; ++i) { - int x = (int)(uv[0]); - int y = (int)(uv[1]); - *(uint32*)(dst_argb) = - *(const uint32*)(src_argb + y * src_argb_stride + - x * 4); - dst_argb += 4; - uv[0] += uv_dudv[2]; - uv[1] += uv_dudv[3]; - } -} - -// Blend 2 rows into 1 for conversions such as I422ToI420. -void HalfRow_C(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix) { - int x; - for (x = 0; x < pix; ++x) { - dst_uv[x] = (src_uv[x] + src_uv[src_uv_stride + x] + 1) >> 1; - } -} - -void HalfRow_16_C(const uint16* src_uv, int src_uv_stride, - uint16* dst_uv, int pix) { - int x; - for (x = 0; x < pix; ++x) { - dst_uv[x] = (src_uv[x] + src_uv[src_uv_stride + x] + 1) >> 1; - } -} - -// C version 2x2 -> 2x1. -void InterpolateRow_C(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, - int width, int source_y_fraction) { - int y1_fraction = source_y_fraction; - int y0_fraction = 256 - y1_fraction; - const uint8* src_ptr1 = src_ptr + src_stride; - int x; - if (source_y_fraction == 0) { - memcpy(dst_ptr, src_ptr, width); - return; - } - if (source_y_fraction == 128) { - HalfRow_C(src_ptr, (int)(src_stride), dst_ptr, width); - return; - } - for (x = 0; x < width - 1; x += 2) { - dst_ptr[0] = (src_ptr[0] * y0_fraction + src_ptr1[0] * y1_fraction) >> 8; - dst_ptr[1] = (src_ptr[1] * y0_fraction + src_ptr1[1] * y1_fraction) >> 8; - src_ptr += 2; - src_ptr1 += 2; - dst_ptr += 2; - } - if (width & 1) { - dst_ptr[0] = (src_ptr[0] * y0_fraction + src_ptr1[0] * y1_fraction) >> 8; - } -} - -void InterpolateRow_16_C(uint16* dst_ptr, const uint16* src_ptr, - ptrdiff_t src_stride, - int width, int source_y_fraction) { - int y1_fraction = source_y_fraction; - int y0_fraction = 256 - y1_fraction; - const uint16* src_ptr1 = src_ptr + src_stride; - int x; - if (source_y_fraction == 0) { - memcpy(dst_ptr, src_ptr, width * 2); - return; - } - if (source_y_fraction == 128) { - HalfRow_16_C(src_ptr, (int)(src_stride), dst_ptr, width); - return; - } - for (x = 0; x < width - 1; x += 2) { - dst_ptr[0] = (src_ptr[0] * y0_fraction + src_ptr1[0] * y1_fraction) >> 8; - dst_ptr[1] = (src_ptr[1] * y0_fraction + src_ptr1[1] * y1_fraction) >> 8; - src_ptr += 2; - src_ptr1 += 2; - dst_ptr += 2; - } - if (width & 1) { - dst_ptr[0] = (src_ptr[0] * y0_fraction + src_ptr1[0] * y1_fraction) >> 8; - } -} - -// Select 2 channels from ARGB on alternating pixels. e.g. BGBGBGBG -void ARGBToBayerRow_C(const uint8* src_argb, - uint8* dst_bayer, uint32 selector, int pix) { - int index0 = selector & 0xff; - int index1 = (selector >> 8) & 0xff; - // Copy a row of Bayer. - int x; - for (x = 0; x < pix - 1; x += 2) { - dst_bayer[0] = src_argb[index0]; - dst_bayer[1] = src_argb[index1]; - src_argb += 8; - dst_bayer += 2; - } - if (pix & 1) { - dst_bayer[0] = src_argb[index0]; - } -} - -// Select G channel from ARGB. e.g. GGGGGGGG -void ARGBToBayerGGRow_C(const uint8* src_argb, - uint8* dst_bayer, uint32 selector, int pix) { - // Copy a row of G. - int x; - for (x = 0; x < pix - 1; x += 2) { - dst_bayer[0] = src_argb[1]; - dst_bayer[1] = src_argb[5]; - src_argb += 8; - dst_bayer += 2; - } - if (pix & 1) { - dst_bayer[0] = src_argb[1]; - } -} - -// Use first 4 shuffler values to reorder ARGB channels. -void ARGBShuffleRow_C(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - int index0 = shuffler[0]; - int index1 = shuffler[1]; - int index2 = shuffler[2]; - int index3 = shuffler[3]; - // Shuffle a row of ARGB. - int x; - for (x = 0; x < pix; ++x) { - // To support in-place conversion. - uint8 b = src_argb[index0]; - uint8 g = src_argb[index1]; - uint8 r = src_argb[index2]; - uint8 a = src_argb[index3]; - dst_argb[0] = b; - dst_argb[1] = g; - dst_argb[2] = r; - dst_argb[3] = a; - src_argb += 4; - dst_argb += 4; - } -} - -void I422ToYUY2Row_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_frame, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - dst_frame[0] = src_y[0]; - dst_frame[1] = src_u[0]; - dst_frame[2] = src_y[1]; - dst_frame[3] = src_v[0]; - dst_frame += 4; - src_y += 2; - src_u += 1; - src_v += 1; - } - if (width & 1) { - dst_frame[0] = src_y[0]; - dst_frame[1] = src_u[0]; - dst_frame[2] = src_y[0]; // duplicate last y - dst_frame[3] = src_v[0]; - } -} - -void I422ToUYVYRow_C(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_frame, int width) { - int x; - for (x = 0; x < width - 1; x += 2) { - dst_frame[0] = src_u[0]; - dst_frame[1] = src_y[0]; - dst_frame[2] = src_v[0]; - dst_frame[3] = src_y[1]; - dst_frame += 4; - src_y += 2; - src_u += 1; - src_v += 1; - } - if (width & 1) { - dst_frame[0] = src_u[0]; - dst_frame[1] = src_y[0]; - dst_frame[2] = src_v[0]; - dst_frame[3] = src_y[0]; // duplicate last y - } -} - -#if !defined(LIBYUV_DISABLE_X86) && defined(HAS_I422TOARGBROW_SSSE3) -// row_win.cc has asm version, but GCC uses 2 step wrapper. -#if !defined(_MSC_VER) && (defined(__x86_64__) || defined(__i386__)) -void I422ToRGB565Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - // Allocate a row of ARGB. - align_buffer_64(row, width * 4); - I422ToARGBRow_SSSE3(src_y, src_u, src_v, row, width); - ARGBToRGB565Row_SSE2(row, rgb_buf, width); - free_aligned_buffer_64(row); -} -#endif // !defined(_MSC_VER) && (defined(__x86_64__) || defined(__i386__)) - -#if defined(_M_IX86) || defined(__x86_64__) || defined(__i386__) -void I422ToARGB1555Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - // Allocate a row of ARGB. - align_buffer_64(row, width * 4); - I422ToARGBRow_SSSE3(src_y, src_u, src_v, row, width); - ARGBToARGB1555Row_SSE2(row, rgb_buf, width); - free_aligned_buffer_64(row); -} - -void I422ToARGB4444Row_SSSE3(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* rgb_buf, - int width) { - // Allocate a row of ARGB. - align_buffer_64(row, width * 4); - I422ToARGBRow_SSSE3(src_y, src_u, src_v, row, width); - ARGBToARGB4444Row_SSE2(row, rgb_buf, width); - free_aligned_buffer_64(row); -} - -void NV12ToRGB565Row_SSSE3(const uint8* src_y, - const uint8* src_uv, - uint8* dst_rgb565, - int width) { - // Allocate a row of ARGB. - align_buffer_64(row, width * 4); - NV12ToARGBRow_SSSE3(src_y, src_uv, row, width); - ARGBToRGB565Row_SSE2(row, dst_rgb565, width); - free_aligned_buffer_64(row); -} - -void NV21ToRGB565Row_SSSE3(const uint8* src_y, - const uint8* src_vu, - uint8* dst_rgb565, - int width) { - // Allocate a row of ARGB. - align_buffer_64(row, width * 4); - NV21ToARGBRow_SSSE3(src_y, src_vu, row, width); - ARGBToRGB565Row_SSE2(row, dst_rgb565, width); - free_aligned_buffer_64(row); -} - -void YUY2ToARGBRow_SSSE3(const uint8* src_yuy2, - uint8* dst_argb, - int width) { - // Allocate a rows of yuv. - align_buffer_64(row_y, ((width + 63) & ~63) * 2); - uint8* row_u = row_y + ((width + 63) & ~63); - uint8* row_v = row_u + ((width + 63) & ~63) / 2; - YUY2ToUV422Row_SSE2(src_yuy2, row_u, row_v, width); - YUY2ToYRow_SSE2(src_yuy2, row_y, width); - I422ToARGBRow_SSSE3(row_y, row_u, row_v, dst_argb, width); - free_aligned_buffer_64(row_y); -} - -void YUY2ToARGBRow_Unaligned_SSSE3(const uint8* src_yuy2, - uint8* dst_argb, - int width) { - // Allocate a rows of yuv. - align_buffer_64(row_y, ((width + 63) & ~63) * 2); - uint8* row_u = row_y + ((width + 63) & ~63); - uint8* row_v = row_u + ((width + 63) & ~63) / 2; - YUY2ToUV422Row_Unaligned_SSE2(src_yuy2, row_u, row_v, width); - YUY2ToYRow_Unaligned_SSE2(src_yuy2, row_y, width); - I422ToARGBRow_Unaligned_SSSE3(row_y, row_u, row_v, dst_argb, width); - free_aligned_buffer_64(row_y); -} - -void UYVYToARGBRow_SSSE3(const uint8* src_uyvy, - uint8* dst_argb, - int width) { - // Allocate a rows of yuv. - align_buffer_64(row_y, ((width + 63) & ~63) * 2); - uint8* row_u = row_y + ((width + 63) & ~63); - uint8* row_v = row_u + ((width + 63) & ~63) / 2; - UYVYToUV422Row_SSE2(src_uyvy, row_u, row_v, width); - UYVYToYRow_SSE2(src_uyvy, row_y, width); - I422ToARGBRow_SSSE3(row_y, row_u, row_v, dst_argb, width); - free_aligned_buffer_64(row_y); -} - -void UYVYToARGBRow_Unaligned_SSSE3(const uint8* src_uyvy, - uint8* dst_argb, - int width) { - // Allocate a rows of yuv. - align_buffer_64(row_y, ((width + 63) & ~63) * 2); - uint8* row_u = row_y + ((width + 63) & ~63); - uint8* row_v = row_u + ((width + 63) & ~63) / 2; - UYVYToUV422Row_Unaligned_SSE2(src_uyvy, row_u, row_v, width); - UYVYToYRow_Unaligned_SSE2(src_uyvy, row_y, width); - I422ToARGBRow_Unaligned_SSSE3(row_y, row_u, row_v, dst_argb, width); - free_aligned_buffer_64(row_y); -} - -#endif // defined(_M_IX86) || defined(__x86_64__) || defined(__i386__) -#endif // !defined(LIBYUV_DISABLE_X86) - -void ARGBPolynomialRow_C(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width) { - int i; - for (i = 0; i < width; ++i) { - float b = (float)(src_argb[0]); - float g = (float)(src_argb[1]); - float r = (float)(src_argb[2]); - float a = (float)(src_argb[3]); - float b2 = b * b; - float g2 = g * g; - float r2 = r * r; - float a2 = a * a; - float db = poly[0] + poly[4] * b; - float dg = poly[1] + poly[5] * g; - float dr = poly[2] + poly[6] * r; - float da = poly[3] + poly[7] * a; - float b3 = b2 * b; - float g3 = g2 * g; - float r3 = r2 * r; - float a3 = a2 * a; - db += poly[8] * b2; - dg += poly[9] * g2; - dr += poly[10] * r2; - da += poly[11] * a2; - db += poly[12] * b3; - dg += poly[13] * g3; - dr += poly[14] * r3; - da += poly[15] * a3; - - dst_argb[0] = Clamp((int32)(db)); - dst_argb[1] = Clamp((int32)(dg)); - dst_argb[2] = Clamp((int32)(dr)); - dst_argb[3] = Clamp((int32)(da)); - src_argb += 4; - dst_argb += 4; - } -} - -void ARGBLumaColorTableRow_C(const uint8* src_argb, uint8* dst_argb, int width, - const uint8* luma, uint32 lumacoeff) { - uint32 bc = lumacoeff & 0xff; - uint32 gc = (lumacoeff >> 8) & 0xff; - uint32 rc = (lumacoeff >> 16) & 0xff; - - int i; - for (i = 0; i < width - 1; i += 2) { - // Luminance in rows, color values in columns. - const uint8* luma0 = ((src_argb[0] * bc + src_argb[1] * gc + - src_argb[2] * rc) & 0x7F00u) + luma; - const uint8* luma1; - dst_argb[0] = luma0[src_argb[0]]; - dst_argb[1] = luma0[src_argb[1]]; - dst_argb[2] = luma0[src_argb[2]]; - dst_argb[3] = src_argb[3]; - luma1 = ((src_argb[4] * bc + src_argb[5] * gc + - src_argb[6] * rc) & 0x7F00u) + luma; - dst_argb[4] = luma1[src_argb[4]]; - dst_argb[5] = luma1[src_argb[5]]; - dst_argb[6] = luma1[src_argb[6]]; - dst_argb[7] = src_argb[7]; - src_argb += 8; - dst_argb += 8; - } - if (width & 1) { - // Luminance in rows, color values in columns. - const uint8* luma0 = ((src_argb[0] * bc + src_argb[1] * gc + - src_argb[2] * rc) & 0x7F00u) + luma; - dst_argb[0] = luma0[src_argb[0]]; - dst_argb[1] = luma0[src_argb[1]]; - dst_argb[2] = luma0[src_argb[2]]; - dst_argb[3] = src_argb[3]; - } -} - -void ARGBCopyAlphaRow_C(const uint8* src, uint8* dst, int width) { - int i; - for (i = 0; i < width - 1; i += 2) { - dst[3] = src[3]; - dst[7] = src[7]; - dst += 8; - src += 8; - } - if (width & 1) { - dst[3] = src[3]; - } -} - -void ARGBCopyYToAlphaRow_C(const uint8* src, uint8* dst, int width) { - int i; - for (i = 0; i < width - 1; i += 2) { - dst[3] = src[0]; - dst[7] = src[1]; - dst += 8; - src += 2; - } - if (width & 1) { - dst[3] = src[0]; - } -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_mips.cc b/src/main/jni/libyuv/source/row_mips.cc deleted file mode 100644 index da7183bc1..000000000 --- a/src/main/jni/libyuv/source/row_mips.cc +++ /dev/null @@ -1,994 +0,0 @@ -/* - * Copyright (c) 2012 The LibYuv project authors. All Rights Reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// The following are available on Mips platforms: -#if !defined(LIBYUV_DISABLE_MIPS) && defined(__mips__) && \ - (_MIPS_SIM == _MIPS_SIM_ABI32) - -#ifdef HAS_COPYROW_MIPS -void CopyRow_MIPS(const uint8* src, uint8* dst, int count) { - __asm__ __volatile__ ( - ".set noreorder \n" - ".set noat \n" - "slti $at, %[count], 8 \n" - "bne $at ,$zero, $last8 \n" - "xor $t8, %[src], %[dst] \n" - "andi $t8, $t8, 0x3 \n" - - "bne $t8, $zero, unaligned \n" - "negu $a3, %[dst] \n" - // make dst/src aligned - "andi $a3, $a3, 0x3 \n" - "beq $a3, $zero, $chk16w \n" - // word-aligned now count is the remining bytes count - "subu %[count], %[count], $a3 \n" - - "lwr $t8, 0(%[src]) \n" - "addu %[src], %[src], $a3 \n" - "swr $t8, 0(%[dst]) \n" - "addu %[dst], %[dst], $a3 \n" - - // Now the dst/src are mutually word-aligned with word-aligned addresses - "$chk16w: \n" - "andi $t8, %[count], 0x3f \n" // whole 64-B chunks? - // t8 is the byte count after 64-byte chunks - "beq %[count], $t8, chk8w \n" - // There will be at most 1 32-byte chunk after it - "subu $a3, %[count], $t8 \n" // the reminder - // Here a3 counts bytes in 16w chunks - "addu $a3, %[dst], $a3 \n" - // Now a3 is the final dst after 64-byte chunks - "addu $t0, %[dst], %[count] \n" - // t0 is the "past the end" address - - // When in the loop we exercise "pref 30,x(a1)", the a1+x should not be past - // the "t0-32" address - // This means: for x=128 the last "safe" a1 address is "t0-160" - // Alternatively, for x=64 the last "safe" a1 address is "t0-96" - // we will use "pref 30,128(a1)", so "t0-160" is the limit - "subu $t9, $t0, 160 \n" - // t9 is the "last safe pref 30,128(a1)" address - "pref 0, 0(%[src]) \n" // first line of src - "pref 0, 32(%[src]) \n" // second line of src - "pref 0, 64(%[src]) \n" - "pref 30, 32(%[dst]) \n" - // In case the a1 > t9 don't use "pref 30" at all - "sgtu $v1, %[dst], $t9 \n" - "bgtz $v1, $loop16w \n" - "nop \n" - // otherwise, start with using pref30 - "pref 30, 64(%[dst]) \n" - "$loop16w: \n" - "pref 0, 96(%[src]) \n" - "lw $t0, 0(%[src]) \n" - "bgtz $v1, $skip_pref30_96 \n" // skip - "lw $t1, 4(%[src]) \n" - "pref 30, 96(%[dst]) \n" // continue - "$skip_pref30_96: \n" - "lw $t2, 8(%[src]) \n" - "lw $t3, 12(%[src]) \n" - "lw $t4, 16(%[src]) \n" - "lw $t5, 20(%[src]) \n" - "lw $t6, 24(%[src]) \n" - "lw $t7, 28(%[src]) \n" - "pref 0, 128(%[src]) \n" - // bring the next lines of src, addr 128 - "sw $t0, 0(%[dst]) \n" - "sw $t1, 4(%[dst]) \n" - "sw $t2, 8(%[dst]) \n" - "sw $t3, 12(%[dst]) \n" - "sw $t4, 16(%[dst]) \n" - "sw $t5, 20(%[dst]) \n" - "sw $t6, 24(%[dst]) \n" - "sw $t7, 28(%[dst]) \n" - "lw $t0, 32(%[src]) \n" - "bgtz $v1, $skip_pref30_128 \n" // skip pref 30,128(a1) - "lw $t1, 36(%[src]) \n" - "pref 30, 128(%[dst]) \n" // set dest, addr 128 - "$skip_pref30_128: \n" - "lw $t2, 40(%[src]) \n" - "lw $t3, 44(%[src]) \n" - "lw $t4, 48(%[src]) \n" - "lw $t5, 52(%[src]) \n" - "lw $t6, 56(%[src]) \n" - "lw $t7, 60(%[src]) \n" - "pref 0, 160(%[src]) \n" - // bring the next lines of src, addr 160 - "sw $t0, 32(%[dst]) \n" - "sw $t1, 36(%[dst]) \n" - "sw $t2, 40(%[dst]) \n" - "sw $t3, 44(%[dst]) \n" - "sw $t4, 48(%[dst]) \n" - "sw $t5, 52(%[dst]) \n" - "sw $t6, 56(%[dst]) \n" - "sw $t7, 60(%[dst]) \n" - - "addiu %[dst], %[dst], 64 \n" // adding 64 to dest - "sgtu $v1, %[dst], $t9 \n" - "bne %[dst], $a3, $loop16w \n" - " addiu %[src], %[src], 64 \n" // adding 64 to src - "move %[count], $t8 \n" - - // Here we have src and dest word-aligned but less than 64-bytes to go - - "chk8w: \n" - "pref 0, 0x0(%[src]) \n" - "andi $t8, %[count], 0x1f \n" // 32-byte chunk? - // the t8 is the reminder count past 32-bytes - "beq %[count], $t8, chk1w \n" - // count=t8,no 32-byte chunk - " nop \n" - - "lw $t0, 0(%[src]) \n" - "lw $t1, 4(%[src]) \n" - "lw $t2, 8(%[src]) \n" - "lw $t3, 12(%[src]) \n" - "lw $t4, 16(%[src]) \n" - "lw $t5, 20(%[src]) \n" - "lw $t6, 24(%[src]) \n" - "lw $t7, 28(%[src]) \n" - "addiu %[src], %[src], 32 \n" - - "sw $t0, 0(%[dst]) \n" - "sw $t1, 4(%[dst]) \n" - "sw $t2, 8(%[dst]) \n" - "sw $t3, 12(%[dst]) \n" - "sw $t4, 16(%[dst]) \n" - "sw $t5, 20(%[dst]) \n" - "sw $t6, 24(%[dst]) \n" - "sw $t7, 28(%[dst]) \n" - "addiu %[dst], %[dst], 32 \n" - - "chk1w: \n" - "andi %[count], $t8, 0x3 \n" - // now count is the reminder past 1w chunks - "beq %[count], $t8, $last8 \n" - " subu $a3, $t8, %[count] \n" - // a3 is count of bytes in 1w chunks - "addu $a3, %[dst], $a3 \n" - // now a3 is the dst address past the 1w chunks - // copying in words (4-byte chunks) - "$wordCopy_loop: \n" - "lw $t3, 0(%[src]) \n" - // the first t3 may be equal t0 ... optimize? - "addiu %[src], %[src],4 \n" - "addiu %[dst], %[dst],4 \n" - "bne %[dst], $a3,$wordCopy_loop \n" - " sw $t3, -4(%[dst]) \n" - - // For the last (<8) bytes - "$last8: \n" - "blez %[count], leave \n" - " addu $a3, %[dst], %[count] \n" // a3 -last dst address - "$last8loop: \n" - "lb $v1, 0(%[src]) \n" - "addiu %[src], %[src], 1 \n" - "addiu %[dst], %[dst], 1 \n" - "bne %[dst], $a3, $last8loop \n" - " sb $v1, -1(%[dst]) \n" - - "leave: \n" - " j $ra \n" - " nop \n" - - // - // UNALIGNED case - // - - "unaligned: \n" - // got here with a3="negu a1" - "andi $a3, $a3, 0x3 \n" // a1 is word aligned? - "beqz $a3, $ua_chk16w \n" - " subu %[count], %[count], $a3 \n" - // bytes left after initial a3 bytes - "lwr $v1, 0(%[src]) \n" - "lwl $v1, 3(%[src]) \n" - "addu %[src], %[src], $a3 \n" // a3 may be 1, 2 or 3 - "swr $v1, 0(%[dst]) \n" - "addu %[dst], %[dst], $a3 \n" - // below the dst will be word aligned (NOTE1) - "$ua_chk16w: \n" - "andi $t8, %[count], 0x3f \n" // whole 64-B chunks? - // t8 is the byte count after 64-byte chunks - "beq %[count], $t8, ua_chk8w \n" - // if a2==t8, no 64-byte chunks - // There will be at most 1 32-byte chunk after it - "subu $a3, %[count], $t8 \n" // the reminder - // Here a3 counts bytes in 16w chunks - "addu $a3, %[dst], $a3 \n" - // Now a3 is the final dst after 64-byte chunks - "addu $t0, %[dst], %[count] \n" // t0 "past the end" - "subu $t9, $t0, 160 \n" - // t9 is the "last safe pref 30,128(a1)" address - "pref 0, 0(%[src]) \n" // first line of src - "pref 0, 32(%[src]) \n" // second line addr 32 - "pref 0, 64(%[src]) \n" - "pref 30, 32(%[dst]) \n" - // safe, as we have at least 64 bytes ahead - // In case the a1 > t9 don't use "pref 30" at all - "sgtu $v1, %[dst], $t9 \n" - "bgtz $v1, $ua_loop16w \n" - // skip "pref 30,64(a1)" for too short arrays - " nop \n" - // otherwise, start with using pref30 - "pref 30, 64(%[dst]) \n" - "$ua_loop16w: \n" - "pref 0, 96(%[src]) \n" - "lwr $t0, 0(%[src]) \n" - "lwl $t0, 3(%[src]) \n" - "lwr $t1, 4(%[src]) \n" - "bgtz $v1, $ua_skip_pref30_96 \n" - " lwl $t1, 7(%[src]) \n" - "pref 30, 96(%[dst]) \n" - // continue setting up the dest, addr 96 - "$ua_skip_pref30_96: \n" - "lwr $t2, 8(%[src]) \n" - "lwl $t2, 11(%[src]) \n" - "lwr $t3, 12(%[src]) \n" - "lwl $t3, 15(%[src]) \n" - "lwr $t4, 16(%[src]) \n" - "lwl $t4, 19(%[src]) \n" - "lwr $t5, 20(%[src]) \n" - "lwl $t5, 23(%[src]) \n" - "lwr $t6, 24(%[src]) \n" - "lwl $t6, 27(%[src]) \n" - "lwr $t7, 28(%[src]) \n" - "lwl $t7, 31(%[src]) \n" - "pref 0, 128(%[src]) \n" - // bring the next lines of src, addr 128 - "sw $t0, 0(%[dst]) \n" - "sw $t1, 4(%[dst]) \n" - "sw $t2, 8(%[dst]) \n" - "sw $t3, 12(%[dst]) \n" - "sw $t4, 16(%[dst]) \n" - "sw $t5, 20(%[dst]) \n" - "sw $t6, 24(%[dst]) \n" - "sw $t7, 28(%[dst]) \n" - "lwr $t0, 32(%[src]) \n" - "lwl $t0, 35(%[src]) \n" - "lwr $t1, 36(%[src]) \n" - "bgtz $v1, ua_skip_pref30_128 \n" - " lwl $t1, 39(%[src]) \n" - "pref 30, 128(%[dst]) \n" - // continue setting up the dest, addr 128 - "ua_skip_pref30_128: \n" - - "lwr $t2, 40(%[src]) \n" - "lwl $t2, 43(%[src]) \n" - "lwr $t3, 44(%[src]) \n" - "lwl $t3, 47(%[src]) \n" - "lwr $t4, 48(%[src]) \n" - "lwl $t4, 51(%[src]) \n" - "lwr $t5, 52(%[src]) \n" - "lwl $t5, 55(%[src]) \n" - "lwr $t6, 56(%[src]) \n" - "lwl $t6, 59(%[src]) \n" - "lwr $t7, 60(%[src]) \n" - "lwl $t7, 63(%[src]) \n" - "pref 0, 160(%[src]) \n" - // bring the next lines of src, addr 160 - "sw $t0, 32(%[dst]) \n" - "sw $t1, 36(%[dst]) \n" - "sw $t2, 40(%[dst]) \n" - "sw $t3, 44(%[dst]) \n" - "sw $t4, 48(%[dst]) \n" - "sw $t5, 52(%[dst]) \n" - "sw $t6, 56(%[dst]) \n" - "sw $t7, 60(%[dst]) \n" - - "addiu %[dst],%[dst],64 \n" // adding 64 to dest - "sgtu $v1,%[dst],$t9 \n" - "bne %[dst],$a3,$ua_loop16w \n" - " addiu %[src],%[src],64 \n" // adding 64 to src - "move %[count],$t8 \n" - - // Here we have src and dest word-aligned but less than 64-bytes to go - - "ua_chk8w: \n" - "pref 0, 0x0(%[src]) \n" - "andi $t8, %[count], 0x1f \n" // 32-byte chunk? - // the t8 is the reminder count - "beq %[count], $t8, $ua_chk1w \n" - // when count==t8, no 32-byte chunk - - "lwr $t0, 0(%[src]) \n" - "lwl $t0, 3(%[src]) \n" - "lwr $t1, 4(%[src]) \n" - "lwl $t1, 7(%[src]) \n" - "lwr $t2, 8(%[src]) \n" - "lwl $t2, 11(%[src]) \n" - "lwr $t3, 12(%[src]) \n" - "lwl $t3, 15(%[src]) \n" - "lwr $t4, 16(%[src]) \n" - "lwl $t4, 19(%[src]) \n" - "lwr $t5, 20(%[src]) \n" - "lwl $t5, 23(%[src]) \n" - "lwr $t6, 24(%[src]) \n" - "lwl $t6, 27(%[src]) \n" - "lwr $t7, 28(%[src]) \n" - "lwl $t7, 31(%[src]) \n" - "addiu %[src], %[src], 32 \n" - - "sw $t0, 0(%[dst]) \n" - "sw $t1, 4(%[dst]) \n" - "sw $t2, 8(%[dst]) \n" - "sw $t3, 12(%[dst]) \n" - "sw $t4, 16(%[dst]) \n" - "sw $t5, 20(%[dst]) \n" - "sw $t6, 24(%[dst]) \n" - "sw $t7, 28(%[dst]) \n" - "addiu %[dst], %[dst], 32 \n" - - "$ua_chk1w: \n" - "andi %[count], $t8, 0x3 \n" - // now count is the reminder past 1w chunks - "beq %[count], $t8, ua_smallCopy \n" - "subu $a3, $t8, %[count] \n" - // a3 is count of bytes in 1w chunks - "addu $a3, %[dst], $a3 \n" - // now a3 is the dst address past the 1w chunks - - // copying in words (4-byte chunks) - "$ua_wordCopy_loop: \n" - "lwr $v1, 0(%[src]) \n" - "lwl $v1, 3(%[src]) \n" - "addiu %[src], %[src], 4 \n" - "addiu %[dst], %[dst], 4 \n" - // note: dst=a1 is word aligned here, see NOTE1 - "bne %[dst], $a3, $ua_wordCopy_loop \n" - " sw $v1,-4(%[dst]) \n" - - // Now less than 4 bytes (value in count) left to copy - "ua_smallCopy: \n" - "beqz %[count], leave \n" - " addu $a3, %[dst], %[count] \n" // a3 = last dst address - "$ua_smallCopy_loop: \n" - "lb $v1, 0(%[src]) \n" - "addiu %[src], %[src], 1 \n" - "addiu %[dst], %[dst], 1 \n" - "bne %[dst],$a3,$ua_smallCopy_loop \n" - " sb $v1, -1(%[dst]) \n" - - "j $ra \n" - " nop \n" - ".set at \n" - ".set reorder \n" - : [dst] "+r" (dst), [src] "+r" (src) - : [count] "r" (count) - : "t0", "t1", "t2", "t3", "t4", "t5", "t6", "t7", - "t8", "t9", "a3", "v1", "at" - ); -} -#endif // HAS_COPYROW_MIPS - -// MIPS DSPR2 functions -#if !defined(LIBYUV_DISABLE_MIPS) && defined(__mips_dsp) && \ - (__mips_dsp_rev >= 2) && \ - (_MIPS_SIM == _MIPS_SIM_ABI32) && (__mips_isa_rev < 6) - -void SplitUVRow_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "srl $t4, %[width], 4 \n" // multiplies of 16 - "blez $t4, 2f \n" - " andi %[width], %[width], 0xf \n" // residual - - ".p2align 2 \n" - "1: \n" - "addiu $t4, $t4, -1 \n" - "lw $t0, 0(%[src_uv]) \n" // V1 | U1 | V0 | U0 - "lw $t1, 4(%[src_uv]) \n" // V3 | U3 | V2 | U2 - "lw $t2, 8(%[src_uv]) \n" // V5 | U5 | V4 | U4 - "lw $t3, 12(%[src_uv]) \n" // V7 | U7 | V6 | U6 - "lw $t5, 16(%[src_uv]) \n" // V9 | U9 | V8 | U8 - "lw $t6, 20(%[src_uv]) \n" // V11 | U11 | V10 | U10 - "lw $t7, 24(%[src_uv]) \n" // V13 | U13 | V12 | U12 - "lw $t8, 28(%[src_uv]) \n" // V15 | U15 | V14 | U14 - "addiu %[src_uv], %[src_uv], 32 \n" - "precrq.qb.ph $t9, $t1, $t0 \n" // V3 | V2 | V1 | V0 - "precr.qb.ph $t0, $t1, $t0 \n" // U3 | U2 | U1 | U0 - "precrq.qb.ph $t1, $t3, $t2 \n" // V7 | V6 | V5 | V4 - "precr.qb.ph $t2, $t3, $t2 \n" // U7 | U6 | U5 | U4 - "precrq.qb.ph $t3, $t6, $t5 \n" // V11 | V10 | V9 | V8 - "precr.qb.ph $t5, $t6, $t5 \n" // U11 | U10 | U9 | U8 - "precrq.qb.ph $t6, $t8, $t7 \n" // V15 | V14 | V13 | V12 - "precr.qb.ph $t7, $t8, $t7 \n" // U15 | U14 | U13 | U12 - "sw $t9, 0(%[dst_v]) \n" - "sw $t0, 0(%[dst_u]) \n" - "sw $t1, 4(%[dst_v]) \n" - "sw $t2, 4(%[dst_u]) \n" - "sw $t3, 8(%[dst_v]) \n" - "sw $t5, 8(%[dst_u]) \n" - "sw $t6, 12(%[dst_v]) \n" - "sw $t7, 12(%[dst_u]) \n" - "addiu %[dst_v], %[dst_v], 16 \n" - "bgtz $t4, 1b \n" - " addiu %[dst_u], %[dst_u], 16 \n" - - "beqz %[width], 3f \n" - " nop \n" - - "2: \n" - "lbu $t0, 0(%[src_uv]) \n" - "lbu $t1, 1(%[src_uv]) \n" - "addiu %[src_uv], %[src_uv], 2 \n" - "addiu %[width], %[width], -1 \n" - "sb $t0, 0(%[dst_u]) \n" - "sb $t1, 0(%[dst_v]) \n" - "addiu %[dst_u], %[dst_u], 1 \n" - "bgtz %[width], 2b \n" - " addiu %[dst_v], %[dst_v], 1 \n" - - "3: \n" - ".set pop \n" - : [src_uv] "+r" (src_uv), - [width] "+r" (width), - [dst_u] "+r" (dst_u), - [dst_v] "+r" (dst_v) - : - : "t0", "t1", "t2", "t3", - "t4", "t5", "t6", "t7", "t8", "t9" - ); -} - -void SplitUVRow_Unaligned_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, - uint8* dst_v, int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "srl $t4, %[width], 4 \n" // multiplies of 16 - "blez $t4, 2f \n" - " andi %[width], %[width], 0xf \n" // residual - - ".p2align 2 \n" - "1: \n" - "addiu $t4, $t4, -1 \n" - "lwr $t0, 0(%[src_uv]) \n" - "lwl $t0, 3(%[src_uv]) \n" // V1 | U1 | V0 | U0 - "lwr $t1, 4(%[src_uv]) \n" - "lwl $t1, 7(%[src_uv]) \n" // V3 | U3 | V2 | U2 - "lwr $t2, 8(%[src_uv]) \n" - "lwl $t2, 11(%[src_uv]) \n" // V5 | U5 | V4 | U4 - "lwr $t3, 12(%[src_uv]) \n" - "lwl $t3, 15(%[src_uv]) \n" // V7 | U7 | V6 | U6 - "lwr $t5, 16(%[src_uv]) \n" - "lwl $t5, 19(%[src_uv]) \n" // V9 | U9 | V8 | U8 - "lwr $t6, 20(%[src_uv]) \n" - "lwl $t6, 23(%[src_uv]) \n" // V11 | U11 | V10 | U10 - "lwr $t7, 24(%[src_uv]) \n" - "lwl $t7, 27(%[src_uv]) \n" // V13 | U13 | V12 | U12 - "lwr $t8, 28(%[src_uv]) \n" - "lwl $t8, 31(%[src_uv]) \n" // V15 | U15 | V14 | U14 - "precrq.qb.ph $t9, $t1, $t0 \n" // V3 | V2 | V1 | V0 - "precr.qb.ph $t0, $t1, $t0 \n" // U3 | U2 | U1 | U0 - "precrq.qb.ph $t1, $t3, $t2 \n" // V7 | V6 | V5 | V4 - "precr.qb.ph $t2, $t3, $t2 \n" // U7 | U6 | U5 | U4 - "precrq.qb.ph $t3, $t6, $t5 \n" // V11 | V10 | V9 | V8 - "precr.qb.ph $t5, $t6, $t5 \n" // U11 | U10 | U9 | U8 - "precrq.qb.ph $t6, $t8, $t7 \n" // V15 | V14 | V13 | V12 - "precr.qb.ph $t7, $t8, $t7 \n" // U15 | U14 | U13 | U12 - "addiu %[src_uv], %[src_uv], 32 \n" - "swr $t9, 0(%[dst_v]) \n" - "swl $t9, 3(%[dst_v]) \n" - "swr $t0, 0(%[dst_u]) \n" - "swl $t0, 3(%[dst_u]) \n" - "swr $t1, 4(%[dst_v]) \n" - "swl $t1, 7(%[dst_v]) \n" - "swr $t2, 4(%[dst_u]) \n" - "swl $t2, 7(%[dst_u]) \n" - "swr $t3, 8(%[dst_v]) \n" - "swl $t3, 11(%[dst_v]) \n" - "swr $t5, 8(%[dst_u]) \n" - "swl $t5, 11(%[dst_u]) \n" - "swr $t6, 12(%[dst_v]) \n" - "swl $t6, 15(%[dst_v]) \n" - "swr $t7, 12(%[dst_u]) \n" - "swl $t7, 15(%[dst_u]) \n" - "addiu %[dst_u], %[dst_u], 16 \n" - "bgtz $t4, 1b \n" - " addiu %[dst_v], %[dst_v], 16 \n" - - "beqz %[width], 3f \n" - " nop \n" - - "2: \n" - "lbu $t0, 0(%[src_uv]) \n" - "lbu $t1, 1(%[src_uv]) \n" - "addiu %[src_uv], %[src_uv], 2 \n" - "addiu %[width], %[width], -1 \n" - "sb $t0, 0(%[dst_u]) \n" - "sb $t1, 0(%[dst_v]) \n" - "addiu %[dst_u], %[dst_u], 1 \n" - "bgtz %[width], 2b \n" - " addiu %[dst_v], %[dst_v], 1 \n" - - "3: \n" - ".set pop \n" - : [src_uv] "+r" (src_uv), - [width] "+r" (width), - [dst_u] "+r" (dst_u), - [dst_v] "+r" (dst_v) - : - : "t0", "t1", "t2", "t3", - "t4", "t5", "t6", "t7", "t8", "t9" - ); -} - -void MirrorRow_MIPS_DSPR2(const uint8* src, uint8* dst, int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - "srl $t4, %[width], 4 \n" // multiplies of 16 - "andi $t5, %[width], 0xf \n" - "blez $t4, 2f \n" - " addu %[src], %[src], %[width] \n" // src += width - - ".p2align 2 \n" - "1: \n" - "lw $t0, -16(%[src]) \n" // |3|2|1|0| - "lw $t1, -12(%[src]) \n" // |7|6|5|4| - "lw $t2, -8(%[src]) \n" // |11|10|9|8| - "lw $t3, -4(%[src]) \n" // |15|14|13|12| - "wsbh $t0, $t0 \n" // |2|3|0|1| - "wsbh $t1, $t1 \n" // |6|7|4|5| - "wsbh $t2, $t2 \n" // |10|11|8|9| - "wsbh $t3, $t3 \n" // |14|15|12|13| - "rotr $t0, $t0, 16 \n" // |0|1|2|3| - "rotr $t1, $t1, 16 \n" // |4|5|6|7| - "rotr $t2, $t2, 16 \n" // |8|9|10|11| - "rotr $t3, $t3, 16 \n" // |12|13|14|15| - "addiu %[src], %[src], -16 \n" - "addiu $t4, $t4, -1 \n" - "sw $t3, 0(%[dst]) \n" // |15|14|13|12| - "sw $t2, 4(%[dst]) \n" // |11|10|9|8| - "sw $t1, 8(%[dst]) \n" // |7|6|5|4| - "sw $t0, 12(%[dst]) \n" // |3|2|1|0| - "bgtz $t4, 1b \n" - " addiu %[dst], %[dst], 16 \n" - "beqz $t5, 3f \n" - " nop \n" - - "2: \n" - "lbu $t0, -1(%[src]) \n" - "addiu $t5, $t5, -1 \n" - "addiu %[src], %[src], -1 \n" - "sb $t0, 0(%[dst]) \n" - "bgez $t5, 2b \n" - " addiu %[dst], %[dst], 1 \n" - - "3: \n" - ".set pop \n" - : [src] "+r" (src), [dst] "+r" (dst) - : [width] "r" (width) - : "t0", "t1", "t2", "t3", "t4", "t5" - ); -} - -void MirrorUVRow_MIPS_DSPR2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width) { - int x = 0; - int y = 0; - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - "addu $t4, %[width], %[width] \n" - "srl %[x], %[width], 4 \n" - "andi %[y], %[width], 0xf \n" - "blez %[x], 2f \n" - " addu %[src_uv], %[src_uv], $t4 \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, -32(%[src_uv]) \n" // |3|2|1|0| - "lw $t1, -28(%[src_uv]) \n" // |7|6|5|4| - "lw $t2, -24(%[src_uv]) \n" // |11|10|9|8| - "lw $t3, -20(%[src_uv]) \n" // |15|14|13|12| - "lw $t4, -16(%[src_uv]) \n" // |19|18|17|16| - "lw $t6, -12(%[src_uv]) \n" // |23|22|21|20| - "lw $t7, -8(%[src_uv]) \n" // |27|26|25|24| - "lw $t8, -4(%[src_uv]) \n" // |31|30|29|28| - - "rotr $t0, $t0, 16 \n" // |1|0|3|2| - "rotr $t1, $t1, 16 \n" // |5|4|7|6| - "rotr $t2, $t2, 16 \n" // |9|8|11|10| - "rotr $t3, $t3, 16 \n" // |13|12|15|14| - "rotr $t4, $t4, 16 \n" // |17|16|19|18| - "rotr $t6, $t6, 16 \n" // |21|20|23|22| - "rotr $t7, $t7, 16 \n" // |25|24|27|26| - "rotr $t8, $t8, 16 \n" // |29|28|31|30| - "precr.qb.ph $t9, $t0, $t1 \n" // |0|2|4|6| - "precrq.qb.ph $t5, $t0, $t1 \n" // |1|3|5|7| - "precr.qb.ph $t0, $t2, $t3 \n" // |8|10|12|14| - "precrq.qb.ph $t1, $t2, $t3 \n" // |9|11|13|15| - "precr.qb.ph $t2, $t4, $t6 \n" // |16|18|20|22| - "precrq.qb.ph $t3, $t4, $t6 \n" // |17|19|21|23| - "precr.qb.ph $t4, $t7, $t8 \n" // |24|26|28|30| - "precrq.qb.ph $t6, $t7, $t8 \n" // |25|27|29|31| - "addiu %[src_uv], %[src_uv], -32 \n" - "addiu %[x], %[x], -1 \n" - "swr $t4, 0(%[dst_u]) \n" - "swl $t4, 3(%[dst_u]) \n" // |30|28|26|24| - "swr $t6, 0(%[dst_v]) \n" - "swl $t6, 3(%[dst_v]) \n" // |31|29|27|25| - "swr $t2, 4(%[dst_u]) \n" - "swl $t2, 7(%[dst_u]) \n" // |22|20|18|16| - "swr $t3, 4(%[dst_v]) \n" - "swl $t3, 7(%[dst_v]) \n" // |23|21|19|17| - "swr $t0, 8(%[dst_u]) \n" - "swl $t0, 11(%[dst_u]) \n" // |14|12|10|8| - "swr $t1, 8(%[dst_v]) \n" - "swl $t1, 11(%[dst_v]) \n" // |15|13|11|9| - "swr $t9, 12(%[dst_u]) \n" - "swl $t9, 15(%[dst_u]) \n" // |6|4|2|0| - "swr $t5, 12(%[dst_v]) \n" - "swl $t5, 15(%[dst_v]) \n" // |7|5|3|1| - "addiu %[dst_v], %[dst_v], 16 \n" - "bgtz %[x], 1b \n" - " addiu %[dst_u], %[dst_u], 16 \n" - "beqz %[y], 3f \n" - " nop \n" - "b 2f \n" - " nop \n" - - "2: \n" - "lbu $t0, -2(%[src_uv]) \n" - "lbu $t1, -1(%[src_uv]) \n" - "addiu %[src_uv], %[src_uv], -2 \n" - "addiu %[y], %[y], -1 \n" - "sb $t0, 0(%[dst_u]) \n" - "sb $t1, 0(%[dst_v]) \n" - "addiu %[dst_u], %[dst_u], 1 \n" - "bgtz %[y], 2b \n" - " addiu %[dst_v], %[dst_v], 1 \n" - - "3: \n" - ".set pop \n" - : [src_uv] "+r" (src_uv), - [dst_u] "+r" (dst_u), - [dst_v] "+r" (dst_v), - [x] "=&r" (x), - [y] "+r" (y) - : [width] "r" (width) - : "t0", "t1", "t2", "t3", "t4", - "t5", "t7", "t8", "t9" - ); -} - -// Convert (4 Y and 2 VU) I422 and arrange RGB values into -// t5 = | 0 | B0 | 0 | b0 | -// t4 = | 0 | B1 | 0 | b1 | -// t9 = | 0 | G0 | 0 | g0 | -// t8 = | 0 | G1 | 0 | g1 | -// t2 = | 0 | R0 | 0 | r0 | -// t1 = | 0 | R1 | 0 | r1 | -#define I422ToTransientMipsRGB \ - "lw $t0, 0(%[y_buf]) \n" \ - "lhu $t1, 0(%[u_buf]) \n" \ - "lhu $t2, 0(%[v_buf]) \n" \ - "preceu.ph.qbr $t1, $t1 \n" \ - "preceu.ph.qbr $t2, $t2 \n" \ - "preceu.ph.qbra $t3, $t0 \n" \ - "preceu.ph.qbla $t0, $t0 \n" \ - "subu.ph $t1, $t1, $s5 \n" \ - "subu.ph $t2, $t2, $s5 \n" \ - "subu.ph $t3, $t3, $s4 \n" \ - "subu.ph $t0, $t0, $s4 \n" \ - "mul.ph $t3, $t3, $s0 \n" \ - "mul.ph $t0, $t0, $s0 \n" \ - "shll.ph $t4, $t1, 0x7 \n" \ - "subu.ph $t4, $t4, $t1 \n" \ - "mul.ph $t6, $t1, $s1 \n" \ - "mul.ph $t1, $t2, $s2 \n" \ - "addq_s.ph $t5, $t4, $t3 \n" \ - "addq_s.ph $t4, $t4, $t0 \n" \ - "shra.ph $t5, $t5, 6 \n" \ - "shra.ph $t4, $t4, 6 \n" \ - "addiu %[u_buf], 2 \n" \ - "addiu %[v_buf], 2 \n" \ - "addu.ph $t6, $t6, $t1 \n" \ - "mul.ph $t1, $t2, $s3 \n" \ - "addu.ph $t9, $t6, $t3 \n" \ - "addu.ph $t8, $t6, $t0 \n" \ - "shra.ph $t9, $t9, 6 \n" \ - "shra.ph $t8, $t8, 6 \n" \ - "addu.ph $t2, $t1, $t3 \n" \ - "addu.ph $t1, $t1, $t0 \n" \ - "shra.ph $t2, $t2, 6 \n" \ - "shra.ph $t1, $t1, 6 \n" \ - "subu.ph $t5, $t5, $s5 \n" \ - "subu.ph $t4, $t4, $s5 \n" \ - "subu.ph $t9, $t9, $s5 \n" \ - "subu.ph $t8, $t8, $s5 \n" \ - "subu.ph $t2, $t2, $s5 \n" \ - "subu.ph $t1, $t1, $s5 \n" \ - "shll_s.ph $t5, $t5, 8 \n" \ - "shll_s.ph $t4, $t4, 8 \n" \ - "shll_s.ph $t9, $t9, 8 \n" \ - "shll_s.ph $t8, $t8, 8 \n" \ - "shll_s.ph $t2, $t2, 8 \n" \ - "shll_s.ph $t1, $t1, 8 \n" \ - "shra.ph $t5, $t5, 8 \n" \ - "shra.ph $t4, $t4, 8 \n" \ - "shra.ph $t9, $t9, 8 \n" \ - "shra.ph $t8, $t8, 8 \n" \ - "shra.ph $t2, $t2, 8 \n" \ - "shra.ph $t1, $t1, 8 \n" \ - "addu.ph $t5, $t5, $s5 \n" \ - "addu.ph $t4, $t4, $s5 \n" \ - "addu.ph $t9, $t9, $s5 \n" \ - "addu.ph $t8, $t8, $s5 \n" \ - "addu.ph $t2, $t2, $s5 \n" \ - "addu.ph $t1, $t1, $s5 \n" - -void I422ToARGBRow_MIPS_DSPR2(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "beqz %[width], 2f \n" - " repl.ph $s0, 74 \n" // |YG|YG| = |74|74| - "repl.ph $s1, -25 \n" // |UG|UG| = |-25|-25| - "repl.ph $s2, -52 \n" // |VG|VG| = |-52|-52| - "repl.ph $s3, 102 \n" // |VR|VR| = |102|102| - "repl.ph $s4, 16 \n" // |0|16|0|16| - "repl.ph $s5, 128 \n" // |128|128| // clipping - "lui $s6, 0xff00 \n" - "ori $s6, 0xff00 \n" // |ff|00|ff|00|ff| - - ".p2align 2 \n" - "1: \n" - I422ToTransientMipsRGB -// Arranging into argb format - "precr.qb.ph $t4, $t8, $t4 \n" // |G1|g1|B1|b1| - "precr.qb.ph $t5, $t9, $t5 \n" // |G0|g0|B0|b0| - "addiu %[width], -4 \n" - "precrq.qb.ph $t8, $t4, $t5 \n" // |G1|B1|G0|B0| - "precr.qb.ph $t9, $t4, $t5 \n" // |g1|b1|g0|b0| - "precr.qb.ph $t2, $t1, $t2 \n" // |R1|r1|R0|r0| - - "addiu %[y_buf], 4 \n" - "preceu.ph.qbla $t1, $t2 \n" // |0 |R1|0 |R0| - "preceu.ph.qbra $t2, $t2 \n" // |0 |r1|0 |r0| - "or $t1, $t1, $s6 \n" // |ff|R1|ff|R0| - "or $t2, $t2, $s6 \n" // |ff|r1|ff|r0| - "precrq.ph.w $t0, $t2, $t9 \n" // |ff|r1|g1|b1| - "precrq.ph.w $t3, $t1, $t8 \n" // |ff|R1|G1|B1| - "sll $t9, $t9, 16 \n" - "sll $t8, $t8, 16 \n" - "packrl.ph $t2, $t2, $t9 \n" // |ff|r0|g0|b0| - "packrl.ph $t1, $t1, $t8 \n" // |ff|R0|G0|B0| -// Store results. - "sw $t2, 0(%[rgb_buf]) \n" - "sw $t0, 4(%[rgb_buf]) \n" - "sw $t1, 8(%[rgb_buf]) \n" - "sw $t3, 12(%[rgb_buf]) \n" - "bnez %[width], 1b \n" - " addiu %[rgb_buf], 16 \n" - "2: \n" - ".set pop \n" - :[y_buf] "+r" (y_buf), - [u_buf] "+r" (u_buf), - [v_buf] "+r" (v_buf), - [width] "+r" (width), - [rgb_buf] "+r" (rgb_buf) - : - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9", - "s0", "s1", "s2", "s3", - "s4", "s5", "s6" - ); -} - -void I422ToABGRRow_MIPS_DSPR2(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "beqz %[width], 2f \n" - " repl.ph $s0, 74 \n" // |YG|YG| = |74|74| - "repl.ph $s1, -25 \n" // |UG|UG| = |-25|-25| - "repl.ph $s2, -52 \n" // |VG|VG| = |-52|-52| - "repl.ph $s3, 102 \n" // |VR|VR| = |102|102| - "repl.ph $s4, 16 \n" // |0|16|0|16| - "repl.ph $s5, 128 \n" // |128|128| - "lui $s6, 0xff00 \n" - "ori $s6, 0xff00 \n" // |ff|00|ff|00| - - ".p2align 2 \n" - "1: \n" - I422ToTransientMipsRGB -// Arranging into abgr format - "precr.qb.ph $t0, $t8, $t1 \n" // |G1|g1|R1|r1| - "precr.qb.ph $t3, $t9, $t2 \n" // |G0|g0|R0|r0| - "precrq.qb.ph $t8, $t0, $t3 \n" // |G1|R1|G0|R0| - "precr.qb.ph $t9, $t0, $t3 \n" // |g1|r1|g0|r0| - - "precr.qb.ph $t2, $t4, $t5 \n" // |B1|b1|B0|b0| - "addiu %[width], -4 \n" - "addiu %[y_buf], 4 \n" - "preceu.ph.qbla $t1, $t2 \n" // |0 |B1|0 |B0| - "preceu.ph.qbra $t2, $t2 \n" // |0 |b1|0 |b0| - "or $t1, $t1, $s6 \n" // |ff|B1|ff|B0| - "or $t2, $t2, $s6 \n" // |ff|b1|ff|b0| - "precrq.ph.w $t0, $t2, $t9 \n" // |ff|b1|g1|r1| - "precrq.ph.w $t3, $t1, $t8 \n" // |ff|B1|G1|R1| - "sll $t9, $t9, 16 \n" - "sll $t8, $t8, 16 \n" - "packrl.ph $t2, $t2, $t9 \n" // |ff|b0|g0|r0| - "packrl.ph $t1, $t1, $t8 \n" // |ff|B0|G0|R0| -// Store results. - "sw $t2, 0(%[rgb_buf]) \n" - "sw $t0, 4(%[rgb_buf]) \n" - "sw $t1, 8(%[rgb_buf]) \n" - "sw $t3, 12(%[rgb_buf]) \n" - "bnez %[width], 1b \n" - " addiu %[rgb_buf], 16 \n" - "2: \n" - ".set pop \n" - :[y_buf] "+r" (y_buf), - [u_buf] "+r" (u_buf), - [v_buf] "+r" (v_buf), - [width] "+r" (width), - [rgb_buf] "+r" (rgb_buf) - : - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9", - "s0", "s1", "s2", "s3", - "s4", "s5", "s6" - ); -} - -void I422ToBGRARow_MIPS_DSPR2(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "beqz %[width], 2f \n" - " repl.ph $s0, 74 \n" // |YG|YG| = |74 |74 | - "repl.ph $s1, -25 \n" // |UG|UG| = |-25|-25| - "repl.ph $s2, -52 \n" // |VG|VG| = |-52|-52| - "repl.ph $s3, 102 \n" // |VR|VR| = |102|102| - "repl.ph $s4, 16 \n" // |0|16|0|16| - "repl.ph $s5, 128 \n" // |128|128| - "lui $s6, 0xff \n" - "ori $s6, 0xff \n" // |00|ff|00|ff| - - ".p2align 2 \n" - "1: \n" - I422ToTransientMipsRGB - // Arranging into bgra format - "precr.qb.ph $t4, $t4, $t8 \n" // |B1|b1|G1|g1| - "precr.qb.ph $t5, $t5, $t9 \n" // |B0|b0|G0|g0| - "precrq.qb.ph $t8, $t4, $t5 \n" // |B1|G1|B0|G0| - "precr.qb.ph $t9, $t4, $t5 \n" // |b1|g1|b0|g0| - - "precr.qb.ph $t2, $t1, $t2 \n" // |R1|r1|R0|r0| - "addiu %[width], -4 \n" - "addiu %[y_buf], 4 \n" - "preceu.ph.qbla $t1, $t2 \n" // |0 |R1|0 |R0| - "preceu.ph.qbra $t2, $t2 \n" // |0 |r1|0 |r0| - "sll $t1, $t1, 8 \n" // |R1|0 |R0|0 | - "sll $t2, $t2, 8 \n" // |r1|0 |r0|0 | - "or $t1, $t1, $s6 \n" // |R1|ff|R0|ff| - "or $t2, $t2, $s6 \n" // |r1|ff|r0|ff| - "precrq.ph.w $t0, $t9, $t2 \n" // |b1|g1|r1|ff| - "precrq.ph.w $t3, $t8, $t1 \n" // |B1|G1|R1|ff| - "sll $t1, $t1, 16 \n" - "sll $t2, $t2, 16 \n" - "packrl.ph $t2, $t9, $t2 \n" // |b0|g0|r0|ff| - "packrl.ph $t1, $t8, $t1 \n" // |B0|G0|R0|ff| -// Store results. - "sw $t2, 0(%[rgb_buf]) \n" - "sw $t0, 4(%[rgb_buf]) \n" - "sw $t1, 8(%[rgb_buf]) \n" - "sw $t3, 12(%[rgb_buf]) \n" - "bnez %[width], 1b \n" - " addiu %[rgb_buf], 16 \n" - "2: \n" - ".set pop \n" - :[y_buf] "+r" (y_buf), - [u_buf] "+r" (u_buf), - [v_buf] "+r" (v_buf), - [width] "+r" (width), - [rgb_buf] "+r" (rgb_buf) - : - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9", - "s0", "s1", "s2", "s3", - "s4", "s5", "s6" - ); -} - -// Bilinear filter 8x2 -> 8x1 -void InterpolateRows_MIPS_DSPR2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - int y0_fraction = 256 - source_y_fraction; - const uint8* src_ptr1 = src_ptr + src_stride; - - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - "replv.ph $t0, %[y0_fraction] \n" - "replv.ph $t1, %[source_y_fraction] \n" - - ".p2align 2 \n" - "1: \n" - "lw $t2, 0(%[src_ptr]) \n" - "lw $t3, 0(%[src_ptr1]) \n" - "lw $t4, 4(%[src_ptr]) \n" - "lw $t5, 4(%[src_ptr1]) \n" - "muleu_s.ph.qbl $t6, $t2, $t0 \n" - "muleu_s.ph.qbr $t7, $t2, $t0 \n" - "muleu_s.ph.qbl $t8, $t3, $t1 \n" - "muleu_s.ph.qbr $t9, $t3, $t1 \n" - "muleu_s.ph.qbl $t2, $t4, $t0 \n" - "muleu_s.ph.qbr $t3, $t4, $t0 \n" - "muleu_s.ph.qbl $t4, $t5, $t1 \n" - "muleu_s.ph.qbr $t5, $t5, $t1 \n" - "addq.ph $t6, $t6, $t8 \n" - "addq.ph $t7, $t7, $t9 \n" - "addq.ph $t2, $t2, $t4 \n" - "addq.ph $t3, $t3, $t5 \n" - "shra.ph $t6, $t6, 8 \n" - "shra.ph $t7, $t7, 8 \n" - "shra.ph $t2, $t2, 8 \n" - "shra.ph $t3, $t3, 8 \n" - "precr.qb.ph $t6, $t6, $t7 \n" - "precr.qb.ph $t2, $t2, $t3 \n" - "addiu %[src_ptr], %[src_ptr], 8 \n" - "addiu %[src_ptr1], %[src_ptr1], 8 \n" - "addiu %[dst_width], %[dst_width], -8 \n" - "sw $t6, 0(%[dst_ptr]) \n" - "sw $t2, 4(%[dst_ptr]) \n" - "bgtz %[dst_width], 1b \n" - " addiu %[dst_ptr], %[dst_ptr], 8 \n" - - ".set pop \n" - : [dst_ptr] "+r" (dst_ptr), - [src_ptr1] "+r" (src_ptr1), - [src_ptr] "+r" (src_ptr), - [dst_width] "+r" (dst_width) - : [source_y_fraction] "r" (source_y_fraction), - [y0_fraction] "r" (y0_fraction), - [src_stride] "r" (src_stride) - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9" - ); -} -#endif // __mips_dsp_rev >= 2 - -#endif // defined(__mips__) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_neon.cc b/src/main/jni/libyuv/source/row_neon.cc deleted file mode 100644 index 1392cf5fc..000000000 --- a/src/main/jni/libyuv/source/row_neon.cc +++ /dev/null @@ -1,3148 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC Neon -#if !defined(LIBYUV_DISABLE_NEON) && defined(__ARM_NEON__) - -// Read 8 Y, 4 U and 4 V from 422 -#define READYUV422 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.32 {d2[0]}, [%1]! \n" \ - MEMACCESS(2) \ - "vld1.32 {d2[1]}, [%2]! \n" - -// Read 8 Y, 2 U and 2 V from 422 -#define READYUV411 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.16 {d2[0]}, [%1]! \n" \ - MEMACCESS(2) \ - "vld1.16 {d2[1]}, [%2]! \n" \ - "vmov.u8 d3, d2 \n" \ - "vzip.u8 d2, d3 \n" - -// Read 8 Y, 8 U and 8 V from 444 -#define READYUV444 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.8 {d2}, [%1]! \n" \ - MEMACCESS(2) \ - "vld1.8 {d3}, [%2]! \n" \ - "vpaddl.u8 q1, q1 \n" \ - "vrshrn.u16 d2, q1, #1 \n" - -// Read 8 Y, and set 4 U and 4 V to 128 -#define READYUV400 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - "vmov.u8 d2, #128 \n" - -// Read 8 Y and 4 UV from NV12 -#define READNV12 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.8 {d2}, [%1]! \n" \ - "vmov.u8 d3, d2 \n"/* split odd/even uv apart */\ - "vuzp.u8 d2, d3 \n" \ - "vtrn.u32 d2, d3 \n" - -// Read 8 Y and 4 VU from NV21 -#define READNV21 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.8 {d2}, [%1]! \n" \ - "vmov.u8 d3, d2 \n"/* split odd/even uv apart */\ - "vuzp.u8 d3, d2 \n" \ - "vtrn.u32 d2, d3 \n" - -// Read 8 YUY2 -#define READYUY2 \ - MEMACCESS(0) \ - "vld2.8 {d0, d2}, [%0]! \n" \ - "vmov.u8 d3, d2 \n" \ - "vuzp.u8 d2, d3 \n" \ - "vtrn.u32 d2, d3 \n" - -// Read 8 UYVY -#define READUYVY \ - MEMACCESS(0) \ - "vld2.8 {d2, d3}, [%0]! \n" \ - "vmov.u8 d0, d3 \n" \ - "vmov.u8 d3, d2 \n" \ - "vuzp.u8 d2, d3 \n" \ - "vtrn.u32 d2, d3 \n" - -#define YUV422TORGB \ - "veor.u8 d2, d26 \n"/*subtract 128 from u and v*/\ - "vmull.s8 q8, d2, d24 \n"/* u/v B/R component */\ - "vmull.s8 q9, d2, d25 \n"/* u/v G component */\ - "vmov.u8 d1, #0 \n"/* split odd/even y apart */\ - "vtrn.u8 d0, d1 \n" \ - "vsub.s16 q0, q0, q15 \n"/* offset y */\ - "vmul.s16 q0, q0, q14 \n" \ - "vadd.s16 d18, d19 \n" \ - "vqadd.s16 d20, d0, d16 \n" /* B */ \ - "vqadd.s16 d21, d1, d16 \n" \ - "vqadd.s16 d22, d0, d17 \n" /* R */ \ - "vqadd.s16 d23, d1, d17 \n" \ - "vqadd.s16 d16, d0, d18 \n" /* G */ \ - "vqadd.s16 d17, d1, d18 \n" \ - "vqshrun.s16 d0, q10, #6 \n" /* B */ \ - "vqshrun.s16 d1, q11, #6 \n" /* G */ \ - "vqshrun.s16 d2, q8, #6 \n" /* R */ \ - "vmovl.u8 q10, d0 \n"/* set up for reinterleave*/\ - "vmovl.u8 q11, d1 \n" \ - "vmovl.u8 q8, d2 \n" \ - "vtrn.u8 d20, d21 \n" \ - "vtrn.u8 d22, d23 \n" \ - "vtrn.u8 d16, d17 \n" \ - "vmov.u8 d21, d16 \n" - -static vec8 kUVToRB = { 127, 127, 127, 127, 102, 102, 102, 102, - 0, 0, 0, 0, 0, 0, 0, 0 }; -static vec8 kUVToG = { -25, -25, -25, -25, -52, -52, -52, -52, - 0, 0, 0, 0, 0, 0, 0, 0 }; - -void I444ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV444 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I422ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I411ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV411 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I422ToBGRARow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vswp.u8 d20, d22 \n" - "vmov.u8 d19, #255 \n" - MEMACCESS(3) - "vst4.8 {d19, d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_bgra), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I422ToABGRRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vswp.u8 d20, d22 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_abgr), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I422ToRGBARow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d19, #255 \n" - MEMACCESS(3) - "vst4.8 {d19, d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_rgba), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I422ToRGB24Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb24, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - MEMACCESS(3) - "vst3.8 {d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_rgb24), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I422ToRAWRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_raw, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vswp.u8 d20, d22 \n" - MEMACCESS(3) - "vst3.8 {d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_raw), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -#define ARGBTORGB565 \ - "vshr.u8 d20, d20, #3 \n" /* B */ \ - "vshr.u8 d21, d21, #2 \n" /* G */ \ - "vshr.u8 d22, d22, #3 \n" /* R */ \ - "vmovl.u8 q8, d20 \n" /* B */ \ - "vmovl.u8 q9, d21 \n" /* G */ \ - "vmovl.u8 q10, d22 \n" /* R */ \ - "vshl.u16 q9, q9, #5 \n" /* G */ \ - "vshl.u16 q10, q10, #11 \n" /* R */ \ - "vorr q0, q8, q9 \n" /* BG */ \ - "vorr q0, q0, q10 \n" /* BGR */ - -void I422ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb565, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - ARGBTORGB565 - MEMACCESS(3) - "vst1.8 {q0}, [%3]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_rgb565), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -#define ARGBTOARGB1555 \ - "vshr.u8 q10, q10, #3 \n" /* B */ \ - "vshr.u8 d22, d22, #3 \n" /* R */ \ - "vshr.u8 d23, d23, #7 \n" /* A */ \ - "vmovl.u8 q8, d20 \n" /* B */ \ - "vmovl.u8 q9, d21 \n" /* G */ \ - "vmovl.u8 q10, d22 \n" /* R */ \ - "vmovl.u8 q11, d23 \n" /* A */ \ - "vshl.u16 q9, q9, #5 \n" /* G */ \ - "vshl.u16 q10, q10, #10 \n" /* R */ \ - "vshl.u16 q11, q11, #15 \n" /* A */ \ - "vorr q0, q8, q9 \n" /* BG */ \ - "vorr q1, q10, q11 \n" /* RA */ \ - "vorr q0, q0, q1 \n" /* BGRA */ - -void I422ToARGB1555Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb1555, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - ARGBTOARGB1555 - MEMACCESS(3) - "vst1.8 {q0}, [%3]! \n" // store 8 pixels ARGB1555. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb1555), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -#define ARGBTOARGB4444 \ - "vshr.u8 d20, d20, #4 \n" /* B */ \ - "vbic.32 d21, d21, d4 \n" /* G */ \ - "vshr.u8 d22, d22, #4 \n" /* R */ \ - "vbic.32 d23, d23, d4 \n" /* A */ \ - "vorr d0, d20, d21 \n" /* BG */ \ - "vorr d1, d22, d23 \n" /* RA */ \ - "vzip.u8 d0, d1 \n" /* BGRA */ - -void I422ToARGB4444Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb4444, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - "vmov.u8 d4, #0x0f \n" // bits to clear with vbic. - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - ARGBTOARGB4444 - MEMACCESS(3) - "vst1.8 {q0}, [%3]! \n" // store 8 pixels ARGB4444. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb4444), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void YToARGBRow_NEON(const uint8* src_y, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {d24}, [%3] \n" - MEMACCESS(4) - "vld1.8 {d25}, [%4] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV400 - YUV422TORGB - "subs %2, %2, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(&kUVToRB), // %3 - "r"(&kUVToG) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void I400ToARGBRow_NEON(const uint8* src_y, - uint8* dst_argb, - int width) { - asm volatile ( - ".p2align 2 \n" - "vmov.u8 d23, #255 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d20}, [%0]! \n" - "vmov d21, d20 \n" - "vmov d22, d20 \n" - "subs %2, %2, #8 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "d20", "d21", "d22", "d23" - ); -} - -void NV12ToARGBRow_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV12 - YUV422TORGB - "subs %3, %3, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(2) - "vst4.8 {d20, d21, d22, d23}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void NV21ToARGBRow_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV21 - YUV422TORGB - "subs %3, %3, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(2) - "vst4.8 {d20, d21, d22, d23}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void NV12ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_rgb565, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV12 - YUV422TORGB - "subs %3, %3, #8 \n" - ARGBTORGB565 - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_rgb565), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void NV21ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_rgb565, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV21 - YUV422TORGB - "subs %3, %3, #8 \n" - ARGBTORGB565 - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_rgb565), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void YUY2ToARGBRow_NEON(const uint8* src_yuy2, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {d24}, [%3] \n" - MEMACCESS(4) - "vld1.8 {d25}, [%4] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUY2 - YUV422TORGB - "subs %2, %2, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(&kUVToRB), // %3 - "r"(&kUVToG) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void UYVYToARGBRow_NEON(const uint8* src_uyvy, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {d24}, [%3] \n" - MEMACCESS(4) - "vld1.8 {d25}, [%4] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READUYVY - YUV422TORGB - "subs %2, %2, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(&kUVToRB), // %3 - "r"(&kUVToG) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// Reads 16 pairs of UV and write even values to dst_u and odd to dst_v. -void SplitUVRow_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld2.8 {q0, q1}, [%0]! \n" // load 16 pairs of UV - "subs %3, %3, #16 \n" // 16 processed per loop - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store U - MEMACCESS(2) - "vst1.8 {q1}, [%2]! \n" // store V - "bgt 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(width) // %3 // Output registers - : // Input registers - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -// Reads 16 U's and V's and writes out 16 pairs of UV. -void MergeUVRow_NEON(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load U - MEMACCESS(1) - "vld1.8 {q1}, [%1]! \n" // load V - "subs %3, %3, #16 \n" // 16 processed per loop - MEMACCESS(2) - "vst2.u8 {q0, q1}, [%2]! \n" // store 16 pairs of UV - "bgt 1b \n" - : - "+r"(src_u), // %0 - "+r"(src_v), // %1 - "+r"(dst_uv), // %2 - "+r"(width) // %3 // Output registers - : // Input registers - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -// Copy multiple of 32. vld4.8 allow unaligned and is fastest on a15. -void CopyRow_NEON(const uint8* src, uint8* dst, int count) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d0, d1, d2, d3}, [%0]! \n" // load 32 - "subs %2, %2, #32 \n" // 32 processed per loop - MEMACCESS(1) - "vst1.8 {d0, d1, d2, d3}, [%1]! \n" // store 32 - "bgt 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(count) // %2 // Output registers - : // Input registers - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -// SetRow8 writes 'count' bytes using a 32 bit value repeated. -void SetRow_NEON(uint8* dst, uint32 v32, int count) { - asm volatile ( - "vdup.u32 q0, %2 \n" // duplicate 4 ints - "1: \n" - "subs %1, %1, #16 \n" // 16 bytes per loop - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" // store - "bgt 1b \n" - : "+r"(dst), // %0 - "+r"(count) // %1 - : "r"(v32) // %2 - : "cc", "memory", "q0" - ); -} - -// TODO(fbarchard): Make fully assembler -// SetRow32 writes 'count' words using a 32 bit value repeated. -void ARGBSetRows_NEON(uint8* dst, uint32 v32, int width, - int dst_stride, int height) { - for (int y = 0; y < height; ++y) { - SetRow_NEON(dst, v32, width << 2); - dst += dst_stride; - } -} - -void MirrorRow_NEON(const uint8* src, uint8* dst, int width) { - asm volatile ( - // Start at end of source row. - "mov r3, #-16 \n" - "add %0, %0, %2 \n" - "sub %0, #16 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0], r3 \n" // src -= 16 - "subs %2, #16 \n" // 16 pixels per loop. - "vrev64.8 q0, q0 \n" - MEMACCESS(1) - "vst1.8 {d1}, [%1]! \n" // dst += 16 - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" - "bgt 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "r3", "q0" - ); -} - -void MirrorUVRow_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width) { - asm volatile ( - // Start at end of source row. - "mov r12, #-16 \n" - "add %0, %0, %3, lsl #1 \n" - "sub %0, #16 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld2.8 {d0, d1}, [%0], r12 \n" // src -= 16 - "subs %3, #8 \n" // 8 pixels per loop. - "vrev64.8 q0, q0 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // dst += 8 - MEMACCESS(2) - "vst1.8 {d1}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "r12", "q0" - ); -} - -void ARGBMirrorRow_NEON(const uint8* src, uint8* dst, int width) { - asm volatile ( - // Start at end of source row. - "mov r3, #-16 \n" - "add %0, %0, %2, lsl #2 \n" - "sub %0, #16 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0], r3 \n" // src -= 16 - "subs %2, #4 \n" // 4 pixels per loop. - "vrev64.32 q0, q0 \n" - MEMACCESS(1) - "vst1.8 {d1}, [%1]! \n" // dst += 16 - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" - "bgt 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "r3", "q0" - ); -} - -void RGB24ToARGBRow_NEON(const uint8* src_rgb24, uint8* dst_argb, int pix) { - asm volatile ( - "vmov.u8 d4, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d1, d2, d3}, [%0]! \n" // load 8 pixels of RGB24. - "subs %2, %2, #8 \n" // 8 processed per loop. - MEMACCESS(1) - "vst4.8 {d1, d2, d3, d4}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d1", "d2", "d3", "d4" // Clobber List - ); -} - -void RAWToARGBRow_NEON(const uint8* src_raw, uint8* dst_argb, int pix) { - asm volatile ( - "vmov.u8 d4, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d1, d2, d3}, [%0]! \n" // load 8 pixels of RAW. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vswp.u8 d1, d3 \n" // swap R, B - MEMACCESS(1) - "vst4.8 {d1, d2, d3, d4}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_raw), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d1", "d2", "d3", "d4" // Clobber List - ); -} - -#define RGB565TOARGB \ - "vshrn.u16 d6, q0, #5 \n" /* G xxGGGGGG */ \ - "vuzp.u8 d0, d1 \n" /* d0 xxxBBBBB RRRRRxxx */ \ - "vshl.u8 d6, d6, #2 \n" /* G GGGGGG00 upper 6 */ \ - "vshr.u8 d1, d1, #3 \n" /* R 000RRRRR lower 5 */ \ - "vshl.u8 q0, q0, #3 \n" /* B,R BBBBB000 upper 5 */ \ - "vshr.u8 q2, q0, #5 \n" /* B,R 00000BBB lower 3 */ \ - "vorr.u8 d0, d0, d4 \n" /* B */ \ - "vshr.u8 d4, d6, #6 \n" /* G 000000GG lower 2 */ \ - "vorr.u8 d2, d1, d5 \n" /* R */ \ - "vorr.u8 d1, d4, d6 \n" /* G */ - -void RGB565ToARGBRow_NEON(const uint8* src_rgb565, uint8* dst_argb, int pix) { - asm volatile ( - "vmov.u8 d3, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 RGB565 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - RGB565TOARGB - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_rgb565), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3" // Clobber List - ); -} - -#define ARGB1555TOARGB \ - "vshrn.u16 d7, q0, #8 \n" /* A Arrrrrxx */ \ - "vshr.u8 d6, d7, #2 \n" /* R xxxRRRRR */ \ - "vshrn.u16 d5, q0, #5 \n" /* G xxxGGGGG */ \ - "vmovn.u16 d4, q0 \n" /* B xxxBBBBB */ \ - "vshr.u8 d7, d7, #7 \n" /* A 0000000A */ \ - "vneg.s8 d7, d7 \n" /* A AAAAAAAA upper 8 */ \ - "vshl.u8 d6, d6, #3 \n" /* R RRRRR000 upper 5 */ \ - "vshr.u8 q1, q3, #5 \n" /* R,A 00000RRR lower 3 */ \ - "vshl.u8 q0, q2, #3 \n" /* B,G BBBBB000 upper 5 */ \ - "vshr.u8 q2, q0, #5 \n" /* B,G 00000BBB lower 3 */ \ - "vorr.u8 q1, q1, q3 \n" /* R,A */ \ - "vorr.u8 q0, q0, q2 \n" /* B,G */ \ - -// RGB555TOARGB is same as ARGB1555TOARGB but ignores alpha. -#define RGB555TOARGB \ - "vshrn.u16 d6, q0, #5 \n" /* G xxxGGGGG */ \ - "vuzp.u8 d0, d1 \n" /* d0 xxxBBBBB xRRRRRxx */ \ - "vshl.u8 d6, d6, #3 \n" /* G GGGGG000 upper 5 */ \ - "vshr.u8 d1, d1, #2 \n" /* R 00xRRRRR lower 5 */ \ - "vshl.u8 q0, q0, #3 \n" /* B,R BBBBB000 upper 5 */ \ - "vshr.u8 q2, q0, #5 \n" /* B,R 00000BBB lower 3 */ \ - "vorr.u8 d0, d0, d4 \n" /* B */ \ - "vshr.u8 d4, d6, #5 \n" /* G 00000GGG lower 3 */ \ - "vorr.u8 d2, d1, d5 \n" /* R */ \ - "vorr.u8 d1, d4, d6 \n" /* G */ - -void ARGB1555ToARGBRow_NEON(const uint8* src_argb1555, uint8* dst_argb, - int pix) { - asm volatile ( - "vmov.u8 d3, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB1555 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB1555TOARGB - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb1555), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3" // Clobber List - ); -} - -#define ARGB4444TOARGB \ - "vuzp.u8 d0, d1 \n" /* d0 BG, d1 RA */ \ - "vshl.u8 q2, q0, #4 \n" /* B,R BBBB0000 */ \ - "vshr.u8 q1, q0, #4 \n" /* G,A 0000GGGG */ \ - "vshr.u8 q0, q2, #4 \n" /* B,R 0000BBBB */ \ - "vorr.u8 q0, q0, q2 \n" /* B,R BBBBBBBB */ \ - "vshl.u8 q2, q1, #4 \n" /* G,A GGGG0000 */ \ - "vorr.u8 q1, q1, q2 \n" /* G,A GGGGGGGG */ \ - "vswp.u8 d1, d2 \n" /* B,R,G,A -> B,G,R,A */ - -void ARGB4444ToARGBRow_NEON(const uint8* src_argb4444, uint8* dst_argb, - int pix) { - asm volatile ( - "vmov.u8 d3, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB4444 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB4444TOARGB - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb4444), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2" // Clobber List - ); -} - -void ARGBToRGB24Row_NEON(const uint8* src_argb, uint8* dst_rgb24, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d1, d2, d3, d4}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - MEMACCESS(1) - "vst3.8 {d1, d2, d3}, [%1]! \n" // store 8 pixels of RGB24. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_rgb24), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d1", "d2", "d3", "d4" // Clobber List - ); -} - -void ARGBToRAWRow_NEON(const uint8* src_argb, uint8* dst_raw, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d1, d2, d3, d4}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vswp.u8 d1, d3 \n" // swap R, B - MEMACCESS(1) - "vst3.8 {d1, d2, d3}, [%1]! \n" // store 8 pixels of RAW. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_raw), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d1", "d2", "d3", "d4" // Clobber List - ); -} - -void YUY2ToYRow_NEON(const uint8* src_yuy2, uint8* dst_y, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld2.8 {q0, q1}, [%0]! \n" // load 16 pixels of YUY2. - "subs %2, %2, #16 \n" // 16 processed per loop. - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 16 pixels of Y. - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -void UYVYToYRow_NEON(const uint8* src_uyvy, uint8* dst_y, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld2.8 {q0, q1}, [%0]! \n" // load 16 pixels of UYVY. - "subs %2, %2, #16 \n" // 16 processed per loop. - MEMACCESS(1) - "vst1.8 {q1}, [%1]! \n" // store 16 pixels of Y. - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -void YUY2ToUV422Row_NEON(const uint8* src_yuy2, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 16 pixels of YUY2. - "subs %3, %3, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "vst1.8 {d1}, [%1]! \n" // store 8 U. - MEMACCESS(2) - "vst1.8 {d3}, [%2]! \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "d0", "d1", "d2", "d3" // Clobber List - ); -} - -void UYVYToUV422Row_NEON(const uint8* src_uyvy, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 16 pixels of UYVY. - "subs %3, %3, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 U. - MEMACCESS(2) - "vst1.8 {d2}, [%2]! \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "d0", "d1", "d2", "d3" // Clobber List - ); -} - -void YUY2ToUVRow_NEON(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // stride + src_yuy2 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 16 pixels of YUY2. - "subs %4, %4, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "vld4.8 {d4, d5, d6, d7}, [%1]! \n" // load next row YUY2. - "vrhadd.u8 d1, d1, d5 \n" // average rows of U - "vrhadd.u8 d3, d3, d7 \n" // average rows of V - MEMACCESS(2) - "vst1.8 {d1}, [%2]! \n" // store 8 U. - MEMACCESS(3) - "vst1.8 {d3}, [%3]! \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(stride_yuy2), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7" // Clobber List - ); -} - -void UYVYToUVRow_NEON(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // stride + src_uyvy - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 16 pixels of UYVY. - "subs %4, %4, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "vld4.8 {d4, d5, d6, d7}, [%1]! \n" // load next row UYVY. - "vrhadd.u8 d0, d0, d4 \n" // average rows of U - "vrhadd.u8 d2, d2, d6 \n" // average rows of V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 U. - MEMACCESS(3) - "vst1.8 {d2}, [%3]! \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(stride_uyvy), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7" // Clobber List - ); -} - -void HalfRow_NEON(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix) { - asm volatile ( - // change the stride to row 2 pointer - "add %1, %0 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load row 1 16 pixels. - "subs %3, %3, #16 \n" // 16 processed per loop - MEMACCESS(1) - "vld1.8 {q1}, [%1]! \n" // load row 2 16 pixels. - "vrhadd.u8 q0, q1 \n" // average row 1 and 2 - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_uv), // %0 - "+r"(src_uv_stride), // %1 - "+r"(dst_uv), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -// Select 2 channels from ARGB on alternating pixels. e.g. BGBGBGBG -void ARGBToBayerRow_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) { - asm volatile ( - "vmov.u32 d6[0], %3 \n" // selector - "1: \n" - MEMACCESS(0) - "vld1.8 {q0, q1}, [%0]! \n" // load row 8 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop - "vtbl.8 d4, {d0, d1}, d6 \n" // look up 4 pixels - "vtbl.8 d5, {d2, d3}, d6 \n" // look up 4 pixels - "vtrn.u32 d4, d5 \n" // combine 8 pixels - MEMACCESS(1) - "vst1.8 {d4}, [%1]! \n" // store 8. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_bayer), // %1 - "+r"(pix) // %2 - : "r"(selector) // %3 - : "cc", "memory", "q0", "q1", "q2", "q3" // Clobber List - ); -} - -// Select G channels from ARGB. e.g. GGGGGGGG -void ARGBToBayerGGRow_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 /*selector*/, int pix) { - asm volatile ( - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load row 8 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop - MEMACCESS(1) - "vst1.8 {d1}, [%1]! \n" // store 8 G's. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_bayer), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -// For BGRAToARGB, ABGRToARGB, RGBAToARGB, and ARGBToRGBA. -void ARGBShuffleRow_NEON(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {q2}, [%3] \n" // shuffler - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 4 pixels. - "subs %2, %2, #4 \n" // 4 processed per loop - "vtbl.8 d2, {d0, d1}, d4 \n" // look up 2 first pixels - "vtbl.8 d3, {d0, d1}, d5 \n" // look up 2 next pixels - MEMACCESS(1) - "vst1.8 {q1}, [%1]! \n" // store 4. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "r"(shuffler) // %3 - : "cc", "memory", "q0", "q1", "q2" // Clobber List - ); -} - -void I422ToYUY2Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld2.8 {d0, d2}, [%0]! \n" // load 16 Ys - MEMACCESS(1) - "vld1.8 {d1}, [%1]! \n" // load 8 Us - MEMACCESS(2) - "vld1.8 {d3}, [%2]! \n" // load 8 Vs - "subs %4, %4, #16 \n" // 16 pixels - MEMACCESS(3) - "vst4.8 {d0, d1, d2, d3}, [%3]! \n" // Store 8 YUY2/16 pixels. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_yuy2), // %3 - "+r"(width) // %4 - : - : "cc", "memory", "d0", "d1", "d2", "d3" - ); -} - -void I422ToUYVYRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld2.8 {d1, d3}, [%0]! \n" // load 16 Ys - MEMACCESS(1) - "vld1.8 {d0}, [%1]! \n" // load 8 Us - MEMACCESS(2) - "vld1.8 {d2}, [%2]! \n" // load 8 Vs - "subs %4, %4, #16 \n" // 16 pixels - MEMACCESS(3) - "vst4.8 {d0, d1, d2, d3}, [%3]! \n" // Store 8 UYVY/16 pixels. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_uyvy), // %3 - "+r"(width) // %4 - : - : "cc", "memory", "d0", "d1", "d2", "d3" - ); -} - -void ARGBToRGB565Row_NEON(const uint8* src_argb, uint8* dst_rgb565, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d21, d22, d23}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGBTORGB565 - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_rgb565), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q8", "q9", "q10", "q11" - ); -} - -void ARGBToARGB1555Row_NEON(const uint8* src_argb, uint8* dst_argb1555, - int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d21, d22, d23}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGBTOARGB1555 - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 8 pixels ARGB1555. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb1555), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q8", "q9", "q10", "q11" - ); -} - -void ARGBToARGB4444Row_NEON(const uint8* src_argb, uint8* dst_argb4444, - int pix) { - asm volatile ( - "vmov.u8 d4, #0x0f \n" // bits to clear with vbic. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d21, d22, d23}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGBTOARGB4444 - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 8 pixels ARGB4444. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb4444), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q8", "q9", "q10", "q11" - ); -} - -void ARGBToYRow_NEON(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q12", "q13" - ); -} - -void ARGBToYJRow_NEON(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #15 \n" // B * 0.11400 coefficient - "vmov.u8 d25, #75 \n" // G * 0.58700 coefficient - "vmov.u8 d26, #38 \n" // R * 0.29900 coefficient - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 15 bit to 8 bit Y - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q12", "q13" - ); -} - -// 8x1 pixels. -void ARGBToUV444Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "vmov.u8 d24, #112 \n" // UB / VR 0.875 coefficient - "vmov.u8 d25, #74 \n" // UG -0.5781 coefficient - "vmov.u8 d26, #38 \n" // UR -0.2969 coefficient - "vmov.u8 d27, #18 \n" // VB -0.1406 coefficient - "vmov.u8 d28, #94 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vmull.u8 q2, d0, d24 \n" // B - "vmlsl.u8 q2, d1, d25 \n" // G - "vmlsl.u8 q2, d2, d26 \n" // R - "vadd.u16 q2, q2, q15 \n" // +128 -> unsigned - - "vmull.u8 q3, d2, d24 \n" // R - "vmlsl.u8 q3, d1, d28 \n" // G - "vmlsl.u8 q3, d0, d27 \n" // B - "vadd.u16 q3, q3, q15 \n" // +128 -> unsigned - - "vqshrn.u16 d0, q2, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q3, #8 \n" // 16 bit to 8 bit V - - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels U. - MEMACCESS(2) - "vst1.8 {d1}, [%2]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q12", "q13", "q14", "q15" - ); -} - -// 16x1 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -void ARGBToUV422Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - - "subs %3, %3, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q0, q10 \n" // B - "vmls.s16 q8, q1, q11 \n" // G - "vmls.s16 q8, q2, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - - "vmul.s16 q9, q2, q10 \n" // R - "vmls.s16 q9, q1, q14 \n" // G - "vmls.s16 q9, q0, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels U. - MEMACCESS(2) - "vst1.8 {d1}, [%2]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// 32x1 pixels -> 8x1. pix is number of argb pixels. e.g. 32. -void ARGBToUV411Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(0) - "vld4.8 {d8, d10, d12, d14}, [%0]! \n" // load 8 more ARGB pixels. - MEMACCESS(0) - "vld4.8 {d9, d11, d13, d15}, [%0]! \n" // load last 8 ARGB pixels. - "vpaddl.u8 q4, q4 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q5, q5 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q6, q6 \n" // R 16 bytes -> 8 shorts. - - "vpadd.u16 d0, d0, d1 \n" // B 16 shorts -> 8 shorts. - "vpadd.u16 d1, d8, d9 \n" // B - "vpadd.u16 d2, d2, d3 \n" // G 16 shorts -> 8 shorts. - "vpadd.u16 d3, d10, d11 \n" // G - "vpadd.u16 d4, d4, d5 \n" // R 16 shorts -> 8 shorts. - "vpadd.u16 d5, d12, d13 \n" // R - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %3, %3, #32 \n" // 32 processed per loop. - "vmul.s16 q8, q0, q10 \n" // B - "vmls.s16 q8, q1, q11 \n" // G - "vmls.s16 q8, q2, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q2, q10 \n" // R - "vmls.s16 q9, q1, q14 \n" // G - "vmls.s16 q9, q0, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels U. - MEMACCESS(2) - "vst1.8 {d1}, [%2]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -#define RGBTOUV(QB, QG, QR) \ - "vmul.s16 q8, " #QB ", q10 \n" /* B */ \ - "vmls.s16 q8, " #QG ", q11 \n" /* G */ \ - "vmls.s16 q8, " #QR ", q12 \n" /* R */ \ - "vadd.u16 q8, q8, q15 \n" /* +128 -> unsigned */ \ - "vmul.s16 q9, " #QR ", q10 \n" /* R */ \ - "vmls.s16 q9, " #QG ", q14 \n" /* G */ \ - "vmls.s16 q9, " #QB ", q13 \n" /* B */ \ - "vadd.u16 q9, q9, q15 \n" /* +128 -> unsigned */ \ - "vqshrn.u16 d0, q8, #8 \n" /* 16 bit to 8 bit U */ \ - "vqshrn.u16 d1, q9, #8 \n" /* 16 bit to 8 bit V */ - -// TODO(fbarchard): Consider vhadd vertical, then vpaddl horizontal, avoid shr. -void ARGBToUVRow_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more ARGB pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 ARGB pixels. - "vpadal.u8 q0, q4 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stride_argb), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// TODO(fbarchard): Subsample match C code. -void ARGBToUVJRow_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #127 / 2 \n" // UB / VR 0.500 coefficient - "vmov.s16 q11, #84 / 2 \n" // UG -0.33126 coefficient - "vmov.s16 q12, #43 / 2 \n" // UR -0.16874 coefficient - "vmov.s16 q13, #20 / 2 \n" // VB -0.08131 coefficient - "vmov.s16 q14, #107 / 2 \n" // VG -0.41869 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more ARGB pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 ARGB pixels. - "vpadal.u8 q0, q4 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stride_argb), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void BGRAToUVRow_NEON(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_bgra - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 BGRA pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 BGRA pixels. - "vpaddl.u8 q3, q3 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more BGRA pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 BGRA pixels. - "vpadal.u8 q3, q7 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q1, q1, #1 \n" // 2x average - "vrshr.u16 q2, q2, #1 \n" - "vrshr.u16 q3, q3, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q3, q2, q1) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_bgra), // %0 - "+r"(src_stride_bgra), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void ABGRToUVRow_NEON(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_abgr - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ABGR pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ABGR pixels. - "vpaddl.u8 q2, q2 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q0, q0 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more ABGR pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 ABGR pixels. - "vpadal.u8 q2, q6 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q0, q4 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q2, q1, q0) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_abgr), // %0 - "+r"(src_stride_abgr), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void RGBAToUVRow_NEON(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_rgba - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 RGBA pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 RGBA pixels. - "vpaddl.u8 q0, q1 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q2 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q3 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more RGBA pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 RGBA pixels. - "vpadal.u8 q0, q5 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q6 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q7 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_rgba), // %0 - "+r"(src_stride_rgba), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void RGB24ToUVRow_NEON(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_rgb24 - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d0, d2, d4}, [%0]! \n" // load 8 RGB24 pixels. - MEMACCESS(0) - "vld3.8 {d1, d3, d5}, [%0]! \n" // load next 8 RGB24 pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld3.8 {d8, d10, d12}, [%1]! \n" // load 8 more RGB24 pixels. - MEMACCESS(1) - "vld3.8 {d9, d11, d13}, [%1]! \n" // load last 8 RGB24 pixels. - "vpadal.u8 q0, q4 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(src_stride_rgb24), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void RAWToUVRow_NEON(const uint8* src_raw, int src_stride_raw, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_raw - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d0, d2, d4}, [%0]! \n" // load 8 RAW pixels. - MEMACCESS(0) - "vld3.8 {d1, d3, d5}, [%0]! \n" // load next 8 RAW pixels. - "vpaddl.u8 q2, q2 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q0, q0 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld3.8 {d8, d10, d12}, [%1]! \n" // load 8 more RAW pixels. - MEMACCESS(1) - "vld3.8 {d9, d11, d13}, [%1]! \n" // load last 8 RAW pixels. - "vpadal.u8 q2, q6 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q0, q4 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q2, q1, q0) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_raw), // %0 - "+r"(src_stride_raw), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -void RGB565ToUVRow_NEON(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 RGB565 pixels. - RGB565TOARGB - "vpaddl.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // next 8 RGB565 pixels. - RGB565TOARGB - "vpaddl.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // load 8 RGB565 pixels. - RGB565TOARGB - "vpadal.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // next 8 RGB565 pixels. - RGB565TOARGB - "vpadal.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - "vrshr.u16 q4, q4, #1 \n" // 2x average - "vrshr.u16 q5, q5, #1 \n" - "vrshr.u16 q6, q6, #1 \n" - - "subs %4, %4, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q4, q10 \n" // B - "vmls.s16 q8, q5, q11 \n" // G - "vmls.s16 q8, q6, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q6, q10 \n" // R - "vmls.s16 q9, q5, q14 \n" // G - "vmls.s16 q9, q4, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_rgb565), // %0 - "+r"(src_stride_rgb565), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -void ARGB1555ToUVRow_NEON(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB1555 pixels. - RGB555TOARGB - "vpaddl.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // next 8 ARGB1555 pixels. - RGB555TOARGB - "vpaddl.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // load 8 ARGB1555 pixels. - RGB555TOARGB - "vpadal.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // next 8 ARGB1555 pixels. - RGB555TOARGB - "vpadal.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - "vrshr.u16 q4, q4, #1 \n" // 2x average - "vrshr.u16 q5, q5, #1 \n" - "vrshr.u16 q6, q6, #1 \n" - - "subs %4, %4, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q4, q10 \n" // B - "vmls.s16 q8, q5, q11 \n" // G - "vmls.s16 q8, q6, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q6, q10 \n" // R - "vmls.s16 q9, q5, q14 \n" // G - "vmls.s16 q9, q4, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb1555), // %0 - "+r"(src_stride_argb1555), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -void ARGB4444ToUVRow_NEON(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpaddl.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // next 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpaddl.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // load 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpadal.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // next 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpadal.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - "vrshr.u16 q4, q4, #1 \n" // 2x average - "vrshr.u16 q5, q5, #1 \n" - "vrshr.u16 q6, q6, #1 \n" - - "subs %4, %4, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q4, q10 \n" // B - "vmls.s16 q8, q5, q11 \n" // G - "vmls.s16 q8, q6, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q6, q10 \n" // R - "vmls.s16 q9, q5, q14 \n" // G - "vmls.s16 q9, q4, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb4444), // %0 - "+r"(src_stride_argb4444), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -void RGB565ToYRow_NEON(const uint8* src_rgb565, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 RGB565 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - RGB565TOARGB - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_rgb565), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q12", "q13" - ); -} - -void ARGB1555ToYRow_NEON(const uint8* src_argb1555, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB1555 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB1555TOARGB - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb1555), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q12", "q13" - ); -} - -void ARGB4444ToYRow_NEON(const uint8* src_argb4444, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB4444 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB4444TOARGB - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb4444), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q12", "q13" - ); -} - -void BGRAToYRow_NEON(const uint8* src_bgra, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d4, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d5, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d6, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d7, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 pixels of BGRA. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q8, d1, d4 \n" // R - "vmlal.u8 q8, d2, d5 \n" // G - "vmlal.u8 q8, d3, d6 \n" // B - "vqrshrun.s16 d0, q8, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d7 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_bgra), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", "q8" - ); -} - -void ABGRToYRow_NEON(const uint8* src_abgr, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d4, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d5, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d6, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d7, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 pixels of ABGR. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q8, d0, d4 \n" // R - "vmlal.u8 q8, d1, d5 \n" // G - "vmlal.u8 q8, d2, d6 \n" // B - "vqrshrun.s16 d0, q8, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d7 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_abgr), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", "q8" - ); -} - -void RGBAToYRow_NEON(const uint8* src_rgba, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d4, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d5, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d6, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d7, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 pixels of RGBA. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q8, d1, d4 \n" // B - "vmlal.u8 q8, d2, d5 \n" // G - "vmlal.u8 q8, d3, d6 \n" // R - "vqrshrun.s16 d0, q8, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d7 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_rgba), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", "q8" - ); -} - -void RGB24ToYRow_NEON(const uint8* src_rgb24, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d4, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d5, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d6, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d7, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d0, d1, d2}, [%0]! \n" // load 8 pixels of RGB24. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q8, d0, d4 \n" // B - "vmlal.u8 q8, d1, d5 \n" // G - "vmlal.u8 q8, d2, d6 \n" // R - "vqrshrun.s16 d0, q8, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d7 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", "q8" - ); -} - -void RAWToYRow_NEON(const uint8* src_raw, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d4, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d5, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d6, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d7, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d0, d1, d2}, [%0]! \n" // load 8 pixels of RAW. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q8, d0, d4 \n" // B - "vmlal.u8 q8, d1, d5 \n" // G - "vmlal.u8 q8, d2, d6 \n" // R - "vqrshrun.s16 d0, q8, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d7 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_raw), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "d0", "d1", "d2", "d3", "d4", "d5", "d6", "d7", "q8" - ); -} - -// Bilinear filter 16x2 -> 16x1 -void InterpolateRow_NEON(uint8* dst_ptr, - const uint8* src_ptr, ptrdiff_t src_stride, - int dst_width, int source_y_fraction) { - asm volatile ( - "cmp %4, #0 \n" - "beq 100f \n" - "add %2, %1 \n" - "cmp %4, #64 \n" - "beq 75f \n" - "cmp %4, #128 \n" - "beq 50f \n" - "cmp %4, #192 \n" - "beq 25f \n" - - "vdup.8 d5, %4 \n" - "rsb %4, #256 \n" - "vdup.8 d4, %4 \n" - // General purpose row blend. - "1: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q1}, [%2]! \n" - "subs %3, %3, #16 \n" - "vmull.u8 q13, d0, d4 \n" - "vmull.u8 q14, d1, d4 \n" - "vmlal.u8 q13, d2, d5 \n" - "vmlal.u8 q14, d3, d5 \n" - "vrshrn.u16 d0, q13, #8 \n" - "vrshrn.u16 d1, q14, #8 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 1b \n" - "b 99f \n" - - // Blend 25 / 75. - "25: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q1}, [%2]! \n" - "subs %3, %3, #16 \n" - "vrhadd.u8 q0, q1 \n" - "vrhadd.u8 q0, q1 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 25b \n" - "b 99f \n" - - // Blend 50 / 50. - "50: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q1}, [%2]! \n" - "subs %3, %3, #16 \n" - "vrhadd.u8 q0, q1 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 50b \n" - "b 99f \n" - - // Blend 75 / 25. - "75: \n" - MEMACCESS(1) - "vld1.8 {q1}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q0}, [%2]! \n" - "subs %3, %3, #16 \n" - "vrhadd.u8 q0, q1 \n" - "vrhadd.u8 q0, q1 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 75b \n" - "b 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - "100: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - "subs %3, %3, #16 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 100b \n" - - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(src_stride), // %2 - "+r"(dst_width), // %3 - "+r"(source_y_fraction) // %4 - : - : "cc", "memory", "q0", "q1", "d4", "d5", "q13", "q14" - ); -} - -// dr * (256 - sa) / 256 + sr = dr - dr * sa / 256 + sr -void ARGBBlendRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - "subs %3, #8 \n" - "blt 89f \n" - // Blend 8 pixels. - "8: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 pixels of ARGB0. - MEMACCESS(1) - "vld4.8 {d4, d5, d6, d7}, [%1]! \n" // load 8 pixels of ARGB1. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vmull.u8 q10, d4, d3 \n" // db * a - "vmull.u8 q11, d5, d3 \n" // dg * a - "vmull.u8 q12, d6, d3 \n" // dr * a - "vqrshrn.u16 d20, q10, #8 \n" // db >>= 8 - "vqrshrn.u16 d21, q11, #8 \n" // dg >>= 8 - "vqrshrn.u16 d22, q12, #8 \n" // dr >>= 8 - "vqsub.u8 q2, q2, q10 \n" // dbg - dbg * a / 256 - "vqsub.u8 d6, d6, d22 \n" // dr - dr * a / 256 - "vqadd.u8 q0, q0, q2 \n" // + sbg - "vqadd.u8 d2, d2, d6 \n" // + sr - "vmov.u8 d3, #255 \n" // a = 255 - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" // store 8 pixels of ARGB. - "bge 8b \n" - - "89: \n" - "adds %3, #8-1 \n" - "blt 99f \n" - - // Blend 1 pixels. - "1: \n" - MEMACCESS(0) - "vld4.8 {d0[0],d1[0],d2[0],d3[0]}, [%0]! \n" // load 1 pixel ARGB0. - MEMACCESS(1) - "vld4.8 {d4[0],d5[0],d6[0],d7[0]}, [%1]! \n" // load 1 pixel ARGB1. - "subs %3, %3, #1 \n" // 1 processed per loop. - "vmull.u8 q10, d4, d3 \n" // db * a - "vmull.u8 q11, d5, d3 \n" // dg * a - "vmull.u8 q12, d6, d3 \n" // dr * a - "vqrshrn.u16 d20, q10, #8 \n" // db >>= 8 - "vqrshrn.u16 d21, q11, #8 \n" // dg >>= 8 - "vqrshrn.u16 d22, q12, #8 \n" // dr >>= 8 - "vqsub.u8 q2, q2, q10 \n" // dbg - dbg * a / 256 - "vqsub.u8 d6, d6, d22 \n" // dr - dr * a / 256 - "vqadd.u8 q0, q0, q2 \n" // + sbg - "vqadd.u8 d2, d2, d6 \n" // + sr - "vmov.u8 d3, #255 \n" // a = 255 - MEMACCESS(2) - "vst4.8 {d0[0],d1[0],d2[0],d3[0]}, [%2]! \n" // store 1 pixel. - "bge 1b \n" - - "99: \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q10", "q11", "q12" - ); -} - -// Attenuate 8 pixels at a time. -void ARGBAttenuateRow_NEON(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - // Attenuate 8 pixels. - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q10, d0, d3 \n" // b * a - "vmull.u8 q11, d1, d3 \n" // g * a - "vmull.u8 q12, d2, d3 \n" // r * a - "vqrshrn.u16 d0, q10, #8 \n" // b >>= 8 - "vqrshrn.u16 d1, q11, #8 \n" // g >>= 8 - "vqrshrn.u16 d2, q12, #8 \n" // r >>= 8 - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "q0", "q1", "q10", "q11", "q12" - ); -} - -// Quantize 8 ARGB pixels (32 bytes). -// dst = (dst * scale >> 16) * interval_size + interval_offset; -void ARGBQuantizeRow_NEON(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width) { - asm volatile ( - "vdup.u16 q8, %2 \n" - "vshr.u16 q8, q8, #1 \n" // scale >>= 1 - "vdup.u16 q9, %3 \n" // interval multiply. - "vdup.u16 q10, %4 \n" // interval add - - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0] \n" // load 8 pixels of ARGB. - "subs %1, %1, #8 \n" // 8 processed per loop. - "vmovl.u8 q0, d0 \n" // b (0 .. 255) - "vmovl.u8 q1, d2 \n" - "vmovl.u8 q2, d4 \n" - "vqdmulh.s16 q0, q0, q8 \n" // b * scale - "vqdmulh.s16 q1, q1, q8 \n" // g - "vqdmulh.s16 q2, q2, q8 \n" // r - "vmul.u16 q0, q0, q9 \n" // b * interval_size - "vmul.u16 q1, q1, q9 \n" // g - "vmul.u16 q2, q2, q9 \n" // r - "vadd.u16 q0, q0, q10 \n" // b + interval_offset - "vadd.u16 q1, q1, q10 \n" // g - "vadd.u16 q2, q2, q10 \n" // r - "vqmovn.u16 d0, q0 \n" - "vqmovn.u16 d2, q1 \n" - "vqmovn.u16 d4, q2 \n" - MEMACCESS(0) - "vst4.8 {d0, d2, d4, d6}, [%0]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(dst_argb), // %0 - "+r"(width) // %1 - : "r"(scale), // %2 - "r"(interval_size), // %3 - "r"(interval_offset) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", "q8", "q9", "q10" - ); -} - -// Shade 8 pixels at a time by specified value. -// NOTE vqrdmulh.s16 q10, q10, d0[0] must use a scaler register from 0 to 8. -// Rounding in vqrdmulh does +1 to high if high bit of low s16 is set. -void ARGBShadeRow_NEON(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value) { - asm volatile ( - "vdup.u32 q0, %3 \n" // duplicate scale value. - "vzip.u8 d0, d1 \n" // d0 aarrggbb. - "vshr.u16 q0, q0, #1 \n" // scale / 2. - - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d22, d24, d26}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmovl.u8 q10, d20 \n" // b (0 .. 255) - "vmovl.u8 q11, d22 \n" - "vmovl.u8 q12, d24 \n" - "vmovl.u8 q13, d26 \n" - "vqrdmulh.s16 q10, q10, d0[0] \n" // b * scale * 2 - "vqrdmulh.s16 q11, q11, d0[1] \n" // g - "vqrdmulh.s16 q12, q12, d0[2] \n" // r - "vqrdmulh.s16 q13, q13, d0[3] \n" // a - "vqmovn.u16 d20, q10 \n" - "vqmovn.u16 d22, q11 \n" - "vqmovn.u16 d24, q12 \n" - "vqmovn.u16 d26, q13 \n" - MEMACCESS(1) - "vst4.8 {d20, d22, d24, d26}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(value) // %3 - : "cc", "memory", "q0", "q10", "q11", "q12", "q13" - ); -} - -// Convert 8 ARGB pixels (64 bytes) to 8 Gray ARGB pixels -// Similar to ARGBToYJ but stores ARGB. -// C code is (15 * b + 75 * g + 38 * r + 64) >> 7; -void ARGBGrayRow_NEON(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - "vmov.u8 d24, #15 \n" // B * 0.11400 coefficient - "vmov.u8 d25, #75 \n" // G * 0.58700 coefficient - "vmov.u8 d26, #38 \n" // R * 0.29900 coefficient - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 15 bit to 8 bit B - "vmov d1, d0 \n" // G - "vmov d2, d0 \n" // R - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q12", "q13" - ); -} - -// Convert 8 ARGB pixels (32 bytes) to 8 Sepia ARGB pixels. -// b = (r * 35 + g * 68 + b * 17) >> 7 -// g = (r * 45 + g * 88 + b * 22) >> 7 -// r = (r * 50 + g * 98 + b * 24) >> 7 -void ARGBSepiaRow_NEON(uint8* dst_argb, int width) { - asm volatile ( - "vmov.u8 d20, #17 \n" // BB coefficient - "vmov.u8 d21, #68 \n" // BG coefficient - "vmov.u8 d22, #35 \n" // BR coefficient - "vmov.u8 d24, #22 \n" // GB coefficient - "vmov.u8 d25, #88 \n" // GG coefficient - "vmov.u8 d26, #45 \n" // GR coefficient - "vmov.u8 d28, #24 \n" // BB coefficient - "vmov.u8 d29, #98 \n" // BG coefficient - "vmov.u8 d30, #50 \n" // BR coefficient - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0] \n" // load 8 ARGB pixels. - "subs %1, %1, #8 \n" // 8 processed per loop. - "vmull.u8 q2, d0, d20 \n" // B to Sepia B - "vmlal.u8 q2, d1, d21 \n" // G - "vmlal.u8 q2, d2, d22 \n" // R - "vmull.u8 q3, d0, d24 \n" // B to Sepia G - "vmlal.u8 q3, d1, d25 \n" // G - "vmlal.u8 q3, d2, d26 \n" // R - "vmull.u8 q8, d0, d28 \n" // B to Sepia R - "vmlal.u8 q8, d1, d29 \n" // G - "vmlal.u8 q8, d2, d30 \n" // R - "vqshrn.u16 d0, q2, #7 \n" // 16 bit to 8 bit B - "vqshrn.u16 d1, q3, #7 \n" // 16 bit to 8 bit G - "vqshrn.u16 d2, q8, #7 \n" // 16 bit to 8 bit R - MEMACCESS(0) - "vst4.8 {d0, d1, d2, d3}, [%0]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(dst_argb), // %0 - "+r"(width) // %1 - : - : "cc", "memory", "q0", "q1", "q2", "q3", - "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// Tranform 8 ARGB pixels (32 bytes) with color matrix. -// TODO(fbarchard): Was same as Sepia except matrix is provided. This function -// needs to saturate. Consider doing a non-saturating version. -void ARGBColorMatrixRow_NEON(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {q2}, [%3] \n" // load 3 ARGB vectors. - "vmovl.s8 q0, d4 \n" // B,G coefficients s16. - "vmovl.s8 q1, d5 \n" // R,A coefficients s16. - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d16, d18, d20, d22}, [%0]! \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "vmovl.u8 q8, d16 \n" // b (0 .. 255) 16 bit - "vmovl.u8 q9, d18 \n" // g - "vmovl.u8 q10, d20 \n" // r - "vmovl.u8 q15, d22 \n" // a - "vmul.s16 q12, q8, d0[0] \n" // B = B * Matrix B - "vmul.s16 q13, q8, d1[0] \n" // G = B * Matrix G - "vmul.s16 q14, q8, d2[0] \n" // R = B * Matrix R - "vmul.s16 q15, q8, d3[0] \n" // A = B * Matrix A - "vmul.s16 q4, q9, d0[1] \n" // B += G * Matrix B - "vmul.s16 q5, q9, d1[1] \n" // G += G * Matrix G - "vmul.s16 q6, q9, d2[1] \n" // R += G * Matrix R - "vmul.s16 q7, q9, d3[1] \n" // A += G * Matrix A - "vqadd.s16 q12, q12, q4 \n" // Accumulate B - "vqadd.s16 q13, q13, q5 \n" // Accumulate G - "vqadd.s16 q14, q14, q6 \n" // Accumulate R - "vqadd.s16 q15, q15, q7 \n" // Accumulate A - "vmul.s16 q4, q10, d0[2] \n" // B += R * Matrix B - "vmul.s16 q5, q10, d1[2] \n" // G += R * Matrix G - "vmul.s16 q6, q10, d2[2] \n" // R += R * Matrix R - "vmul.s16 q7, q10, d3[2] \n" // A += R * Matrix A - "vqadd.s16 q12, q12, q4 \n" // Accumulate B - "vqadd.s16 q13, q13, q5 \n" // Accumulate G - "vqadd.s16 q14, q14, q6 \n" // Accumulate R - "vqadd.s16 q15, q15, q7 \n" // Accumulate A - "vmul.s16 q4, q15, d0[3] \n" // B += A * Matrix B - "vmul.s16 q5, q15, d1[3] \n" // G += A * Matrix G - "vmul.s16 q6, q15, d2[3] \n" // R += A * Matrix R - "vmul.s16 q7, q15, d3[3] \n" // A += A * Matrix A - "vqadd.s16 q12, q12, q4 \n" // Accumulate B - "vqadd.s16 q13, q13, q5 \n" // Accumulate G - "vqadd.s16 q14, q14, q6 \n" // Accumulate R - "vqadd.s16 q15, q15, q7 \n" // Accumulate A - "vqshrun.s16 d16, q12, #6 \n" // 16 bit to 8 bit B - "vqshrun.s16 d18, q13, #6 \n" // 16 bit to 8 bit G - "vqshrun.s16 d20, q14, #6 \n" // 16 bit to 8 bit R - "vqshrun.s16 d22, q15, #6 \n" // 16 bit to 8 bit A - MEMACCESS(1) - "vst4.8 {d16, d18, d20, d22}, [%1]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(matrix_argb) // %3 - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", "q8", "q9", - "q10", "q11", "q12", "q13", "q14", "q15" - ); -} - -// TODO(fbarchard): fix vqshrun in ARGBMultiplyRow_NEON and reenable. -#ifdef HAS_ARGBMULTIPLYROW_NEON -// Multiply 2 rows of ARGB pixels together, 8 pixels at a time. -void ARGBMultiplyRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(1) - "vld4.8 {d1, d3, d5, d7}, [%1]! \n" // load 8 more ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vmull.u8 q0, d0, d1 \n" // multiply B - "vmull.u8 q1, d2, d3 \n" // multiply G - "vmull.u8 q2, d4, d5 \n" // multiply R - "vmull.u8 q3, d6, d7 \n" // multiply A - "vrshrn.u16 d0, q0, #8 \n" // 16 bit to 8 bit B - "vrshrn.u16 d1, q1, #8 \n" // 16 bit to 8 bit G - "vrshrn.u16 d2, q2, #8 \n" // 16 bit to 8 bit R - "vrshrn.u16 d3, q3, #8 \n" // 16 bit to 8 bit A - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3" - ); -} -#endif // HAS_ARGBMULTIPLYROW_NEON - -// Add 2 rows of ARGB pixels together, 8 pixels at a time. -void ARGBAddRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(1) - "vld4.8 {d4, d5, d6, d7}, [%1]! \n" // load 8 more ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vqadd.u8 q0, q0, q2 \n" // add B, G - "vqadd.u8 q1, q1, q3 \n" // add R, A - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3" - ); -} - -// Subtract 2 rows of ARGB pixels, 8 pixels at a time. -void ARGBSubtractRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(1) - "vld4.8 {d4, d5, d6, d7}, [%1]! \n" // load 8 more ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vqsub.u8 q0, q0, q2 \n" // subtract B, G - "vqsub.u8 q1, q1, q3 \n" // subtract R, A - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1", "q2", "q3" - ); -} - -// Adds Sobel X and Sobel Y and stores Sobel into ARGB. -// A = 255 -// R = Sobel -// G = Sobel -// B = Sobel -void SobelRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - asm volatile ( - "vmov.u8 d3, #255 \n" // alpha - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d0}, [%0]! \n" // load 8 sobelx. - MEMACCESS(1) - "vld1.8 {d1}, [%1]! \n" // load 8 sobely. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vqadd.u8 d0, d0, d1 \n" // add - "vmov.u8 d1, d0 \n" - "vmov.u8 d2, d0 \n" - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1" - ); -} - -// Adds Sobel X and Sobel Y and stores Sobel into plane. -void SobelToPlaneRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width) { - asm volatile ( - // 16 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 16 sobelx. - MEMACCESS(1) - "vld1.8 {q1}, [%1]! \n" // load 16 sobely. - "subs %3, %3, #16 \n" // 16 processed per loop. - "vqadd.u8 q0, q0, q1 \n" // add - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" // store 16 pixels. - "bgt 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_y), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1" - ); -} - -// Mixes Sobel X, Sobel Y and Sobel into ARGB. -// A = 255 -// R = Sobel X -// G = Sobel -// B = Sobel Y -void SobelXYRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - asm volatile ( - "vmov.u8 d3, #255 \n" // alpha - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d2}, [%0]! \n" // load 8 sobelx. - MEMACCESS(1) - "vld1.8 {d0}, [%1]! \n" // load 8 sobely. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vqadd.u8 d1, d0, d2 \n" // add - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "q0", "q1" - ); -} - -// SobelX as a matrix is -// -1 0 1 -// -2 0 2 -// -1 0 1 -void SobelXRow_NEON(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobelx, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d0}, [%0],%5 \n" // top - MEMACCESS(0) - "vld1.8 {d1}, [%0],%6 \n" - "vsubl.u8 q0, d0, d1 \n" - MEMACCESS(1) - "vld1.8 {d2}, [%1],%5 \n" // center * 2 - MEMACCESS(1) - "vld1.8 {d3}, [%1],%6 \n" - "vsubl.u8 q1, d2, d3 \n" - "vadd.s16 q0, q0, q1 \n" - "vadd.s16 q0, q0, q1 \n" - MEMACCESS(2) - "vld1.8 {d2}, [%2],%5 \n" // bottom - MEMACCESS(2) - "vld1.8 {d3}, [%2],%6 \n" - "subs %4, %4, #8 \n" // 8 pixels - "vsubl.u8 q1, d2, d3 \n" - "vadd.s16 q0, q0, q1 \n" - "vabs.s16 q0, q0 \n" - "vqmovn.u16 d0, q0 \n" - MEMACCESS(3) - "vst1.8 {d0}, [%3]! \n" // store 8 sobelx - "bgt 1b \n" - : "+r"(src_y0), // %0 - "+r"(src_y1), // %1 - "+r"(src_y2), // %2 - "+r"(dst_sobelx), // %3 - "+r"(width) // %4 - : "r"(2), // %5 - "r"(6) // %6 - : "cc", "memory", "q0", "q1" // Clobber List - ); -} - -// SobelY as a matrix is -// -1 -2 -1 -// 0 0 0 -// 1 2 1 -void SobelYRow_NEON(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d0}, [%0],%4 \n" // left - MEMACCESS(1) - "vld1.8 {d1}, [%1],%4 \n" - "vsubl.u8 q0, d0, d1 \n" - MEMACCESS(0) - "vld1.8 {d2}, [%0],%4 \n" // center * 2 - MEMACCESS(1) - "vld1.8 {d3}, [%1],%4 \n" - "vsubl.u8 q1, d2, d3 \n" - "vadd.s16 q0, q0, q1 \n" - "vadd.s16 q0, q0, q1 \n" - MEMACCESS(0) - "vld1.8 {d2}, [%0],%5 \n" // right - MEMACCESS(1) - "vld1.8 {d3}, [%1],%5 \n" - "subs %3, %3, #8 \n" // 8 pixels - "vsubl.u8 q1, d2, d3 \n" - "vadd.s16 q0, q0, q1 \n" - "vabs.s16 q0, q0 \n" - "vqmovn.u16 d0, q0 \n" - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 sobely - "bgt 1b \n" - : "+r"(src_y0), // %0 - "+r"(src_y1), // %1 - "+r"(dst_sobely), // %2 - "+r"(width) // %3 - : "r"(1), // %4 - "r"(6) // %5 - : "cc", "memory", "q0", "q1" // Clobber List - ); -} -#endif // __ARM_NEON__ - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_neon64.cc b/src/main/jni/libyuv/source/row_neon64.cc deleted file mode 100644 index 952e10d73..000000000 --- a/src/main/jni/libyuv/source/row_neon64.cc +++ /dev/null @@ -1,3327 +0,0 @@ -/* - * Copyright 2014 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC Neon -#if !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__) - -// Read 8 Y, 4 U and 4 V from 422 -#define READYUV422 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.32 {d2[0]}, [%1]! \n" \ - MEMACCESS(2) \ - "vld1.32 {d2[1]}, [%2]! \n" - -// Read 8 Y, 2 U and 2 V from 422 -#define READYUV411 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.16 {d2[0]}, [%1]! \n" \ - MEMACCESS(2) \ - "vld1.16 {d2[1]}, [%2]! \n" \ - "vmov.u8 d3, d2 \n" \ - "vzip.u8 d2, d3 \n" - -// Read 8 Y, 8 U and 8 V from 444 -#define READYUV444 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.8 {d2}, [%1]! \n" \ - MEMACCESS(2) \ - "vld1.8 {d3}, [%2]! \n" \ - "vpaddl.u8 q1, q1 \n" \ - "vrshrn.u16 d2, q1, #1 \n" - -// Read 8 Y, and set 4 U and 4 V to 128 -#define READYUV400 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - "vmov.u8 d2, #128 \n" - -// Read 8 Y and 4 UV from NV12 -#define READNV12 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.8 {d2}, [%1]! \n" \ - "vmov.u8 d3, d2 \n"/* split odd/even uv apart */\ - "vuzp.u8 d2, d3 \n" \ - "vtrn.u32 d2, d3 \n" - -// Read 8 Y and 4 VU from NV21 -#define READNV21 \ - MEMACCESS(0) \ - "vld1.8 {d0}, [%0]! \n" \ - MEMACCESS(1) \ - "vld1.8 {d2}, [%1]! \n" \ - "vmov.u8 d3, d2 \n"/* split odd/even uv apart */\ - "vuzp.u8 d3, d2 \n" \ - "vtrn.u32 d2, d3 \n" - -// Read 8 YUY2 -#define READYUY2 \ - MEMACCESS(0) \ - "vld2.8 {d0, d2}, [%0]! \n" \ - "vmov.u8 d3, d2 \n" \ - "vuzp.u8 d2, d3 \n" \ - "vtrn.u32 d2, d3 \n" - -// Read 8 UYVY -#define READUYVY \ - MEMACCESS(0) \ - "vld2.8 {d2, d3}, [%0]! \n" \ - "vmov.u8 d0, d3 \n" \ - "vmov.u8 d3, d2 \n" \ - "vuzp.u8 d2, d3 \n" \ - "vtrn.u32 d2, d3 \n" - -#define YUV422TORGB \ - "veor.u8 d2, d26 \n"/*subtract 128 from u and v*/\ - "vmull.s8 q8, d2, d24 \n"/* u/v B/R component */\ - "vmull.s8 q9, d2, d25 \n"/* u/v G component */\ - "vmov.u8 d1, #0 \n"/* split odd/even y apart */\ - "vtrn.u8 d0, d1 \n" \ - "vsub.s16 q0, q0, q15 \n"/* offset y */\ - "vmul.s16 q0, q0, q14 \n" \ - "vadd.s16 d18, d19 \n" \ - "vqadd.s16 d20, d0, d16 \n" /* B */ \ - "vqadd.s16 d21, d1, d16 \n" \ - "vqadd.s16 d22, d0, d17 \n" /* R */ \ - "vqadd.s16 d23, d1, d17 \n" \ - "vqadd.s16 d16, d0, d18 \n" /* G */ \ - "vqadd.s16 d17, d1, d18 \n" \ - "vqshrun.s16 d0, q10, #6 \n" /* B */ \ - "vqshrun.s16 d1, q11, #6 \n" /* G */ \ - "vqshrun.s16 d2, q8, #6 \n" /* R */ \ - "vmovl.u8 q10, d0 \n"/* set up for reinterleave*/\ - "vmovl.u8 q11, d1 \n" \ - "vmovl.u8 q8, d2 \n" \ - "vtrn.u8 d20, d21 \n" \ - "vtrn.u8 d22, d23 \n" \ - "vtrn.u8 d16, d17 \n" \ - "vmov.u8 d21, d16 \n" - -static vec8 kUVToRB = { 127, 127, 127, 127, 102, 102, 102, 102, - 0, 0, 0, 0, 0, 0, 0, 0 }; -static vec8 kUVToG = { -25, -25, -25, -25, -52, -52, -52, -52, - 0, 0, 0, 0, 0, 0, 0, 0 }; - -#ifdef HAS_I444TOARGBROW_NEON -void I444ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV444 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I444TOARGBROW_NEON - -#ifdef HAS_I422TOARGBROW_NEON -void I422ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TOARGBROW_NEON - -#ifdef HAS_I411TOARGBROW_NEON -void I411ToARGBRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV411 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I411TOARGBROW_NEON - -#ifdef HAS_I422TOBGRAROW_NEON -void I422ToBGRARow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_bgra, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vswp.u8 d20, d22 \n" - "vmov.u8 d19, #255 \n" - MEMACCESS(3) - "vst4.8 {d19, d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_bgra), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TOBGRAROW_NEON - -#ifdef HAS_I422TOABGRROW_NEON -void I422ToABGRRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_abgr, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vswp.u8 d20, d22 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(3) - "vst4.8 {d20, d21, d22, d23}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_abgr), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TOABGRROW_NEON - -#ifdef HAS_I422TORGBAROW_NEON -void I422ToRGBARow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgba, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d19, #255 \n" - MEMACCESS(3) - "vst4.8 {d19, d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_rgba), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TORGBAROW_NEON - -#ifdef HAS_I422TORGB24ROW_NEON -void I422ToRGB24Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb24, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - MEMACCESS(3) - "vst3.8 {d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_rgb24), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TORGB24ROW_NEON - -#ifdef HAS_I422TORAWROW_NEON -void I422ToRAWRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_raw, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vswp.u8 d20, d22 \n" - MEMACCESS(3) - "vst3.8 {d20, d21, d22}, [%3]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_raw), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TORAWROW_NEON - -#define ARGBTORGB565 \ - "vshr.u8 d20, d20, #3 \n" /* B */ \ - "vshr.u8 d21, d21, #2 \n" /* G */ \ - "vshr.u8 d22, d22, #3 \n" /* R */ \ - "vmovl.u8 q8, d20 \n" /* B */ \ - "vmovl.u8 q9, d21 \n" /* G */ \ - "vmovl.u8 q10, d22 \n" /* R */ \ - "vshl.u16 q9, q9, #5 \n" /* G */ \ - "vshl.u16 q10, q10, #11 \n" /* R */ \ - "vorr q0, q8, q9 \n" /* BG */ \ - "vorr q0, q0, q10 \n" /* BGR */ - -#ifdef HAS_I422TORGB565ROW_NEON -void I422ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_rgb565, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - ARGBTORGB565 - MEMACCESS(3) - "vst1.8 {q0}, [%3]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_rgb565), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TORGB565ROW_NEON - -#define ARGBTOARGB1555 \ - "vshr.u8 q10, q10, #3 \n" /* B */ \ - "vshr.u8 d22, d22, #3 \n" /* R */ \ - "vshr.u8 d23, d23, #7 \n" /* A */ \ - "vmovl.u8 q8, d20 \n" /* B */ \ - "vmovl.u8 q9, d21 \n" /* G */ \ - "vmovl.u8 q10, d22 \n" /* R */ \ - "vmovl.u8 q11, d23 \n" /* A */ \ - "vshl.u16 q9, q9, #5 \n" /* G */ \ - "vshl.u16 q10, q10, #10 \n" /* R */ \ - "vshl.u16 q11, q11, #15 \n" /* A */ \ - "vorr q0, q8, q9 \n" /* BG */ \ - "vorr q1, q10, q11 \n" /* RA */ \ - "vorr q0, q0, q1 \n" /* BGRA */ - -#ifdef HAS_I422TOARGB1555ROW_NEON -void I422ToARGB1555Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb1555, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - ARGBTOARGB1555 - MEMACCESS(3) - "vst1.8 {q0}, [%3]! \n" // store 8 pixels ARGB1555. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb1555), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TOARGB1555ROW_NEON - -#define ARGBTOARGB4444 \ - "vshr.u8 d20, d20, #4 \n" /* B */ \ - "vbic.32 d21, d21, d4 \n" /* G */ \ - "vshr.u8 d22, d22, #4 \n" /* R */ \ - "vbic.32 d23, d23, d4 \n" /* A */ \ - "vorr d0, d20, d21 \n" /* BG */ \ - "vorr d1, d22, d23 \n" /* RA */ \ - "vzip.u8 d0, d1 \n" /* BGRA */ - -#ifdef HAS_I422TOARGB4444ROW_NEON -void I422ToARGB4444Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb4444, - int width) { - asm volatile ( - MEMACCESS(5) - "vld1.8 {d24}, [%5] \n" - MEMACCESS(6) - "vld1.8 {d25}, [%6] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - "vmov.u8 d4, #0x0f \n" // bits to clear with vbic. - ".p2align 2 \n" - "1: \n" - READYUV422 - YUV422TORGB - "subs %4, %4, #8 \n" - "vmov.u8 d23, #255 \n" - ARGBTOARGB4444 - MEMACCESS(3) - "vst1.8 {q0}, [%3]! \n" // store 8 pixels ARGB4444. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_argb4444), // %3 - "+r"(width) // %4 - : "r"(&kUVToRB), // %5 - "r"(&kUVToG) // %6 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_I422TOARGB4444ROW_NEON - -#ifdef HAS_YTOARGBROW_NEON -void YToARGBRow_NEON(const uint8* src_y, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {d24}, [%3] \n" - MEMACCESS(4) - "vld1.8 {d25}, [%4] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUV400 - YUV422TORGB - "subs %2, %2, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(&kUVToRB), // %3 - "r"(&kUVToG) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_YTOARGBROW_NEON - -#ifdef HAS_I400TOARGBROW_NEON -void I400ToARGBRow_NEON(const uint8* src_y, - uint8* dst_argb, - int width) { - asm volatile ( - ".p2align 2 \n" - "vmov.u8 d23, #255 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d20}, [%0]! \n" - "vmov d21, d20 \n" - "vmov d22, d20 \n" - "subs %2, %2, #8 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "d20", "d21", "d22", "d23" - ); -} -#endif // HAS_I400TOARGBROW_NEON - -#ifdef HAS_NV12TOARGBROW_NEON -void NV12ToARGBRow_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV12 - YUV422TORGB - "subs %3, %3, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(2) - "vst4.8 {d20, d21, d22, d23}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_NV12TOARGBROW_NEON - -#ifdef HAS_NV21TOARGBROW_NEON -void NV21ToARGBRow_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV21 - YUV422TORGB - "subs %3, %3, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(2) - "vst4.8 {d20, d21, d22, d23}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_NV21TOARGBROW_NEON - -#ifdef HAS_NV12TORGB565ROW_NEON -void NV12ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_rgb565, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV12 - YUV422TORGB - "subs %3, %3, #8 \n" - ARGBTORGB565 - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_rgb565), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_NV12TORGB565ROW_NEON - -#ifdef HAS_NV21TORGB565ROW_NEON -void NV21ToRGB565Row_NEON(const uint8* src_y, - const uint8* src_uv, - uint8* dst_rgb565, - int width) { - asm volatile ( - MEMACCESS(4) - "vld1.8 {d24}, [%4] \n" - MEMACCESS(5) - "vld1.8 {d25}, [%5] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READNV21 - YUV422TORGB - "subs %3, %3, #8 \n" - ARGBTORGB565 - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_uv), // %1 - "+r"(dst_rgb565), // %2 - "+r"(width) // %3 - : "r"(&kUVToRB), // %4 - "r"(&kUVToG) // %5 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_NV21TORGB565ROW_NEON - -#ifdef HAS_YUY2TOARGBROW_NEON -void YUY2ToARGBRow_NEON(const uint8* src_yuy2, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {d24}, [%3] \n" - MEMACCESS(4) - "vld1.8 {d25}, [%4] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READYUY2 - YUV422TORGB - "subs %2, %2, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(&kUVToRB), // %3 - "r"(&kUVToG) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_YUY2TOARGBROW_NEON - -#ifdef HAS_UYVYTOARGBROW_NEON -void UYVYToARGBRow_NEON(const uint8* src_uyvy, - uint8* dst_argb, - int width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {d24}, [%3] \n" - MEMACCESS(4) - "vld1.8 {d25}, [%4] \n" - "vmov.u8 d26, #128 \n" - "vmov.u16 q14, #74 \n" - "vmov.u16 q15, #16 \n" - ".p2align 2 \n" - "1: \n" - READUYVY - YUV422TORGB - "subs %2, %2, #8 \n" - "vmov.u8 d23, #255 \n" - MEMACCESS(1) - "vst4.8 {d20, d21, d22, d23}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(&kUVToRB), // %3 - "r"(&kUVToG) // %4 - : "cc", "memory", "q0", "q1", "q2", "q3", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_UYVYTOARGBROW_NEON - -// Reads 16 pairs of UV and write even values to dst_u and odd to dst_v. -#ifdef HAS_SPLITUVROW_NEON -void SplitUVRow_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld2 {v0.16b, v1.16b}, [%0], #32 \n" // load 16 pairs of UV - "subs %3, %3, #16 \n" // 16 processed per loop - MEMACCESS(1) - "st1 {v0.16b}, [%1], #16 \n" // store U - MEMACCESS(2) - "st1 {v1.16b}, [%2], #16 \n" // store V - "bgt 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(width) // %3 // Output registers - : // Input registers - : "cc", "memory", "v0", "v1" // Clobber List - ); -} -#endif // HAS_SPLITUVROW_NEON - -// Reads 16 U's and V's and writes out 16 pairs of UV. -#ifdef HAS_MERGEUVROW_NEON -void MergeUVRow_NEON(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], #16 \n" // load U - MEMACCESS(1) - "ld1 {v1.16b}, [%1], #16 \n" // load V - "subs %3, %3, #16 \n" // 16 processed per loop - MEMACCESS(2) - "st2 {v0.16b, v1.16b}, [%2], #32 \n" // store 16 pairs of UV - "bgt 1b \n" - : - "+r"(src_u), // %0 - "+r"(src_v), // %1 - "+r"(dst_uv), // %2 - "+r"(width) // %3 // Output registers - : // Input registers - : "cc", "memory", "v0", "v1" // Clobber List - ); -} -#endif // HAS_MERGEUVROW_NEON - -// Copy multiple of 32. vld4.8 allow unaligned and is fastest on a15. -#ifdef HAS_COPYROW_NEON -void CopyRow_NEON(const uint8* src, uint8* dst, int count) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.8b-v3.8b}, [%0], #32 \n" // load 32 - "subs %2, %2, #32 \n" // 32 processed per loop - MEMACCESS(1) - "st1 {v0.8b-v3.8b}, [%1], #32 \n" // store 32 - "bgt 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(count) // %2 // Output registers - : // Input registers - : "cc", "memory", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif // HAS_COPYROW_NEON - -// SetRow8 writes 'count' bytes using a 32 bit value repeated. -#ifdef HAS_SETROW_NEON -void SetRow_NEON(uint8* dst, uint32 v32, int count) { - asm volatile ( - "dup v0.4s, %w2 \n" // duplicate 4 ints - "1: \n" - "subs %1, %1, #16 \n" // 16 bytes per loop - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" // store - "bgt 1b \n" - : "+r"(dst), // %0 - "+r"(count) // %1 - : "r"(v32) // %2 - : "cc", "memory", "v0" - ); -} -#endif // HAS_SETROW_NEON - -// TODO(fbarchard): Make fully assembler -// SetRow32 writes 'count' words using a 32 bit value repeated. -#ifdef HAS_ARGBSETROWS_NEON -void ARGBSetRows_NEON(uint8* dst, uint32 v32, int width, - int dst_stride, int height) { - for (int y = 0; y < height; ++y) { - SetRow_NEON(dst, v32, width << 2); - dst += dst_stride; - } -} -#endif // HAS_ARGBSETROWS_NEON - -#ifdef HAS_MIRRORROW_NEON -void MirrorRow_NEON(const uint8* src, uint8* dst, int width) { - asm volatile ( - // Start at end of source row. - "add %0, %0, %2 \n" - "sub %0, %0, #16 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], %3 \n" // src -= 16 - "subs %2, %2, #16 \n" // 16 pixels per loop. - "rev64 v0.16b, v0.16b \n" - MEMACCESS(1) - "st1 {v0.D}[1], [%1], #8 \n" // dst += 16 - MEMACCESS(1) - "st1 {v0.D}[0], [%1], #8 \n" - "bgt 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : "r"((ptrdiff_t)-16) // %3 - : "cc", "memory", "v0" - ); -} -#endif // HAS_MIRRORROW_NEON - -#ifdef HAS_MIRRORUVROW_NEON -void MirrorUVRow_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int width) { - asm volatile ( - // Start at end of source row. - "add %0, %0, %3, lsl #1 \n" - "sub %0, %0, #16 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld2 {v0.8b, v1.8b}, [%0], %4 \n" // src -= 16 - "subs %3, %3, #8 \n" // 8 pixels per loop. - "rev64 v0.8b, v0.8b \n" - "rev64 v1.8b, v1.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // dst += 8 - MEMACCESS(2) - "st1 {v1.8b}, [%2], #8 \n" - "bgt 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(width) // %3 - : "r"((ptrdiff_t)-16) // %4 - : "cc", "memory", "v0", "v1" - ); -} -#endif // HAS_MIRRORUVROW_NEON - -#ifdef HAS_ARGBMIRRORROW_NEON -void ARGBMirrorRow_NEON(const uint8* src, uint8* dst, int width) { - asm volatile ( - // Start at end of source row. - "add %0, %0, %2, lsl #2 \n" - "sub %0, %0, #16 \n" - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], %3 \n" // src -= 16 - "subs %2, %2, #4 \n" // 4 pixels per loop. - "rev64 v0.4s, v0.4s \n" - MEMACCESS(1) - "st1 {v0.D}[1], [%1], #8 \n" // dst += 16 - MEMACCESS(1) - "st1 {v0.D}[0], [%1], #8 \n" - "bgt 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : "r"((ptrdiff_t)-16) // %3 - : "cc", "memory", "v0" - ); -} -#endif // HAS_ARGBMIRRORROW_NEON - -#ifdef HAS_RGB24TOARGBROW_NEON -void RGB24ToARGBRow_NEON(const uint8* src_rgb24, uint8* dst_argb, int pix) { - asm volatile ( - "movi v4.8b, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld3 {v1.8b-v3.8b}, [%0], #24 \n" // load 8 pixels of RGB24. - "subs %2, %2, #8 \n" // 8 processed per loop. - MEMACCESS(1) - "st4 {v1.8b-v4.8b}, [%1], #32 \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v1", "v2", "v3", "v4" // Clobber List - ); -} -#endif // HAS_RGB24TOARGBROW_NEON - -#ifdef HAS_RAWTOARGBROW_NEON -void RAWToARGBRow_NEON(const uint8* src_raw, uint8* dst_argb, int pix) { - asm volatile ( - "movi v5.8b, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld3 {v0.8b-v2.8b}, [%0], #24 \n" // read r g b - "subs %2, %2, #8 \n" // 8 processed per loop. - "mov v3.8b, v1.8b \n" // move g - "mov v4.8b, v0.8b \n" // move r - MEMACCESS(1) - "st4 {v2.8b-v5.8b}, [%1], #32 \n" // store b g r a - "bgt 1b \n" - : "+r"(src_raw), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5" // Clobber List - ); -} -#endif // HAS_RAWTOARGBROW_NEON - -#define RGB565TOARGB \ - "vshrn.u16 d6, q0, #5 \n" /* G xxGGGGGG */ \ - "vuzp.u8 d0, d1 \n" /* d0 xxxBBBBB RRRRRxxx */ \ - "vshl.u8 d6, d6, #2 \n" /* G GGGGGG00 upper 6 */ \ - "vshr.u8 d1, d1, #3 \n" /* R 000RRRRR lower 5 */ \ - "vshl.u8 q0, q0, #3 \n" /* B,R BBBBB000 upper 5 */ \ - "vshr.u8 q2, q0, #5 \n" /* B,R 00000BBB lower 3 */ \ - "vorr.u8 d0, d0, d4 \n" /* B */ \ - "vshr.u8 d4, d6, #6 \n" /* G 000000GG lower 2 */ \ - "vorr.u8 d2, d1, d5 \n" /* R */ \ - "vorr.u8 d1, d4, d6 \n" /* G */ - -#ifdef HAS_RGB565TOARGBROW_NEON -void RGB565ToARGBRow_NEON(const uint8* src_rgb565, uint8* dst_argb, int pix) { - asm volatile ( - "vmov.u8 d3, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 RGB565 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - RGB565TOARGB - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_rgb565), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3" // Clobber List - ); -} -#endif // HAS_RGB565TOARGBROW_NEON - -#define ARGB1555TOARGB \ - "vshrn.u16 d7, q0, #8 \n" /* A Arrrrrxx */ \ - "vshr.u8 d6, d7, #2 \n" /* R xxxRRRRR */ \ - "vshrn.u16 d5, q0, #5 \n" /* G xxxGGGGG */ \ - "vmovn.u16 d4, q0 \n" /* B xxxBBBBB */ \ - "vshr.u8 d7, d7, #7 \n" /* A 0000000A */ \ - "vneg.s8 d7, d7 \n" /* A AAAAAAAA upper 8 */ \ - "vshl.u8 d6, d6, #3 \n" /* R RRRRR000 upper 5 */ \ - "vshr.u8 q1, q3, #5 \n" /* R,A 00000RRR lower 3 */ \ - "vshl.u8 q0, q2, #3 \n" /* B,G BBBBB000 upper 5 */ \ - "vshr.u8 q2, q0, #5 \n" /* B,G 00000BBB lower 3 */ \ - "vorr.u8 q1, q1, q3 \n" /* R,A */ \ - "vorr.u8 q0, q0, q2 \n" /* B,G */ \ - -// RGB555TOARGB is same as ARGB1555TOARGB but ignores alpha. -#define RGB555TOARGB \ - "vshrn.u16 d6, q0, #5 \n" /* G xxxGGGGG */ \ - "vuzp.u8 d0, d1 \n" /* d0 xxxBBBBB xRRRRRxx */ \ - "vshl.u8 d6, d6, #3 \n" /* G GGGGG000 upper 5 */ \ - "vshr.u8 d1, d1, #2 \n" /* R 00xRRRRR lower 5 */ \ - "vshl.u8 q0, q0, #3 \n" /* B,R BBBBB000 upper 5 */ \ - "vshr.u8 q2, q0, #5 \n" /* B,R 00000BBB lower 3 */ \ - "vorr.u8 d0, d0, d4 \n" /* B */ \ - "vshr.u8 d4, d6, #5 \n" /* G 00000GGG lower 3 */ \ - "vorr.u8 d2, d1, d5 \n" /* R */ \ - "vorr.u8 d1, d4, d6 \n" /* G */ - -#ifdef HAS_ARGB1555TOARGBROW_NEON -void ARGB1555ToARGBRow_NEON(const uint8* src_argb1555, uint8* dst_argb, - int pix) { - asm volatile ( - "vmov.u8 d3, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB1555 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB1555TOARGB - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb1555), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3" // Clobber List - ); -} -#endif // HAS_ARGB1555TOARGBROW_NEON - -#define ARGB4444TOARGB \ - "vuzp.u8 d0, d1 \n" /* d0 BG, d1 RA */ \ - "vshl.u8 q2, q0, #4 \n" /* B,R BBBB0000 */ \ - "vshr.u8 q1, q0, #4 \n" /* G,A 0000GGGG */ \ - "vshr.u8 q0, q2, #4 \n" /* B,R 0000BBBB */ \ - "vorr.u8 q0, q0, q2 \n" /* B,R BBBBBBBB */ \ - "vshl.u8 q2, q1, #4 \n" /* G,A GGGG0000 */ \ - "vorr.u8 q1, q1, q2 \n" /* G,A GGGGGGGG */ \ - "vswp.u8 d1, d2 \n" /* B,R,G,A -> B,G,R,A */ - -#ifdef HAS_ARGB4444TOARGBROW_NEON -void ARGB4444ToARGBRow_NEON(const uint8* src_argb4444, uint8* dst_argb, - int pix) { - asm volatile ( - "vmov.u8 d3, #255 \n" // Alpha - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB4444 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB4444TOARGB - MEMACCESS(1) - "vst4.8 {d0, d1, d2, d3}, [%1]! \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb4444), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2" // Clobber List - ); -} -#endif // HAS_ARGB4444TOARGBROW_NEON - -#ifdef HAS_ARGBTORGB24ROW_NEON -void ARGBToRGB24Row_NEON(const uint8* src_argb, uint8* dst_rgb24, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v1.8b-v4.8b}, [%0], #32 \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - MEMACCESS(1) - "st3 {v1.8b-v3.8b}, [%1], #24 \n" // store 8 pixels of RGB24. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_rgb24), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v1", "v2", "v3", "v4" // Clobber List - ); -} -#endif // HAS_ARGBTORGB24ROW_NEON - -#ifdef HAS_ARGBTORAWROW_NEON -void ARGBToRAWRow_NEON(const uint8* src_argb, uint8* dst_raw, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v1.8b-v4.8b}, [%0], #32 \n" // load b g r a - "subs %2, %2, #8 \n" // 8 processed per loop. - "mov v4.8b, v2.8b \n" // mov g - "mov v5.8b, v1.8b \n" // mov b - MEMACCESS(1) - "st3 {v3.8b-v5.8b}, [%1], #24 \n" // store r g b - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_raw), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v1", "v2", "v3", "v4", "v5" // Clobber List - ); -} -#endif // HAS_ARGBTORAWROW_NEON - -#ifdef HAS_YUY2TOYROW_NEON -void YUY2ToYRow_NEON(const uint8* src_yuy2, uint8* dst_y, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld2 {v0.16b, v1.16b}, [%0], #32 \n" // load 16 pixels of YUY2. - "subs %2, %2, #16 \n" // 16 processed per loop. - MEMACCESS(1) - "st1 {v0.16b}, [%1], #16 \n" // store 16 pixels of Y. - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1" // Clobber List - ); -} -#endif // HAS_YUY2TOYROW_NEON - -#ifdef HAS_UYVYTOYROW_NEON -void UYVYToYRow_NEON(const uint8* src_uyvy, uint8* dst_y, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld2 {v0.16b, v1.16b}, [%0], #32 \n" // load 16 pixels of UYVY. - "subs %2, %2, #16 \n" // 16 processed per loop. - MEMACCESS(1) - "st1 {v1.16b}, [%1], #16 \n" // store 16 pixels of Y. - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1" // Clobber List - ); -} -#endif // HAS_UYVYTOYROW_NEON - -#ifdef HAS_YUY2TOUV422ROW_NEON -void YUY2ToUV422Row_NEON(const uint8* src_yuy2, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 16 pixels of YUY2. - "subs %3, %3, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "st1 {v1.8b}, [%1], #8 \n" // store 8 U. - MEMACCESS(2) - "st1 {v3.8b}, [%2], #8 \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif // HAS_YUY2TOUV422ROW_NEON - -#ifdef HAS_UYVYTOUV422ROW_NEON -void UYVYToUV422Row_NEON(const uint8* src_uyvy, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 16 pixels of UYVY. - "subs %3, %3, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 U. - MEMACCESS(2) - "st1 {v2.8b}, [%2], #8 \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif // HAS_UYVYTOUV422ROW_NEON - -#ifdef HAS_YUY2TOUVROW_NEON -void YUY2ToUVRow_NEON(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %x1, %x0, %w1, sxtw \n" // stride + src_yuy2 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 16 pixels of YUY2. - "subs %4, %4, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "ld4 {v4.8b-v7.8b}, [%1], #32 \n" // load next row YUY2. - "urhadd v1.8b, v1.8b, v5.8b \n" // average rows of U - "urhadd v3.8b, v3.8b, v7.8b \n" // average rows of V - MEMACCESS(2) - "st1 {v1.8b}, [%2], #8 \n" // store 8 U. - MEMACCESS(3) - "st1 {v3.8b}, [%3], #8 \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(stride_yuy2), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7" // Clobber List - ); -} -#endif // HAS_YUY2TOUVROW_NEON - -#ifdef HAS_UYVYTOUVROW_NEON -void UYVYToUVRow_NEON(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %x1, %x0, %w1, sxtw \n" // stride + src_uyvy - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 16 pixels of UYVY. - "subs %4, %4, #16 \n" // 16 pixels = 8 UVs. - MEMACCESS(1) - "ld4 {v4.8b-v7.8b}, [%1], #32 \n" // load next row UYVY. - "urhadd v0.8b, v0.8b, v4.8b \n" // average rows of U - "urhadd v2.8b, v2.8b, v6.8b \n" // average rows of V - MEMACCESS(2) - "st1 {v0.8b}, [%2], #8 \n" // store 8 U. - MEMACCESS(3) - "st1 {v2.8b}, [%3], #8 \n" // store 8 V. - "bgt 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(stride_uyvy), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7" // Clobber List - ); -} -#endif // HAS_UYVYTOUVROW_NEON - -#ifdef HAS_HALFROW_NEON -void HalfRow_NEON(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix) { - asm volatile ( - // change the stride to row 2 pointer - "add %x1, %x0, %w1, sxtw \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], #16 \n" // load row 1 16 pixels. - "subs %3, %3, #16 \n" // 16 processed per loop - MEMACCESS(1) - "ld1 {v1.16b}, [%1], #16 \n" // load row 2 16 pixels. - "urhadd v0.16b, v0.16b, v1.16b \n" // average row 1 and 2 - MEMACCESS(2) - "st1 {v0.16b}, [%2], #16 \n" - "bgt 1b \n" - : "+r"(src_uv), // %0 - "+r"(src_uv_stride), // %1 - "+r"(dst_uv), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "v0", "v1" // Clobber List - ); -} -#endif // HAS_HALFROW_NEON - -// Select 2 channels from ARGB on alternating pixels. e.g. BGBGBGBG -#ifdef HAS_ARGBTOBAYERROW_NEON -void ARGBToBayerRow_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) { - asm volatile ( - "mov v2.s[0], %w3 \n" // selector - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b, v1.16b}, [%0], 32 \n" // load row 8 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop - "tbl v4.8b, {v0.16b}, v2.8b \n" // look up 4 pixels - "tbl v5.8b, {v1.16b}, v2.8b \n" // look up 4 pixels - "trn1 v4.4s, v4.4s, v5.4s \n" // combine 8 pixels - MEMACCESS(1) - "st1 {v4.8b}, [%1], #8 \n" // store 8. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_bayer), // %1 - "+r"(pix) // %2 - : "r"(selector) // %3 - : "cc", "memory", "v0", "v1", "v2", "v4", "v5" // Clobber List - ); -} -#endif // HAS_ARGBTOBAYERROW_NEON - -// Select G channels from ARGB. e.g. GGGGGGGG -#ifdef HAS_ARGBTOBAYERGGROW_NEON -void ARGBToBayerGGRow_NEON(const uint8* src_argb, uint8* dst_bayer, - uint32 /*selector*/, int pix) { - asm volatile ( - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load row 8 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop - MEMACCESS(1) - "st1 {v1.8b}, [%1], #8 \n" // store 8 G's. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_bayer), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif // HAS_ARGBTOBAYERGGROW_NEON - -// For BGRAToARGB, ABGRToARGB, RGBAToARGB, and ARGBToRGBA. -#ifdef HAS_ARGBSHUFFLEROW_NEON -void ARGBShuffleRow_NEON(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - asm volatile ( - MEMACCESS(3) - "ld1 {v2.16b}, [%3] \n" // shuffler - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], #16 \n" // load 4 pixels. - "subs %2, %2, #4 \n" // 4 processed per loop - "tbl v1.16b, {v0.16b}, v2.16b \n" // look up 4 pixels - MEMACCESS(1) - "st1 {v1.16b}, [%1], #16 \n" // store 4. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "r"(shuffler) // %3 - : "cc", "memory", "v0", "v1", "v2" // Clobber List - ); -} -#endif // HAS_ARGBSHUFFLEROW_NEON - -#ifdef HAS_I422TOYUY2ROW_NEON -void I422ToYUY2Row_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_yuy2, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld2 {v0.8b, v1.8b}, [%0], #16 \n" // load 16 Ys - "mov v2.8b, v1.8b \n" - MEMACCESS(1) - "ld1 {v1.8b}, [%1], #8 \n" // load 8 Us - MEMACCESS(2) - "ld1 {v3.8b}, [%2], #8 \n" // load 8 Vs - "subs %4, %4, #16 \n" // 16 pixels - MEMACCESS(3) - "st4 {v0.8b-v3.8b}, [%3], #32 \n" // Store 8 YUY2/16 pixels. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_yuy2), // %3 - "+r"(width) // %4 - : - : "cc", "memory", "v0", "v1", "v2", "v3" - ); -} -#endif // HAS_I422TOYUY2ROW_NEON - -#ifdef HAS_I422TOUYVYROW_NEON -void I422ToUYVYRow_NEON(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_uyvy, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld2 {v1.8b, v2.8b}, [%0], #16 \n" // load 16 Ys - "mov v3.8b, v2.8b \n" - MEMACCESS(1) - "ld1 {v0.8b}, [%1], #8 \n" // load 8 Us - MEMACCESS(2) - "ld1 {v2.8b}, [%2], #8 \n" // load 8 Vs - "subs %4, %4, #16 \n" // 16 pixels - MEMACCESS(3) - "st4 {v0.8b-v3.8b}, [%3], #32 \n" // Store 8 UYVY/16 pixels. - "bgt 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_uyvy), // %3 - "+r"(width) // %4 - : - : "cc", "memory", "v0", "v1", "v2", "v3" - ); -} -#endif // HAS_I422TOUYVYROW_NEON - -#ifdef HAS_ARGBTORGB565ROW_NEON -void ARGBToRGB565Row_NEON(const uint8* src_argb, uint8* dst_rgb565, int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d21, d22, d23}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGBTORGB565 - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 8 pixels RGB565. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_rgb565), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q8", "q9", "q10", "q11" - ); -} -#endif // HAS_ARGBTORGB565ROW_NEON - -#ifdef HAS_ARGBTOARGB1555ROW_NEON -void ARGBToARGB1555Row_NEON(const uint8* src_argb, uint8* dst_argb1555, - int pix) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d21, d22, d23}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGBTOARGB1555 - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 8 pixels ARGB1555. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb1555), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q8", "q9", "q10", "q11" - ); -} -#endif // HAS_ARGBTOARGB1555ROW_NEON - -#ifdef HAS_ARGBTOARGB4444ROW_NEON -void ARGBToARGB4444Row_NEON(const uint8* src_argb, uint8* dst_argb4444, - int pix) { - asm volatile ( - "vmov.u8 d4, #0x0f \n" // bits to clear with vbic. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d20, d21, d22, d23}, [%0]! \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGBTOARGB4444 - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" // store 8 pixels ARGB4444. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb4444), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q8", "q9", "q10", "q11" - ); -} -#endif // HAS_ARGBTOARGB4444ROW_NEON - -#ifdef HAS_ARGBTOYROW_NEON -void ARGBToYRow_NEON(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #13 \n" // B * 0.1016 coefficient - "movi v5.8b, #65 \n" // G * 0.5078 coefficient - "movi v6.8b, #33 \n" // R * 0.2578 coefficient - "movi v7.8b, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v3.8h, v0.8b, v4.8b \n" // B - "umlal v3.8h, v1.8b, v5.8b \n" // G - "umlal v3.8h, v2.8b, v6.8b \n" // R - "sqrshrun v0.8b, v3.8h, #7 \n" // 16 bit to 8 bit Y - "uqadd v0.8b, v0.8b, v7.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7" - ); -} -#endif // HAS_ARGBTOYROW_NEON - -#ifdef HAS_ARGBTOYJROW_NEON -void ARGBToYJRow_NEON(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #15 \n" // B * 0.11400 coefficient - "movi v5.8b, #75 \n" // G * 0.58700 coefficient - "movi v6.8b, #38 \n" // R * 0.29900 coefficient - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v3.8h, v0.8b, v4.8b \n" // B - "umlal v3.8h, v1.8b, v5.8b \n" // G - "umlal v3.8h, v2.8b, v6.8b \n" // R - "sqrshrun v0.8b, v3.8h, #7 \n" // 15 bit to 8 bit Y - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6" - ); -} -#endif // HAS_ARGBTOYJROW_NEON - -// 8x1 pixels. -#ifdef HAS_ARGBTOUV444ROW_NEON -void ARGBToUV444Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "movi v24.8b, #112 \n" // UB / VR 0.875 coefficient - "movi v25.8b, #74 \n" // UG -0.5781 coefficient - "movi v26.8b, #38 \n" // UR -0.2969 coefficient - "movi v27.8b, #18 \n" // VB -0.1406 coefficient - "movi v28.8b, #94 \n" // VG -0.7344 coefficient - "movi v29.16b,#0x80 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "umull v4.8h, v0.8b, v24.8b \n" // B - "umlsl v4.8h, v1.8b, v25.8b \n" // G - "umlsl v4.8h, v2.8b, v26.8b \n" // R - "add v4.8h, v4.8h, v29.8h \n" // +128 -> unsigned - - "umull v3.8h, v2.8b, v24.8b \n" // R - "umlsl v3.8h, v1.8b, v28.8b \n" // G - "umlsl v3.8h, v0.8b, v27.8b \n" // B - "add v3.8h, v3.8h, v29.8h \n" // +128 -> unsigned - - "uqshrn v0.8b, v4.8h, #8 \n" // 16 bit to 8 bit U - "uqshrn v1.8b, v3.8h, #8 \n" // 16 bit to 8 bit V - - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels U. - MEMACCESS(2) - "st1 {v1.8b}, [%2], #8 \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", - "v24", "v25", "v26", "v27", "v28", "v29" - ); -} -#endif // HAS_ARGBTOUV444ROW_NEON - -// 16x1 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -#ifdef HAS_ARGBTOUV422ROW_NEON -void ARGBToUV422Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "movi v20.8h, #112 / 2 \n" // UB / VR 0.875 coefficient - "movi v21.8h, #74 / 2 \n" // UG -0.5781 coefficient - "movi v22.8h, #38 / 2 \n" // UR -0.2969 coefficient - "movi v23.8h, #18 / 2 \n" // VB -0.1406 coefficient - "movi v24.8h, #94 / 2 \n" // VG -0.7344 coefficient - "movi v25.16b, #0x80 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.16b-v3.16b}, [%0], #64 \n" // load 16 ARGB pixels. - - "uaddlp v0.8h, v0.16b \n" // B 16 bytes -> 8 shorts. - "uaddlp v1.8h, v1.16b \n" // G 16 bytes -> 8 shorts. - "uaddlp v2.8h, v2.16b \n" // R 16 bytes -> 8 shorts. - - "subs %3, %3, #16 \n" // 16 processed per loop. - "mul v3.8h, v0.8h, v20.8h \n" // B - "mls v3.8h, v1.8h, v21.8h \n" // G - "mls v3.8h, v2.8h, v22.8h \n" // R - "add v3.8h, v3.8h, v25.8h \n" // +128 -> unsigned - - "mul v4.8h, v2.8h, v20.8h \n" // R - "mls v4.8h, v1.8h, v24.8h \n" // G - "mls v4.8h, v0.8h, v23.8h \n" // B - "add v4.8h, v4.8h, v25.8h \n" // +128 -> unsigned - - "uqshrn v0.8b, v3.8h, #8 \n" // 16 bit to 8 bit U - "uqshrn v1.8b, v4.8h, #8 \n" // 16 bit to 8 bit V - - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels U. - MEMACCESS(2) - "st1 {v1.8b}, [%2], #8 \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", - "v20", "v21", "v22", "v23", "v24", "v25" - ); -} -#endif // HAS_ARGBTOUV422ROW_NEON - -// 32x1 pixels -> 8x1. pix is number of argb pixels. e.g. 32. -#ifdef HAS_ARGBTOUV411ROW_NEON -void ARGBToUV411Row_NEON(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "movi v20.8h, #112 / 2 \n" // UB / VR 0.875 coefficient - "movi v21.8h, #74 / 2 \n" // UG -0.5781 coefficient - "movi v22.8h, #38 / 2 \n" // UR -0.2969 coefficient - "movi v23.8h, #18 / 2 \n" // VB -0.1406 coefficient - "movi v24.8h, #94 / 2 \n" // VG -0.7344 coefficient - "movi v25.16b, #0x80 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.16b-v3.16b}, [%0], #64 \n" // load 16 ARGB pixels. - "uaddlp v0.8h, v0.16b \n" // B 16 bytes -> 8 shorts. - "uaddlp v1.8h, v1.16b \n" // G 16 bytes -> 8 shorts. - "uaddlp v2.8h, v2.16b \n" // R 16 bytes -> 8 shorts. - MEMACCESS(0) - "ld4 {v4.16b-v7.16b}, [%0], #64 \n" // load next 16 ARGB pixels. - "uaddlp v4.8h, v4.16b \n" // B 16 bytes -> 8 shorts. - "uaddlp v5.8h, v5.16b \n" // G 16 bytes -> 8 shorts. - "uaddlp v6.8h, v6.16b \n" // R 16 bytes -> 8 shorts. - - "addp v0.8h, v0.8h, v4.8h \n" // B 16 shorts -> 8 shorts. - "addp v1.8h, v1.8h, v5.8h \n" // G 16 shorts -> 8 shorts. - "addp v2.8h, v2.8h, v6.8h \n" // R 16 shorts -> 8 shorts. - - "urshr v0.8h, v0.8h, #1 \n" // 2x average - "urshr v1.8h, v1.8h, #1 \n" - "urshr v2.8h, v2.8h, #1 \n" - - "subs %3, %3, #32 \n" // 32 processed per loop. - "mul v3.8h, v0.8h, v20.8h \n" // B - "mls v3.8h, v1.8h, v21.8h \n" // G - "mls v3.8h, v2.8h, v22.8h \n" // R - "add v3.8h, v3.8h, v25.8h \n" // +128 -> unsigned - "mul v4.8h, v2.8h, v20.8h \n" // R - "mls v4.8h, v1.8h, v24.8h \n" // G - "mls v4.8h, v0.8h, v23.8h \n" // B - "add v4.8h, v4.8h, v25.8h \n" // +128 -> unsigned - "uqshrn v0.8b, v3.8h, #8 \n" // 16 bit to 8 bit U - "uqshrn v1.8b, v4.8h, #8 \n" // 16 bit to 8 bit V - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels U. - MEMACCESS(2) - "st1 {v1.8b}, [%2], #8 \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", - "v20", "v21", "v22", "v23", "v24", "v25" - ); -} -#endif // HAS_ARGBTOUV411ROW_NEON - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -#define RGBTOUV(QB, QG, QR) \ - "vmul.s16 q8, " #QB ", q10 \n" /* B */ \ - "vmls.s16 q8, " #QG ", q11 \n" /* G */ \ - "vmls.s16 q8, " #QR ", q12 \n" /* R */ \ - "vadd.u16 q8, q8, q15 \n" /* +128 -> unsigned */ \ - "vmul.s16 q9, " #QR ", q10 \n" /* R */ \ - "vmls.s16 q9, " #QG ", q14 \n" /* G */ \ - "vmls.s16 q9, " #QB ", q13 \n" /* B */ \ - "vadd.u16 q9, q9, q15 \n" /* +128 -> unsigned */ \ - "vqshrn.u16 d0, q8, #8 \n" /* 16 bit to 8 bit U */ \ - "vqshrn.u16 d1, q9, #8 \n" /* 16 bit to 8 bit V */ - -// TODO(fbarchard): Consider vhadd vertical, then vpaddl horizontal, avoid shr. -#ifdef HAS_ARGBTOUVROW_NEON -void ARGBToUVRow_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more ARGB pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 ARGB pixels. - "vpadal.u8 q0, q4 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stride_argb), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_ARGBTOUVROW_NEON - -// TODO(fbarchard): Subsample match C code. -#ifdef HAS_ARGBTOUVJROW_NEON -void ARGBToUVJRow_NEON(const uint8* src_argb, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #127 / 2 \n" // UB / VR 0.500 coefficient - "vmov.s16 q11, #84 / 2 \n" // UG -0.33126 coefficient - "vmov.s16 q12, #43 / 2 \n" // UR -0.16874 coefficient - "vmov.s16 q13, #20 / 2 \n" // VB -0.08131 coefficient - "vmov.s16 q14, #107 / 2 \n" // VG -0.41869 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more ARGB pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 ARGB pixels. - "vpadal.u8 q0, q4 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stride_argb), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_ARGBTOUVJROW_NEON - -#ifdef HAS_BGRATOUVROW_NEON -void BGRAToUVRow_NEON(const uint8* src_bgra, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_bgra - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 BGRA pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 BGRA pixels. - "vpaddl.u8 q3, q3 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more BGRA pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 BGRA pixels. - "vpadal.u8 q3, q7 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q1, q1, #1 \n" // 2x average - "vrshr.u16 q2, q2, #1 \n" - "vrshr.u16 q3, q3, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q3, q2, q1) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_bgra), // %0 - "+r"(src_stride_bgra), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_BGRATOUVROW_NEON - -#ifdef HAS_ABGRTOUVROW_NEON -void ABGRToUVRow_NEON(const uint8* src_abgr, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_abgr - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ABGR pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ABGR pixels. - "vpaddl.u8 q2, q2 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q0, q0 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more ABGR pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 ABGR pixels. - "vpadal.u8 q2, q6 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q0, q4 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q2, q1, q0) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_abgr), // %0 - "+r"(src_stride_abgr), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_ABGRTOUVROW_NEON - -#ifdef HAS_RGBATOUVROW_NEON -void RGBAToUVRow_NEON(const uint8* src_rgba, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_rgba - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 RGBA pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 RGBA pixels. - "vpaddl.u8 q0, q1 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q2 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q3 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d8, d10, d12, d14}, [%1]! \n" // load 8 more RGBA pixels. - MEMACCESS(1) - "vld4.8 {d9, d11, d13, d15}, [%1]! \n" // load last 8 RGBA pixels. - "vpadal.u8 q0, q5 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q6 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q7 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_rgba), // %0 - "+r"(src_stride_rgba), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_RGBATOUVROW_NEON - -#ifdef HAS_RGB24TOUVROW_NEON -void RGB24ToUVRow_NEON(const uint8* src_rgb24, int src_stride_rgb24, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_rgb24 - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d0, d2, d4}, [%0]! \n" // load 8 RGB24 pixels. - MEMACCESS(0) - "vld3.8 {d1, d3, d5}, [%0]! \n" // load next 8 RGB24 pixels. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld3.8 {d8, d10, d12}, [%1]! \n" // load 8 more RGB24 pixels. - MEMACCESS(1) - "vld3.8 {d9, d11, d13}, [%1]! \n" // load last 8 RGB24 pixels. - "vpadal.u8 q0, q4 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q6 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q0, q1, q2) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(src_stride_rgb24), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_RGB24TOUVROW_NEON - -#ifdef HAS_RAWTOUVROW_NEON -void RAWToUVRow_NEON(const uint8* src_raw, int src_stride_raw, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_raw - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld3.8 {d0, d2, d4}, [%0]! \n" // load 8 RAW pixels. - MEMACCESS(0) - "vld3.8 {d1, d3, d5}, [%0]! \n" // load next 8 RAW pixels. - "vpaddl.u8 q2, q2 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q0, q0 \n" // R 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld3.8 {d8, d10, d12}, [%1]! \n" // load 8 more RAW pixels. - MEMACCESS(1) - "vld3.8 {d9, d11, d13}, [%1]! \n" // load last 8 RAW pixels. - "vpadal.u8 q2, q6 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q5 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q0, q4 \n" // R 16 bytes -> 8 shorts. - - "vrshr.u16 q0, q0, #1 \n" // 2x average - "vrshr.u16 q1, q1, #1 \n" - "vrshr.u16 q2, q2, #1 \n" - - "subs %4, %4, #16 \n" // 32 processed per loop. - RGBTOUV(q2, q1, q0) - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_raw), // %0 - "+r"(src_stride_raw), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_RAWTOUVROW_NEON - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -#ifdef HAS_RGB565TOUVROW_NEON -void RGB565ToUVRow_NEON(const uint8* src_rgb565, int src_stride_rgb565, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 RGB565 pixels. - RGB565TOARGB - "vpaddl.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // next 8 RGB565 pixels. - RGB565TOARGB - "vpaddl.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // load 8 RGB565 pixels. - RGB565TOARGB - "vpadal.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // next 8 RGB565 pixels. - RGB565TOARGB - "vpadal.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - "vrshr.u16 q4, q4, #1 \n" // 2x average - "vrshr.u16 q5, q5, #1 \n" - "vrshr.u16 q6, q6, #1 \n" - - "subs %4, %4, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q4, q10 \n" // B - "vmls.s16 q8, q5, q11 \n" // G - "vmls.s16 q8, q6, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q6, q10 \n" // R - "vmls.s16 q9, q5, q14 \n" // G - "vmls.s16 q9, q4, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_rgb565), // %0 - "+r"(src_stride_rgb565), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_RGB565TOUVROW_NEON - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -#ifdef HAS_ARGB1555TOUVROW_NEON -void ARGB1555ToUVRow_NEON(const uint8* src_argb1555, int src_stride_argb1555, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB1555 pixels. - RGB555TOARGB - "vpaddl.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // next 8 ARGB1555 pixels. - RGB555TOARGB - "vpaddl.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // load 8 ARGB1555 pixels. - RGB555TOARGB - "vpadal.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // next 8 ARGB1555 pixels. - RGB555TOARGB - "vpadal.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - "vrshr.u16 q4, q4, #1 \n" // 2x average - "vrshr.u16 q5, q5, #1 \n" - "vrshr.u16 q6, q6, #1 \n" - - "subs %4, %4, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q4, q10 \n" // B - "vmls.s16 q8, q5, q11 \n" // G - "vmls.s16 q8, q6, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q6, q10 \n" // R - "vmls.s16 q9, q5, q14 \n" // G - "vmls.s16 q9, q4, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb1555), // %0 - "+r"(src_stride_argb1555), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_ARGB1555TOUVROW_NEON - -// 16x2 pixels -> 8x1. pix is number of argb pixels. e.g. 16. -#ifdef HAS_ARGB4444TOUVROW_NEON -void ARGB4444ToUVRow_NEON(const uint8* src_argb4444, int src_stride_argb4444, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "add %1, %0, %1 \n" // src_stride + src_argb - "vmov.s16 q10, #112 / 2 \n" // UB / VR 0.875 coefficient - "vmov.s16 q11, #74 / 2 \n" // UG -0.5781 coefficient - "vmov.s16 q12, #38 / 2 \n" // UR -0.2969 coefficient - "vmov.s16 q13, #18 / 2 \n" // VB -0.1406 coefficient - "vmov.s16 q14, #94 / 2 \n" // VG -0.7344 coefficient - "vmov.u16 q15, #0x8080 \n" // 128.5 - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpaddl.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // next 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpaddl.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpaddl.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpaddl.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // load 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpadal.u8 d8, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d10, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d12, d2 \n" // R 8 bytes -> 4 shorts. - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" // next 8 ARGB4444 pixels. - ARGB4444TOARGB - "vpadal.u8 d9, d0 \n" // B 8 bytes -> 4 shorts. - "vpadal.u8 d11, d1 \n" // G 8 bytes -> 4 shorts. - "vpadal.u8 d13, d2 \n" // R 8 bytes -> 4 shorts. - - "vrshr.u16 q4, q4, #1 \n" // 2x average - "vrshr.u16 q5, q5, #1 \n" - "vrshr.u16 q6, q6, #1 \n" - - "subs %4, %4, #16 \n" // 16 processed per loop. - "vmul.s16 q8, q4, q10 \n" // B - "vmls.s16 q8, q5, q11 \n" // G - "vmls.s16 q8, q6, q12 \n" // R - "vadd.u16 q8, q8, q15 \n" // +128 -> unsigned - "vmul.s16 q9, q6, q10 \n" // R - "vmls.s16 q9, q5, q14 \n" // G - "vmls.s16 q9, q4, q13 \n" // B - "vadd.u16 q9, q9, q15 \n" // +128 -> unsigned - "vqshrn.u16 d0, q8, #8 \n" // 16 bit to 8 bit U - "vqshrn.u16 d1, q9, #8 \n" // 16 bit to 8 bit V - MEMACCESS(2) - "vst1.8 {d0}, [%2]! \n" // store 8 pixels U. - MEMACCESS(3) - "vst1.8 {d1}, [%3]! \n" // store 8 pixels V. - "bgt 1b \n" - : "+r"(src_argb4444), // %0 - "+r"(src_stride_argb4444), // %1 - "+r"(dst_u), // %2 - "+r"(dst_v), // %3 - "+r"(pix) // %4 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7", - "q8", "q9", "q10", "q11", "q12", "q13", "q14", "q15" - ); -} -#endif // HAS_ARGB4444TOUVROW_NEON - -#ifdef HAS_RGB565TOYROW_NEON -void RGB565ToYRow_NEON(const uint8* src_rgb565, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 RGB565 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - RGB565TOARGB - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_rgb565), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q12", "q13" - ); -} -#endif // HAS_RGB565TOYROW_NEON - -#ifdef HAS_ARGB1555TOYROW_NEON -void ARGB1555ToYRow_NEON(const uint8* src_argb1555, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB1555 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB1555TOARGB - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb1555), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q12", "q13" - ); -} -#endif // HAS_ARGB1555TOYROW_NEON - -#ifdef HAS_ARGB4444TOYROW_NEON -void ARGB4444ToYRow_NEON(const uint8* src_argb4444, uint8* dst_y, int pix) { - asm volatile ( - "vmov.u8 d24, #13 \n" // B * 0.1016 coefficient - "vmov.u8 d25, #65 \n" // G * 0.5078 coefficient - "vmov.u8 d26, #33 \n" // R * 0.2578 coefficient - "vmov.u8 d27, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load 8 ARGB4444 pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - ARGB4444TOARGB - "vmull.u8 q2, d0, d24 \n" // B - "vmlal.u8 q2, d1, d25 \n" // G - "vmlal.u8 q2, d2, d26 \n" // R - "vqrshrun.s16 d0, q2, #7 \n" // 16 bit to 8 bit Y - "vqadd.u8 d0, d27 \n" - MEMACCESS(1) - "vst1.8 {d0}, [%1]! \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_argb4444), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "q0", "q1", "q2", "q3", "q12", "q13" - ); -} -#endif // HAS_ARGB4444TOYROW_NEON - -#ifdef HAS_BGRATOYROW_NEON -void BGRAToYRow_NEON(const uint8* src_bgra, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #33 \n" // R * 0.2578 coefficient - "movi v5.8b, #65 \n" // G * 0.5078 coefficient - "movi v6.8b, #13 \n" // B * 0.1016 coefficient - "movi v7.8b, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 pixels of BGRA. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v16.8h, v1.8b, v4.8b \n" // R - "umlal v16.8h, v2.8b, v5.8b \n" // G - "umlal v16.8h, v3.8b, v6.8b \n" // B - "sqrshrun v0.8b, v16.8h, #7 \n" // 16 bit to 8 bit Y - "uqadd v0.8b, v0.8b, v7.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_bgra), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16" - ); -} -#endif // HAS_BGRATOYROW_NEON - -#ifdef HAS_ABGRTOYROW_NEON -void ABGRToYRow_NEON(const uint8* src_abgr, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #33 \n" // R * 0.2578 coefficient - "movi v5.8b, #65 \n" // G * 0.5078 coefficient - "movi v6.8b, #13 \n" // B * 0.1016 coefficient - "movi v7.8b, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 pixels of ABGR. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v16.8h, v0.8b, v4.8b \n" // R - "umlal v16.8h, v1.8b, v5.8b \n" // G - "umlal v16.8h, v2.8b, v6.8b \n" // B - "sqrshrun v0.8b, v16.8h, #7 \n" // 16 bit to 8 bit Y - "uqadd v0.8b, v0.8b, v7.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_abgr), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16" - ); -} -#endif // HAS_ABGRTOYROW_NEON - -#ifdef HAS_RGBATOYROW_NEON -void RGBAToYRow_NEON(const uint8* src_rgba, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #13 \n" // B * 0.1016 coefficient - "movi v5.8b, #65 \n" // G * 0.5078 coefficient - "movi v6.8b, #33 \n" // R * 0.2578 coefficient - "movi v7.8b, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 pixels of RGBA. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v16.8h, v1.8b, v4.8b \n" // B - "umlal v16.8h, v2.8b, v5.8b \n" // G - "umlal v16.8h, v3.8b, v6.8b \n" // R - "sqrshrun v0.8b, v16.8h, #7 \n" // 16 bit to 8 bit Y - "uqadd v0.8b, v0.8b, v7.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_rgba), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16" - ); -} -#endif // HAS_RGBATOYROW_NEON - -#ifdef HAS_RGB24TOYROW_NEON -void RGB24ToYRow_NEON(const uint8* src_rgb24, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #13 \n" // B * 0.1016 coefficient - "movi v5.8b, #65 \n" // G * 0.5078 coefficient - "movi v6.8b, #33 \n" // R * 0.2578 coefficient - "movi v7.8b, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld3 {v0.8b-v2.8b}, [%0], #24 \n" // load 8 pixels of RGB24. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v16.8h, v0.8b, v4.8b \n" // B - "umlal v16.8h, v1.8b, v5.8b \n" // G - "umlal v16.8h, v2.8b, v6.8b \n" // R - "sqrshrun v0.8b, v16.8h, #7 \n" // 16 bit to 8 bit Y - "uqadd v0.8b, v0.8b, v7.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16" - ); -} -#endif // HAS_RGB24TOYROW_NEON - -#ifdef HAS_RAWTOYROW_NEON -void RAWToYRow_NEON(const uint8* src_raw, uint8* dst_y, int pix) { - asm volatile ( - "movi v4.8b, #33 \n" // R * 0.2578 coefficient - "movi v5.8b, #65 \n" // G * 0.5078 coefficient - "movi v6.8b, #13 \n" // B * 0.1016 coefficient - "movi v7.8b, #16 \n" // Add 16 constant - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld3 {v0.8b-v2.8b}, [%0], #24 \n" // load 8 pixels of RAW. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v16.8h, v0.8b, v4.8b \n" // B - "umlal v16.8h, v1.8b, v5.8b \n" // G - "umlal v16.8h, v2.8b, v6.8b \n" // R - "sqrshrun v0.8b, v16.8h, #7 \n" // 16 bit to 8 bit Y - "uqadd v0.8b, v0.8b, v7.8b \n" - MEMACCESS(1) - "st1 {v0.8b}, [%1], #8 \n" // store 8 pixels Y. - "bgt 1b \n" - : "+r"(src_raw), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16" - ); -} -#endif // HAS_RAWTOYROW_NEON - -// Bilinear filter 16x2 -> 16x1 -#ifdef HAS_INTERPOLATEROW_NEON -void InterpolateRow_NEON(uint8* dst_ptr, - const uint8* src_ptr, ptrdiff_t src_stride, - int dst_width, int source_y_fraction) { - int y1_fraction = source_y_fraction; - int y0_fraction = 256 - y1_fraction; - const uint8* src_ptr1 = src_ptr + src_stride; - asm volatile ( - "cmp %4, #0 \n" - "beq 100f \n" - "cmp %4, #64 \n" - "beq 75f \n" - "cmp %4, #128 \n" - "beq 50f \n" - "cmp %4, #192 \n" - "beq 25f \n" - - "dup v5.16b, %w4 \n" - "dup v4.16b, %w5 \n" - // General purpose row blend. - "1: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v1.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "umull v2.8h, v0.8b, v4.8b \n" - "umull2 v3.8h, v0.16b, v4.16b \n" - "umlal v2.8h, v1.8b, v5.8b \n" - "umlal2 v3.8h, v1.16b, v5.16b \n" - "rshrn v0.8b, v2.8h, #8 \n" - "rshrn2 v0.16b, v3.8h, #8 \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 1b \n" - "b 99f \n" - - // Blend 25 / 75. - "25: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v1.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 25b \n" - "b 99f \n" - - // Blend 50 / 50. - "50: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v1.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 50b \n" - "b 99f \n" - - // Blend 75 / 25. - "75: \n" - MEMACCESS(1) - "ld1 {v1.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v0.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 75b \n" - "b 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - "100: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - "subs %3, %3, #16 \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 100b \n" - - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(src_ptr1), // %2 - "+r"(dst_width), // %3 - "+r"(y1_fraction), // %4 - "+r"(y0_fraction) // %5 - : - : "cc", "memory", "v0", "v1", "v3", "v4", "v5" - ); -} -#endif // HAS_INTERPOLATEROW_NEON - -// dr * (256 - sa) / 256 + sr = dr - dr * sa / 256 + sr -#ifdef HAS_ARGBBLENDROW_NEON -void ARGBBlendRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - "subs %3, %3, #8 \n" - "blt 89f \n" - // Blend 8 pixels. - "8: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 pixels of ARGB0. - MEMACCESS(1) - "ld4 {v4.8b-v7.8b}, [%1], #32 \n" // load 8 pixels of ARGB1. - "subs %3, %3, #8 \n" // 8 processed per loop. - "umull v16.8h, v4.8b, v3.8b \n" // db * a - "umull v17.8h, v5.8b, v3.8b \n" // dg * a - "umull v18.8h, v6.8b, v3.8b \n" // dr * a - "uqrshrn v16.8b, v16.8h, #8 \n" // db >>= 8 - "uqrshrn v17.8b, v17.8h, #8 \n" // dg >>= 8 - "uqrshrn v18.8b, v18.8h, #8 \n" // dr >>= 8 - "uqsub v4.8b, v4.8b, v16.8b \n" // db - (db * a / 256) - "uqsub v5.8b, v5.8b, v17.8b \n" // dg - (dg * a / 256) - "uqsub v6.8b, v6.8b, v18.8b \n" // dr - (dr * a / 256) - "uqadd v0.8b, v0.8b, v4.8b \n" // + sb - "uqadd v1.8b, v1.8b, v5.8b \n" // + sg - "uqadd v2.8b, v2.8b, v6.8b \n" // + sr - "movi v3.8b, #255 \n" // a = 255 - MEMACCESS(2) - "st4 {v0.8b-v3.8b}, [%2], #32 \n" // store 8 pixels of ARGB. - "bge 8b \n" - - "89: \n" - "adds %3, %3, #8-1 \n" - "blt 99f \n" - - // Blend 1 pixels. - "1: \n" - MEMACCESS(0) - "ld4 {v0.b-v3.b}[0], [%0], #4 \n" // load 1 pixel ARGB0. - MEMACCESS(1) - "ld4 {v4.b-v7.b}[0], [%1], #4 \n" // load 1 pixel ARGB1. - "subs %3, %3, #1 \n" // 1 processed per loop. - "umull v16.8h, v4.8b, v3.8b \n" // db * a - "umull v17.8h, v5.8b, v3.8b \n" // dg * a - "umull v18.8h, v6.8b, v3.8b \n" // dr * a - "uqrshrn v16.8b, v16.8h, #8 \n" // db >>= 8 - "uqrshrn v17.8b, v17.8h, #8 \n" // dg >>= 8 - "uqrshrn v18.8b, v18.8h, #8 \n" // dr >>= 8 - "uqsub v4.8b, v4.8b, v16.8b \n" // db - (db * a / 256) - "uqsub v5.8b, v5.8b, v17.8b \n" // dg - (dg * a / 256) - "uqsub v6.8b, v6.8b, v18.8b \n" // dr - (dr * a / 256) - "uqadd v0.8b, v0.8b, v4.8b \n" // + sb - "uqadd v1.8b, v1.8b, v5.8b \n" // + sg - "uqadd v2.8b, v2.8b, v6.8b \n" // + sr - "movi v3.8b, #255 \n" // a = 255 - MEMACCESS(2) - "st4 {v0.b-v3.b}[0], [%2], #4 \n" // store 1 pixel. - "bge 1b \n" - - "99: \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", - "v16", "v17", "v18" - ); -} -#endif // HAS_ARGBBLENDROW_NEON - -// Attenuate 8 pixels at a time. -#ifdef HAS_ARGBATTENUATEROW_NEON -void ARGBAttenuateRow_NEON(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - // Attenuate 8 pixels. - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v4.8h, v0.8b, v3.8b \n" // b * a - "umull v5.8h, v1.8b, v3.8b \n" // g * a - "umull v6.8h, v2.8b, v3.8b \n" // r * a - "uqrshrn v0.8b, v4.8h, #8 \n" // b >>= 8 - "uqrshrn v1.8b, v5.8h, #8 \n" // g >>= 8 - "uqrshrn v2.8b, v6.8h, #8 \n" // r >>= 8 - MEMACCESS(1) - "st4 {v0.8b-v3.8b}, [%1], #32 \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6" - ); -} -#endif // HAS_ARGBATTENUATEROW_NEON - -// Quantize 8 ARGB pixels (32 bytes). -// dst = (dst * scale >> 16) * interval_size + interval_offset; -#ifdef HAS_ARGBQUANTIZEROW_NEON -void ARGBQuantizeRow_NEON(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width) { - asm volatile ( - "dup v4.8h, %w2 \n" - "ushr v4.8h, v4.8h, #1 \n" // scale >>= 1 - "dup v5.8h, %w3 \n" // interval multiply. - "dup v6.8h, %w4 \n" // interval add - - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0] \n" // load 8 pixels of ARGB. - "subs %1, %1, #8 \n" // 8 processed per loop. - "uxtl v0.8h, v0.8b \n" // b (0 .. 255) - "uxtl v1.8h, v1.8b \n" - "uxtl v2.8h, v2.8b \n" - "sqdmulh v0.8h, v0.8h, v4.8h \n" // b * scale - "sqdmulh v1.8h, v1.8h, v4.8h \n" // g - "sqdmulh v2.8h, v2.8h, v4.8h \n" // r - "mul v0.8h, v0.8h, v5.8h \n" // b * interval_size - "mul v1.8h, v1.8h, v5.8h \n" // g - "mul v2.8h, v2.8h, v5.8h \n" // r - "add v0.8h, v0.8h, v6.8h \n" // b + interval_offset - "add v1.8h, v1.8h, v6.8h \n" // g - "add v2.8h, v2.8h, v6.8h \n" // r - "uqxtn v0.8b, v0.8h \n" - "uqxtn v1.8b, v1.8h \n" - "uqxtn v2.8b, v2.8h \n" - MEMACCESS(0) - "st4 {v0.8b-v3.8b}, [%0], #32 \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(dst_argb), // %0 - "+r"(width) // %1 - : "r"(scale), // %2 - "r"(interval_size), // %3 - "r"(interval_offset) // %4 - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6" - ); -} -#endif // HAS_ARGBQUANTIZEROW_NEON - -// Shade 8 pixels at a time by specified value. -// NOTE vqrdmulh.s16 q10, q10, d0[0] must use a scaler register from 0 to 8. -// Rounding in vqrdmulh does +1 to high if high bit of low s16 is set. -#ifdef HAS_ARGBSHADEROW_NEON -void ARGBShadeRow_NEON(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value) { - asm volatile ( - "dup v0.4s, %w3 \n" // duplicate scale value. - "zip1 v0.8b, v0.8b, v0.8b \n" // v0.8b aarrggbb. - "ushr v0.8h, v0.8h, #1 \n" // scale / 2. - - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v4.8b-v7.8b}, [%0], #32 \n" // load 8 pixels of ARGB. - "subs %2, %2, #8 \n" // 8 processed per loop. - "uxtl v4.8h, v4.8b \n" // b (0 .. 255) - "uxtl v5.8h, v5.8b \n" - "uxtl v6.8h, v6.8b \n" - "uxtl v7.8h, v7.8b \n" - "sqrdmulh v4.8h, v4.8h, v0.h[0] \n" // b * scale * 2 - "sqrdmulh v5.8h, v5.8h, v0.h[1] \n" // g - "sqrdmulh v6.8h, v6.8h, v0.h[2] \n" // r - "sqrdmulh v7.8h, v7.8h, v0.h[3] \n" // a - "uqxtn v4.8b, v4.8h \n" - "uqxtn v5.8b, v5.8h \n" - "uqxtn v6.8b, v6.8h \n" - "uqxtn v7.8b, v7.8h \n" - MEMACCESS(1) - "st4 {v4.8b-v7.8b}, [%1], #32 \n" // store 8 pixels of ARGB. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(value) // %3 - : "cc", "memory", "v0", "v4", "v5", "v6", "v7" - ); -} -#endif // HAS_ARGBSHADEROW_NEON - -// Convert 8 ARGB pixels (64 bytes) to 8 Gray ARGB pixels -// Similar to ARGBToYJ but stores ARGB. -// C code is (15 * b + 75 * g + 38 * r + 64) >> 7; -#ifdef HAS_ARGBGRAYROW_NEON -void ARGBGrayRow_NEON(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - "movi v24.8b, #15 \n" // B * 0.11400 coefficient - "movi v25.8b, #75 \n" // G * 0.58700 coefficient - "movi v26.8b, #38 \n" // R * 0.29900 coefficient - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "umull v4.8h, v0.8b, v24.8b \n" // B - "umlal v4.8h, v1.8b, v25.8b \n" // G - "umlal v4.8h, v2.8b, v26.8b \n" // R - "sqrshrun v0.8b, v4.8h, #7 \n" // 15 bit to 8 bit B - "mov v1.8b, v0.8b \n" // G - "mov v2.8b, v0.8b \n" // R - MEMACCESS(1) - "st4 {v0.8b-v3.8b}, [%1], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v24", "v25", "v26" - ); -} -#endif // HAS_ARGBGRAYROW_NEON - -// Convert 8 ARGB pixels (32 bytes) to 8 Sepia ARGB pixels. -// b = (r * 35 + g * 68 + b * 17) >> 7 -// g = (r * 45 + g * 88 + b * 22) >> 7 -// r = (r * 50 + g * 98 + b * 24) >> 7 - -#ifdef HAS_ARGBSEPIAROW_NEON -void ARGBSepiaRow_NEON(uint8* dst_argb, int width) { - asm volatile ( - "movi v20.8b, #17 \n" // BB coefficient - "movi v21.8b, #68 \n" // BG coefficient - "movi v22.8b, #35 \n" // BR coefficient - "movi v24.8b, #22 \n" // GB coefficient - "movi v25.8b, #88 \n" // GG coefficient - "movi v26.8b, #45 \n" // GR coefficient - "movi v28.8b, #24 \n" // BB coefficient - "movi v29.8b, #98 \n" // BG coefficient - "movi v30.8b, #50 \n" // BR coefficient - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0] \n" // load 8 ARGB pixels. - "subs %1, %1, #8 \n" // 8 processed per loop. - "umull v4.8h, v0.8b, v20.8b \n" // B to Sepia B - "umlal v4.8h, v1.8b, v21.8b \n" // G - "umlal v4.8h, v2.8b, v22.8b \n" // R - "umull v5.8h, v0.8b, v24.8b \n" // B to Sepia G - "umlal v5.8h, v1.8b, v25.8b \n" // G - "umlal v5.8h, v2.8b, v26.8b \n" // R - "umull v6.8h, v0.8b, v28.8b \n" // B to Sepia R - "umlal v6.8h, v1.8b, v29.8b \n" // G - "umlal v6.8h, v2.8b, v30.8b \n" // R - "uqshrn v0.8b, v4.8h, #7 \n" // 16 bit to 8 bit B - "uqshrn v1.8b, v5.8h, #7 \n" // 16 bit to 8 bit G - "uqshrn v2.8b, v6.8h, #7 \n" // 16 bit to 8 bit R - MEMACCESS(0) - "st4 {v0.8b-v3.8b}, [%0], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(dst_argb), // %0 - "+r"(width) // %1 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", - "v20", "v21", "v22", "v24", "v25", "v26", "v28", "v29", "v30" - ); -} -#endif // HAS_ARGBSEPIAROW_NEON - -// Tranform 8 ARGB pixels (32 bytes) with color matrix. -// TODO(fbarchard): Was same as Sepia except matrix is provided. This function -// needs to saturate. Consider doing a non-saturating version. -#ifdef HAS_ARGBCOLORMATRIXROW_NEON -void ARGBColorMatrixRow_NEON(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width) { - asm volatile ( - MEMACCESS(3) - "ld1 {v2.16b}, [%3] \n" // load 3 ARGB vectors. - "sxtl v0.8h, v2.8b \n" // B,G coefficients s16. - "sxtl2 v1.8h, v2.16b \n" // R,A coefficients s16. - - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v16.8b-v19.8b}, [%0], #32 \n" // load 8 ARGB pixels. - "subs %2, %2, #8 \n" // 8 processed per loop. - "uxtl v16.8h, v16.8b \n" // b (0 .. 255) 16 bit - "uxtl v17.8h, v17.8b \n" // g - "uxtl v18.8h, v18.8b \n" // r - "uxtl v19.8h, v19.8b \n" // a - "mul v22.8h, v16.8h, v0.h[0] \n" // B = B * Matrix B - "mul v23.8h, v16.8h, v0.h[4] \n" // G = B * Matrix G - "mul v24.8h, v16.8h, v1.h[0] \n" // R = B * Matrix R - "mul v25.8h, v16.8h, v1.h[4] \n" // A = B * Matrix A - "mul v4.8h, v17.8h, v0.h[1] \n" // B += G * Matrix B - "mul v5.8h, v17.8h, v0.h[5] \n" // G += G * Matrix G - "mul v6.8h, v17.8h, v1.h[1] \n" // R += G * Matrix R - "mul v7.8h, v17.8h, v1.h[5] \n" // A += G * Matrix A - "sqadd v22.8h, v22.8h, v4.8h \n" // Accumulate B - "sqadd v23.8h, v23.8h, v5.8h \n" // Accumulate G - "sqadd v24.8h, v24.8h, v6.8h \n" // Accumulate R - "sqadd v25.8h, v25.8h, v7.8h \n" // Accumulate A - "mul v4.8h, v18.8h, v0.h[2] \n" // B += R * Matrix B - "mul v5.8h, v18.8h, v0.h[6] \n" // G += R * Matrix G - "mul v6.8h, v18.8h, v1.h[2] \n" // R += R * Matrix R - "mul v7.8h, v18.8h, v1.h[6] \n" // A += R * Matrix A - "sqadd v22.8h, v22.8h, v4.8h \n" // Accumulate B - "sqadd v23.8h, v23.8h, v5.8h \n" // Accumulate G - "sqadd v24.8h, v24.8h, v6.8h \n" // Accumulate R - "sqadd v25.8h, v25.8h, v7.8h \n" // Accumulate A - "mul v4.8h, v19.8h, v0.h[3] \n" // B += A * Matrix B - "mul v5.8h, v19.8h, v0.h[7] \n" // G += A * Matrix G - "mul v6.8h, v19.8h, v1.h[3] \n" // R += A * Matrix R - "mul v7.8h, v19.8h, v1.h[7] \n" // A += A * Matrix A - "sqadd v22.8h, v22.8h, v4.8h \n" // Accumulate B - "sqadd v23.8h, v23.8h, v5.8h \n" // Accumulate G - "sqadd v24.8h, v24.8h, v6.8h \n" // Accumulate R - "sqadd v25.8h, v25.8h, v7.8h \n" // Accumulate A - "sqshrun v16.8b, v22.8h, #6 \n" // 16 bit to 8 bit B - "sqshrun v17.8b, v23.8h, #6 \n" // 16 bit to 8 bit G - "sqshrun v18.8b, v24.8h, #6 \n" // 16 bit to 8 bit R - "sqshrun v19.8b, v25.8h, #6 \n" // 16 bit to 8 bit A - MEMACCESS(1) - "st4 {v16.8b-v19.8b}, [%1], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(matrix_argb) // %3 - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16", "v17", - "v18", "v19", "v22", "v23", "v24", "v25" - ); -} -#endif // HAS_ARGBCOLORMATRIXROW_NEON - -// TODO(fbarchard): fix vqshrun in ARGBMultiplyRow_NEON and reenable. -// Multiply 2 rows of ARGB pixels together, 8 pixels at a time. -#ifdef HAS_ARGBMULTIPLYROW_NEON -void ARGBMultiplyRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - MEMACCESS(1) - "ld4 {v4.8b-v7.8b}, [%1], #32 \n" // load 8 more ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "umull v0.8h, v0.8b, v4.8b \n" // multiply B - "umull v1.8h, v1.8b, v5.8b \n" // multiply G - "umull v2.8h, v2.8b, v6.8b \n" // multiply R - "umull v3.8h, v3.8b, v7.8b \n" // multiply A - "rshrn v0.8b, v0.8h, #8 \n" // 16 bit to 8 bit B - "rshrn v1.8b, v1.8h, #8 \n" // 16 bit to 8 bit G - "rshrn v2.8b, v2.8h, #8 \n" // 16 bit to 8 bit R - "rshrn v3.8b, v3.8h, #8 \n" // 16 bit to 8 bit A - MEMACCESS(2) - "st4 {v0.8b-v3.8b}, [%2], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7" - ); -} -#endif // HAS_ARGBMULTIPLYROW_NEON - -// Add 2 rows of ARGB pixels together, 8 pixels at a time. -#ifdef HAS_ARGBADDROW_NEON -void ARGBAddRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - MEMACCESS(1) - "ld4 {v4.8b-v7.8b}, [%1], #32 \n" // load 8 more ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "uqadd v0.8b, v0.8b, v4.8b \n" - "uqadd v1.8b, v1.8b, v5.8b \n" - "uqadd v2.8b, v2.8b, v6.8b \n" - "uqadd v3.8b, v3.8b, v7.8b \n" - MEMACCESS(2) - "st4 {v0.8b-v3.8b}, [%2], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7" - ); -} -#endif // HAS_ARGBADDROW_NEON - -// Subtract 2 rows of ARGB pixels, 8 pixels at a time. -#ifdef HAS_ARGBSUBTRACTROW_NEON -void ARGBSubtractRow_NEON(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // load 8 ARGB pixels. - MEMACCESS(1) - "ld4 {v4.8b-v7.8b}, [%1], #32 \n" // load 8 more ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "uqsub v0.8b, v0.8b, v4.8b \n" - "uqsub v1.8b, v1.8b, v5.8b \n" - "uqsub v2.8b, v2.8b, v6.8b \n" - "uqsub v3.8b, v3.8b, v7.8b \n" - MEMACCESS(2) - "st4 {v0.8b-v3.8b}, [%2], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7" - ); -} -#endif // HAS_ARGBSUBTRACTROW_NEON - -// Adds Sobel X and Sobel Y and stores Sobel into ARGB. -// A = 255 -// R = Sobel -// G = Sobel -// B = Sobel -#ifdef HAS_SOBELROW_NEON -void SobelRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - asm volatile ( - "movi v3.8b, #255 \n" // alpha - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.8b}, [%0], #8 \n" // load 8 sobelx. - MEMACCESS(1) - "ld1 {v1.8b}, [%1], #8 \n" // load 8 sobely. - "subs %3, %3, #8 \n" // 8 processed per loop. - "uqadd v0.8b, v0.8b, v1.8b \n" // add - "mov v1.8b, v0.8b \n" - "mov v2.8b, v0.8b \n" - MEMACCESS(2) - "st4 {v0.8b-v3.8b}, [%2], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3" - ); -} -#endif // HAS_SOBELROW_NEON - -// Adds Sobel X and Sobel Y and stores Sobel into plane. -#ifdef HAS_SOBELTOPLANEROW_NEON -void SobelToPlaneRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width) { - asm volatile ( - // 16 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], #16 \n" // load 16 sobelx. - MEMACCESS(1) - "ld1 {v1.16b}, [%1], #16 \n" // load 16 sobely. - "subs %3, %3, #16 \n" // 16 processed per loop. - "uqadd v0.16b, v0.16b, v1.16b \n" // add - MEMACCESS(2) - "st1 {v0.16b}, [%2], #16 \n" // store 16 pixels. - "bgt 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_y), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1" - ); -} -#endif // HAS_SOBELTOPLANEROW_NEON - -// Mixes Sobel X, Sobel Y and Sobel into ARGB. -// A = 255 -// R = Sobel X -// G = Sobel -// B = Sobel Y -#ifdef HAS_SOBELXYROW_NEON -void SobelXYRow_NEON(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - asm volatile ( - "movi v3.8b, #255 \n" // alpha - // 8 pixel loop. - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v2.8b}, [%0], #8 \n" // load 8 sobelx. - MEMACCESS(1) - "ld1 {v0.8b}, [%1], #8 \n" // load 8 sobely. - "subs %3, %3, #8 \n" // 8 processed per loop. - "uqadd v1.8b, v0.8b, v2.8b \n" // add - MEMACCESS(2) - "st4 {v0.8b-v3.8b}, [%2], #32 \n" // store 8 ARGB pixels. - "bgt 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "cc", "memory", "v0", "v1", "v2", "v3" - ); -} -#endif // HAS_SOBELXYROW_NEON - -// SobelX as a matrix is -// -1 0 1 -// -2 0 2 -// -1 0 1 -#ifdef HAS_SOBELXROW_NEON -void SobelXRow_NEON(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobelx, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.8b}, [%0],%5 \n" // top - MEMACCESS(0) - "ld1 {v1.8b}, [%0],%6 \n" - "usubl v0.8h, v0.8b, v1.8b \n" - MEMACCESS(1) - "ld1 {v2.8b}, [%1],%5 \n" // center * 2 - MEMACCESS(1) - "ld1 {v3.8b}, [%1],%6 \n" - "usubl v1.8h, v2.8b, v3.8b \n" - "add v0.8h, v0.8h, v1.8h \n" - "add v0.8h, v0.8h, v1.8h \n" - MEMACCESS(2) - "ld1 {v2.8b}, [%2],%5 \n" // bottom - MEMACCESS(2) - "ld1 {v3.8b}, [%2],%6 \n" - "subs %4, %4, #8 \n" // 8 pixels - "usubl v1.8h, v2.8b, v3.8b \n" - "add v0.8h, v0.8h, v1.8h \n" - "abs v0.8h, v0.8h \n" - "uqxtn v0.8b, v0.8h \n" - MEMACCESS(3) - "st1 {v0.8b}, [%3], #8 \n" // store 8 sobelx - "bgt 1b \n" - : "+r"(src_y0), // %0 - "+r"(src_y1), // %1 - "+r"(src_y2), // %2 - "+r"(dst_sobelx), // %3 - "+r"(width) // %4 - : "r"(2), // %5 - "r"(6) // %6 - : "cc", "memory", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif // HAS_SOBELXROW_NEON - -// SobelY as a matrix is -// -1 -2 -1 -// 0 0 0 -// 1 2 1 -#ifdef HAS_SOBELYROW_NEON -void SobelYRow_NEON(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.8b}, [%0],%4 \n" // left - MEMACCESS(1) - "ld1 {v1.8b}, [%1],%4 \n" - "usubl v0.8h, v0.8b, v1.8b \n" - MEMACCESS(0) - "ld1 {v2.8b}, [%0],%4 \n" // center * 2 - MEMACCESS(1) - "ld1 {v3.8b}, [%1],%4 \n" - "usubl v1.8h, v2.8b, v3.8b \n" - "add v0.8h, v0.8h, v1.8h \n" - "add v0.8h, v0.8h, v1.8h \n" - MEMACCESS(0) - "ld1 {v2.8b}, [%0],%5 \n" // right - MEMACCESS(1) - "ld1 {v3.8b}, [%1],%5 \n" - "subs %3, %3, #8 \n" // 8 pixels - "usubl v1.8h, v2.8b, v3.8b \n" - "add v0.8h, v0.8h, v1.8h \n" - "abs v0.8h, v0.8h \n" - "uqxtn v0.8b, v0.8h \n" - MEMACCESS(2) - "st1 {v0.8b}, [%2], #8 \n" // store 8 sobely - "bgt 1b \n" - : "+r"(src_y0), // %0 - "+r"(src_y1), // %1 - "+r"(dst_sobely), // %2 - "+r"(width) // %3 - : "r"(1), // %4 - "r"(6) // %5 - : "cc", "memory", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif // HAS_SOBELYROW_NEON -#endif // !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_posix.cc b/src/main/jni/libyuv/source/row_posix.cc deleted file mode 100644 index 106fda568..000000000 --- a/src/main/jni/libyuv/source/row_posix.cc +++ /dev/null @@ -1,6443 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC x86 and x64. -#if !defined(LIBYUV_DISABLE_X86) && (defined(__x86_64__) || defined(__i386__)) - -#if defined(HAS_ARGBTOYROW_SSSE3) || defined(HAS_ARGBGRAYROW_SSSE3) - -// Constants for ARGB -static vec8 kARGBToY = { - 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33, 0 -}; - -// JPeg full range. -static vec8 kARGBToYJ = { - 15, 75, 38, 0, 15, 75, 38, 0, 15, 75, 38, 0, 15, 75, 38, 0 -}; -#endif // defined(HAS_ARGBTOYROW_SSSE3) || defined(HAS_ARGBGRAYROW_SSSE3) - -#if defined(HAS_ARGBTOYROW_SSSE3) || defined(HAS_I422TOARGBROW_SSSE3) - -static vec8 kARGBToU = { - 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38, 0 -}; - -static vec8 kARGBToUJ = { - 127, -84, -43, 0, 127, -84, -43, 0, 127, -84, -43, 0, 127, -84, -43, 0 -}; - -static vec8 kARGBToV = { - -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112, 0, -}; - -static vec8 kARGBToVJ = { - -20, -107, 127, 0, -20, -107, 127, 0, -20, -107, 127, 0, -20, -107, 127, 0 -}; - -// Constants for BGRA -static vec8 kBGRAToY = { - 0, 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13 -}; - -static vec8 kBGRAToU = { - 0, -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112 -}; - -static vec8 kBGRAToV = { - 0, 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18 -}; - -// Constants for ABGR -static vec8 kABGRToY = { - 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13, 0 -}; - -static vec8 kABGRToU = { - -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112, 0 -}; - -static vec8 kABGRToV = { - 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18, 0 -}; - -// Constants for RGBA. -static vec8 kRGBAToY = { - 0, 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33 -}; - -static vec8 kRGBAToU = { - 0, 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38 -}; - -static vec8 kRGBAToV = { - 0, -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112 -}; - -static uvec8 kAddY16 = { - 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u -}; - -static vec16 kAddYJ64 = { - 64, 64, 64, 64, 64, 64, 64, 64 -}; - -static uvec8 kAddUV128 = { - 128u, 128u, 128u, 128u, 128u, 128u, 128u, 128u, - 128u, 128u, 128u, 128u, 128u, 128u, 128u, 128u -}; - -static uvec16 kAddUVJ128 = { - 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u -}; -#endif // defined(HAS_ARGBTOYROW_SSSE3) || defined(HAS_I422TOARGBROW_SSSE3) - -#ifdef HAS_RGB24TOARGBROW_SSSE3 - -// Shuffle table for converting RGB24 to ARGB. -static uvec8 kShuffleMaskRGB24ToARGB = { - 0u, 1u, 2u, 12u, 3u, 4u, 5u, 13u, 6u, 7u, 8u, 14u, 9u, 10u, 11u, 15u -}; - -// Shuffle table for converting RAW to ARGB. -static uvec8 kShuffleMaskRAWToARGB = { - 2u, 1u, 0u, 12u, 5u, 4u, 3u, 13u, 8u, 7u, 6u, 14u, 11u, 10u, 9u, 15u -}; - -// Shuffle table for converting ARGB to RGB24. -static uvec8 kShuffleMaskARGBToRGB24 = { - 0u, 1u, 2u, 4u, 5u, 6u, 8u, 9u, 10u, 12u, 13u, 14u, 128u, 128u, 128u, 128u -}; - -// Shuffle table for converting ARGB to RAW. -static uvec8 kShuffleMaskARGBToRAW = { - 2u, 1u, 0u, 6u, 5u, 4u, 10u, 9u, 8u, 14u, 13u, 12u, 128u, 128u, 128u, 128u -}; - -// Shuffle table for converting ARGBToRGB24 for I422ToRGB24. First 8 + next 4 -static uvec8 kShuffleMaskARGBToRGB24_0 = { - 0u, 1u, 2u, 4u, 5u, 6u, 8u, 9u, 128u, 128u, 128u, 128u, 10u, 12u, 13u, 14u -}; - -// Shuffle table for converting ARGB to RAW. -static uvec8 kShuffleMaskARGBToRAW_0 = { - 2u, 1u, 0u, 6u, 5u, 4u, 10u, 9u, 128u, 128u, 128u, 128u, 8u, 14u, 13u, 12u -}; -#endif // HAS_RGB24TOARGBROW_SSSE3 - -#if defined(TESTING) && defined(__x86_64__) -void TestRow_SSE2(const uint8* src_y, uint8* dst_argb, int pix) { - asm volatile ( - ".p2align 5 \n" - "mov %%eax,%%eax \n" - "mov %%ebx,%%ebx \n" - "mov %%ecx,%%ecx \n" - "mov %%edx,%%edx \n" - "mov %%esi,%%esi \n" - "mov %%edi,%%edi \n" - "mov %%ebp,%%ebp \n" - "mov %%esp,%%esp \n" - ".p2align 5 \n" - "mov %%r8d,%%r8d \n" - "mov %%r9d,%%r9d \n" - "mov %%r10d,%%r10d \n" - "mov %%r11d,%%r11d \n" - "mov %%r12d,%%r12d \n" - "mov %%r13d,%%r13d \n" - "mov %%r14d,%%r14d \n" - "mov %%r15d,%%r15d \n" - ".p2align 5 \n" - "lea (%%rax),%%eax \n" - "lea (%%rbx),%%ebx \n" - "lea (%%rcx),%%ecx \n" - "lea (%%rdx),%%edx \n" - "lea (%%rsi),%%esi \n" - "lea (%%rdi),%%edi \n" - "lea (%%rbp),%%ebp \n" - "lea (%%rsp),%%esp \n" - ".p2align 5 \n" - "lea (%%r8),%%r8d \n" - "lea (%%r9),%%r9d \n" - "lea (%%r10),%%r10d \n" - "lea (%%r11),%%r11d \n" - "lea (%%r12),%%r12d \n" - "lea (%%r13),%%r13d \n" - "lea (%%r14),%%r14d \n" - "lea (%%r15),%%r15d \n" - - ".p2align 5 \n" - "lea 0x10(%%rax),%%eax \n" - "lea 0x10(%%rbx),%%ebx \n" - "lea 0x10(%%rcx),%%ecx \n" - "lea 0x10(%%rdx),%%edx \n" - "lea 0x10(%%rsi),%%esi \n" - "lea 0x10(%%rdi),%%edi \n" - "lea 0x10(%%rbp),%%ebp \n" - "lea 0x10(%%rsp),%%esp \n" - ".p2align 5 \n" - "lea 0x10(%%r8),%%r8d \n" - "lea 0x10(%%r9),%%r9d \n" - "lea 0x10(%%r10),%%r10d \n" - "lea 0x10(%%r11),%%r11d \n" - "lea 0x10(%%r12),%%r12d \n" - "lea 0x10(%%r13),%%r13d \n" - "lea 0x10(%%r14),%%r14d \n" - "lea 0x10(%%r15),%%r15d \n" - - ".p2align 5 \n" - "add 0x10,%%eax \n" - "add 0x10,%%ebx \n" - "add 0x10,%%ecx \n" - "add 0x10,%%edx \n" - "add 0x10,%%esi \n" - "add 0x10,%%edi \n" - "add 0x10,%%ebp \n" - "add 0x10,%%esp \n" - ".p2align 5 \n" - "add 0x10,%%r8d \n" - "add 0x10,%%r9d \n" - "add 0x10,%%r10d \n" - "add 0x10,%%r11d \n" - "add 0x10,%%r12d \n" - "add 0x10,%%r13d \n" - "add 0x10,%%r14d \n" - "add 0x10,%%r15d \n" - - ".p2align 2 \n" - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // TESTING - -#ifdef HAS_I400TOARGBROW_SSE2 -void I400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "pslld $0x18,%%xmm5 \n" - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm0,%%xmm0 \n" - "punpckhwd %%xmm1,%%xmm1 \n" - "por %%xmm5,%%xmm0 \n" - "por %%xmm5,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void I400ToARGBRow_Unaligned_SSE2(const uint8* src_y, uint8* dst_argb, - int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "pslld $0x18,%%xmm5 \n" - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm0,%%xmm0 \n" - "punpckhwd %%xmm1,%%xmm1 \n" - "por %%xmm5,%%xmm0 \n" - "por %%xmm5,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "movdqu %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src_y), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_I400TOARGBROW_SSE2 - -#ifdef HAS_RGB24TOARGBROW_SSSE3 -void RGB24ToARGBRow_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" // generate mask 0xff000000 - "pslld $0x18,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm3 \n" - "lea " MEMLEA(0x30,0) ",%0 \n" - "movdqa %%xmm3,%%xmm2 \n" - "palignr $0x8,%%xmm1,%%xmm2 \n" - "pshufb %%xmm4,%%xmm2 \n" - "por %%xmm5,%%xmm2 \n" - "palignr $0xc,%%xmm0,%%xmm1 \n" - "pshufb %%xmm4,%%xmm0 \n" - "movdqa %%xmm2," MEMACCESS2(0x20,1) " \n" - "por %%xmm5,%%xmm0 \n" - "pshufb %%xmm4,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "por %%xmm5,%%xmm1 \n" - "palignr $0x4,%%xmm3,%%xmm3 \n" - "pshufb %%xmm4,%%xmm3 \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "por %%xmm5,%%xmm3 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm3," MEMACCESS2(0x30,1) " \n" - "lea " MEMLEA(0x40,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_rgb24), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "m"(kShuffleMaskRGB24ToARGB) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void RAWToARGBRow_SSSE3(const uint8* src_raw, uint8* dst_argb, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" // generate mask 0xff000000 - "pslld $0x18,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm3 \n" - "lea " MEMLEA(0x30,0) ",%0 \n" - "movdqa %%xmm3,%%xmm2 \n" - "palignr $0x8,%%xmm1,%%xmm2 \n" - "pshufb %%xmm4,%%xmm2 \n" - "por %%xmm5,%%xmm2 \n" - "palignr $0xc,%%xmm0,%%xmm1 \n" - "pshufb %%xmm4,%%xmm0 \n" - "movdqa %%xmm2," MEMACCESS2(0x20,1) " \n" - "por %%xmm5,%%xmm0 \n" - "pshufb %%xmm4,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "por %%xmm5,%%xmm1 \n" - "palignr $0x4,%%xmm3,%%xmm3 \n" - "pshufb %%xmm4,%%xmm3 \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "por %%xmm5,%%xmm3 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm3," MEMACCESS2(0x30,1) " \n" - "lea " MEMLEA(0x40,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_raw), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "m"(kShuffleMaskRAWToARGB) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void RGB565ToARGBRow_SSE2(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "mov $0x1080108,%%eax \n" - "movd %%eax,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - "mov $0x20802080,%%eax \n" - "movd %%eax,%%xmm6 \n" - "pshufd $0x0,%%xmm6,%%xmm6 \n" - "pcmpeqb %%xmm3,%%xmm3 \n" - "psllw $0xb,%%xmm3 \n" - "pcmpeqb %%xmm4,%%xmm4 \n" - "psllw $0xa,%%xmm4 \n" - "psrlw $0x5,%%xmm4 \n" - "pcmpeqb %%xmm7,%%xmm7 \n" - "psllw $0x8,%%xmm7 \n" - "sub %0,%1 \n" - "sub %0,%1 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "pand %%xmm3,%%xmm1 \n" - "psllw $0xb,%%xmm2 \n" - "pmulhuw %%xmm5,%%xmm1 \n" - "pmulhuw %%xmm5,%%xmm2 \n" - "psllw $0x8,%%xmm1 \n" - "por %%xmm2,%%xmm1 \n" - "pand %%xmm4,%%xmm0 \n" - "pmulhuw %%xmm6,%%xmm0 \n" - "por %%xmm7,%%xmm0 \n" - "movdqa %%xmm1,%%xmm2 \n" - "punpcklbw %%xmm0,%%xmm1 \n" - "punpckhbw %%xmm0,%%xmm2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm1,0x00,1,0,2) // movdqa %%xmm1,(%1,%0,2) - MEMOPMEM(movdqa,xmm2,0x10,1,0,2) // movdqa %%xmm2,0x10(%1,%0,2) - "lea " MEMLEA(0x10,0) ",%0 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : - : "memory", "cc", "eax" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} - -void ARGB1555ToARGBRow_SSE2(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "mov $0x1080108,%%eax \n" - "movd %%eax,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - "mov $0x42004200,%%eax \n" - "movd %%eax,%%xmm6 \n" - "pshufd $0x0,%%xmm6,%%xmm6 \n" - "pcmpeqb %%xmm3,%%xmm3 \n" - "psllw $0xb,%%xmm3 \n" - "movdqa %%xmm3,%%xmm4 \n" - "psrlw $0x6,%%xmm4 \n" - "pcmpeqb %%xmm7,%%xmm7 \n" - "psllw $0x8,%%xmm7 \n" - "sub %0,%1 \n" - "sub %0,%1 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psllw $0x1,%%xmm1 \n" - "psllw $0xb,%%xmm2 \n" - "pand %%xmm3,%%xmm1 \n" - "pmulhuw %%xmm5,%%xmm2 \n" - "pmulhuw %%xmm5,%%xmm1 \n" - "psllw $0x8,%%xmm1 \n" - "por %%xmm2,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "pand %%xmm4,%%xmm0 \n" - "psraw $0x8,%%xmm2 \n" - "pmulhuw %%xmm6,%%xmm0 \n" - "pand %%xmm7,%%xmm2 \n" - "por %%xmm2,%%xmm0 \n" - "movdqa %%xmm1,%%xmm2 \n" - "punpcklbw %%xmm0,%%xmm1 \n" - "punpckhbw %%xmm0,%%xmm2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm1,0x00,1,0,2) // movdqa %%xmm1,(%1,%0,2) - MEMOPMEM(movdqa,xmm2,0x10,1,0,2) // movdqa %%xmm2,0x10(%1,%0,2) - "lea " MEMLEA(0x10,0) ",%0 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : - : "memory", "cc", "eax" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} - -void ARGB4444ToARGBRow_SSE2(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "mov $0xf0f0f0f,%%eax \n" - "movd %%eax,%%xmm4 \n" - "pshufd $0x0,%%xmm4,%%xmm4 \n" - "movdqa %%xmm4,%%xmm5 \n" - "pslld $0x4,%%xmm5 \n" - "sub %0,%1 \n" - "sub %0,%1 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "pand %%xmm4,%%xmm0 \n" - "pand %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm3 \n" - "psllw $0x4,%%xmm1 \n" - "psrlw $0x4,%%xmm3 \n" - "por %%xmm1,%%xmm0 \n" - "por %%xmm3,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm2,%%xmm0 \n" - "punpckhbw %%xmm2,%%xmm1 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,2) // movdqa %%xmm0,(%1,%0,2) - MEMOPMEM(movdqa,xmm1,0x10,1,0,2) // movdqa %%xmm1,0x10(%1,%0,2) - "lea " MEMLEA(0x10,0) ",%0 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : - : "memory", "cc", "eax" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ARGBToRGB24Row_SSSE3(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "movdqa %3,%%xmm6 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "pshufb %%xmm6,%%xmm0 \n" - "pshufb %%xmm6,%%xmm1 \n" - "pshufb %%xmm6,%%xmm2 \n" - "pshufb %%xmm6,%%xmm3 \n" - "movdqa %%xmm1,%%xmm4 \n" - "psrldq $0x4,%%xmm1 \n" - "pslldq $0xc,%%xmm4 \n" - "movdqa %%xmm2,%%xmm5 \n" - "por %%xmm4,%%xmm0 \n" - "pslldq $0x8,%%xmm5 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "por %%xmm5,%%xmm1 \n" - "psrldq $0x8,%%xmm2 \n" - "pslldq $0x4,%%xmm3 \n" - "por %%xmm3,%%xmm2 \n" - "movdqu %%xmm1," MEMACCESS2(0x10,1) " \n" - "movdqu %%xmm2," MEMACCESS2(0x20,1) " \n" - "lea " MEMLEA(0x30,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : "m"(kShuffleMaskARGBToRGB24) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -void ARGBToRAWRow_SSSE3(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "movdqa %3,%%xmm6 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "pshufb %%xmm6,%%xmm0 \n" - "pshufb %%xmm6,%%xmm1 \n" - "pshufb %%xmm6,%%xmm2 \n" - "pshufb %%xmm6,%%xmm3 \n" - "movdqa %%xmm1,%%xmm4 \n" - "psrldq $0x4,%%xmm1 \n" - "pslldq $0xc,%%xmm4 \n" - "movdqa %%xmm2,%%xmm5 \n" - "por %%xmm4,%%xmm0 \n" - "pslldq $0x8,%%xmm5 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "por %%xmm5,%%xmm1 \n" - "psrldq $0x8,%%xmm2 \n" - "pslldq $0x4,%%xmm3 \n" - "por %%xmm3,%%xmm2 \n" - "movdqu %%xmm1," MEMACCESS2(0x10,1) " \n" - "movdqu %%xmm2," MEMACCESS2(0x20,1) " \n" - "lea " MEMLEA(0x30,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : "m"(kShuffleMaskARGBToRAW) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -void ARGBToRGB565Row_SSE2(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "pcmpeqb %%xmm3,%%xmm3 \n" - "psrld $0x1b,%%xmm3 \n" - "pcmpeqb %%xmm4,%%xmm4 \n" - "psrld $0x1a,%%xmm4 \n" - "pslld $0x5,%%xmm4 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pslld $0xb,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "pslld $0x8,%%xmm0 \n" - "psrld $0x3,%%xmm1 \n" - "psrld $0x5,%%xmm2 \n" - "psrad $0x10,%%xmm0 \n" - "pand %%xmm3,%%xmm1 \n" - "pand %%xmm4,%%xmm2 \n" - "pand %%xmm5,%%xmm0 \n" - "por %%xmm2,%%xmm1 \n" - "por %%xmm1,%%xmm0 \n" - "packssdw %%xmm0,%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x4,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ARGBToARGB1555Row_SSE2(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "pcmpeqb %%xmm4,%%xmm4 \n" - "psrld $0x1b,%%xmm4 \n" - "movdqa %%xmm4,%%xmm5 \n" - "pslld $0x5,%%xmm5 \n" - "movdqa %%xmm4,%%xmm6 \n" - "pslld $0xa,%%xmm6 \n" - "pcmpeqb %%xmm7,%%xmm7 \n" - "pslld $0xf,%%xmm7 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "movdqa %%xmm0,%%xmm3 \n" - "psrad $0x10,%%xmm0 \n" - "psrld $0x3,%%xmm1 \n" - "psrld $0x6,%%xmm2 \n" - "psrld $0x9,%%xmm3 \n" - "pand %%xmm7,%%xmm0 \n" - "pand %%xmm4,%%xmm1 \n" - "pand %%xmm5,%%xmm2 \n" - "pand %%xmm6,%%xmm3 \n" - "por %%xmm1,%%xmm0 \n" - "por %%xmm3,%%xmm2 \n" - "por %%xmm2,%%xmm0 \n" - "packssdw %%xmm0,%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMACCESS2(0x8,1) ",%1 \n" - "sub $0x4,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} - -void ARGBToARGB4444Row_SSE2(const uint8* src, uint8* dst, int pix) { - asm volatile ( - "pcmpeqb %%xmm4,%%xmm4 \n" - "psllw $0xc,%%xmm4 \n" - "movdqa %%xmm4,%%xmm3 \n" - "psrlw $0x8,%%xmm3 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm3,%%xmm0 \n" - "pand %%xmm4,%%xmm1 \n" - "psrlq $0x4,%%xmm0 \n" - "psrlq $0x8,%%xmm1 \n" - "por %%xmm1,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x4,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4" -#endif - ); -} -#endif // HAS_RGB24TOARGBROW_SSSE3 - -#ifdef HAS_ARGBTOYROW_SSSE3 -void ARGBToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kARGBToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ARGBToYRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kARGBToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBTOYROW_SSSE3 - -#ifdef HAS_ARGBTOYJROW_SSSE3 -void ARGBToYJRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %3,%%xmm4 \n" - "movdqa %4,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "paddw %%xmm5,%%xmm0 \n" - "paddw %%xmm5,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kARGBToYJ), // %3 - "m"(kAddYJ64) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ARGBToYJRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %3,%%xmm4 \n" - "movdqa %4,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "paddw %%xmm5,%%xmm0 \n" - "paddw %%xmm5,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kARGBToYJ), // %3 - "m"(kAddYJ64) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBTOYJROW_SSSE3 - -#ifdef HAS_ARGBTOUVROW_SSSE3 -// TODO(fbarchard): pass xmm constants to single block of assembly. -// fpic on GCC 4.2 for OSX runs out of GPR registers. "m" effectively takes -// 3 registers - ebx, ebp and eax. "m" can be passed with 3 normal registers, -// or 4 if stack frame is disabled. Doing 2 assembly blocks is a work around -// and considered unsafe. -void ARGBToUVRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToU), // %0 - "m"(kARGBToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(pavgb,0x00,0,4,1,xmm0) // pavgb (%0,%4,1),%%xmm0 - MEMOPREG(pavgb,0x10,0,4,1,xmm1) // pavgb 0x10(%0,%4,1),%%xmm1 - MEMOPREG(pavgb,0x20,0,4,1,xmm2) // pavgb 0x20(%0,%4,1),%%xmm2 - MEMOPREG(pavgb,0x30,0,4,1,xmm6) // pavgb 0x30(%0,%4,1),%%xmm6 - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_argb)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -// TODO(fbarchard): Share code with ARGBToUVRow_SSSE3. -void ARGBToUVJRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToUJ), // %0 - "m"(kARGBToVJ), // %1 - "m"(kAddUVJ128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(pavgb,0x00,0,4,1,xmm0) // pavgb (%0,%4,1),%%xmm0 - MEMOPREG(pavgb,0x10,0,4,1,xmm1) // pavgb 0x10(%0,%4,1),%%xmm1 - MEMOPREG(pavgb,0x20,0,4,1,xmm2) // pavgb 0x20(%0,%4,1),%%xmm2 - MEMOPREG(pavgb,0x30,0,4,1,xmm6) // pavgb 0x30(%0,%4,1),%%xmm6 - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "paddw %%xmm5,%%xmm0 \n" - "paddw %%xmm5,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_argb)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void ARGBToUVRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToU), // %0 - "m"(kARGBToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm7) // movdqu (%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm0 \n" - MEMOPREG(movdqu,0x10,0,4,1,xmm7) // movdqu 0x10(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm1 \n" - MEMOPREG(movdqu,0x20,0,4,1,xmm7) // movdqu 0x20(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm2 \n" - MEMOPREG(movdqu,0x30,0,4,1,xmm7) // movdqu 0x30(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_argb)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void ARGBToUVJRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToUJ), // %0 - "m"(kARGBToVJ), // %1 - "m"(kAddUVJ128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm7) // movdqu (%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm0 \n" - MEMOPREG(movdqu,0x10,0,4,1,xmm7) // movdqu 0x10(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm1 \n" - MEMOPREG(movdqu,0x20,0,4,1,xmm7) // movdqu 0x20(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm2 \n" - MEMOPREG(movdqu,0x30,0,4,1,xmm7) // movdqu 0x30(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "paddw %%xmm5,%%xmm0 \n" - "paddw %%xmm5,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_argb)) - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void ARGBToUV444Row_SSSE3(const uint8* src_argb, uint8* dst_u, uint8* dst_v, - int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToU), // %0 - "m"(kARGBToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm6 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm6,%%xmm2 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm2 \n" - "packsswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - "pmaddubsw %%xmm3,%%xmm0 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm6,%%xmm2 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm2 \n" - "packsswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,2,1) // movdqa %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6" -#endif - ); -} - -void ARGBToUV444Row_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_u, - uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToU), // %0 - "m"(kARGBToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm6 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm6,%%xmm2 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm2 \n" - "packsswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - "pmaddubsw %%xmm3,%%xmm0 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm6,%%xmm2 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm2 \n" - "packsswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,2,1) // movdqu %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6" -#endif - ); -} - -void ARGBToUV422Row_SSSE3(const uint8* src_argb0, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToU), // %0 - "m"(kARGBToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void ARGBToUV422Row_Unaligned_SSSE3(const uint8* src_argb0, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kARGBToU), // %0 - "m"(kARGBToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void BGRAToYRow_SSSE3(const uint8* src_bgra, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_bgra), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kBGRAToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void BGRAToYRow_Unaligned_SSSE3(const uint8* src_bgra, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_bgra), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kBGRAToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void BGRAToUVRow_SSSE3(const uint8* src_bgra0, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kBGRAToU), // %0 - "m"(kBGRAToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(pavgb,0x00,0,4,1,xmm0) // pavgb (%0,%4,1),%%xmm0 - MEMOPREG(pavgb,0x10,0,4,1,xmm1) // pavgb 0x10(%0,%4,1),%%xmm1 - MEMOPREG(pavgb,0x20,0,4,1,xmm2) // pavgb 0x20(%0,%4,1),%%xmm2 - MEMOPREG(pavgb,0x30,0,4,1,xmm6) // pavgb 0x30(%0,%4,1),%%xmm6 - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_bgra0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_bgra)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void BGRAToUVRow_Unaligned_SSSE3(const uint8* src_bgra0, int src_stride_bgra, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kBGRAToU), // %0 - "m"(kBGRAToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm7) // movdqu (%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm0 \n" - MEMOPREG(movdqu,0x10,0,4,1,xmm7) // movdqu 0x10(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm1 \n" - MEMOPREG(movdqu,0x20,0,4,1,xmm7) // movdqu 0x20(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm2 \n" - MEMOPREG(movdqu,0x30,0,4,1,xmm7) // movdqu 0x30(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_bgra0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_bgra)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void ABGRToYRow_SSSE3(const uint8* src_abgr, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_abgr), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kABGRToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ABGRToYRow_Unaligned_SSSE3(const uint8* src_abgr, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_abgr), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kABGRToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void RGBAToYRow_SSSE3(const uint8* src_rgba, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_rgba), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kRGBAToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void RGBAToYRow_Unaligned_SSSE3(const uint8* src_rgba, uint8* dst_y, int pix) { - asm volatile ( - "movdqa %4,%%xmm5 \n" - "movdqa %3,%%xmm4 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm3 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm4,%%xmm3 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "phaddw %%xmm1,%%xmm0 \n" - "phaddw %%xmm3,%%xmm2 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_rgba), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : "m"(kRGBAToY), // %3 - "m"(kAddY16) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ABGRToUVRow_SSSE3(const uint8* src_abgr0, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kABGRToU), // %0 - "m"(kABGRToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(pavgb,0x00,0,4,1,xmm0) // pavgb (%0,%4,1),%%xmm0 - MEMOPREG(pavgb,0x10,0,4,1,xmm1) // pavgb 0x10(%0,%4,1),%%xmm1 - MEMOPREG(pavgb,0x20,0,4,1,xmm2) // pavgb 0x20(%0,%4,1),%%xmm2 - MEMOPREG(pavgb,0x30,0,4,1,xmm6) // pavgb 0x30(%0,%4,1),%%xmm6 - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_abgr0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_abgr)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void ABGRToUVRow_Unaligned_SSSE3(const uint8* src_abgr0, int src_stride_abgr, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kABGRToU), // %0 - "m"(kABGRToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm7) // movdqu (%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm0 \n" - MEMOPREG(movdqu,0x10,0,4,1,xmm7) // movdqu 0x10(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm1 \n" - MEMOPREG(movdqu,0x20,0,4,1,xmm7) // movdqu 0x20(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm2 \n" - MEMOPREG(movdqu,0x30,0,4,1,xmm7) // movdqu 0x30(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_abgr0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_abgr)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void RGBAToUVRow_SSSE3(const uint8* src_rgba0, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kRGBAToU), // %0 - "m"(kRGBAToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(pavgb,0x00,0,4,1,xmm0) // pavgb (%0,%4,1),%%xmm0 - MEMOPREG(pavgb,0x10,0,4,1,xmm1) // pavgb 0x10(%0,%4,1),%%xmm1 - MEMOPREG(pavgb,0x20,0,4,1,xmm2) // pavgb 0x20(%0,%4,1),%%xmm2 - MEMOPREG(pavgb,0x30,0,4,1,xmm6) // pavgb 0x30(%0,%4,1),%%xmm6 - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_rgba0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_rgba)) - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} - -void RGBAToUVRow_Unaligned_SSSE3(const uint8* src_rgba0, int src_stride_rgba, - uint8* dst_u, uint8* dst_v, int width) { - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kRGBAToU), // %0 - "m"(kRGBAToV), // %1 - "m"(kAddUV128) // %2 - ); - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqu " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqu " MEMACCESS2(0x30,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm7) // movdqu (%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm0 \n" - MEMOPREG(movdqu,0x10,0,4,1,xmm7) // movdqu 0x10(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm1 \n" - MEMOPREG(movdqu,0x20,0,4,1,xmm7) // movdqu 0x20(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm2 \n" - MEMOPREG(movdqu,0x30,0,4,1,xmm7) // movdqu 0x30(%0,%4,1),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "movdqa %%xmm0,%%xmm7 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm7 \n" - "pavgb %%xmm7,%%xmm0 \n" - "movdqa %%xmm2,%%xmm7 \n" - "shufps $0x88,%%xmm6,%%xmm2 \n" - "shufps $0xdd,%%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm2 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "phaddw %%xmm2,%%xmm0 \n" - "phaddw %%xmm6,%%xmm1 \n" - "psraw $0x8,%%xmm0 \n" - "psraw $0x8,%%xmm1 \n" - "packsswb %%xmm1,%%xmm0 \n" - "paddb %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movlps %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhps,xmm0,0x00,1,2,1) // movhps %%xmm0,(%1,%2,1) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_rgba0), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+rm"(width) // %3 - : "r"((intptr_t)(src_stride_rgba)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBTOUVROW_SSSE3 - -#ifdef HAS_I422TOARGBROW_SSSE3 -#define UB 127 /* min(63,(int8)(2.018 * 64)) */ -#define UG -25 /* (int8)(-0.391 * 64 - 0.5) */ -#define UR 0 - -#define VB 0 -#define VG -52 /* (int8)(-0.813 * 64 - 0.5) */ -#define VR 102 /* (int8)(1.596 * 64 + 0.5) */ - -// Bias -#define BB UB * 128 + VB * 128 -#define BG UG * 128 + VG * 128 -#define BR UR * 128 + VR * 128 - -#define YG 74 /* (int8)(1.164 * 64 + 0.5) */ - -struct { - vec8 kUVToB; // 0 - vec8 kUVToG; // 16 - vec8 kUVToR; // 32 - vec16 kUVBiasB; // 48 - vec16 kUVBiasG; // 64 - vec16 kUVBiasR; // 80 - vec16 kYSub16; // 96 - vec16 kYToRgb; // 112 - vec8 kVUToB; // 128 - vec8 kVUToG; // 144 - vec8 kVUToR; // 160 -} static SIMD_ALIGNED(kYuvConstants) = { - { UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB }, - { UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG }, - { UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR }, - { BB, BB, BB, BB, BB, BB, BB, BB }, - { BG, BG, BG, BG, BG, BG, BG, BG }, - { BR, BR, BR, BR, BR, BR, BR, BR }, - { 16, 16, 16, 16, 16, 16, 16, 16 }, - { YG, YG, YG, YG, YG, YG, YG, YG }, - { VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB }, - { VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG }, - { VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR } -}; - - -// Read 8 UV from 411 -#define READYUV444 \ - "movq " MEMACCESS([u_buf]) ",%%xmm0 \n" \ - BUNDLEALIGN \ - MEMOPREG(movq, 0x00, [u_buf], [v_buf], 1, xmm1) \ - "lea " MEMLEA(0x8, [u_buf]) ",%[u_buf] \n" \ - "punpcklbw %%xmm1,%%xmm0 \n" - -// Read 4 UV from 422, upsample to 8 UV -#define READYUV422 \ - "movd " MEMACCESS([u_buf]) ",%%xmm0 \n" \ - BUNDLEALIGN \ - MEMOPREG(movd, 0x00, [u_buf], [v_buf], 1, xmm1) \ - "lea " MEMLEA(0x4, [u_buf]) ",%[u_buf] \n" \ - "punpcklbw %%xmm1,%%xmm0 \n" \ - "punpcklwd %%xmm0,%%xmm0 \n" - -// Read 2 UV from 411, upsample to 8 UV -#define READYUV411 \ - "movd " MEMACCESS([u_buf]) ",%%xmm0 \n" \ - BUNDLEALIGN \ - MEMOPREG(movd, 0x00, [u_buf], [v_buf], 1, xmm1) \ - "lea " MEMLEA(0x2, [u_buf]) ",%[u_buf] \n" \ - "punpcklbw %%xmm1,%%xmm0 \n" \ - "punpcklwd %%xmm0,%%xmm0 \n" \ - "punpckldq %%xmm0,%%xmm0 \n" - -// Read 4 UV from NV12, upsample to 8 UV -#define READNV12 \ - "movq " MEMACCESS([uv_buf]) ",%%xmm0 \n" \ - "lea " MEMLEA(0x8, [uv_buf]) ",%[uv_buf] \n" \ - "punpcklwd %%xmm0,%%xmm0 \n" - -// Convert 8 pixels: 8 UV and 8 Y -#define YUVTORGB \ - "movdqa %%xmm0,%%xmm1 \n" \ - "movdqa %%xmm0,%%xmm2 \n" \ - "pmaddubsw " MEMACCESS([kYuvConstants]) ",%%xmm0 \n" \ - "pmaddubsw " MEMACCESS2(16, [kYuvConstants]) ",%%xmm1 \n" \ - "pmaddubsw " MEMACCESS2(32, [kYuvConstants]) ",%%xmm2 \n" \ - "psubw " MEMACCESS2(48, [kYuvConstants]) ",%%xmm0 \n" \ - "psubw " MEMACCESS2(64, [kYuvConstants]) ",%%xmm1 \n" \ - "psubw " MEMACCESS2(80, [kYuvConstants]) ",%%xmm2 \n" \ - "movq " MEMACCESS([y_buf]) ",%%xmm3 \n" \ - "lea " MEMLEA(0x8, [y_buf]) ",%[y_buf] \n" \ - "punpcklbw %%xmm4,%%xmm3 \n" \ - "psubsw " MEMACCESS2(96, [kYuvConstants]) ",%%xmm3 \n" \ - "pmullw " MEMACCESS2(112, [kYuvConstants]) ",%%xmm3 \n" \ - "paddsw %%xmm3,%%xmm0 \n" \ - "paddsw %%xmm3,%%xmm1 \n" \ - "paddsw %%xmm3,%%xmm2 \n" \ - "psraw $0x6,%%xmm0 \n" \ - "psraw $0x6,%%xmm1 \n" \ - "psraw $0x6,%%xmm2 \n" \ - "packuswb %%xmm0,%%xmm0 \n" \ - "packuswb %%xmm1,%%xmm1 \n" \ - "packuswb %%xmm2,%%xmm2 \n" - -// Convert 8 pixels: 8 VU and 8 Y -#define YVUTORGB \ - "movdqa %%xmm0,%%xmm1 \n" \ - "movdqa %%xmm0,%%xmm2 \n" \ - "pmaddubsw " MEMACCESS2(128, [kYuvConstants]) ",%%xmm0 \n" \ - "pmaddubsw " MEMACCESS2(144, [kYuvConstants]) ",%%xmm1 \n" \ - "pmaddubsw " MEMACCESS2(160, [kYuvConstants]) ",%%xmm2 \n" \ - "psubw " MEMACCESS2(48, [kYuvConstants]) ",%%xmm0 \n" \ - "psubw " MEMACCESS2(64, [kYuvConstants]) ",%%xmm1 \n" \ - "psubw " MEMACCESS2(80, [kYuvConstants]) ",%%xmm2 \n" \ - "movq " MEMACCESS([y_buf]) ",%%xmm3 \n" \ - "lea " MEMLEA(0x8, [y_buf]) ",%[y_buf] \n" \ - "punpcklbw %%xmm4,%%xmm3 \n" \ - "psubsw " MEMACCESS2(96, [kYuvConstants]) ",%%xmm3 \n" \ - "pmullw " MEMACCESS2(112, [kYuvConstants]) ",%%xmm3 \n" \ - "paddsw %%xmm3,%%xmm0 \n" \ - "paddsw %%xmm3,%%xmm1 \n" \ - "paddsw %%xmm3,%%xmm2 \n" \ - "psraw $0x6,%%xmm0 \n" \ - "psraw $0x6,%%xmm1 \n" \ - "psraw $0x6,%%xmm2 \n" \ - "packuswb %%xmm0,%%xmm0 \n" \ - "packuswb %%xmm1,%%xmm1 \n" \ - "packuswb %%xmm2,%%xmm2 \n" - -void OMITFP I444ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV444 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS([dst_argb]) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,[dst_argb]) " \n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToRGB24Row_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_rgb24, - int width) { -// fpic 32 bit gcc 4.2 on OSX runs out of GPR regs. -#if defined(__i386__) - asm volatile ( - "movdqa %[kShuffleMaskARGBToRGB24_0],%%xmm5 \n" - "movdqa %[kShuffleMaskARGBToRGB24],%%xmm6 \n" - :: [kShuffleMaskARGBToRGB24_0]"m"(kShuffleMaskARGBToRGB24_0), - [kShuffleMaskARGBToRGB24]"m"(kShuffleMaskARGBToRGB24)); -#endif - - asm volatile ( -#if !defined(__i386__) - "movdqa %[kShuffleMaskARGBToRGB24_0],%%xmm5 \n" - "movdqa %[kShuffleMaskARGBToRGB24],%%xmm6 \n" -#endif - "sub %[u_buf],%[v_buf] \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm2,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "pshufb %%xmm5,%%xmm0 \n" - "pshufb %%xmm6,%%xmm1 \n" - "palignr $0xc,%%xmm0,%%xmm1 \n" - "movq %%xmm0," MEMACCESS([dst_rgb24]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x8,[dst_rgb24]) "\n" - "lea " MEMLEA(0x18,[dst_rgb24]) ",%[dst_rgb24] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_rgb24]"+r"(dst_rgb24), // %[dst_rgb24] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) -#if !defined(__i386__) - , [kShuffleMaskARGBToRGB24_0]"m"(kShuffleMaskARGBToRGB24_0), - [kShuffleMaskARGBToRGB24]"m"(kShuffleMaskARGBToRGB24) -#endif - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -void OMITFP I422ToRAWRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_raw, - int width) { -// fpic 32 bit gcc 4.2 on OSX runs out of GPR regs. -#if defined(__i386__) - asm volatile ( - "movdqa %[kShuffleMaskARGBToRAW_0],%%xmm5 \n" - "movdqa %[kShuffleMaskARGBToRAW],%%xmm6 \n" - :: [kShuffleMaskARGBToRAW_0]"m"(kShuffleMaskARGBToRAW_0), - [kShuffleMaskARGBToRAW]"m"(kShuffleMaskARGBToRAW)); -#endif - - asm volatile ( -#if !defined(__i386__) - "movdqa %[kShuffleMaskARGBToRAW_0],%%xmm5 \n" - "movdqa %[kShuffleMaskARGBToRAW],%%xmm6 \n" -#endif - "sub %[u_buf],%[v_buf] \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm2,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "pshufb %%xmm5,%%xmm0 \n" - "pshufb %%xmm6,%%xmm1 \n" - "palignr $0xc,%%xmm0,%%xmm1 \n" - "movq %%xmm0," MEMACCESS([dst_raw]) " \n" - "movdqu %%xmm1," MEMACCESS2(0x8,[dst_raw]) "\n" - "lea " MEMLEA(0x18,[dst_raw]) ",%[dst_raw] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_raw]"+r"(dst_raw), // %[dst_raw] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) -#if !defined(__i386__) - , [kShuffleMaskARGBToRAW_0]"m"(kShuffleMaskARGBToRAW_0), - [kShuffleMaskARGBToRAW]"m"(kShuffleMaskARGBToRAW) -#endif - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -void OMITFP I422ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqa %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I411ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV411 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqa %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP NV12ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READNV12 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqa %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [uv_buf]"+r"(uv_buf), // %[uv_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" - // Does not use r14. -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP NV21ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READNV12 - YVUTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqa %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [uv_buf]"+r"(uv_buf), // %[uv_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" - // Does not use r14. -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I444ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV444 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I411ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV411 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP NV12ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READNV12 - YUVTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [uv_buf]"+r"(uv_buf), // %[uv_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" - // Does not use r14. -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP NV21ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READNV12 - YVUTORGB - "punpcklbw %%xmm1,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm0 \n" - "punpckhwd %%xmm2,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS([dst_argb]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x10,[dst_argb]) "\n" - "lea " MEMLEA(0x20,[dst_argb]) ",%[dst_argb] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [uv_buf]"+r"(uv_buf), // %[uv_buf] - [dst_argb]"+r"(dst_argb), // %[dst_argb] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" - // Does not use r14. -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToBGRARow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_bgra, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "pcmpeqb %%xmm5,%%xmm5 \n" - "punpcklbw %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm2,%%xmm5 \n" - "movdqa %%xmm5,%%xmm0 \n" - "punpcklwd %%xmm1,%%xmm5 \n" - "punpckhwd %%xmm1,%%xmm0 \n" - "movdqa %%xmm5," MEMACCESS([dst_bgra]) "\n" - "movdqa %%xmm0," MEMACCESS2(0x10,[dst_bgra]) "\n" - "lea " MEMLEA(0x20,[dst_bgra]) ",%[dst_bgra] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_bgra]"+r"(dst_bgra), // %[dst_bgra] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToABGRRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_abgr, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "punpcklbw %%xmm1,%%xmm2 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "movdqa %%xmm2,%%xmm1 \n" - "punpcklwd %%xmm0,%%xmm2 \n" - "punpckhwd %%xmm0,%%xmm1 \n" - "movdqa %%xmm2," MEMACCESS([dst_abgr]) "\n" - "movdqa %%xmm1," MEMACCESS2(0x10,[dst_abgr]) "\n" - "lea " MEMLEA(0x20,[dst_abgr]) ",%[dst_abgr] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_abgr]"+r"(dst_abgr), // %[dst_abgr] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToRGBARow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_rgba, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "pcmpeqb %%xmm5,%%xmm5 \n" - "punpcklbw %%xmm2,%%xmm1 \n" - "punpcklbw %%xmm0,%%xmm5 \n" - "movdqa %%xmm5,%%xmm0 \n" - "punpcklwd %%xmm1,%%xmm5 \n" - "punpckhwd %%xmm1,%%xmm0 \n" - "movdqa %%xmm5," MEMACCESS([dst_rgba]) "\n" - "movdqa %%xmm0," MEMACCESS2(0x10,[dst_rgba]) "\n" - "lea " MEMLEA(0x20,[dst_rgba]) ",%[dst_rgba] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_rgba]"+r"(dst_rgba), // %[dst_rgba] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToBGRARow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_bgra, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "pcmpeqb %%xmm5,%%xmm5 \n" - "punpcklbw %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm2,%%xmm5 \n" - "movdqa %%xmm5,%%xmm0 \n" - "punpcklwd %%xmm1,%%xmm5 \n" - "punpckhwd %%xmm1,%%xmm0 \n" - "movdqu %%xmm5," MEMACCESS([dst_bgra]) "\n" - "movdqu %%xmm0," MEMACCESS2(0x10,[dst_bgra]) "\n" - "lea " MEMLEA(0x20,[dst_bgra]) ",%[dst_bgra] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_bgra]"+r"(dst_bgra), // %[dst_bgra] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToABGRRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_abgr, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "punpcklbw %%xmm1,%%xmm2 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "movdqa %%xmm2,%%xmm1 \n" - "punpcklwd %%xmm0,%%xmm2 \n" - "punpckhwd %%xmm0,%%xmm1 \n" - "movdqu %%xmm2," MEMACCESS([dst_abgr]) "\n" - "movdqu %%xmm1," MEMACCESS2(0x10,[dst_abgr]) "\n" - "lea " MEMLEA(0x20,[dst_abgr]) ",%[dst_abgr] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_abgr]"+r"(dst_abgr), // %[dst_abgr] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void OMITFP I422ToRGBARow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_rgba, - int width) { - asm volatile ( - "sub %[u_buf],%[v_buf] \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - LABELALIGN - "1: \n" - READYUV422 - YUVTORGB - "pcmpeqb %%xmm5,%%xmm5 \n" - "punpcklbw %%xmm2,%%xmm1 \n" - "punpcklbw %%xmm0,%%xmm5 \n" - "movdqa %%xmm5,%%xmm0 \n" - "punpcklwd %%xmm1,%%xmm5 \n" - "punpckhwd %%xmm1,%%xmm0 \n" - "movdqu %%xmm5," MEMACCESS([dst_rgba]) "\n" - "movdqu %%xmm0," MEMACCESS2(0x10,[dst_rgba]) "\n" - "lea " MEMLEA(0x20,[dst_rgba]) ",%[dst_rgba] \n" - "sub $0x8,%[width] \n" - "jg 1b \n" - : [y_buf]"+r"(y_buf), // %[y_buf] - [u_buf]"+r"(u_buf), // %[u_buf] - [v_buf]"+r"(v_buf), // %[v_buf] - [dst_rgba]"+r"(dst_rgba), // %[dst_rgba] - [width]"+rm"(width) // %[width] - : [kYuvConstants]"r"(&kYuvConstants.kUVToB) // %[kYuvConstants] - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -#endif // HAS_I422TOARGBROW_SSSE3 - -#ifdef HAS_YTOARGBROW_SSE2 -void YToARGBRow_SSE2(const uint8* y_buf, - uint8* dst_argb, - int width) { - asm volatile ( - "pxor %%xmm5,%%xmm5 \n" - "pcmpeqb %%xmm4,%%xmm4 \n" - "pslld $0x18,%%xmm4 \n" - "mov $0x00100010,%%eax \n" - "movd %%eax,%%xmm3 \n" - "pshufd $0x0,%%xmm3,%%xmm3 \n" - "mov $0x004a004a,%%eax \n" - "movd %%eax,%%xmm2 \n" - "pshufd $0x0,%%xmm2,%%xmm2 \n" - LABELALIGN - "1: \n" - // Step 1: Scale Y contribution to 8 G values. G = (y - 16) * 1.164 - "movq " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "psubusw %%xmm3,%%xmm0 \n" - "pmullw %%xmm2,%%xmm0 \n" - "psrlw $6, %%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - - // Step 2: Weave into ARGB - "punpcklbw %%xmm0,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm0,%%xmm0 \n" - "punpckhwd %%xmm1,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "por %%xmm4,%%xmm1 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(y_buf), // %0 - "+r"(dst_argb), // %1 - "+rm"(width) // %2 - : - : "memory", "cc", "eax" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4" -#endif - ); -} -#endif // HAS_YTOARGBROW_SSE2 - -#ifdef HAS_MIRRORROW_SSSE3 -// Shuffle table for reversing the bytes. -static uvec8 kShuffleMirror = { - 15u, 14u, 13u, 12u, 11u, 10u, 9u, 8u, 7u, 6u, 5u, 4u, 3u, 2u, 1u, 0u -}; - -void MirrorRow_SSSE3(const uint8* src, uint8* dst, int width) { - intptr_t temp_width = (intptr_t)(width); - asm volatile ( - "movdqa %3,%%xmm5 \n" - "lea " MEMLEA(-0x10,0) ",%0 \n" - LABELALIGN - "1: \n" - MEMOPREG(movdqa,0x00,0,2,1,xmm0) // movdqa (%0,%2),%%xmm0 - "pshufb %%xmm5,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(temp_width) // %2 - : "m"(kShuffleMirror) // %3 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm5" -#endif - ); -} -#endif // HAS_MIRRORROW_SSSE3 - -#ifdef HAS_MIRRORROW_SSE2 -void MirrorRow_SSE2(const uint8* src, uint8* dst, int width) { - intptr_t temp_width = (intptr_t)(width); - asm volatile ( - "lea " MEMLEA(-0x10,0) ",%0 \n" - LABELALIGN - "1: \n" - MEMOPREG(movdqu,0x00,0,2,1,xmm0) // movdqu (%0,%2),%%xmm0 - "movdqa %%xmm0,%%xmm1 \n" - "psllw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm1,%%xmm0 \n" - "pshuflw $0x1b,%%xmm0,%%xmm0 \n" - "pshufhw $0x1b,%%xmm0,%%xmm0 \n" - "pshufd $0x4e,%%xmm0,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1)",%1 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(temp_width) // %2 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} -#endif // HAS_MIRRORROW_SSE2 - -#ifdef HAS_MIRRORROW_UV_SSSE3 -// Shuffle table for reversing the bytes of UV channels. -static uvec8 kShuffleMirrorUV = { - 14u, 12u, 10u, 8u, 6u, 4u, 2u, 0u, 15u, 13u, 11u, 9u, 7u, 5u, 3u, 1u -}; -void MirrorUVRow_SSSE3(const uint8* src, uint8* dst_u, uint8* dst_v, - int width) { - intptr_t temp_width = (intptr_t)(width); - asm volatile ( - "movdqa %4,%%xmm1 \n" - "lea " MEMLEA4(-0x10,0,3,2) ",%0 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(-0x10,0) ",%0 \n" - "pshufb %%xmm1,%%xmm0 \n" - "sub $8,%3 \n" - "movlpd %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movhpd,xmm0,0x00,1,2,1) // movhpd %%xmm0,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(temp_width) // %3 - : "m"(kShuffleMirrorUV) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} -#endif // HAS_MIRRORROW_UV_SSSE3 - -#ifdef HAS_ARGBMIRRORROW_SSSE3 -// Shuffle table for reversing the bytes. -static uvec8 kARGBShuffleMirror = { - 12u, 13u, 14u, 15u, 8u, 9u, 10u, 11u, 4u, 5u, 6u, 7u, 0u, 1u, 2u, 3u -}; - -void ARGBMirrorRow_SSSE3(const uint8* src, uint8* dst, int width) { - intptr_t temp_width = (intptr_t)(width); - asm volatile ( - "lea " MEMLEA4(-0x10,0,2,4) ",%0 \n" - "movdqa %3,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "pshufb %%xmm5,%%xmm0 \n" - "lea " MEMLEA(-0x10,0) ",%0 \n" - "sub $0x4,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(temp_width) // %2 - : "m"(kARGBShuffleMirror) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm5" -#endif - ); -} -#endif // HAS_ARGBMIRRORROW_SSSE3 - -#ifdef HAS_SPLITUVROW_SSE2 -void SplitUVRow_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "movdqa %%xmm1,%%xmm3 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "psrlw $0x8,%%xmm2 \n" - "psrlw $0x8,%%xmm3 \n" - "packuswb %%xmm3,%%xmm2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - MEMOPMEM(movdqa,xmm2,0x00,1,2,1) // movdqa %%xmm2,(%1,%2) - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void SplitUVRow_Unaligned_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "movdqa %%xmm1,%%xmm3 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "psrlw $0x8,%%xmm2 \n" - "psrlw $0x8,%%xmm3 \n" - "packuswb %%xmm3,%%xmm2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - MEMOPMEM(movdqu,xmm2,0x00,1,2,1) // movdqu %%xmm2,(%1,%2) - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} -#endif // HAS_SPLITUVROW_SSE2 - -#ifdef HAS_MERGEUVROW_SSE2 -void MergeUVRow_SSE2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) { - asm volatile ( - "sub %0,%1 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,0,1,1,xmm1) // movdqa (%0,%1,1),%%xmm1 - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "punpcklbw %%xmm1,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm2 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "movdqa %%xmm2," MEMACCESS2(0x10,2) " \n" - "lea " MEMLEA(0x20,2) ",%2 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_u), // %0 - "+r"(src_v), // %1 - "+r"(dst_uv), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2" -#endif - ); -} - -void MergeUVRow_Unaligned_SSE2(const uint8* src_u, const uint8* src_v, - uint8* dst_uv, int width) { - asm volatile ( - "sub %0,%1 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,0,1,1,xmm1) // movdqu (%0,%1,1),%%xmm1 - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "punpcklbw %%xmm1,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm2 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "movdqu %%xmm2," MEMACCESS2(0x10,2) " \n" - "lea " MEMLEA(0x20,2) ",%2 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_u), // %0 - "+r"(src_v), // %1 - "+r"(dst_uv), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2" -#endif - ); -} -#endif // HAS_MERGEUVROW_SSE2 - -#ifdef HAS_COPYROW_SSE2 -void CopyRow_SSE2(const uint8* src, uint8* dst, int count) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x20,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(count) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} -#endif // HAS_COPYROW_SSE2 - -#ifdef HAS_COPYROW_X86 -void CopyRow_X86(const uint8* src, uint8* dst, int width) { - size_t width_tmp = (size_t)(width); - asm volatile ( - "shr $0x2,%2 \n" - "rep movsl " MEMMOVESTRING(0,1) " \n" - : "+S"(src), // %0 - "+D"(dst), // %1 - "+c"(width_tmp) // %2 - : - : "memory", "cc" - ); -} -#endif // HAS_COPYROW_X86 - -#ifdef HAS_COPYROW_ERMS -// Unaligned Multiple of 1. -void CopyRow_ERMS(const uint8* src, uint8* dst, int width) { - size_t width_tmp = (size_t)(width); - asm volatile ( - "rep movsb " MEMMOVESTRING(0,1) " \n" - : "+S"(src), // %0 - "+D"(dst), // %1 - "+c"(width_tmp) // %2 - : - : "memory", "cc" - ); -} -#endif // HAS_COPYROW_ERMS - -#ifdef HAS_ARGBCOPYALPHAROW_SSE2 -// width in pixels -void ARGBCopyAlphaRow_SSE2(const uint8* src, uint8* dst, int width) { - asm volatile ( - "pcmpeqb %%xmm0,%%xmm0 \n" - "pslld $0x18,%%xmm0 \n" - "pcmpeqb %%xmm1,%%xmm1 \n" - "psrld $0x8,%%xmm1 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm3 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa " MEMACCESS(1) ",%%xmm4 \n" - "movdqa " MEMACCESS2(0x10,1) ",%%xmm5 \n" - "pand %%xmm0,%%xmm2 \n" - "pand %%xmm0,%%xmm3 \n" - "pand %%xmm1,%%xmm4 \n" - "pand %%xmm1,%%xmm5 \n" - "por %%xmm4,%%xmm2 \n" - "por %%xmm5,%%xmm3 \n" - "movdqa %%xmm2," MEMACCESS(1) " \n" - "movdqa %%xmm3," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBCOPYALPHAROW_SSE2 - -#ifdef HAS_ARGBCOPYALPHAROW_AVX2 -// width in pixels -void ARGBCopyAlphaRow_AVX2(const uint8* src, uint8* dst, int width) { - asm volatile ( - "vpcmpeqb %%ymm0,%%ymm0,%%ymm0 \n" - "vpsrld $0x8,%%ymm0,%%ymm0 \n" - LABELALIGN - "1: \n" - "vmovdqu " MEMACCESS(0) ",%%ymm1 \n" - "vmovdqu " MEMACCESS2(0x20,0) ",%%ymm2 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "vpblendvb %%ymm0," MEMACCESS(1) ",%%ymm1,%%ymm1 \n" - "vpblendvb %%ymm0," MEMACCESS2(0x20,1) ",%%ymm2,%%ymm2 \n" - "vmovdqu %%ymm1," MEMACCESS(1) " \n" - "vmovdqu %%ymm2," MEMACCESS2(0x20,1) " \n" - "lea " MEMLEA(0x40,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - "vzeroupper \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2" -#endif - ); -} -#endif // HAS_ARGBCOPYALPHAROW_AVX2 - -#ifdef HAS_ARGBCOPYYTOALPHAROW_SSE2 -// width in pixels -void ARGBCopyYToAlphaRow_SSE2(const uint8* src, uint8* dst, int width) { - asm volatile ( - "pcmpeqb %%xmm0,%%xmm0 \n" - "pslld $0x18,%%xmm0 \n" - "pcmpeqb %%xmm1,%%xmm1 \n" - "psrld $0x8,%%xmm1 \n" - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm2 \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "punpcklbw %%xmm2,%%xmm2 \n" - "punpckhwd %%xmm2,%%xmm3 \n" - "punpcklwd %%xmm2,%%xmm2 \n" - "movdqa " MEMACCESS(1) ",%%xmm4 \n" - "movdqa " MEMACCESS2(0x10,1) ",%%xmm5 \n" - "pand %%xmm0,%%xmm2 \n" - "pand %%xmm0,%%xmm3 \n" - "pand %%xmm1,%%xmm4 \n" - "pand %%xmm1,%%xmm5 \n" - "por %%xmm4,%%xmm2 \n" - "por %%xmm5,%%xmm3 \n" - "movdqa %%xmm2," MEMACCESS(1) " \n" - "movdqa %%xmm3," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBCOPYYTOALPHAROW_SSE2 - -#ifdef HAS_ARGBCOPYYTOALPHAROW_AVX2 -// width in pixels -void ARGBCopyYToAlphaRow_AVX2(const uint8* src, uint8* dst, int width) { - asm volatile ( - "vpcmpeqb %%ymm0,%%ymm0,%%ymm0 \n" - "vpsrld $0x8,%%ymm0,%%ymm0 \n" - LABELALIGN - "1: \n" - "vpmovzxbd " MEMACCESS(0) ",%%ymm1 \n" - "vpmovzxbd " MEMACCESS2(0x8,0) ",%%ymm2 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "vpslld $0x18,%%ymm1,%%ymm1 \n" - "vpslld $0x18,%%ymm2,%%ymm2 \n" - "vpblendvb %%ymm0," MEMACCESS(1) ",%%ymm1,%%ymm1 \n" - "vpblendvb %%ymm0," MEMACCESS2(0x20,1) ",%%ymm2,%%ymm2 \n" - "vmovdqu %%ymm1," MEMACCESS(1) " \n" - "vmovdqu %%ymm2," MEMACCESS2(0x20,1) " \n" - "lea " MEMLEA(0x40,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - "vzeroupper \n" - : "+r"(src), // %0 - "+r"(dst), // %1 - "+r"(width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2" -#endif - ); -} -#endif // HAS_ARGBCOPYYTOALPHAROW_AVX2 - -#ifdef HAS_SETROW_X86 -void SetRow_X86(uint8* dst, uint32 v32, int width) { - size_t width_tmp = (size_t)(width); - asm volatile ( - "shr $0x2,%1 \n" - "rep stosl " MEMSTORESTRING(eax,0) " \n" - : "+D"(dst), // %0 - "+c"(width_tmp) // %1 - : "a"(v32) // %2 - : "memory", "cc"); -} - -void ARGBSetRows_X86(uint8* dst, uint32 v32, int width, - int dst_stride, int height) { - for (int y = 0; y < height; ++y) { - size_t width_tmp = (size_t)(width); - uint32* d = (uint32*)(dst); - asm volatile ( - "rep stosl " MEMSTORESTRING(eax,0) " \n" - : "+D"(d), // %0 - "+c"(width_tmp) // %1 - : "a"(v32) // %2 - : "memory", "cc"); - dst += dst_stride; - } -} -#endif // HAS_SETROW_X86 - -#ifdef HAS_YUY2TOYROW_SSE2 -void YUY2ToYRow_SSE2(const uint8* src_yuy2, uint8* dst_y, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void YUY2ToUVRow_SSE2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movdqa,0x00,0,4,1,xmm2) // movdqa (%0,%4,1),%%xmm2 - MEMOPREG(movdqa,0x10,0,4,1,xmm3) // movdqa 0x10(%0,%4,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : "r"((intptr_t)(stride_yuy2)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void YUY2ToUV422Row_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void YUY2ToYRow_Unaligned_SSE2(const uint8* src_yuy2, - uint8* dst_y, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void YUY2ToUVRow_Unaligned_SSE2(const uint8* src_yuy2, - int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm2) // movdqu (%0,%4,1),%%xmm2 - MEMOPREG(movdqu,0x10,0,4,1,xmm3) // movdqu 0x10(%0,%4,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : "r"((intptr_t)(stride_yuy2)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void YUY2ToUV422Row_Unaligned_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_yuy2), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void UYVYToYRow_SSE2(const uint8* src_uyvy, uint8* dst_y, int pix) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void UYVYToUVRow_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movdqa,0x00,0,4,1,xmm2) // movdqa (%0,%4,1),%%xmm2 - MEMOPREG(movdqa,0x10,0,4,1,xmm3) // movdqa 0x10(%0,%4,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : "r"((intptr_t)(stride_uyvy)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void UYVYToUV422Row_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void UYVYToYRow_Unaligned_SSE2(const uint8* src_uyvy, - uint8* dst_y, int pix) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_y), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void UYVYToUVRow_Unaligned_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movdqu,0x00,0,4,1,xmm2) // movdqu (%0,%4,1),%%xmm2 - MEMOPREG(movdqu,0x10,0,4,1,xmm3) // movdqu 0x10(%0,%4,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : "r"((intptr_t)(stride_uyvy)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void UYVYToUV422Row_Unaligned_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm1,0x00,1,2,1) // movq %%xmm1,(%1,%2) - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x10,%3 \n" - "jg 1b \n" - : "+r"(src_uyvy), // %0 - "+r"(dst_u), // %1 - "+r"(dst_v), // %2 - "+r"(pix) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_YUY2TOYROW_SSE2 - -#ifdef HAS_ARGBBLENDROW_SSE2 -// Blend 8 pixels at a time. -void ARGBBlendRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - "pcmpeqb %%xmm7,%%xmm7 \n" - "psrlw $0xf,%%xmm7 \n" - "pcmpeqb %%xmm6,%%xmm6 \n" - "psrlw $0x8,%%xmm6 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "psllw $0x8,%%xmm5 \n" - "pcmpeqb %%xmm4,%%xmm4 \n" - "pslld $0x18,%%xmm4 \n" - "sub $0x1,%3 \n" - "je 91f \n" - "jl 99f \n" - - // 1 pixel loop until destination pointer is aligned. - "10: \n" - "test $0xf,%2 \n" - "je 19f \n" - "movd " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movd " MEMACCESS(1) ",%%xmm2 \n" - "psrlw $0x8,%%xmm3 \n" - "pshufhw $0xf5,%%xmm3,%%xmm3 \n" - "pshuflw $0xf5,%%xmm3,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movd " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x4,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x1,%3 \n" - "movd %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x4,2) ",%2 \n" - "jge 10b \n" - - "19: \n" - "add $1-4,%3 \n" - "jl 49f \n" - - // 4 pixel loop. - LABELALIGN - "41: \n" - "movdqu " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movdqu " MEMACCESS(1) ",%%xmm2 \n" - "psrlw $0x8,%%xmm3 \n" - "pshufhw $0xf5,%%xmm3,%%xmm3 \n" - "pshuflw $0xf5,%%xmm3,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movdqu " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jge 41b \n" - - "49: \n" - "add $0x3,%3 \n" - "jl 99f \n" - - // 1 pixel loop. - "91: \n" - "movd " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movd " MEMACCESS(1) ",%%xmm2 \n" - "psrlw $0x8,%%xmm3 \n" - "pshufhw $0xf5,%%xmm3,%%xmm3 \n" - "pshuflw $0xf5,%%xmm3,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movd " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x4,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x1,%3 \n" - "movd %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x4,2) ",%2 \n" - "jge 91b \n" - "99: \n" - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBBLENDROW_SSE2 - -#ifdef HAS_ARGBBLENDROW_SSSE3 -// Shuffle table for isolating alpha. -static uvec8 kShuffleAlpha = { - 3u, 0x80, 3u, 0x80, 7u, 0x80, 7u, 0x80, - 11u, 0x80, 11u, 0x80, 15u, 0x80, 15u, 0x80 -}; - -// Blend 8 pixels at a time -// Shuffle table for reversing the bytes. - -// Same as SSE2, but replaces -// psrlw xmm3, 8 // alpha -// pshufhw xmm3, xmm3,0F5h // 8 alpha words -// pshuflw xmm3, xmm3,0F5h -// with.. -// pshufb xmm3, kShuffleAlpha // alpha - -void ARGBBlendRow_SSSE3(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - "pcmpeqb %%xmm7,%%xmm7 \n" - "psrlw $0xf,%%xmm7 \n" - "pcmpeqb %%xmm6,%%xmm6 \n" - "psrlw $0x8,%%xmm6 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "psllw $0x8,%%xmm5 \n" - "pcmpeqb %%xmm4,%%xmm4 \n" - "pslld $0x18,%%xmm4 \n" - "sub $0x1,%3 \n" - "je 91f \n" - "jl 99f \n" - - // 1 pixel loop until destination pointer is aligned. - "10: \n" - "test $0xf,%2 \n" - "je 19f \n" - "movd " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movd " MEMACCESS(1) ",%%xmm2 \n" - "pshufb %4,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movd " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x4,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x1,%3 \n" - "movd %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x4,2) ",%2 \n" - "jge 10b \n" - - "19: \n" - "add $1-4,%3 \n" - "jl 49f \n" - "test $0xf,%0 \n" - "jne 41f \n" - "test $0xf,%1 \n" - "jne 41f \n" - - // 4 pixel loop. - LABELALIGN - "40: \n" - "movdqa " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movdqa " MEMACCESS(1) ",%%xmm2 \n" - "pshufb %4,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movdqa " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jge 40b \n" - "jmp 49f \n" - - // 4 pixel unaligned loop. - LABELALIGN - "41: \n" - "movdqu " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movdqu " MEMACCESS(1) ",%%xmm2 \n" - "pshufb %4,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movdqu " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jge 41b \n" - - "49: \n" - "add $0x3,%3 \n" - "jl 99f \n" - - // 1 pixel loop. - "91: \n" - "movd " MEMACCESS(0) ",%%xmm3 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - "movdqa %%xmm3,%%xmm0 \n" - "pxor %%xmm4,%%xmm3 \n" - "movd " MEMACCESS(1) ",%%xmm2 \n" - "pshufb %4,%%xmm3 \n" - "pand %%xmm6,%%xmm2 \n" - "paddw %%xmm7,%%xmm3 \n" - "pmullw %%xmm3,%%xmm2 \n" - "movd " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x4,1) ",%1 \n" - "psrlw $0x8,%%xmm1 \n" - "por %%xmm4,%%xmm0 \n" - "pmullw %%xmm3,%%xmm1 \n" - "psrlw $0x8,%%xmm2 \n" - "paddusb %%xmm2,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x1,%3 \n" - "movd %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x4,2) ",%2 \n" - "jge 91b \n" - "99: \n" - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : "m"(kShuffleAlpha) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBBLENDROW_SSSE3 - -#ifdef HAS_ARGBATTENUATEROW_SSE2 -// Attenuate 4 pixels at a time. -// aligned to 16 bytes -void ARGBAttenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - "pcmpeqb %%xmm4,%%xmm4 \n" - "pslld $0x18,%%xmm4 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrld $0x8,%%xmm5 \n" - - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "pshufhw $0xff,%%xmm0,%%xmm2 \n" - "pshuflw $0xff,%%xmm2,%%xmm2 \n" - "pmulhuw %%xmm2,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm1 \n" - "punpckhbw %%xmm1,%%xmm1 \n" - "pshufhw $0xff,%%xmm1,%%xmm2 \n" - "pshuflw $0xff,%%xmm2,%%xmm2 \n" - "pmulhuw %%xmm2,%%xmm1 \n" - "movdqa " MEMACCESS(0) ",%%xmm2 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "pand %%xmm4,%%xmm2 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "pand %%xmm5,%%xmm0 \n" - "por %%xmm2,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBATTENUATEROW_SSE2 - -#ifdef HAS_ARGBATTENUATEROW_SSSE3 -// Shuffle table duplicating alpha -static uvec8 kShuffleAlpha0 = { - 3u, 3u, 3u, 3u, 3u, 3u, 128u, 128u, 7u, 7u, 7u, 7u, 7u, 7u, 128u, 128u, -}; -static uvec8 kShuffleAlpha1 = { - 11u, 11u, 11u, 11u, 11u, 11u, 128u, 128u, - 15u, 15u, 15u, 15u, 15u, 15u, 128u, 128u, -}; -// Attenuate 4 pixels at a time. -// aligned to 16 bytes -void ARGBAttenuateRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - "pcmpeqb %%xmm3,%%xmm3 \n" - "pslld $0x18,%%xmm3 \n" - "movdqa %3,%%xmm4 \n" - "movdqa %4,%%xmm5 \n" - - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "pshufb %%xmm4,%%xmm0 \n" - "movdqu " MEMACCESS(0) ",%%xmm1 \n" - "punpcklbw %%xmm1,%%xmm1 \n" - "pmulhuw %%xmm1,%%xmm0 \n" - "movdqu " MEMACCESS(0) ",%%xmm1 \n" - "pshufb %%xmm5,%%xmm1 \n" - "movdqu " MEMACCESS(0) ",%%xmm2 \n" - "punpckhbw %%xmm2,%%xmm2 \n" - "pmulhuw %%xmm2,%%xmm1 \n" - "movdqu " MEMACCESS(0) ",%%xmm2 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "pand %%xmm3,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "por %%xmm2,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "m"(kShuffleAlpha0), // %3 - "m"(kShuffleAlpha1) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBATTENUATEROW_SSSE3 - -#ifdef HAS_ARGBUNATTENUATEROW_SSE2 -// Unattenuate 4 pixels at a time. -// aligned to 16 bytes -void ARGBUnattenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, - int width) { - uintptr_t alpha = 0; - asm volatile ( - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movzb " MEMACCESS2(0x03,0) ",%3 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - MEMOPREG(movd,0x00,4,3,4,xmm2) // movd 0x0(%4,%3,4),%%xmm2 - "movzb " MEMACCESS2(0x07,0) ",%3 \n" - MEMOPREG(movd,0x00,4,3,4,xmm3) // movd 0x0(%4,%3,4),%%xmm3 - "pshuflw $0x40,%%xmm2,%%xmm2 \n" - "pshuflw $0x40,%%xmm3,%%xmm3 \n" - "movlhps %%xmm3,%%xmm2 \n" - "pmulhuw %%xmm2,%%xmm0 \n" - "movdqu " MEMACCESS(0) ",%%xmm1 \n" - "movzb " MEMACCESS2(0x0b,0) ",%3 \n" - "punpckhbw %%xmm1,%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movd,0x00,4,3,4,xmm2) // movd 0x0(%4,%3,4),%%xmm2 - "movzb " MEMACCESS2(0x0f,0) ",%3 \n" - MEMOPREG(movd,0x00,4,3,4,xmm3) // movd 0x0(%4,%3,4),%%xmm3 - "pshuflw $0x40,%%xmm2,%%xmm2 \n" - "pshuflw $0x40,%%xmm3,%%xmm3 \n" - "movlhps %%xmm3,%%xmm2 \n" - "pmulhuw %%xmm2,%%xmm1 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width), // %2 - "+r"(alpha) // %3 - : "r"(fixed_invtbl8) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBUNATTENUATEROW_SSE2 - -#ifdef HAS_ARGBGRAYROW_SSSE3 -// Convert 8 ARGB pixels (64 bytes) to 8 Gray ARGB pixels -void ARGBGrayRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width) { - asm volatile ( - "movdqa %3,%%xmm4 \n" - "movdqa %4,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm0 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "phaddw %%xmm1,%%xmm0 \n" - "paddw %%xmm5,%%xmm0 \n" - "psrlw $0x7,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm3 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrld $0x18,%%xmm2 \n" - "psrld $0x18,%%xmm3 \n" - "packuswb %%xmm3,%%xmm2 \n" - "packuswb %%xmm2,%%xmm2 \n" - "movdqa %%xmm0,%%xmm3 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "punpcklbw %%xmm2,%%xmm3 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm3,%%xmm0 \n" - "punpckhwd %%xmm3,%%xmm1 \n" - "sub $0x8,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "m"(kARGBToYJ), // %3 - "m"(kAddYJ64) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBGRAYROW_SSSE3 - -#ifdef HAS_ARGBSEPIAROW_SSSE3 -// b = (r * 35 + g * 68 + b * 17) >> 7 -// g = (r * 45 + g * 88 + b * 22) >> 7 -// r = (r * 50 + g * 98 + b * 24) >> 7 -// Constant for ARGB color to sepia tone -static vec8 kARGBToSepiaB = { - 17, 68, 35, 0, 17, 68, 35, 0, 17, 68, 35, 0, 17, 68, 35, 0 -}; - -static vec8 kARGBToSepiaG = { - 22, 88, 45, 0, 22, 88, 45, 0, 22, 88, 45, 0, 22, 88, 45, 0 -}; - -static vec8 kARGBToSepiaR = { - 24, 98, 50, 0, 24, 98, 50, 0, 24, 98, 50, 0, 24, 98, 50, 0 -}; - -// Convert 8 ARGB pixels (32 bytes) to 8 Sepia ARGB pixels. -void ARGBSepiaRow_SSSE3(uint8* dst_argb, int width) { - asm volatile ( - "movdqa %2,%%xmm2 \n" - "movdqa %3,%%xmm3 \n" - "movdqa %4,%%xmm4 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm6 \n" - "pmaddubsw %%xmm2,%%xmm0 \n" - "pmaddubsw %%xmm2,%%xmm6 \n" - "phaddw %%xmm6,%%xmm0 \n" - "psrlw $0x7,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm5 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm5 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "phaddw %%xmm1,%%xmm5 \n" - "psrlw $0x7,%%xmm5 \n" - "packuswb %%xmm5,%%xmm5 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm5 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm5 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "phaddw %%xmm1,%%xmm5 \n" - "psrlw $0x7,%%xmm5 \n" - "packuswb %%xmm5,%%xmm5 \n" - "movdqa " MEMACCESS(0) ",%%xmm6 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "psrld $0x18,%%xmm6 \n" - "psrld $0x18,%%xmm1 \n" - "packuswb %%xmm1,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "punpcklbw %%xmm6,%%xmm5 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklwd %%xmm5,%%xmm0 \n" - "punpckhwd %%xmm5,%%xmm1 \n" - "sub $0x8,%1 \n" - "movdqa %%xmm0," MEMACCESS(0) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,0) " \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "jg 1b \n" - : "+r"(dst_argb), // %0 - "+r"(width) // %1 - : "m"(kARGBToSepiaB), // %2 - "m"(kARGBToSepiaG), // %3 - "m"(kARGBToSepiaR) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} -#endif // HAS_ARGBSEPIAROW_SSSE3 - -#ifdef HAS_ARGBCOLORMATRIXROW_SSSE3 -// Tranform 8 ARGB pixels (32 bytes) with color matrix. -// Same as Sepia except matrix is provided. -void ARGBColorMatrixRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width) { - asm volatile ( - "movdqu " MEMACCESS(3) ",%%xmm5 \n" - "pshufd $0x00,%%xmm5,%%xmm2 \n" - "pshufd $0x55,%%xmm5,%%xmm3 \n" - "pshufd $0xaa,%%xmm5,%%xmm4 \n" - "pshufd $0xff,%%xmm5,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm7 \n" - "pmaddubsw %%xmm2,%%xmm0 \n" - "pmaddubsw %%xmm2,%%xmm7 \n" - "movdqa " MEMACCESS(0) ",%%xmm6 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "pmaddubsw %%xmm3,%%xmm6 \n" - "pmaddubsw %%xmm3,%%xmm1 \n" - "phaddsw %%xmm7,%%xmm0 \n" - "phaddsw %%xmm1,%%xmm6 \n" - "psraw $0x6,%%xmm0 \n" - "psraw $0x6,%%xmm6 \n" - "packuswb %%xmm0,%%xmm0 \n" - "packuswb %%xmm6,%%xmm6 \n" - "punpcklbw %%xmm6,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm7 \n" - "pmaddubsw %%xmm4,%%xmm1 \n" - "pmaddubsw %%xmm4,%%xmm7 \n" - "phaddsw %%xmm7,%%xmm1 \n" - "movdqa " MEMACCESS(0) ",%%xmm6 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm7 \n" - "pmaddubsw %%xmm5,%%xmm6 \n" - "pmaddubsw %%xmm5,%%xmm7 \n" - "phaddsw %%xmm7,%%xmm6 \n" - "psraw $0x6,%%xmm1 \n" - "psraw $0x6,%%xmm6 \n" - "packuswb %%xmm1,%%xmm1 \n" - "packuswb %%xmm6,%%xmm6 \n" - "punpcklbw %%xmm6,%%xmm1 \n" - "movdqa %%xmm0,%%xmm6 \n" - "punpcklwd %%xmm1,%%xmm0 \n" - "punpckhwd %%xmm1,%%xmm6 \n" - "sub $0x8,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm6," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(matrix_argb) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBCOLORMATRIXROW_SSSE3 - -#ifdef HAS_ARGBQUANTIZEROW_SSE2 -// Quantize 4 ARGB pixels (16 bytes). -// aligned to 16 bytes -void ARGBQuantizeRow_SSE2(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width) { - asm volatile ( - "movd %2,%%xmm2 \n" - "movd %3,%%xmm3 \n" - "movd %4,%%xmm4 \n" - "pshuflw $0x40,%%xmm2,%%xmm2 \n" - "pshufd $0x44,%%xmm2,%%xmm2 \n" - "pshuflw $0x40,%%xmm3,%%xmm3 \n" - "pshufd $0x44,%%xmm3,%%xmm3 \n" - "pshuflw $0x40,%%xmm4,%%xmm4 \n" - "pshufd $0x44,%%xmm4,%%xmm4 \n" - "pxor %%xmm5,%%xmm5 \n" - "pcmpeqb %%xmm6,%%xmm6 \n" - "pslld $0x18,%%xmm6 \n" - - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "pmulhuw %%xmm2,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm1 \n" - "punpckhbw %%xmm5,%%xmm1 \n" - "pmulhuw %%xmm2,%%xmm1 \n" - "pmullw %%xmm3,%%xmm0 \n" - "movdqa " MEMACCESS(0) ",%%xmm7 \n" - "pmullw %%xmm3,%%xmm1 \n" - "pand %%xmm6,%%xmm7 \n" - "paddw %%xmm4,%%xmm0 \n" - "paddw %%xmm4,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "por %%xmm7,%%xmm0 \n" - "sub $0x4,%1 \n" - "movdqa %%xmm0," MEMACCESS(0) " \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "jg 1b \n" - : "+r"(dst_argb), // %0 - "+r"(width) // %1 - : "r"(scale), // %2 - "r"(interval_size), // %3 - "r"(interval_offset) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBQUANTIZEROW_SSE2 - -#ifdef HAS_ARGBSHADEROW_SSE2 -// Shade 4 pixels at a time by specified value. -// Aligned to 16 bytes. -void ARGBShadeRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value) { - asm volatile ( - "movd %3,%%xmm2 \n" - "punpcklbw %%xmm2,%%xmm2 \n" - "punpcklqdq %%xmm2,%%xmm2 \n" - - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm1 \n" - "pmulhuw %%xmm2,%%xmm0 \n" - "pmulhuw %%xmm2,%%xmm1 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(value) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2" -#endif - ); -} -#endif // HAS_ARGBSHADEROW_SSE2 - -#ifdef HAS_ARGBMULTIPLYROW_SSE2 -// Multiply 2 rows of ARGB pixels together, 4 pixels at a time. -void ARGBMultiplyRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - "pxor %%xmm5,%%xmm5 \n" - - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqu " MEMACCESS(1) ",%%xmm2 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "movdqu %%xmm0,%%xmm1 \n" - "movdqu %%xmm2,%%xmm3 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "punpckhbw %%xmm5,%%xmm3 \n" - "pmulhuw %%xmm2,%%xmm0 \n" - "pmulhuw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} -#endif // HAS_ARGBMULTIPLYROW_SSE2 - -#ifdef HAS_ARGBADDROW_SSE2 -// Add 2 rows of ARGB pixels together, 4 pixels at a time. -void ARGBAddRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqu " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} -#endif // HAS_ARGBADDROW_SSE2 - -#ifdef HAS_ARGBSUBTRACTROW_SSE2 -// Subtract 2 rows of ARGB pixels, 4 pixels at a time. -void ARGBSubtractRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - asm volatile ( - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqu " MEMACCESS(1) ",%%xmm1 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "psubusb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_argb0), // %0 - "+r"(src_argb1), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} -#endif // HAS_ARGBSUBTRACTROW_SSE2 - -#ifdef HAS_SOBELXROW_SSE2 -// SobelX as a matrix is -// -1 0 1 -// -2 0 2 -// -1 0 1 -void SobelXRow_SSE2(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobelx, int width) { - asm volatile ( - "sub %0,%1 \n" - "sub %0,%2 \n" - "sub %0,%3 \n" - "pxor %%xmm5,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - "movq " MEMACCESS2(0x2,0) ",%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm1 \n" - "psubw %%xmm1,%%xmm0 \n" - BUNDLEALIGN - MEMOPREG(movq,0x00,0,1,1,xmm1) // movq (%0,%1,1),%%xmm1 - MEMOPREG(movq,0x02,0,1,1,xmm2) // movq 0x2(%0,%1,1),%%xmm2 - "punpcklbw %%xmm5,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "psubw %%xmm2,%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movq,0x00,0,2,1,xmm2) // movq (%0,%2,1),%%xmm2 - MEMOPREG(movq,0x02,0,2,1,xmm3) // movq 0x2(%0,%2,1),%%xmm3 - "punpcklbw %%xmm5,%%xmm2 \n" - "punpcklbw %%xmm5,%%xmm3 \n" - "psubw %%xmm3,%%xmm2 \n" - "paddw %%xmm2,%%xmm0 \n" - "paddw %%xmm1,%%xmm0 \n" - "paddw %%xmm1,%%xmm0 \n" - "pxor %%xmm1,%%xmm1 \n" - "psubw %%xmm0,%%xmm1 \n" - "pmaxsw %%xmm1,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "sub $0x8,%4 \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm0,0x00,0,3,1) // movq %%xmm0,(%0,%3,1) - "lea " MEMLEA(0x8,0) ",%0 \n" - "jg 1b \n" - : "+r"(src_y0), // %0 - "+r"(src_y1), // %1 - "+r"(src_y2), // %2 - "+r"(dst_sobelx), // %3 - "+r"(width) // %4 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} -#endif // HAS_SOBELXROW_SSE2 - -#ifdef HAS_SOBELYROW_SSE2 -// SobelY as a matrix is -// -1 -2 -1 -// 0 0 0 -// 1 2 1 -void SobelYRow_SSE2(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width) { - asm volatile ( - "sub %0,%1 \n" - "sub %0,%2 \n" - "pxor %%xmm5,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movq,0x00,0,1,1,xmm1) // movq (%0,%1,1),%%xmm1 - "punpcklbw %%xmm5,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm1 \n" - "psubw %%xmm1,%%xmm0 \n" - BUNDLEALIGN - "movq " MEMACCESS2(0x1,0) ",%%xmm1 \n" - MEMOPREG(movq,0x01,0,1,1,xmm2) // movq 0x1(%0,%1,1),%%xmm2 - "punpcklbw %%xmm5,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm2 \n" - "psubw %%xmm2,%%xmm1 \n" - BUNDLEALIGN - "movq " MEMACCESS2(0x2,0) ",%%xmm2 \n" - MEMOPREG(movq,0x02,0,1,1,xmm3) // movq 0x2(%0,%1,1),%%xmm3 - "punpcklbw %%xmm5,%%xmm2 \n" - "punpcklbw %%xmm5,%%xmm3 \n" - "psubw %%xmm3,%%xmm2 \n" - "paddw %%xmm2,%%xmm0 \n" - "paddw %%xmm1,%%xmm0 \n" - "paddw %%xmm1,%%xmm0 \n" - "pxor %%xmm1,%%xmm1 \n" - "psubw %%xmm0,%%xmm1 \n" - "pmaxsw %%xmm1,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "sub $0x8,%3 \n" - BUNDLEALIGN - MEMOPMEM(movq,xmm0,0x00,0,2,1) // movq %%xmm0,(%0,%2,1) - "lea " MEMLEA(0x8,0) ",%0 \n" - "jg 1b \n" - : "+r"(src_y0), // %0 - "+r"(src_y1), // %1 - "+r"(dst_sobely), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} -#endif // HAS_SOBELYROW_SSE2 - -#ifdef HAS_SOBELROW_SSE2 -// Adds Sobel X and Sobel Y and stores Sobel into ARGB. -// A = 255 -// R = Sobel -// G = Sobel -// B = Sobel -void SobelRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - asm volatile ( - "sub %0,%1 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pslld $0x18,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,0,1,1,xmm1) // movdqa (%0,%1,1),%%xmm1 - "lea " MEMLEA(0x10,0) ",%0 \n" - "paddusb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "punpcklbw %%xmm0,%%xmm2 \n" - "punpckhbw %%xmm0,%%xmm0 \n" - "movdqa %%xmm2,%%xmm1 \n" - "punpcklwd %%xmm2,%%xmm1 \n" - "punpckhwd %%xmm2,%%xmm2 \n" - "por %%xmm5,%%xmm1 \n" - "por %%xmm5,%%xmm2 \n" - "movdqa %%xmm0,%%xmm3 \n" - "punpcklwd %%xmm0,%%xmm3 \n" - "punpckhwd %%xmm0,%%xmm0 \n" - "por %%xmm5,%%xmm3 \n" - "por %%xmm5,%%xmm0 \n" - "sub $0x10,%3 \n" - "movdqa %%xmm1," MEMACCESS(2) " \n" - "movdqa %%xmm2," MEMACCESS2(0x10,2) " \n" - "movdqa %%xmm3," MEMACCESS2(0x20,2) " \n" - "movdqa %%xmm0," MEMACCESS2(0x30,2) " \n" - "lea " MEMLEA(0x40,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} -#endif // HAS_SOBELROW_SSE2 - -#ifdef HAS_SOBELTOPLANEROW_SSE2 -// Adds Sobel X and Sobel Y and stores Sobel into a plane. -void SobelToPlaneRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width) { - asm volatile ( - "sub %0,%1 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - "pslld $0x18,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,0,1,1,xmm1) // movdqa (%0,%1,1),%%xmm1 - "lea " MEMLEA(0x10,0) ",%0 \n" - "paddusb %%xmm1,%%xmm0 \n" - "sub $0x10,%3 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_y), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} -#endif // HAS_SOBELTOPLANEROW_SSE2 - -#ifdef HAS_SOBELXYROW_SSE2 -// Mixes Sobel X, Sobel Y and Sobel into ARGB. -// A = 255 -// R = Sobel X -// G = Sobel -// B = Sobel Y -void SobelXYRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - asm volatile ( - "sub %0,%1 \n" - "pcmpeqb %%xmm5,%%xmm5 \n" - - // 8 pixel loop. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,0,1,1,xmm1) // movdqa (%0,%1,1),%%xmm1 - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "paddusb %%xmm1,%%xmm2 \n" - "movdqa %%xmm0,%%xmm3 \n" - "punpcklbw %%xmm5,%%xmm3 \n" - "punpckhbw %%xmm5,%%xmm0 \n" - "movdqa %%xmm1,%%xmm4 \n" - "punpcklbw %%xmm2,%%xmm4 \n" - "punpckhbw %%xmm2,%%xmm1 \n" - "movdqa %%xmm4,%%xmm6 \n" - "punpcklwd %%xmm3,%%xmm6 \n" - "punpckhwd %%xmm3,%%xmm4 \n" - "movdqa %%xmm1,%%xmm7 \n" - "punpcklwd %%xmm0,%%xmm7 \n" - "punpckhwd %%xmm0,%%xmm1 \n" - "sub $0x10,%3 \n" - "movdqa %%xmm6," MEMACCESS(2) " \n" - "movdqa %%xmm4," MEMACCESS2(0x10,2) " \n" - "movdqa %%xmm7," MEMACCESS2(0x20,2) " \n" - "movdqa %%xmm1," MEMACCESS2(0x30,2) " \n" - "lea " MEMLEA(0x40,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_sobelx), // %0 - "+r"(src_sobely), // %1 - "+r"(dst_argb), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_SOBELXYROW_SSE2 - -#ifdef HAS_COMPUTECUMULATIVESUMROW_SSE2 -// Creates a table of cumulative sums where each value is a sum of all values -// above and to the left of the value, inclusive of the value. -void ComputeCumulativeSumRow_SSE2(const uint8* row, int32* cumsum, - const int32* previous_cumsum, int width) { - asm volatile ( - "pxor %%xmm0,%%xmm0 \n" - "pxor %%xmm1,%%xmm1 \n" - "sub $0x4,%3 \n" - "jl 49f \n" - "test $0xf,%1 \n" - "jne 49f \n" - - // 4 pixel loop \n" - LABELALIGN - "40: \n" - "movdqu " MEMACCESS(0) ",%%xmm2 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm2,%%xmm4 \n" - "punpcklbw %%xmm1,%%xmm2 \n" - "movdqa %%xmm2,%%xmm3 \n" - "punpcklwd %%xmm1,%%xmm2 \n" - "punpckhwd %%xmm1,%%xmm3 \n" - "punpckhbw %%xmm1,%%xmm4 \n" - "movdqa %%xmm4,%%xmm5 \n" - "punpcklwd %%xmm1,%%xmm4 \n" - "punpckhwd %%xmm1,%%xmm5 \n" - "paddd %%xmm2,%%xmm0 \n" - "movdqa " MEMACCESS(2) ",%%xmm2 \n" - "paddd %%xmm0,%%xmm2 \n" - "paddd %%xmm3,%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,2) ",%%xmm3 \n" - "paddd %%xmm0,%%xmm3 \n" - "paddd %%xmm4,%%xmm0 \n" - "movdqa " MEMACCESS2(0x20,2) ",%%xmm4 \n" - "paddd %%xmm0,%%xmm4 \n" - "paddd %%xmm5,%%xmm0 \n" - "movdqa " MEMACCESS2(0x30,2) ",%%xmm5 \n" - "lea " MEMLEA(0x40,2) ",%2 \n" - "paddd %%xmm0,%%xmm5 \n" - "movdqa %%xmm2," MEMACCESS(1) " \n" - "movdqa %%xmm3," MEMACCESS2(0x10,1) " \n" - "movdqa %%xmm4," MEMACCESS2(0x20,1) " \n" - "movdqa %%xmm5," MEMACCESS2(0x30,1) " \n" - "lea " MEMLEA(0x40,1) ",%1 \n" - "sub $0x4,%3 \n" - "jge 40b \n" - - "49: \n" - "add $0x3,%3 \n" - "jl 19f \n" - - // 1 pixel loop \n" - LABELALIGN - "10: \n" - "movd " MEMACCESS(0) ",%%xmm2 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - "punpcklbw %%xmm1,%%xmm2 \n" - "punpcklwd %%xmm1,%%xmm2 \n" - "paddd %%xmm2,%%xmm0 \n" - "movdqu " MEMACCESS(2) ",%%xmm2 \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "paddd %%xmm0,%%xmm2 \n" - "movdqu %%xmm2," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x1,%3 \n" - "jge 10b \n" - - "19: \n" - : "+r"(row), // %0 - "+r"(cumsum), // %1 - "+r"(previous_cumsum), // %2 - "+r"(width) // %3 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_COMPUTECUMULATIVESUMROW_SSE2 - -#ifdef HAS_CUMULATIVESUMTOAVERAGEROW_SSE2 -void CumulativeSumToAverageRow_SSE2(const int32* topleft, const int32* botleft, - int width, int area, uint8* dst, - int count) { - asm volatile ( - "movd %5,%%xmm5 \n" - "cvtdq2ps %%xmm5,%%xmm5 \n" - "rcpss %%xmm5,%%xmm4 \n" - "pshufd $0x0,%%xmm4,%%xmm4 \n" - "sub $0x4,%3 \n" - "jl 49f \n" - "cmpl $0x80,%5 \n" - "ja 40f \n" - - "pshufd $0x0,%%xmm5,%%xmm5 \n" - "pcmpeqb %%xmm6,%%xmm6 \n" - "psrld $0x10,%%xmm6 \n" - "cvtdq2ps %%xmm6,%%xmm6 \n" - "addps %%xmm6,%%xmm5 \n" - "mulps %%xmm4,%%xmm5 \n" - "cvtps2dq %%xmm5,%%xmm5 \n" - "packssdw %%xmm5,%%xmm5 \n" - - // 4 pixel small loop \n" - LABELALIGN - "4: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - BUNDLEALIGN - MEMOPREG(psubd,0x00,0,4,4,xmm0) // psubd 0x00(%0,%4,4),%%xmm0 - MEMOPREG(psubd,0x10,0,4,4,xmm1) // psubd 0x10(%0,%4,4),%%xmm1 - MEMOPREG(psubd,0x20,0,4,4,xmm2) // psubd 0x20(%0,%4,4),%%xmm2 - MEMOPREG(psubd,0x30,0,4,4,xmm3) // psubd 0x30(%0,%4,4),%%xmm3 - "lea " MEMLEA(0x40,0) ",%0 \n" - "psubd " MEMACCESS(1) ",%%xmm0 \n" - "psubd " MEMACCESS2(0x10,1) ",%%xmm1 \n" - "psubd " MEMACCESS2(0x20,1) ",%%xmm2 \n" - "psubd " MEMACCESS2(0x30,1) ",%%xmm3 \n" - BUNDLEALIGN - MEMOPREG(paddd,0x00,1,4,4,xmm0) // paddd 0x00(%1,%4,4),%%xmm0 - MEMOPREG(paddd,0x10,1,4,4,xmm1) // paddd 0x10(%1,%4,4),%%xmm1 - MEMOPREG(paddd,0x20,1,4,4,xmm2) // paddd 0x20(%1,%4,4),%%xmm2 - MEMOPREG(paddd,0x30,1,4,4,xmm3) // paddd 0x30(%1,%4,4),%%xmm3 - "lea " MEMLEA(0x40,1) ",%1 \n" - "packssdw %%xmm1,%%xmm0 \n" - "packssdw %%xmm3,%%xmm2 \n" - "pmulhuw %%xmm5,%%xmm0 \n" - "pmulhuw %%xmm5,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "sub $0x4,%3 \n" - "jge 4b \n" - "jmp 49f \n" - - // 4 pixel loop \n" - LABELALIGN - "40: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "movdqa " MEMACCESS2(0x20,0) ",%%xmm2 \n" - "movdqa " MEMACCESS2(0x30,0) ",%%xmm3 \n" - BUNDLEALIGN - MEMOPREG(psubd,0x00,0,4,4,xmm0) // psubd 0x00(%0,%4,4),%%xmm0 - MEMOPREG(psubd,0x10,0,4,4,xmm1) // psubd 0x10(%0,%4,4),%%xmm1 - MEMOPREG(psubd,0x20,0,4,4,xmm2) // psubd 0x20(%0,%4,4),%%xmm2 - MEMOPREG(psubd,0x30,0,4,4,xmm3) // psubd 0x30(%0,%4,4),%%xmm3 - "lea " MEMLEA(0x40,0) ",%0 \n" - "psubd " MEMACCESS(1) ",%%xmm0 \n" - "psubd " MEMACCESS2(0x10,1) ",%%xmm1 \n" - "psubd " MEMACCESS2(0x20,1) ",%%xmm2 \n" - "psubd " MEMACCESS2(0x30,1) ",%%xmm3 \n" - BUNDLEALIGN - MEMOPREG(paddd,0x00,1,4,4,xmm0) // paddd 0x00(%1,%4,4),%%xmm0 - MEMOPREG(paddd,0x10,1,4,4,xmm1) // paddd 0x10(%1,%4,4),%%xmm1 - MEMOPREG(paddd,0x20,1,4,4,xmm2) // paddd 0x20(%1,%4,4),%%xmm2 - MEMOPREG(paddd,0x30,1,4,4,xmm3) // paddd 0x30(%1,%4,4),%%xmm3 - "lea " MEMLEA(0x40,1) ",%1 \n" - "cvtdq2ps %%xmm0,%%xmm0 \n" - "cvtdq2ps %%xmm1,%%xmm1 \n" - "mulps %%xmm4,%%xmm0 \n" - "mulps %%xmm4,%%xmm1 \n" - "cvtdq2ps %%xmm2,%%xmm2 \n" - "cvtdq2ps %%xmm3,%%xmm3 \n" - "mulps %%xmm4,%%xmm2 \n" - "mulps %%xmm4,%%xmm3 \n" - "cvtps2dq %%xmm0,%%xmm0 \n" - "cvtps2dq %%xmm1,%%xmm1 \n" - "cvtps2dq %%xmm2,%%xmm2 \n" - "cvtps2dq %%xmm3,%%xmm3 \n" - "packssdw %%xmm1,%%xmm0 \n" - "packssdw %%xmm3,%%xmm2 \n" - "packuswb %%xmm2,%%xmm0 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "sub $0x4,%3 \n" - "jge 40b \n" - - "49: \n" - "add $0x3,%3 \n" - "jl 19f \n" - - // 1 pixel loop \n" - LABELALIGN - "10: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(psubd,0x00,0,4,4,xmm0) // psubd 0x00(%0,%4,4),%%xmm0 - "lea " MEMLEA(0x10,0) ",%0 \n" - "psubd " MEMACCESS(1) ",%%xmm0 \n" - BUNDLEALIGN - MEMOPREG(paddd,0x00,1,4,4,xmm0) // paddd 0x00(%1,%4,4),%%xmm0 - "lea " MEMLEA(0x10,1) ",%1 \n" - "cvtdq2ps %%xmm0,%%xmm0 \n" - "mulps %%xmm4,%%xmm0 \n" - "cvtps2dq %%xmm0,%%xmm0 \n" - "packssdw %%xmm0,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movd %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x4,2) ",%2 \n" - "sub $0x1,%3 \n" - "jge 10b \n" - "19: \n" - : "+r"(topleft), // %0 - "+r"(botleft), // %1 - "+r"(dst), // %2 - "+rm"(count) // %3 - : "r"((intptr_t)(width)), // %4 - "rm"(area) // %5 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} -#endif // HAS_CUMULATIVESUMTOAVERAGEROW_SSE2 - -#ifdef HAS_ARGBAFFINEROW_SSE2 -// Copy ARGB pixels from source image with slope to a row of destination. -LIBYUV_API -void ARGBAffineRow_SSE2(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* src_dudv, int width) { - intptr_t src_argb_stride_temp = src_argb_stride; - intptr_t temp = 0; - asm volatile ( - "movq " MEMACCESS(3) ",%%xmm2 \n" - "movq " MEMACCESS2(0x08,3) ",%%xmm7 \n" - "shl $0x10,%1 \n" - "add $0x4,%1 \n" - "movd %1,%%xmm5 \n" - "sub $0x4,%4 \n" - "jl 49f \n" - - "pshufd $0x44,%%xmm7,%%xmm7 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - "movdqa %%xmm2,%%xmm0 \n" - "addps %%xmm7,%%xmm0 \n" - "movlhps %%xmm0,%%xmm2 \n" - "movdqa %%xmm7,%%xmm4 \n" - "addps %%xmm4,%%xmm4 \n" - "movdqa %%xmm2,%%xmm3 \n" - "addps %%xmm4,%%xmm3 \n" - "addps %%xmm4,%%xmm4 \n" - - // 4 pixel loop \n" - LABELALIGN - "40: \n" - "cvttps2dq %%xmm2,%%xmm0 \n" // x, y float to int first 2 - "cvttps2dq %%xmm3,%%xmm1 \n" // x, y float to int next 2 - "packssdw %%xmm1,%%xmm0 \n" // x, y as 8 shorts - "pmaddwd %%xmm5,%%xmm0 \n" // off = x * 4 + y * stride - "movd %%xmm0,%k1 \n" - "pshufd $0x39,%%xmm0,%%xmm0 \n" - "movd %%xmm0,%k5 \n" - "pshufd $0x39,%%xmm0,%%xmm0 \n" - BUNDLEALIGN - MEMOPREG(movd,0x00,0,1,1,xmm1) // movd (%0,%1,1),%%xmm1 - MEMOPREG(movd,0x00,0,5,1,xmm6) // movd (%0,%5,1),%%xmm6 - "punpckldq %%xmm6,%%xmm1 \n" - "addps %%xmm4,%%xmm2 \n" - "movq %%xmm1," MEMACCESS(2) " \n" - "movd %%xmm0,%k1 \n" - "pshufd $0x39,%%xmm0,%%xmm0 \n" - "movd %%xmm0,%k5 \n" - BUNDLEALIGN - MEMOPREG(movd,0x00,0,1,1,xmm0) // movd (%0,%1,1),%%xmm0 - MEMOPREG(movd,0x00,0,5,1,xmm6) // movd (%0,%5,1),%%xmm6 - "punpckldq %%xmm6,%%xmm0 \n" - "addps %%xmm4,%%xmm3 \n" - "sub $0x4,%4 \n" - "movq %%xmm0," MEMACCESS2(0x08,2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jge 40b \n" - - "49: \n" - "add $0x3,%4 \n" - "jl 19f \n" - - // 1 pixel loop \n" - LABELALIGN - "10: \n" - "cvttps2dq %%xmm2,%%xmm0 \n" - "packssdw %%xmm0,%%xmm0 \n" - "pmaddwd %%xmm5,%%xmm0 \n" - "addps %%xmm7,%%xmm2 \n" - "movd %%xmm0,%k1 \n" - BUNDLEALIGN - MEMOPREG(movd,0x00,0,1,1,xmm0) // movd (%0,%1,1),%%xmm0 - "sub $0x1,%4 \n" - "movd %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x04,2) ",%2 \n" - "jge 10b \n" - "19: \n" - : "+r"(src_argb), // %0 - "+r"(src_argb_stride_temp), // %1 - "+r"(dst_argb), // %2 - "+r"(src_dudv), // %3 - "+rm"(width), // %4 - "+r"(temp) // %5 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBAFFINEROW_SSE2 - -#ifdef HAS_INTERPOLATEROW_SSSE3 -// Bilinear filter 16x2 -> 16x1 -void InterpolateRow_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - asm volatile ( - "sub %1,%0 \n" - "shr %3 \n" - "cmp $0x0,%3 \n" - "je 100f \n" - "cmp $0x20,%3 \n" - "je 75f \n" - "cmp $0x40,%3 \n" - "je 50f \n" - "cmp $0x60,%3 \n" - "je 25f \n" - - "movd %3,%%xmm0 \n" - "neg %3 \n" - "add $0x80,%3 \n" - "movd %3,%%xmm5 \n" - "punpcklbw %%xmm0,%%xmm5 \n" - "punpcklwd %%xmm5,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - - // General purpose row blend. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm2) - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm2,%%xmm0 \n" - "punpckhbw %%xmm2,%%xmm1 \n" - "pmaddubsw %%xmm5,%%xmm0 \n" - "pmaddubsw %%xmm5,%%xmm1 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - "jmp 99f \n" - - // Blend 25 / 75. - LABELALIGN - "25: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm1) - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 25b \n" - "jmp 99f \n" - - // Blend 50 / 50. - LABELALIGN - "50: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm1) - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 50b \n" - "jmp 99f \n" - - // Blend 75 / 25. - LABELALIGN - "75: \n" - "movdqa " MEMACCESS(1) ",%%xmm1 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm0) - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 75b \n" - "jmp 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - LABELALIGN - "100: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - "sub $0x10,%2 \n" - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 100b \n" - - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(source_y_fraction) // %3 - : "r"((intptr_t)(src_stride)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm5" -#endif - ); -} -#endif // HAS_INTERPOLATEROW_SSSE3 - -#ifdef HAS_INTERPOLATEROW_SSE2 -// Bilinear filter 16x2 -> 16x1 -void InterpolateRow_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - asm volatile ( - "sub %1,%0 \n" - "shr %3 \n" - "cmp $0x0,%3 \n" - "je 100f \n" - "cmp $0x20,%3 \n" - "je 75f \n" - "cmp $0x40,%3 \n" - "je 50f \n" - "cmp $0x60,%3 \n" - "je 25f \n" - - "movd %3,%%xmm0 \n" - "neg %3 \n" - "add $0x80,%3 \n" - "movd %3,%%xmm5 \n" - "punpcklbw %%xmm0,%%xmm5 \n" - "punpcklwd %%xmm5,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - - // General purpose row blend. - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm2) // movdqa (%1,%4,1),%%xmm2 - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm2,%%xmm3 \n" - "punpcklbw %%xmm4,%%xmm2 \n" - "punpckhbw %%xmm4,%%xmm3 \n" - "punpcklbw %%xmm4,%%xmm0 \n" - "punpckhbw %%xmm4,%%xmm1 \n" - "psubw %%xmm0,%%xmm2 \n" - "psubw %%xmm1,%%xmm3 \n" - "paddw %%xmm2,%%xmm2 \n" - "paddw %%xmm3,%%xmm3 \n" - "pmulhw %%xmm5,%%xmm2 \n" - "pmulhw %%xmm5,%%xmm3 \n" - "paddw %%xmm2,%%xmm0 \n" - "paddw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) // movdqa %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - "jmp 99f \n" - - // Blend 25 / 75. - LABELALIGN - "25: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm1) // movdqa (%1,%4,1),%%xmm1 - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) // movdqa %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 25b \n" - "jmp 99f \n" - - // Blend 50 / 50. - LABELALIGN - "50: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm1) // movdqa (%1,%4,1),%%xmm1 - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) // movdqa %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 50b \n" - "jmp 99f \n" - - // Blend 75 / 25. - LABELALIGN - "75: \n" - "movdqa " MEMACCESS(1) ",%%xmm1 \n" - MEMOPREG(movdqa,0x00,1,4,1,xmm0) // movdqa (%1,%4,1),%%xmm0 - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) // movdqa %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 75b \n" - "jmp 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - LABELALIGN - "100: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - "sub $0x10,%2 \n" - MEMOPMEM(movdqa,xmm0,0x00,1,0,1) // movdqa %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 100b \n" - - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(source_y_fraction) // %3 - : "r"((intptr_t)(src_stride)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_INTERPOLATEROW_SSE2 - -#ifdef HAS_INTERPOLATEROW_SSSE3 -// Bilinear filter 16x2 -> 16x1 -void InterpolateRow_Unaligned_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - asm volatile ( - "sub %1,%0 \n" - "shr %3 \n" - "cmp $0x0,%3 \n" - "je 100f \n" - "cmp $0x20,%3 \n" - "je 75f \n" - "cmp $0x40,%3 \n" - "je 50f \n" - "cmp $0x60,%3 \n" - "je 25f \n" - - "movd %3,%%xmm0 \n" - "neg %3 \n" - "add $0x80,%3 \n" - "movd %3,%%xmm5 \n" - "punpcklbw %%xmm0,%%xmm5 \n" - "punpcklwd %%xmm5,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - - // General purpose row blend. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm2) - "movdqu %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm2,%%xmm0 \n" - "punpckhbw %%xmm2,%%xmm1 \n" - "pmaddubsw %%xmm5,%%xmm0 \n" - "pmaddubsw %%xmm5,%%xmm1 \n" - "psrlw $0x7,%%xmm0 \n" - "psrlw $0x7,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - "jmp 99f \n" - - // Blend 25 / 75. - LABELALIGN - "25: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm1) - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 25b \n" - "jmp 99f \n" - - // Blend 50 / 50. - LABELALIGN - "50: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm1) - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 50b \n" - "jmp 99f \n" - - // Blend 75 / 25. - LABELALIGN - "75: \n" - "movdqu " MEMACCESS(1) ",%%xmm1 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm0) - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 75b \n" - "jmp 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - LABELALIGN - "100: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - "sub $0x10,%2 \n" - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 100b \n" - - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(source_y_fraction) // %3 - : "r"((intptr_t)(src_stride)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm5" -#endif - ); -} -#endif // HAS_INTERPOLATEROW_SSSE3 - -#ifdef HAS_INTERPOLATEROW_SSE2 -// Bilinear filter 16x2 -> 16x1 -void InterpolateRow_Unaligned_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - asm volatile ( - "sub %1,%0 \n" - "shr %3 \n" - "cmp $0x0,%3 \n" - "je 100f \n" - "cmp $0x20,%3 \n" - "je 75f \n" - "cmp $0x40,%3 \n" - "je 50f \n" - "cmp $0x60,%3 \n" - "je 25f \n" - - "movd %3,%%xmm0 \n" - "neg %3 \n" - "add $0x80,%3 \n" - "movd %3,%%xmm5 \n" - "punpcklbw %%xmm0,%%xmm5 \n" - "punpcklwd %%xmm5,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - "pxor %%xmm4,%%xmm4 \n" - - // General purpose row blend. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm2) // movdqu (%1,%4,1),%%xmm2 - "movdqu %%xmm0,%%xmm1 \n" - "movdqu %%xmm2,%%xmm3 \n" - "punpcklbw %%xmm4,%%xmm2 \n" - "punpckhbw %%xmm4,%%xmm3 \n" - "punpcklbw %%xmm4,%%xmm0 \n" - "punpckhbw %%xmm4,%%xmm1 \n" - "psubw %%xmm0,%%xmm2 \n" - "psubw %%xmm1,%%xmm3 \n" - "paddw %%xmm2,%%xmm2 \n" - "paddw %%xmm3,%%xmm3 \n" - "pmulhw %%xmm5,%%xmm2 \n" - "pmulhw %%xmm5,%%xmm3 \n" - "paddw %%xmm2,%%xmm0 \n" - "paddw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) // movdqu %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - "jmp 99f \n" - - // Blend 25 / 75. - LABELALIGN - "25: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm1) // movdqu (%1,%4,1),%%xmm1 - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) // movdqu %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 25b \n" - "jmp 99f \n" - - // Blend 50 / 50. - LABELALIGN - "50: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm1) // movdqu (%1,%4,1),%%xmm1 - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) // movdqu %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 50b \n" - "jmp 99f \n" - - // Blend 75 / 25. - LABELALIGN - "75: \n" - "movdqu " MEMACCESS(1) ",%%xmm1 \n" - MEMOPREG(movdqu,0x00,1,4,1,xmm0) // movdqu (%1,%4,1),%%xmm0 - "pavgb %%xmm1,%%xmm0 \n" - "pavgb %%xmm1,%%xmm0 \n" - "sub $0x10,%2 \n" - BUNDLEALIGN - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) // movdqu %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 75b \n" - "jmp 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - LABELALIGN - "100: \n" - "movdqu " MEMACCESS(1) ",%%xmm0 \n" - "sub $0x10,%2 \n" - MEMOPMEM(movdqu,xmm0,0x00,1,0,1) // movdqu %%xmm0,(%1,%0,1) - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 100b \n" - - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(source_y_fraction) // %3 - : "r"((intptr_t)(src_stride)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_INTERPOLATEROW_SSE2 - -#ifdef HAS_HALFROW_SSE2 -void HalfRow_SSE2(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix) { - asm volatile ( - "sub %0,%1 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(pavgb,0x00,0,3,1,xmm0) // pavgb (%0,%3),%%xmm0 - "sub $0x10,%2 \n" - MEMOPMEM(movdqa,xmm0,0x00,0,1,1) // movdqa %%xmm0,(%0,%1) - "lea " MEMLEA(0x10,0) ",%0 \n" - "jg 1b \n" - : "+r"(src_uv), // %0 - "+r"(dst_uv), // %1 - "+r"(pix) // %2 - : "r"((intptr_t)(src_uv_stride)) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0" -#endif - ); -} -#endif // HAS_HALFROW_SSE2 - -#ifdef HAS_ARGBTOBAYERROW_SSSE3 -void ARGBToBayerRow_SSSE3(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) { - asm volatile ( - // NaCL caveat - assumes movd is from GPR - "movd %3,%%xmm5 \n" - "pshufd $0x0,%%xmm5,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pshufb %%xmm5,%%xmm0 \n" - "pshufb %%xmm5,%%xmm1 \n" - "punpckldq %%xmm1,%%xmm0 \n" - "sub $0x8,%2 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_bayer), // %1 - "+r"(pix) // %2 - : "g"(selector) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_ARGBTOBAYERROW_SSSE3 - -#ifdef HAS_ARGBTOBAYERGGROW_SSE2 -void ARGBToBayerGGRow_SSE2(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrld $0x18,%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrld $0x8,%%xmm0 \n" - "psrld $0x8,%%xmm1 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packssdw %%xmm1,%%xmm0 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x8,%2 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_bayer), // %1 - "+r"(pix) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_ARGBTOBAYERGGROW_SSE2 - -#ifdef HAS_ARGBSHUFFLEROW_SSSE3 -// For BGRAToARGB, ABGRToARGB, RGBAToARGB, and ARGBToRGBA. -void ARGBShuffleRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - asm volatile ( - "movdqa " MEMACCESS(3) ",%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pshufb %%xmm5,%%xmm0 \n" - "pshufb %%xmm5,%%xmm1 \n" - "sub $0x8,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "r"(shuffler) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void ARGBShuffleRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - asm volatile ( - "movdqa " MEMACCESS(3) ",%%xmm5 \n" - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pshufb %%xmm5,%%xmm0 \n" - "pshufb %%xmm5,%%xmm1 \n" - "sub $0x8,%2 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "movdqu %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "r"(shuffler) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_ARGBSHUFFLEROW_SSSE3 - -#ifdef HAS_ARGBSHUFFLEROW_AVX2 -// For BGRAToARGB, ABGRToARGB, RGBAToARGB, and ARGBToRGBA. -void ARGBShuffleRow_AVX2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - asm volatile ( - "vbroadcastf128 " MEMACCESS(3) ",%%ymm5 \n" - LABELALIGN - "1: \n" - "vmovdqu " MEMACCESS(0) ",%%ymm0 \n" - "vmovdqu " MEMACCESS2(0x20,0) ",%%ymm1 \n" - "lea " MEMLEA(0x40,0) ",%0 \n" - "vpshufb %%ymm5,%%ymm0,%%ymm0 \n" - "vpshufb %%ymm5,%%ymm1,%%ymm1 \n" - "sub $0x10,%2 \n" - "vmovdqu %%ymm0," MEMACCESS(1) " \n" - "vmovdqu %%ymm1," MEMACCESS2(0x20,1) " \n" - "lea " MEMLEA(0x40,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(pix) // %2 - : "r"(shuffler) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_ARGBSHUFFLEROW_AVX2 - -#ifdef HAS_ARGBSHUFFLEROW_SSE2 -// For BGRAToARGB, ABGRToARGB, RGBAToARGB, and ARGBToRGBA. -void ARGBShuffleRow_SSE2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - uintptr_t pixel_temp = 0u; - asm volatile ( - "pxor %%xmm5,%%xmm5 \n" - "mov " MEMACCESS(4) ",%k2 \n" - "cmp $0x3000102,%k2 \n" - "je 3012f \n" - "cmp $0x10203,%k2 \n" - "je 123f \n" - "cmp $0x30201,%k2 \n" - "je 321f \n" - "cmp $0x2010003,%k2 \n" - "je 2103f \n" - - LABELALIGN - "1: \n" - "movzb " MEMACCESS(4) ",%2 \n" - MEMOPARG(movzb,0x00,0,2,1,2) " \n" // movzb (%0,%2,1),%2 - "mov %b2," MEMACCESS(1) " \n" - "movzb " MEMACCESS2(0x1,4) ",%2 \n" - MEMOPARG(movzb,0x00,0,2,1,2) " \n" // movzb (%0,%2,1),%2 - "mov %b2," MEMACCESS2(0x1,1) " \n" - BUNDLEALIGN - "movzb " MEMACCESS2(0x2,4) ",%2 \n" - MEMOPARG(movzb,0x00,0,2,1,2) " \n" // movzb (%0,%2,1),%2 - "mov %b2," MEMACCESS2(0x2,1) " \n" - "movzb " MEMACCESS2(0x3,4) ",%2 \n" - MEMOPARG(movzb,0x00,0,2,1,2) " \n" // movzb (%0,%2,1),%2 - "mov %b2," MEMACCESS2(0x3,1) " \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - "lea " MEMLEA(0x4,1) ",%1 \n" - "sub $0x1,%3 \n" - "jg 1b \n" - "jmp 99f \n" - - LABELALIGN - "123: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "punpckhbw %%xmm5,%%xmm1 \n" - "pshufhw $0x1b,%%xmm0,%%xmm0 \n" - "pshuflw $0x1b,%%xmm0,%%xmm0 \n" - "pshufhw $0x1b,%%xmm1,%%xmm1 \n" - "pshuflw $0x1b,%%xmm1,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 123b \n" - "jmp 99f \n" - - LABELALIGN - "321: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "punpckhbw %%xmm5,%%xmm1 \n" - "pshufhw $0x39,%%xmm0,%%xmm0 \n" - "pshuflw $0x39,%%xmm0,%%xmm0 \n" - "pshufhw $0x39,%%xmm1,%%xmm1 \n" - "pshuflw $0x39,%%xmm1,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 321b \n" - "jmp 99f \n" - - LABELALIGN - "2103: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "punpckhbw %%xmm5,%%xmm1 \n" - "pshufhw $0x93,%%xmm0,%%xmm0 \n" - "pshuflw $0x93,%%xmm0,%%xmm0 \n" - "pshufhw $0x93,%%xmm1,%%xmm1 \n" - "pshuflw $0x93,%%xmm1,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 2103b \n" - "jmp 99f \n" - - LABELALIGN - "3012: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "punpckhbw %%xmm5,%%xmm1 \n" - "pshufhw $0xc6,%%xmm0,%%xmm0 \n" - "pshuflw $0xc6,%%xmm0,%%xmm0 \n" - "pshufhw $0xc6,%%xmm1,%%xmm1 \n" - "pshuflw $0xc6,%%xmm1,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 3012b \n" - - "99: \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+d"(pixel_temp), // %2 - "+r"(pix) // %3 - : "r"(shuffler) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} -#endif // HAS_ARGBSHUFFLEROW_SSE2 - -#ifdef HAS_I422TOYUY2ROW_SSE2 -void I422ToYUY2Row_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_frame, int width) { - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movq " MEMACCESS(1) ",%%xmm2 \n" - MEMOPREG(movq,0x00,1,2,1,xmm3) // movq (%1,%2,1),%%xmm3 - "lea " MEMLEA(0x8,1) ",%1 \n" - "punpcklbw %%xmm3,%%xmm2 \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm2,%%xmm0 \n" - "punpckhbw %%xmm2,%%xmm1 \n" - "movdqu %%xmm0," MEMACCESS(3) " \n" - "movdqu %%xmm1," MEMACCESS2(0x10,3) " \n" - "lea " MEMLEA(0x20,3) ",%3 \n" - "sub $0x10,%4 \n" - "jg 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_frame), // %3 - "+rm"(width) // %4 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3" -#endif - ); -} -#endif // HAS_I422TOYUY2ROW_SSE2 - -#ifdef HAS_I422TOUYVYROW_SSE2 -void I422ToUYVYRow_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_frame, int width) { - asm volatile ( - "sub %1,%2 \n" - LABELALIGN - "1: \n" - "movq " MEMACCESS(1) ",%%xmm2 \n" - MEMOPREG(movq,0x00,1,2,1,xmm3) // movq (%1,%2,1),%%xmm3 - "lea " MEMLEA(0x8,1) ",%1 \n" - "punpcklbw %%xmm3,%%xmm2 \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqa %%xmm2,%%xmm1 \n" - "lea " MEMLEA(0x10,0) ",%0 \n" - "punpcklbw %%xmm0,%%xmm1 \n" - "punpckhbw %%xmm0,%%xmm2 \n" - "movdqu %%xmm1," MEMACCESS(3) " \n" - "movdqu %%xmm2," MEMACCESS2(0x10,3) " \n" - "lea " MEMLEA(0x20,3) ",%3 \n" - "sub $0x10,%4 \n" - "jg 1b \n" - : "+r"(src_y), // %0 - "+r"(src_u), // %1 - "+r"(src_v), // %2 - "+r"(dst_frame), // %3 - "+rm"(width) // %4 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3" -#endif - ); -} -#endif // HAS_I422TOUYVYROW_SSE2 - -#ifdef HAS_ARGBPOLYNOMIALROW_SSE2 -void ARGBPolynomialRow_SSE2(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width) { - asm volatile ( - "pxor %%xmm3,%%xmm3 \n" - - // 2 pixel loop. - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "punpcklbw %%xmm3,%%xmm0 \n" - "movdqa %%xmm0,%%xmm4 \n" - "punpcklwd %%xmm3,%%xmm0 \n" - "punpckhwd %%xmm3,%%xmm4 \n" - "cvtdq2ps %%xmm0,%%xmm0 \n" - "cvtdq2ps %%xmm4,%%xmm4 \n" - "movdqa %%xmm0,%%xmm1 \n" - "movdqa %%xmm4,%%xmm5 \n" - "mulps " MEMACCESS2(0x10,3) ",%%xmm0 \n" - "mulps " MEMACCESS2(0x10,3) ",%%xmm4 \n" - "addps " MEMACCESS(3) ",%%xmm0 \n" - "addps " MEMACCESS(3) ",%%xmm4 \n" - "movdqa %%xmm1,%%xmm2 \n" - "movdqa %%xmm5,%%xmm6 \n" - "mulps %%xmm1,%%xmm2 \n" - "mulps %%xmm5,%%xmm6 \n" - "mulps %%xmm2,%%xmm1 \n" - "mulps %%xmm6,%%xmm5 \n" - "mulps " MEMACCESS2(0x20,3) ",%%xmm2 \n" - "mulps " MEMACCESS2(0x20,3) ",%%xmm6 \n" - "mulps " MEMACCESS2(0x30,3) ",%%xmm1 \n" - "mulps " MEMACCESS2(0x30,3) ",%%xmm5 \n" - "addps %%xmm2,%%xmm0 \n" - "addps %%xmm6,%%xmm4 \n" - "addps %%xmm1,%%xmm0 \n" - "addps %%xmm5,%%xmm4 \n" - "cvttps2dq %%xmm0,%%xmm0 \n" - "cvttps2dq %%xmm4,%%xmm4 \n" - "packuswb %%xmm4,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "sub $0x2,%2 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(poly) // %3 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} -#endif // HAS_ARGBPOLYNOMIALROW_SSE2 - -#ifdef HAS_ARGBPOLYNOMIALROW_AVX2 -void ARGBPolynomialRow_AVX2(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width) { - asm volatile ( - "vbroadcastf128 " MEMACCESS(3) ",%%ymm4 \n" - "vbroadcastf128 " MEMACCESS2(0x10,3) ",%%ymm5 \n" - "vbroadcastf128 " MEMACCESS2(0x20,3) ",%%ymm6 \n" - "vbroadcastf128 " MEMACCESS2(0x30,3) ",%%ymm7 \n" - - // 2 pixel loop. - LABELALIGN - "1: \n" - "vpmovzxbd " MEMACCESS(0) ",%%ymm0 \n" // 2 ARGB pixels - "lea " MEMLEA(0x8,0) ",%0 \n" - "vcvtdq2ps %%ymm0,%%ymm0 \n" // X 8 floats - "vmulps %%ymm0,%%ymm0,%%ymm2 \n" // X * X - "vmulps %%ymm7,%%ymm0,%%ymm3 \n" // C3 * X - "vfmadd132ps %%ymm5,%%ymm4,%%ymm0 \n" // result = C0 + C1 * X - "vfmadd231ps %%ymm6,%%ymm2,%%ymm0 \n" // result += C2 * X * X - "vfmadd231ps %%ymm3,%%ymm2,%%ymm0 \n" // result += C3 * X * X * X - "vcvttps2dq %%ymm0,%%ymm0 \n" - "vpackusdw %%ymm0,%%ymm0,%%ymm0 \n" - "vpermq $0xd8,%%ymm0,%%ymm0 \n" - "vpackuswb %%xmm0,%%xmm0,%%xmm0 \n" - "sub $0x2,%2 \n" - "vmovq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "jg 1b \n" - "vzeroupper \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(width) // %2 - : "r"(poly) // %3 - : "memory", "cc" -#if defined(__SSE2__) -// TODO(fbarchard): declare ymm usage when applicable. - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} -#endif // HAS_ARGBPOLYNOMIALROW_AVX2 - -#ifdef HAS_ARGBCOLORTABLEROW_X86 -// Tranform ARGB pixels with color table. -void ARGBColorTableRow_X86(uint8* dst_argb, const uint8* table_argb, - int width) { - uintptr_t pixel_temp = 0u; - asm volatile ( - // 1 pixel loop. - LABELALIGN - "1: \n" - "movzb " MEMACCESS(0) ",%1 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - MEMOPARG(movzb,0x00,3,1,4,1) " \n" // movzb (%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x4,0) " \n" - "movzb " MEMACCESS2(-0x3,0) ",%1 \n" - MEMOPARG(movzb,0x01,3,1,4,1) " \n" // movzb 0x1(%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x3,0) " \n" - "movzb " MEMACCESS2(-0x2,0) ",%1 \n" - MEMOPARG(movzb,0x02,3,1,4,1) " \n" // movzb 0x2(%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x2,0) " \n" - "movzb " MEMACCESS2(-0x1,0) ",%1 \n" - MEMOPARG(movzb,0x03,3,1,4,1) " \n" // movzb 0x3(%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x1,0) " \n" - "dec %2 \n" - "jg 1b \n" - : "+r"(dst_argb), // %0 - "+d"(pixel_temp), // %1 - "+r"(width) // %2 - : "r"(table_argb) // %3 - : "memory", "cc"); -} -#endif // HAS_ARGBCOLORTABLEROW_X86 - -#ifdef HAS_RGBCOLORTABLEROW_X86 -// Tranform RGB pixels with color table. -void RGBColorTableRow_X86(uint8* dst_argb, const uint8* table_argb, int width) { - uintptr_t pixel_temp = 0u; - asm volatile ( - // 1 pixel loop. - LABELALIGN - "1: \n" - "movzb " MEMACCESS(0) ",%1 \n" - "lea " MEMLEA(0x4,0) ",%0 \n" - MEMOPARG(movzb,0x00,3,1,4,1) " \n" // movzb (%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x4,0) " \n" - "movzb " MEMACCESS2(-0x3,0) ",%1 \n" - MEMOPARG(movzb,0x01,3,1,4,1) " \n" // movzb 0x1(%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x3,0) " \n" - "movzb " MEMACCESS2(-0x2,0) ",%1 \n" - MEMOPARG(movzb,0x02,3,1,4,1) " \n" // movzb 0x2(%3,%1,4),%1 - "mov %b1," MEMACCESS2(-0x2,0) " \n" - "dec %2 \n" - "jg 1b \n" - : "+r"(dst_argb), // %0 - "+d"(pixel_temp), // %1 - "+r"(width) // %2 - : "r"(table_argb) // %3 - : "memory", "cc"); -} -#endif // HAS_RGBCOLORTABLEROW_X86 - -#ifdef HAS_ARGBLUMACOLORTABLEROW_SSSE3 -// Tranform RGB pixels with luma table. -void ARGBLumaColorTableRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - int width, - const uint8* luma, uint32 lumacoeff) { - uintptr_t pixel_temp = 0u; - uintptr_t table_temp = 0u; - asm volatile ( - "movd %6,%%xmm3 \n" - "pshufd $0x0,%%xmm3,%%xmm3 \n" - "pcmpeqb %%xmm4,%%xmm4 \n" - "psllw $0x8,%%xmm4 \n" - "pxor %%xmm5,%%xmm5 \n" - - // 4 pixel loop. - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(2) ",%%xmm0 \n" - "pmaddubsw %%xmm3,%%xmm0 \n" - "phaddw %%xmm0,%%xmm0 \n" - "pand %%xmm4,%%xmm0 \n" - "punpcklwd %%xmm5,%%xmm0 \n" - "movd %%xmm0,%k1 \n" // 32 bit offset - "add %5,%1 \n" - "pshufd $0x39,%%xmm0,%%xmm0 \n" - - "movzb " MEMACCESS(2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS(3) " \n" - "movzb " MEMACCESS2(0x1,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x1,3) " \n" - "movzb " MEMACCESS2(0x2,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x2,3) " \n" - "movzb " MEMACCESS2(0x3,2) ",%0 \n" - "mov %b0," MEMACCESS2(0x3,3) " \n" - - "movd %%xmm0,%k1 \n" // 32 bit offset - "add %5,%1 \n" - "pshufd $0x39,%%xmm0,%%xmm0 \n" - - "movzb " MEMACCESS2(0x4,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x4,3) " \n" - BUNDLEALIGN - "movzb " MEMACCESS2(0x5,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x5,3) " \n" - "movzb " MEMACCESS2(0x6,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x6,3) " \n" - "movzb " MEMACCESS2(0x7,2) ",%0 \n" - "mov %b0," MEMACCESS2(0x7,3) " \n" - - "movd %%xmm0,%k1 \n" // 32 bit offset - "add %5,%1 \n" - "pshufd $0x39,%%xmm0,%%xmm0 \n" - - "movzb " MEMACCESS2(0x8,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x8,3) " \n" - "movzb " MEMACCESS2(0x9,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0x9,3) " \n" - "movzb " MEMACCESS2(0xa,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0xa,3) " \n" - "movzb " MEMACCESS2(0xb,2) ",%0 \n" - "mov %b0," MEMACCESS2(0xb,3) " \n" - - "movd %%xmm0,%k1 \n" // 32 bit offset - "add %5,%1 \n" - - "movzb " MEMACCESS2(0xc,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0xc,3) " \n" - "movzb " MEMACCESS2(0xd,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0xd,3) " \n" - "movzb " MEMACCESS2(0xe,2) ",%0 \n" - MEMOPARG(movzb,0x00,1,0,1,0) " \n" // movzb (%1,%0,1),%0 - "mov %b0," MEMACCESS2(0xe,3) " \n" - "movzb " MEMACCESS2(0xf,2) ",%0 \n" - "mov %b0," MEMACCESS2(0xf,3) " \n" - "sub $0x4,%4 \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "lea " MEMLEA(0x10,3) ",%3 \n" - "jg 1b \n" - : "+d"(pixel_temp), // %0 - "+a"(table_temp), // %1 - "+r"(src_argb), // %2 - "+r"(dst_argb), // %3 - "+rm"(width) // %4 - : "r"(luma), // %5 - "rm"(lumacoeff) // %6 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm3", "xmm4", "xmm5" -#endif - ); -} -#endif // HAS_ARGBLUMACOLORTABLEROW_SSSE3 - -#endif // defined(__x86_64__) || defined(__i386__) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_win.cc b/src/main/jni/libyuv/source/row_win.cc deleted file mode 100644 index f58fc5138..000000000 --- a/src/main/jni/libyuv/source/row_win.cc +++ /dev/null @@ -1,7402 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#if defined (_M_X64) && !defined(LIBYUV_DISABLE_X86) && defined(_MSC_VER) -#include <emmintrin.h> -#include <tmmintrin.h> // For _mm_maddubs_epi16 -#endif - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for Visual C. -#if !defined(LIBYUV_DISABLE_X86) && defined(_MSC_VER) - -#define YG 74 /* (int8)(1.164 * 64 + 0.5) */ - -#define UB 127 /* min(127,(int8)(2.018 * 64)) */ -#define UG -25 /* (int8)(-0.391 * 64 - 0.5) */ -#define UR 0 - -#define VB 0 -#define VG -52 /* (int8)(-0.813 * 64 - 0.5) */ -#define VR 102 /* (int8)(1.596 * 64 + 0.5) */ - -// Bias -#define BB UB * 128 + VB * 128 -#define BG UG * 128 + VG * 128 -#define BR UR * 128 + VR * 128 - -static const vec8 kUVToB = { - UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB -}; - -static const vec8 kUVToR = { - UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR -}; - -static const vec8 kUVToG = { - UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG -}; - -static const vec8 kVUToB = { - VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, -}; - -static const vec8 kVUToR = { - VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, -}; - -static const vec8 kVUToG = { - VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, -}; - -static const vec16 kYToRgb = { YG, YG, YG, YG, YG, YG, YG, YG }; -static const vec16 kYSub16 = { 16, 16, 16, 16, 16, 16, 16, 16 }; -static const vec16 kUVBiasB = { BB, BB, BB, BB, BB, BB, BB, BB }; -static const vec16 kUVBiasG = { BG, BG, BG, BG, BG, BG, BG, BG }; -static const vec16 kUVBiasR = { BR, BR, BR, BR, BR, BR, BR, BR }; - -// 64 bit -#if defined(_M_X64) - -// Aligned destination version. -__declspec(align(16)) -void I422ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __m128i xmm0, xmm1, xmm2, xmm3; - const __m128i xmm5 = _mm_set1_epi8(-1); - const __m128i xmm4 = _mm_setzero_si128(); - const ptrdiff_t offset = (uint8*)v_buf - (uint8*)u_buf; - - while (width > 0) { - xmm0 = _mm_cvtsi32_si128(*(uint32*)u_buf); - xmm1 = _mm_cvtsi32_si128(*(uint32*)(u_buf + offset)); - xmm0 = _mm_unpacklo_epi8(xmm0, xmm1); - xmm0 = _mm_unpacklo_epi16(xmm0, xmm0); - xmm1 = _mm_load_si128(&xmm0); - xmm2 = _mm_load_si128(&xmm0); - xmm0 = _mm_maddubs_epi16(xmm0, *(__m128i*)kUVToB); - xmm1 = _mm_maddubs_epi16(xmm1, *(__m128i*)kUVToG); - xmm2 = _mm_maddubs_epi16(xmm2, *(__m128i*)kUVToR); - xmm0 = _mm_sub_epi16(xmm0, *(__m128i*)kUVBiasB); - xmm1 = _mm_sub_epi16(xmm1, *(__m128i*)kUVBiasG); - xmm2 = _mm_sub_epi16(xmm2, *(__m128i*)kUVBiasR); - xmm3 = _mm_loadl_epi64((__m128i*)y_buf); - xmm3 = _mm_unpacklo_epi8(xmm3, xmm4); - xmm3 = _mm_subs_epi16(xmm3, *(__m128i*)kYSub16); - xmm3 = _mm_mullo_epi16(xmm3, *(__m128i*)kYToRgb); - xmm0 = _mm_adds_epi16(xmm0, xmm3); - xmm1 = _mm_adds_epi16(xmm1, xmm3); - xmm2 = _mm_adds_epi16(xmm2, xmm3); - xmm0 = _mm_srai_epi16(xmm0, 6); - xmm1 = _mm_srai_epi16(xmm1, 6); - xmm2 = _mm_srai_epi16(xmm2, 6); - xmm0 = _mm_packus_epi16(xmm0, xmm0); - xmm1 = _mm_packus_epi16(xmm1, xmm1); - xmm2 = _mm_packus_epi16(xmm2, xmm2); - xmm0 = _mm_unpacklo_epi8(xmm0, xmm1); - xmm2 = _mm_unpacklo_epi8(xmm2, xmm5); - xmm1 = _mm_load_si128(&xmm0); - xmm0 = _mm_unpacklo_epi16(xmm0, xmm2); - xmm1 = _mm_unpackhi_epi16(xmm1, xmm2); - - _mm_store_si128((__m128i *)dst_argb, xmm0); - _mm_store_si128((__m128i *)(dst_argb + 16), xmm1); - - y_buf += 8; - u_buf += 4; - dst_argb += 32; - width -= 8; - } -} - -// Unaligned destination version. -void I422ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __m128i xmm0, xmm1, xmm2, xmm3; - const __m128i xmm5 = _mm_set1_epi8(-1); - const __m128i xmm4 = _mm_setzero_si128(); - const ptrdiff_t offset = (uint8*)v_buf - (uint8*)u_buf; - - while (width > 0) { - xmm0 = _mm_cvtsi32_si128(*(uint32*)u_buf); - xmm1 = _mm_cvtsi32_si128(*(uint32*)(u_buf + offset)); - xmm0 = _mm_unpacklo_epi8(xmm0, xmm1); - xmm0 = _mm_unpacklo_epi16(xmm0, xmm0); - xmm1 = _mm_load_si128(&xmm0); - xmm2 = _mm_load_si128(&xmm0); - xmm0 = _mm_maddubs_epi16(xmm0, *(__m128i*)kUVToB); - xmm1 = _mm_maddubs_epi16(xmm1, *(__m128i*)kUVToG); - xmm2 = _mm_maddubs_epi16(xmm2, *(__m128i*)kUVToR); - xmm0 = _mm_sub_epi16(xmm0, *(__m128i*)kUVBiasB); - xmm1 = _mm_sub_epi16(xmm1, *(__m128i*)kUVBiasG); - xmm2 = _mm_sub_epi16(xmm2, *(__m128i*)kUVBiasR); - xmm3 = _mm_loadl_epi64((__m128i*)y_buf); - xmm3 = _mm_unpacklo_epi8(xmm3, xmm4); - xmm3 = _mm_subs_epi16(xmm3, *(__m128i*)kYSub16); - xmm3 = _mm_mullo_epi16(xmm3, *(__m128i*)kYToRgb); - xmm0 = _mm_adds_epi16(xmm0, xmm3); - xmm1 = _mm_adds_epi16(xmm1, xmm3); - xmm2 = _mm_adds_epi16(xmm2, xmm3); - xmm0 = _mm_srai_epi16(xmm0, 6); - xmm1 = _mm_srai_epi16(xmm1, 6); - xmm2 = _mm_srai_epi16(xmm2, 6); - xmm0 = _mm_packus_epi16(xmm0, xmm0); - xmm1 = _mm_packus_epi16(xmm1, xmm1); - xmm2 = _mm_packus_epi16(xmm2, xmm2); - xmm0 = _mm_unpacklo_epi8(xmm0, xmm1); - xmm2 = _mm_unpacklo_epi8(xmm2, xmm5); - xmm1 = _mm_load_si128(&xmm0); - xmm0 = _mm_unpacklo_epi16(xmm0, xmm2); - xmm1 = _mm_unpackhi_epi16(xmm1, xmm2); - - _mm_storeu_si128((__m128i *)dst_argb, xmm0); - _mm_storeu_si128((__m128i *)(dst_argb + 16), xmm1); - - y_buf += 8; - u_buf += 4; - dst_argb += 32; - width -= 8; - } -} -// 32 bit -#else // defined(_M_X64) - -#ifdef HAS_ARGBTOYROW_SSSE3 - -// Constants for ARGB. -static const vec8 kARGBToY = { - 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33, 0 -}; - -// JPeg full range. -static const vec8 kARGBToYJ = { - 15, 75, 38, 0, 15, 75, 38, 0, 15, 75, 38, 0, 15, 75, 38, 0 -}; - -static const vec8 kARGBToU = { - 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38, 0 -}; - -static const vec8 kARGBToUJ = { - 127, -84, -43, 0, 127, -84, -43, 0, 127, -84, -43, 0, 127, -84, -43, 0 -}; - -static const vec8 kARGBToV = { - -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112, 0, -}; - -static const vec8 kARGBToVJ = { - -20, -107, 127, 0, -20, -107, 127, 0, -20, -107, 127, 0, -20, -107, 127, 0 -}; - -// vpermd for vphaddw + vpackuswb vpermd. -static const lvec32 kPermdARGBToY_AVX = { - 0, 4, 1, 5, 2, 6, 3, 7 -}; - -// vpshufb for vphaddw + vpackuswb packed to shorts. -static const lvec8 kShufARGBToUV_AVX = { - 0, 1, 8, 9, 2, 3, 10, 11, 4, 5, 12, 13, 6, 7, 14, 15, - 0, 1, 8, 9, 2, 3, 10, 11, 4, 5, 12, 13, 6, 7, 14, 15, -}; - -// Constants for BGRA. -static const vec8 kBGRAToY = { - 0, 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13 -}; - -static const vec8 kBGRAToU = { - 0, -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112 -}; - -static const vec8 kBGRAToV = { - 0, 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18 -}; - -// Constants for ABGR. -static const vec8 kABGRToY = { - 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13, 0, 33, 65, 13, 0 -}; - -static const vec8 kABGRToU = { - -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112, 0, -38, -74, 112, 0 -}; - -static const vec8 kABGRToV = { - 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18, 0, 112, -94, -18, 0 -}; - -// Constants for RGBA. -static const vec8 kRGBAToY = { - 0, 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33, 0, 13, 65, 33 -}; - -static const vec8 kRGBAToU = { - 0, 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38, 0, 112, -74, -38 -}; - -static const vec8 kRGBAToV = { - 0, -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112, 0, -18, -94, 112 -}; - -static const uvec8 kAddY16 = { - 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u, 16u -}; - -static const vec16 kAddYJ64 = { - 64, 64, 64, 64, 64, 64, 64, 64 -}; - -static const uvec8 kAddUV128 = { - 128u, 128u, 128u, 128u, 128u, 128u, 128u, 128u, - 128u, 128u, 128u, 128u, 128u, 128u, 128u, 128u -}; - -static const uvec16 kAddUVJ128 = { - 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u, 0x8080u -}; - -// Shuffle table for converting RGB24 to ARGB. -static const uvec8 kShuffleMaskRGB24ToARGB = { - 0u, 1u, 2u, 12u, 3u, 4u, 5u, 13u, 6u, 7u, 8u, 14u, 9u, 10u, 11u, 15u -}; - -// Shuffle table for converting RAW to ARGB. -static const uvec8 kShuffleMaskRAWToARGB = { - 2u, 1u, 0u, 12u, 5u, 4u, 3u, 13u, 8u, 7u, 6u, 14u, 11u, 10u, 9u, 15u -}; - -// Shuffle table for converting ARGB to RGB24. -static const uvec8 kShuffleMaskARGBToRGB24 = { - 0u, 1u, 2u, 4u, 5u, 6u, 8u, 9u, 10u, 12u, 13u, 14u, 128u, 128u, 128u, 128u -}; - -// Shuffle table for converting ARGB to RAW. -static const uvec8 kShuffleMaskARGBToRAW = { - 2u, 1u, 0u, 6u, 5u, 4u, 10u, 9u, 8u, 14u, 13u, 12u, 128u, 128u, 128u, 128u -}; - -// Shuffle table for converting ARGBToRGB24 for I422ToRGB24. First 8 + next 4 -static const uvec8 kShuffleMaskARGBToRGB24_0 = { - 0u, 1u, 2u, 4u, 5u, 6u, 8u, 9u, 128u, 128u, 128u, 128u, 10u, 12u, 13u, 14u -}; - -// Shuffle table for converting ARGB to RAW. -static const uvec8 kShuffleMaskARGBToRAW_0 = { - 2u, 1u, 0u, 6u, 5u, 4u, 10u, 9u, 128u, 128u, 128u, 128u, 8u, 14u, 13u, 12u -}; - -// Duplicates gray value 3 times and fills in alpha opaque. -__declspec(naked) __declspec(align(16)) -void I400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int pix) { - __asm { - mov eax, [esp + 4] // src_y - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - pcmpeqb xmm5, xmm5 // generate mask 0xff000000 - pslld xmm5, 24 - - align 4 - convertloop: - movq xmm0, qword ptr [eax] - lea eax, [eax + 8] - punpcklbw xmm0, xmm0 - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm0 - punpckhwd xmm1, xmm1 - por xmm0, xmm5 - por xmm1, xmm5 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I400ToARGBRow_Unaligned_SSE2(const uint8* src_y, uint8* dst_argb, - int pix) { - __asm { - mov eax, [esp + 4] // src_y - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - pcmpeqb xmm5, xmm5 // generate mask 0xff000000 - pslld xmm5, 24 - - align 4 - convertloop: - movq xmm0, qword ptr [eax] - lea eax, [eax + 8] - punpcklbw xmm0, xmm0 - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm0 - punpckhwd xmm1, xmm1 - por xmm0, xmm5 - por xmm1, xmm5 - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void RGB24ToARGBRow_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int pix) { - __asm { - mov eax, [esp + 4] // src_rgb24 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - pcmpeqb xmm5, xmm5 // generate mask 0xff000000 - pslld xmm5, 24 - movdqa xmm4, kShuffleMaskRGB24ToARGB - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm3, [eax + 32] - lea eax, [eax + 48] - movdqa xmm2, xmm3 - palignr xmm2, xmm1, 8 // xmm2 = { xmm3[0:3] xmm1[8:15]} - pshufb xmm2, xmm4 - por xmm2, xmm5 - palignr xmm1, xmm0, 12 // xmm1 = { xmm3[0:7] xmm0[12:15]} - pshufb xmm0, xmm4 - movdqa [edx + 32], xmm2 - por xmm0, xmm5 - pshufb xmm1, xmm4 - movdqa [edx], xmm0 - por xmm1, xmm5 - palignr xmm3, xmm3, 4 // xmm3 = { xmm3[4:15]} - pshufb xmm3, xmm4 - movdqa [edx + 16], xmm1 - por xmm3, xmm5 - sub ecx, 16 - movdqa [edx + 48], xmm3 - lea edx, [edx + 64] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void RAWToARGBRow_SSSE3(const uint8* src_raw, uint8* dst_argb, - int pix) { - __asm { - mov eax, [esp + 4] // src_raw - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - pcmpeqb xmm5, xmm5 // generate mask 0xff000000 - pslld xmm5, 24 - movdqa xmm4, kShuffleMaskRAWToARGB - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm3, [eax + 32] - lea eax, [eax + 48] - movdqa xmm2, xmm3 - palignr xmm2, xmm1, 8 // xmm2 = { xmm3[0:3] xmm1[8:15]} - pshufb xmm2, xmm4 - por xmm2, xmm5 - palignr xmm1, xmm0, 12 // xmm1 = { xmm3[0:7] xmm0[12:15]} - pshufb xmm0, xmm4 - movdqa [edx + 32], xmm2 - por xmm0, xmm5 - pshufb xmm1, xmm4 - movdqa [edx], xmm0 - por xmm1, xmm5 - palignr xmm3, xmm3, 4 // xmm3 = { xmm3[4:15]} - pshufb xmm3, xmm4 - movdqa [edx + 16], xmm1 - por xmm3, xmm5 - sub ecx, 16 - movdqa [edx + 48], xmm3 - lea edx, [edx + 64] - jg convertloop - ret - } -} - -// pmul method to replicate bits. -// Math to replicate bits: -// (v << 8) | (v << 3) -// v * 256 + v * 8 -// v * (256 + 8) -// G shift of 5 is incorporated, so shift is 5 + 8 and 5 + 3 -// 20 instructions. -__declspec(naked) __declspec(align(16)) -void RGB565ToARGBRow_SSE2(const uint8* src_rgb565, uint8* dst_argb, - int pix) { - __asm { - mov eax, 0x01080108 // generate multiplier to repeat 5 bits - movd xmm5, eax - pshufd xmm5, xmm5, 0 - mov eax, 0x20802080 // multiplier shift by 5 and then repeat 6 bits - movd xmm6, eax - pshufd xmm6, xmm6, 0 - pcmpeqb xmm3, xmm3 // generate mask 0xf800f800 for Red - psllw xmm3, 11 - pcmpeqb xmm4, xmm4 // generate mask 0x07e007e0 for Green - psllw xmm4, 10 - psrlw xmm4, 5 - pcmpeqb xmm7, xmm7 // generate mask 0xff00ff00 for Alpha - psllw xmm7, 8 - - mov eax, [esp + 4] // src_rgb565 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - sub edx, eax - sub edx, eax - - align 4 - convertloop: - movdqu xmm0, [eax] // fetch 8 pixels of bgr565 - movdqa xmm1, xmm0 - movdqa xmm2, xmm0 - pand xmm1, xmm3 // R in upper 5 bits - psllw xmm2, 11 // B in upper 5 bits - pmulhuw xmm1, xmm5 // * (256 + 8) - pmulhuw xmm2, xmm5 // * (256 + 8) - psllw xmm1, 8 - por xmm1, xmm2 // RB - pand xmm0, xmm4 // G in middle 6 bits - pmulhuw xmm0, xmm6 // << 5 * (256 + 4) - por xmm0, xmm7 // AG - movdqa xmm2, xmm1 - punpcklbw xmm1, xmm0 - punpckhbw xmm2, xmm0 - movdqa [eax * 2 + edx], xmm1 // store 4 pixels of ARGB - movdqa [eax * 2 + edx + 16], xmm2 // store next 4 pixels of ARGB - lea eax, [eax + 16] - sub ecx, 8 - jg convertloop - ret - } -} - -// 24 instructions -__declspec(naked) __declspec(align(16)) -void ARGB1555ToARGBRow_SSE2(const uint8* src_argb1555, uint8* dst_argb, - int pix) { - __asm { - mov eax, 0x01080108 // generate multiplier to repeat 5 bits - movd xmm5, eax - pshufd xmm5, xmm5, 0 - mov eax, 0x42004200 // multiplier shift by 6 and then repeat 5 bits - movd xmm6, eax - pshufd xmm6, xmm6, 0 - pcmpeqb xmm3, xmm3 // generate mask 0xf800f800 for Red - psllw xmm3, 11 - movdqa xmm4, xmm3 // generate mask 0x03e003e0 for Green - psrlw xmm4, 6 - pcmpeqb xmm7, xmm7 // generate mask 0xff00ff00 for Alpha - psllw xmm7, 8 - - mov eax, [esp + 4] // src_argb1555 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - sub edx, eax - sub edx, eax - - align 4 - convertloop: - movdqu xmm0, [eax] // fetch 8 pixels of 1555 - movdqa xmm1, xmm0 - movdqa xmm2, xmm0 - psllw xmm1, 1 // R in upper 5 bits - psllw xmm2, 11 // B in upper 5 bits - pand xmm1, xmm3 - pmulhuw xmm2, xmm5 // * (256 + 8) - pmulhuw xmm1, xmm5 // * (256 + 8) - psllw xmm1, 8 - por xmm1, xmm2 // RB - movdqa xmm2, xmm0 - pand xmm0, xmm4 // G in middle 5 bits - psraw xmm2, 8 // A - pmulhuw xmm0, xmm6 // << 6 * (256 + 8) - pand xmm2, xmm7 - por xmm0, xmm2 // AG - movdqa xmm2, xmm1 - punpcklbw xmm1, xmm0 - punpckhbw xmm2, xmm0 - movdqa [eax * 2 + edx], xmm1 // store 4 pixels of ARGB - movdqa [eax * 2 + edx + 16], xmm2 // store next 4 pixels of ARGB - lea eax, [eax + 16] - sub ecx, 8 - jg convertloop - ret - } -} - -// 18 instructions. -__declspec(naked) __declspec(align(16)) -void ARGB4444ToARGBRow_SSE2(const uint8* src_argb4444, uint8* dst_argb, - int pix) { - __asm { - mov eax, 0x0f0f0f0f // generate mask 0x0f0f0f0f - movd xmm4, eax - pshufd xmm4, xmm4, 0 - movdqa xmm5, xmm4 // 0xf0f0f0f0 for high nibbles - pslld xmm5, 4 - mov eax, [esp + 4] // src_argb4444 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // pix - sub edx, eax - sub edx, eax - - align 4 - convertloop: - movdqu xmm0, [eax] // fetch 8 pixels of bgra4444 - movdqa xmm2, xmm0 - pand xmm0, xmm4 // mask low nibbles - pand xmm2, xmm5 // mask high nibbles - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - psllw xmm1, 4 - psrlw xmm3, 4 - por xmm0, xmm1 - por xmm2, xmm3 - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm2 - punpckhbw xmm1, xmm2 - movdqa [eax * 2 + edx], xmm0 // store 4 pixels of ARGB - movdqa [eax * 2 + edx + 16], xmm1 // store next 4 pixels of ARGB - lea eax, [eax + 16] - sub ecx, 8 - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToRGB24Row_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_rgb - mov ecx, [esp + 12] // pix - movdqa xmm6, kShuffleMaskARGBToRGB24 - - align 4 - convertloop: - movdqu xmm0, [eax] // fetch 16 pixels of argb - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - lea eax, [eax + 64] - pshufb xmm0, xmm6 // pack 16 bytes of ARGB to 12 bytes of RGB - pshufb xmm1, xmm6 - pshufb xmm2, xmm6 - pshufb xmm3, xmm6 - movdqa xmm4, xmm1 // 4 bytes from 1 for 0 - psrldq xmm1, 4 // 8 bytes from 1 - pslldq xmm4, 12 // 4 bytes from 1 for 0 - movdqa xmm5, xmm2 // 8 bytes from 2 for 1 - por xmm0, xmm4 // 4 bytes from 1 for 0 - pslldq xmm5, 8 // 8 bytes from 2 for 1 - movdqu [edx], xmm0 // store 0 - por xmm1, xmm5 // 8 bytes from 2 for 1 - psrldq xmm2, 8 // 4 bytes from 2 - pslldq xmm3, 4 // 12 bytes from 3 for 2 - por xmm2, xmm3 // 12 bytes from 3 for 2 - movdqu [edx + 16], xmm1 // store 1 - movdqu [edx + 32], xmm2 // store 2 - lea edx, [edx + 48] - sub ecx, 16 - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToRAWRow_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_rgb - mov ecx, [esp + 12] // pix - movdqa xmm6, kShuffleMaskARGBToRAW - - align 4 - convertloop: - movdqu xmm0, [eax] // fetch 16 pixels of argb - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - lea eax, [eax + 64] - pshufb xmm0, xmm6 // pack 16 bytes of ARGB to 12 bytes of RGB - pshufb xmm1, xmm6 - pshufb xmm2, xmm6 - pshufb xmm3, xmm6 - movdqa xmm4, xmm1 // 4 bytes from 1 for 0 - psrldq xmm1, 4 // 8 bytes from 1 - pslldq xmm4, 12 // 4 bytes from 1 for 0 - movdqa xmm5, xmm2 // 8 bytes from 2 for 1 - por xmm0, xmm4 // 4 bytes from 1 for 0 - pslldq xmm5, 8 // 8 bytes from 2 for 1 - movdqu [edx], xmm0 // store 0 - por xmm1, xmm5 // 8 bytes from 2 for 1 - psrldq xmm2, 8 // 4 bytes from 2 - pslldq xmm3, 4 // 12 bytes from 3 for 2 - por xmm2, xmm3 // 12 bytes from 3 for 2 - movdqu [edx + 16], xmm1 // store 1 - movdqu [edx + 32], xmm2 // store 2 - lea edx, [edx + 48] - sub ecx, 16 - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToRGB565Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_rgb - mov ecx, [esp + 12] // pix - pcmpeqb xmm3, xmm3 // generate mask 0x0000001f - psrld xmm3, 27 - pcmpeqb xmm4, xmm4 // generate mask 0x000007e0 - psrld xmm4, 26 - pslld xmm4, 5 - pcmpeqb xmm5, xmm5 // generate mask 0xfffff800 - pslld xmm5, 11 - - align 4 - convertloop: - movdqa xmm0, [eax] // fetch 4 pixels of argb - movdqa xmm1, xmm0 // B - movdqa xmm2, xmm0 // G - pslld xmm0, 8 // R - psrld xmm1, 3 // B - psrld xmm2, 5 // G - psrad xmm0, 16 // R - pand xmm1, xmm3 // B - pand xmm2, xmm4 // G - pand xmm0, xmm5 // R - por xmm1, xmm2 // BG - por xmm0, xmm1 // BGR - packssdw xmm0, xmm0 - lea eax, [eax + 16] - movq qword ptr [edx], xmm0 // store 4 pixels of RGB565 - lea edx, [edx + 8] - sub ecx, 4 - jg convertloop - ret - } -} - -// TODO(fbarchard): Improve sign extension/packing. -__declspec(naked) __declspec(align(16)) -void ARGBToARGB1555Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_rgb - mov ecx, [esp + 12] // pix - pcmpeqb xmm4, xmm4 // generate mask 0x0000001f - psrld xmm4, 27 - movdqa xmm5, xmm4 // generate mask 0x000003e0 - pslld xmm5, 5 - movdqa xmm6, xmm4 // generate mask 0x00007c00 - pslld xmm6, 10 - pcmpeqb xmm7, xmm7 // generate mask 0xffff8000 - pslld xmm7, 15 - - align 4 - convertloop: - movdqa xmm0, [eax] // fetch 4 pixels of argb - movdqa xmm1, xmm0 // B - movdqa xmm2, xmm0 // G - movdqa xmm3, xmm0 // R - psrad xmm0, 16 // A - psrld xmm1, 3 // B - psrld xmm2, 6 // G - psrld xmm3, 9 // R - pand xmm0, xmm7 // A - pand xmm1, xmm4 // B - pand xmm2, xmm5 // G - pand xmm3, xmm6 // R - por xmm0, xmm1 // BA - por xmm2, xmm3 // GR - por xmm0, xmm2 // BGRA - packssdw xmm0, xmm0 - lea eax, [eax + 16] - movq qword ptr [edx], xmm0 // store 4 pixels of ARGB1555 - lea edx, [edx + 8] - sub ecx, 4 - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToARGB4444Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_rgb - mov ecx, [esp + 12] // pix - pcmpeqb xmm4, xmm4 // generate mask 0xf000f000 - psllw xmm4, 12 - movdqa xmm3, xmm4 // generate mask 0x00f000f0 - psrlw xmm3, 8 - - align 4 - convertloop: - movdqa xmm0, [eax] // fetch 4 pixels of argb - movdqa xmm1, xmm0 - pand xmm0, xmm3 // low nibble - pand xmm1, xmm4 // high nibble - psrl xmm0, 4 - psrl xmm1, 8 - por xmm0, xmm1 - packuswb xmm0, xmm0 - lea eax, [eax + 16] - movq qword ptr [edx], xmm0 // store 4 pixels of ARGB4444 - lea edx, [edx + 8] - sub ecx, 4 - jg convertloop - ret - } -} - -// Convert 16 ARGB pixels (64 bytes) to 16 Y values. -__declspec(naked) __declspec(align(16)) -void ARGBToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kARGBToY - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -// Convert 16 ARGB pixels (64 bytes) to 16 Y values. -__declspec(naked) __declspec(align(16)) -void ARGBToYJRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm4, kARGBToYJ - movdqa xmm5, kAddYJ64 - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - paddw xmm0, xmm5 // Add .5 for rounding. - paddw xmm2, xmm5 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -#ifdef HAS_ARGBTOYROW_AVX2 -// Convert 32 ARGB pixels (128 bytes) to 32 Y values. -__declspec(naked) __declspec(align(32)) -void ARGBToYRow_AVX2(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - vbroadcastf128 ymm4, kARGBToY - vbroadcastf128 ymm5, kAddY16 - vmovdqa ymm6, kPermdARGBToY_AVX - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - vmovdqu ymm2, [eax + 64] - vmovdqu ymm3, [eax + 96] - vpmaddubsw ymm0, ymm0, ymm4 - vpmaddubsw ymm1, ymm1, ymm4 - vpmaddubsw ymm2, ymm2, ymm4 - vpmaddubsw ymm3, ymm3, ymm4 - lea eax, [eax + 128] - vphaddw ymm0, ymm0, ymm1 // mutates. - vphaddw ymm2, ymm2, ymm3 - vpsrlw ymm0, ymm0, 7 - vpsrlw ymm2, ymm2, 7 - vpackuswb ymm0, ymm0, ymm2 // mutates. - vpermd ymm0, ymm6, ymm0 // For vphaddw + vpackuswb mutation. - vpaddb ymm0, ymm0, ymm5 - sub ecx, 32 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - jg convertloop - vzeroupper - ret - } -} -#endif // HAS_ARGBTOYROW_AVX2 - -#ifdef HAS_ARGBTOYROW_AVX2 -// Convert 32 ARGB pixels (128 bytes) to 32 Y values. -__declspec(naked) __declspec(align(32)) -void ARGBToYJRow_AVX2(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - vbroadcastf128 ymm4, kARGBToYJ - vbroadcastf128 ymm5, kAddYJ64 - vmovdqa ymm6, kPermdARGBToY_AVX - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - vmovdqu ymm2, [eax + 64] - vmovdqu ymm3, [eax + 96] - vpmaddubsw ymm0, ymm0, ymm4 - vpmaddubsw ymm1, ymm1, ymm4 - vpmaddubsw ymm2, ymm2, ymm4 - vpmaddubsw ymm3, ymm3, ymm4 - lea eax, [eax + 128] - vphaddw ymm0, ymm0, ymm1 // mutates. - vphaddw ymm2, ymm2, ymm3 - vpaddw ymm0, ymm0, ymm5 // Add .5 for rounding. - vpaddw ymm2, ymm2, ymm5 - vpsrlw ymm0, ymm0, 7 - vpsrlw ymm2, ymm2, 7 - vpackuswb ymm0, ymm0, ymm2 // mutates. - vpermd ymm0, ymm6, ymm0 // For vphaddw + vpackuswb mutation. - sub ecx, 32 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - jg convertloop - - vzeroupper - ret - } -} -#endif // HAS_ARGBTOYJROW_AVX2 - -__declspec(naked) __declspec(align(16)) -void ARGBToYRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kARGBToY - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToYJRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm4, kARGBToYJ - movdqa xmm5, kAddYJ64 - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - paddw xmm0, xmm5 - paddw xmm2, xmm5 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void BGRAToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kBGRAToY - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void BGRAToYRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kBGRAToY - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ABGRToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kABGRToY - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ABGRToYRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kABGRToY - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void RGBAToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kRGBAToY - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void RGBAToYRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_y */ - mov ecx, [esp + 12] /* pix */ - movdqa xmm5, kAddY16 - movdqa xmm4, kRGBAToY - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - pmaddubsw xmm2, xmm4 - pmaddubsw xmm3, xmm4 - lea eax, [eax + 64] - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psrlw xmm0, 7 - psrlw xmm2, 7 - packuswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUVRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kARGBToU - movdqa xmm6, kARGBToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pavgb xmm0, [eax + esi] - pavgb xmm1, [eax + esi + 16] - pavgb xmm2, [eax + esi + 32] - pavgb xmm3, [eax + esi + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUVJRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kARGBToUJ - movdqa xmm6, kARGBToVJ - movdqa xmm5, kAddUVJ128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pavgb xmm0, [eax + esi] - pavgb xmm1, [eax + esi + 16] - pavgb xmm2, [eax + esi + 32] - pavgb xmm3, [eax + esi + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - paddw xmm0, xmm5 // +.5 rounding -> unsigned - paddw xmm1, xmm5 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -#ifdef HAS_ARGBTOUVROW_AVX2 -__declspec(naked) __declspec(align(32)) -void ARGBToUVRow_AVX2(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - vbroadcastf128 ymm5, kAddUV128 - vbroadcastf128 ymm6, kARGBToV - vbroadcastf128 ymm7, kARGBToU - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 32x2 argb pixels to 16x1 */ - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - vmovdqu ymm2, [eax + 64] - vmovdqu ymm3, [eax + 96] - vpavgb ymm0, ymm0, [eax + esi] - vpavgb ymm1, ymm1, [eax + esi + 32] - vpavgb ymm2, ymm2, [eax + esi + 64] - vpavgb ymm3, ymm3, [eax + esi + 96] - lea eax, [eax + 128] - vshufps ymm4, ymm0, ymm1, 0x88 - vshufps ymm0, ymm0, ymm1, 0xdd - vpavgb ymm0, ymm0, ymm4 // mutated by vshufps - vshufps ymm4, ymm2, ymm3, 0x88 - vshufps ymm2, ymm2, ymm3, 0xdd - vpavgb ymm2, ymm2, ymm4 // mutated by vshufps - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 32 different pixels, its 16 pixels of U and 16 of V - vpmaddubsw ymm1, ymm0, ymm7 // U - vpmaddubsw ymm3, ymm2, ymm7 - vpmaddubsw ymm0, ymm0, ymm6 // V - vpmaddubsw ymm2, ymm2, ymm6 - vphaddw ymm1, ymm1, ymm3 // mutates - vphaddw ymm0, ymm0, ymm2 - vpsraw ymm1, ymm1, 8 - vpsraw ymm0, ymm0, 8 - vpacksswb ymm0, ymm1, ymm0 // mutates - vpermq ymm0, ymm0, 0xd8 // For vpacksswb - vpshufb ymm0, ymm0, kShufARGBToUV_AVX // For vshufps + vphaddw - vpaddb ymm0, ymm0, ymm5 // -> unsigned - - // step 3 - store 16 U and 16 V values - sub ecx, 32 - vextractf128 [edx], ymm0, 0 // U - vextractf128 [edx + edi], ymm0, 1 // V - lea edx, [edx + 16] - jg convertloop - - pop edi - pop esi - vzeroupper - ret - } -} -#endif // HAS_ARGBTOUVROW_AVX2 - -__declspec(naked) __declspec(align(16)) -void ARGBToUVRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kARGBToU - movdqa xmm6, kARGBToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - movdqu xmm4, [eax + esi] - pavgb xmm0, xmm4 - movdqu xmm4, [eax + esi + 16] - pavgb xmm1, xmm4 - movdqu xmm4, [eax + esi + 32] - pavgb xmm2, xmm4 - movdqu xmm4, [eax + esi + 48] - pavgb xmm3, xmm4 - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUVJRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kARGBToUJ - movdqa xmm6, kARGBToVJ - movdqa xmm5, kAddUVJ128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - movdqu xmm4, [eax + esi] - pavgb xmm0, xmm4 - movdqu xmm4, [eax + esi + 16] - pavgb xmm1, xmm4 - movdqu xmm4, [eax + esi + 32] - pavgb xmm2, xmm4 - movdqu xmm4, [eax + esi + 48] - pavgb xmm3, xmm4 - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - paddw xmm0, xmm5 // +.5 rounding -> unsigned - paddw xmm1, xmm5 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUV444Row_SSSE3(const uint8* src_argb0, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_argb - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - movdqa xmm7, kARGBToU - movdqa xmm6, kARGBToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* convert to U and V */ - movdqa xmm0, [eax] // U - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm7 - pmaddubsw xmm1, xmm7 - pmaddubsw xmm2, xmm7 - pmaddubsw xmm3, xmm7 - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psraw xmm0, 8 - psraw xmm2, 8 - packsswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm0 - - movdqa xmm0, [eax] // V - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pmaddubsw xmm0, xmm6 - pmaddubsw xmm1, xmm6 - pmaddubsw xmm2, xmm6 - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psraw xmm0, 8 - psraw xmm2, 8 - packsswb xmm0, xmm2 - paddb xmm0, xmm5 - lea eax, [eax + 64] - movdqa [edx + edi], xmm0 - lea edx, [edx + 16] - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUV444Row_Unaligned_SSSE3(const uint8* src_argb0, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_argb - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - movdqa xmm7, kARGBToU - movdqa xmm6, kARGBToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* convert to U and V */ - movdqu xmm0, [eax] // U - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm7 - pmaddubsw xmm1, xmm7 - pmaddubsw xmm2, xmm7 - pmaddubsw xmm3, xmm7 - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psraw xmm0, 8 - psraw xmm2, 8 - packsswb xmm0, xmm2 - paddb xmm0, xmm5 - sub ecx, 16 - movdqu [edx], xmm0 - - movdqu xmm0, [eax] // V - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - pmaddubsw xmm0, xmm6 - pmaddubsw xmm1, xmm6 - pmaddubsw xmm2, xmm6 - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm1 - phaddw xmm2, xmm3 - psraw xmm0, 8 - psraw xmm2, 8 - packsswb xmm0, xmm2 - paddb xmm0, xmm5 - lea eax, [eax + 64] - movdqu [edx + edi], xmm0 - lea edx, [edx + 16] - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUV422Row_SSSE3(const uint8* src_argb0, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_argb - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - movdqa xmm7, kARGBToU - movdqa xmm6, kARGBToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBToUV422Row_Unaligned_SSSE3(const uint8* src_argb0, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_argb - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - movdqa xmm7, kARGBToU - movdqa xmm6, kARGBToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void BGRAToUVRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kBGRAToU - movdqa xmm6, kBGRAToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pavgb xmm0, [eax + esi] - pavgb xmm1, [eax + esi + 16] - pavgb xmm2, [eax + esi + 32] - pavgb xmm3, [eax + esi + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void BGRAToUVRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kBGRAToU - movdqa xmm6, kBGRAToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - movdqu xmm4, [eax + esi] - pavgb xmm0, xmm4 - movdqu xmm4, [eax + esi + 16] - pavgb xmm1, xmm4 - movdqu xmm4, [eax + esi + 32] - pavgb xmm2, xmm4 - movdqu xmm4, [eax + esi + 48] - pavgb xmm3, xmm4 - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ABGRToUVRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kABGRToU - movdqa xmm6, kABGRToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pavgb xmm0, [eax + esi] - pavgb xmm1, [eax + esi + 16] - pavgb xmm2, [eax + esi + 32] - pavgb xmm3, [eax + esi + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ABGRToUVRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kABGRToU - movdqa xmm6, kABGRToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - movdqu xmm4, [eax + esi] - pavgb xmm0, xmm4 - movdqu xmm4, [eax + esi + 16] - pavgb xmm1, xmm4 - movdqu xmm4, [eax + esi + 32] - pavgb xmm2, xmm4 - movdqu xmm4, [eax + esi + 48] - pavgb xmm3, xmm4 - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void RGBAToUVRow_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kRGBAToU - movdqa xmm6, kRGBAToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - pavgb xmm0, [eax + esi] - pavgb xmm1, [eax + esi + 16] - pavgb xmm2, [eax + esi + 32] - pavgb xmm3, [eax + esi + 48] - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void RGBAToUVRow_Unaligned_SSSE3(const uint8* src_argb0, int src_stride_argb, - uint8* dst_u, uint8* dst_v, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb - mov esi, [esp + 8 + 8] // src_stride_argb - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - movdqa xmm7, kRGBAToU - movdqa xmm6, kRGBAToV - movdqa xmm5, kAddUV128 - sub edi, edx // stride from u to v - - align 4 - convertloop: - /* step 1 - subsample 16x2 argb pixels to 8x1 */ - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + 32] - movdqu xmm3, [eax + 48] - movdqu xmm4, [eax + esi] - pavgb xmm0, xmm4 - movdqu xmm4, [eax + esi + 16] - pavgb xmm1, xmm4 - movdqu xmm4, [eax + esi + 32] - pavgb xmm2, xmm4 - movdqu xmm4, [eax + esi + 48] - pavgb xmm3, xmm4 - lea eax, [eax + 64] - movdqa xmm4, xmm0 - shufps xmm0, xmm1, 0x88 - shufps xmm4, xmm1, 0xdd - pavgb xmm0, xmm4 - movdqa xmm4, xmm2 - shufps xmm2, xmm3, 0x88 - shufps xmm4, xmm3, 0xdd - pavgb xmm2, xmm4 - - // step 2 - convert to U and V - // from here down is very similar to Y code except - // instead of 16 different pixels, its 8 pixels of U and 8 of V - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - pmaddubsw xmm0, xmm7 // U - pmaddubsw xmm2, xmm7 - pmaddubsw xmm1, xmm6 // V - pmaddubsw xmm3, xmm6 - phaddw xmm0, xmm2 - phaddw xmm1, xmm3 - psraw xmm0, 8 - psraw xmm1, 8 - packsswb xmm0, xmm1 - paddb xmm0, xmm5 // -> unsigned - - // step 3 - store 8 U and 8 V values - sub ecx, 16 - movlps qword ptr [edx], xmm0 // U - movhps qword ptr [edx + edi], xmm0 // V - lea edx, [edx + 8] - jg convertloop - - pop edi - pop esi - ret - } -} -#endif // HAS_ARGBTOYROW_SSSE3 - -#ifdef HAS_I422TOARGBROW_AVX2 - -static const lvec8 kUVToB_AVX = { - UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, - UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB, UB, VB -}; -static const lvec8 kUVToR_AVX = { - UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, - UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR, UR, VR -}; -static const lvec8 kUVToG_AVX = { - UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, - UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG, UG, VG -}; -static const lvec16 kYToRgb_AVX = { - YG, YG, YG, YG, YG, YG, YG, YG, YG, YG, YG, YG, YG, YG, YG, YG -}; -static const lvec16 kYSub16_AVX = { - 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16 -}; -static const lvec16 kUVBiasB_AVX = { - BB, BB, BB, BB, BB, BB, BB, BB, BB, BB, BB, BB, BB, BB, BB, BB -}; -static const lvec16 kUVBiasG_AVX = { - BG, BG, BG, BG, BG, BG, BG, BG, BG, BG, BG, BG, BG, BG, BG, BG -}; -static const lvec16 kUVBiasR_AVX = { - BR, BR, BR, BR, BR, BR, BR, BR, BR, BR, BR, BR, BR, BR, BR, BR -}; - -// 16 pixels -// 8 UV values upsampled to 16 UV, mixed with 16 Y producing 16 ARGB (64 bytes). -__declspec(naked) __declspec(align(16)) -void I422ToARGBRow_AVX2(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // argb - mov ecx, [esp + 8 + 20] // width - sub edi, esi - vpcmpeqb ymm5, ymm5, ymm5 // generate 0xffffffffffffffff for alpha - vpxor ymm4, ymm4, ymm4 - - align 4 - convertloop: - vmovq xmm0, qword ptr [esi] // U - vmovq xmm1, qword ptr [esi + edi] // V - lea esi, [esi + 8] - vpunpcklbw ymm0, ymm0, ymm1 // UV - vpermq ymm0, ymm0, 0xd8 - vpunpcklwd ymm0, ymm0, ymm0 // UVUV - vpmaddubsw ymm2, ymm0, kUVToB_AVX // scale B UV - vpmaddubsw ymm1, ymm0, kUVToG_AVX // scale G UV - vpmaddubsw ymm0, ymm0, kUVToR_AVX // scale R UV - vpsubw ymm2, ymm2, kUVBiasB_AVX // unbias back to signed - vpsubw ymm1, ymm1, kUVBiasG_AVX - vpsubw ymm0, ymm0, kUVBiasR_AVX - - // Step 2: Find Y contribution to 16 R,G,B values - vmovdqu xmm3, [eax] // NOLINT - lea eax, [eax + 16] - vpermq ymm3, ymm3, 0xd8 - vpunpcklbw ymm3, ymm3, ymm4 - vpsubsw ymm3, ymm3, kYSub16_AVX - vpmullw ymm3, ymm3, kYToRgb_AVX - vpaddsw ymm2, ymm2, ymm3 // B += Y - vpaddsw ymm1, ymm1, ymm3 // G += Y - vpaddsw ymm0, ymm0, ymm3 // R += Y - vpsraw ymm2, ymm2, 6 - vpsraw ymm1, ymm1, 6 - vpsraw ymm0, ymm0, 6 - vpackuswb ymm2, ymm2, ymm2 // B - vpackuswb ymm1, ymm1, ymm1 // G - vpackuswb ymm0, ymm0, ymm0 // R - - // Step 3: Weave into ARGB - vpunpcklbw ymm2, ymm2, ymm1 // BG - vpermq ymm2, ymm2, 0xd8 - vpunpcklbw ymm0, ymm0, ymm5 // RA - vpermq ymm0, ymm0, 0xd8 - vpunpcklwd ymm1, ymm2, ymm0 // BGRA first 8 pixels - vpunpckhwd ymm2, ymm2, ymm0 // BGRA next 8 pixels - vmovdqu [edx], ymm1 - vmovdqu [edx + 32], ymm2 - lea edx, [edx + 64] - sub ecx, 16 - jg convertloop - vzeroupper - - pop edi - pop esi - ret - } -} -#endif // HAS_I422TOARGBROW_AVX2 - -#ifdef HAS_I422TOARGBROW_SSSE3 - -// TODO(fbarchard): Read that does half size on Y and treats 420 as 444. - -// Read 8 UV from 444. -#define READYUV444 __asm { \ - __asm movq xmm0, qword ptr [esi] /* U */ /* NOLINT */ \ - __asm movq xmm1, qword ptr [esi + edi] /* V */ /* NOLINT */ \ - __asm lea esi, [esi + 8] \ - __asm punpcklbw xmm0, xmm1 /* UV */ \ - } - -// Read 4 UV from 422, upsample to 8 UV. -#define READYUV422 __asm { \ - __asm movd xmm0, [esi] /* U */ \ - __asm movd xmm1, [esi + edi] /* V */ \ - __asm lea esi, [esi + 4] \ - __asm punpcklbw xmm0, xmm1 /* UV */ \ - __asm punpcklwd xmm0, xmm0 /* UVUV (upsample) */ \ - } - -// Read 2 UV from 411, upsample to 8 UV. -#define READYUV411 __asm { \ - __asm movzx ebx, word ptr [esi] /* U */ /* NOLINT */ \ - __asm movd xmm0, ebx \ - __asm movzx ebx, word ptr [esi + edi] /* V */ /* NOLINT */ \ - __asm movd xmm1, ebx \ - __asm lea esi, [esi + 2] \ - __asm punpcklbw xmm0, xmm1 /* UV */ \ - __asm punpcklwd xmm0, xmm0 /* UVUV (upsample) */ \ - __asm punpckldq xmm0, xmm0 /* UVUV (upsample) */ \ - } - -// Read 4 UV from NV12, upsample to 8 UV. -#define READNV12 __asm { \ - __asm movq xmm0, qword ptr [esi] /* UV */ /* NOLINT */ \ - __asm lea esi, [esi + 8] \ - __asm punpcklwd xmm0, xmm0 /* UVUV (upsample) */ \ - } - -// Convert 8 pixels: 8 UV and 8 Y. -#define YUVTORGB __asm { \ - /* Step 1: Find 4 UV contributions to 8 R,G,B values */ \ - __asm movdqa xmm1, xmm0 \ - __asm movdqa xmm2, xmm0 \ - __asm pmaddubsw xmm0, kUVToB /* scale B UV */ \ - __asm pmaddubsw xmm1, kUVToG /* scale G UV */ \ - __asm pmaddubsw xmm2, kUVToR /* scale R UV */ \ - __asm psubw xmm0, kUVBiasB /* unbias back to signed */ \ - __asm psubw xmm1, kUVBiasG \ - __asm psubw xmm2, kUVBiasR \ - /* Step 2: Find Y contribution to 8 R,G,B values */ \ - __asm movq xmm3, qword ptr [eax] /* NOLINT */ \ - __asm lea eax, [eax + 8] \ - __asm punpcklbw xmm3, xmm4 \ - __asm psubsw xmm3, kYSub16 \ - __asm pmullw xmm3, kYToRgb \ - __asm paddsw xmm0, xmm3 /* B += Y */ \ - __asm paddsw xmm1, xmm3 /* G += Y */ \ - __asm paddsw xmm2, xmm3 /* R += Y */ \ - __asm psraw xmm0, 6 \ - __asm psraw xmm1, 6 \ - __asm psraw xmm2, 6 \ - __asm packuswb xmm0, xmm0 /* B */ \ - __asm packuswb xmm1, xmm1 /* G */ \ - __asm packuswb xmm2, xmm2 /* R */ \ - } - -// Convert 8 pixels: 8 VU and 8 Y. -#define YVUTORGB __asm { \ - /* Step 1: Find 4 UV contributions to 8 R,G,B values */ \ - __asm movdqa xmm1, xmm0 \ - __asm movdqa xmm2, xmm0 \ - __asm pmaddubsw xmm0, kVUToB /* scale B UV */ \ - __asm pmaddubsw xmm1, kVUToG /* scale G UV */ \ - __asm pmaddubsw xmm2, kVUToR /* scale R UV */ \ - __asm psubw xmm0, kUVBiasB /* unbias back to signed */ \ - __asm psubw xmm1, kUVBiasG \ - __asm psubw xmm2, kUVBiasR \ - /* Step 2: Find Y contribution to 8 R,G,B values */ \ - __asm movq xmm3, qword ptr [eax] /* NOLINT */ \ - __asm lea eax, [eax + 8] \ - __asm punpcklbw xmm3, xmm4 \ - __asm psubsw xmm3, kYSub16 \ - __asm pmullw xmm3, kYToRgb \ - __asm paddsw xmm0, xmm3 /* B += Y */ \ - __asm paddsw xmm1, xmm3 /* G += Y */ \ - __asm paddsw xmm2, xmm3 /* R += Y */ \ - __asm psraw xmm0, 6 \ - __asm psraw xmm1, 6 \ - __asm psraw xmm2, 6 \ - __asm packuswb xmm0, xmm0 /* B */ \ - __asm packuswb xmm1, xmm1 /* G */ \ - __asm packuswb xmm2, xmm2 /* R */ \ - } - -// 8 pixels, dest aligned 16. -// 8 UV values, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I444ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // argb - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV444 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I422ToRGB24Row_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_rgb24, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // rgb24 - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - movdqa xmm5, kShuffleMaskARGBToRGB24_0 - movdqa xmm6, kShuffleMaskARGBToRGB24 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into RRGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm2 // RR - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRR first 4 pixels - punpckhwd xmm1, xmm2 // BGRR next 4 pixels - pshufb xmm0, xmm5 // Pack into first 8 and last 4 bytes. - pshufb xmm1, xmm6 // Pack into first 12 bytes. - palignr xmm1, xmm0, 12 // last 4 bytes of xmm0 + 12 from xmm1 - movq qword ptr [edx], xmm0 // First 8 bytes - movdqu [edx + 8], xmm1 // Last 16 bytes. = 24 bytes, 8 RGB pixels. - lea edx, [edx + 24] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I422ToRAWRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_raw, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // raw - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - movdqa xmm5, kShuffleMaskARGBToRAW_0 - movdqa xmm6, kShuffleMaskARGBToRAW - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into RRGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm2 // RR - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRR first 4 pixels - punpckhwd xmm1, xmm2 // BGRR next 4 pixels - pshufb xmm0, xmm5 // Pack into first 8 and last 4 bytes. - pshufb xmm1, xmm6 // Pack into first 12 bytes. - palignr xmm1, xmm0, 12 // last 4 bytes of xmm0 + 12 from xmm1 - movq qword ptr [edx], xmm0 // First 8 bytes - movdqu [edx + 8], xmm1 // Last 16 bytes. = 24 bytes, 8 RGB pixels. - lea edx, [edx + 24] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, dest unaligned. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I422ToRGB565Row_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb565_buf, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // rgb565 - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - pcmpeqb xmm5, xmm5 // generate mask 0x0000001f - psrld xmm5, 27 - pcmpeqb xmm6, xmm6 // generate mask 0x000007e0 - psrld xmm6, 26 - pslld xmm6, 5 - pcmpeqb xmm7, xmm7 // generate mask 0xfffff800 - pslld xmm7, 11 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into RRGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm2 // RR - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRR first 4 pixels - punpckhwd xmm1, xmm2 // BGRR next 4 pixels - - // Step 3b: RRGB -> RGB565 - movdqa xmm3, xmm0 // B first 4 pixels of argb - movdqa xmm2, xmm0 // G - pslld xmm0, 8 // R - psrld xmm3, 3 // B - psrld xmm2, 5 // G - psrad xmm0, 16 // R - pand xmm3, xmm5 // B - pand xmm2, xmm6 // G - pand xmm0, xmm7 // R - por xmm3, xmm2 // BG - por xmm0, xmm3 // BGR - movdqa xmm3, xmm1 // B next 4 pixels of argb - movdqa xmm2, xmm1 // G - pslld xmm1, 8 // R - psrld xmm3, 3 // B - psrld xmm2, 5 // G - psrad xmm1, 16 // R - pand xmm3, xmm5 // B - pand xmm2, xmm6 // G - pand xmm1, xmm7 // R - por xmm3, xmm2 // BG - por xmm1, xmm3 // BGR - packssdw xmm0, xmm1 - sub ecx, 8 - movdqu [edx], xmm0 // store 8 pixels of RGB565 - lea edx, [edx + 16] - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I422ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // argb - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, dest aligned 16. -// 2 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -// Similar to I420 but duplicate UV once more. -__declspec(naked) __declspec(align(16)) -void I411ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push ebx - push esi - push edi - mov eax, [esp + 12 + 4] // Y - mov esi, [esp + 12 + 8] // U - mov edi, [esp + 12 + 12] // V - mov edx, [esp + 12 + 16] // argb - mov ecx, [esp + 12 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV411 // modifies EBX - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - pop ebx - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void NV12ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // Y - mov esi, [esp + 4 + 8] // UV - mov edx, [esp + 4 + 12] // argb - mov ecx, [esp + 4 + 16] // width - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READNV12 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void NV21ToARGBRow_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // Y - mov esi, [esp + 4 + 8] // VU - mov edx, [esp + 4 + 12] // argb - mov ecx, [esp + 4 + 16] // width - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READNV12 - YVUTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - ret - } -} - -// 8 pixels, unaligned. -// 8 UV values, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I444ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // argb - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV444 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, unaligned. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void I422ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // argb - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -// 8 pixels, unaligned. -// 2 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -// Similar to I420 but duplicate UV once more. -__declspec(naked) __declspec(align(16)) -void I411ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_argb, - int width) { - __asm { - push ebx - push esi - push edi - mov eax, [esp + 12 + 4] // Y - mov esi, [esp + 12 + 8] // U - mov edi, [esp + 12 + 12] // V - mov edx, [esp + 12 + 16] // argb - mov ecx, [esp + 12 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV411 // modifies EBX - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - pop ebx - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void NV12ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // Y - mov esi, [esp + 4 + 8] // UV - mov edx, [esp + 4 + 12] // argb - mov ecx, [esp + 4 + 16] // width - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READNV12 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - ret - } -} - -// 8 pixels, dest aligned 16. -// 4 UV values upsampled to 8 UV, mixed with 8 Y producing 8 ARGB (32 bytes). -__declspec(naked) __declspec(align(16)) -void NV21ToARGBRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* uv_buf, - uint8* dst_argb, - int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // Y - mov esi, [esp + 4 + 8] // VU - mov edx, [esp + 4 + 12] // argb - mov ecx, [esp + 4 + 16] // width - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READNV12 - YVUTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm0, xmm1 // BG - punpcklbw xmm2, xmm5 // RA - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm2 // BGRA first 4 pixels - punpckhwd xmm1, xmm2 // BGRA next 4 pixels - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToBGRARow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_bgra, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // bgra - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into BGRA - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - punpcklbw xmm1, xmm0 // GB - punpcklbw xmm5, xmm2 // AR - movdqa xmm0, xmm5 - punpcklwd xmm5, xmm1 // BGRA first 4 pixels - punpckhwd xmm0, xmm1 // BGRA next 4 pixels - movdqa [edx], xmm5 - movdqa [edx + 16], xmm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToBGRARow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_bgra, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // bgra - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into BGRA - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - punpcklbw xmm1, xmm0 // GB - punpcklbw xmm5, xmm2 // AR - movdqa xmm0, xmm5 - punpcklwd xmm5, xmm1 // BGRA first 4 pixels - punpckhwd xmm0, xmm1 // BGRA next 4 pixels - movdqu [edx], xmm5 - movdqu [edx + 16], xmm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToABGRRow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_abgr, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // abgr - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm2, xmm1 // RG - punpcklbw xmm0, xmm5 // BA - movdqa xmm1, xmm2 - punpcklwd xmm2, xmm0 // RGBA first 4 pixels - punpckhwd xmm1, xmm0 // RGBA next 4 pixels - movdqa [edx], xmm2 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToABGRRow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_abgr, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // abgr - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into ARGB - punpcklbw xmm2, xmm1 // RG - punpcklbw xmm0, xmm5 // BA - movdqa xmm1, xmm2 - punpcklwd xmm2, xmm0 // RGBA first 4 pixels - punpckhwd xmm1, xmm0 // RGBA next 4 pixels - movdqu [edx], xmm2 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToRGBARow_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_rgba, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // rgba - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into RGBA - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - punpcklbw xmm1, xmm2 // GR - punpcklbw xmm5, xmm0 // AB - movdqa xmm0, xmm5 - punpcklwd xmm5, xmm1 // RGBA first 4 pixels - punpckhwd xmm0, xmm1 // RGBA next 4 pixels - movdqa [edx], xmm5 - movdqa [edx + 16], xmm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToRGBARow_Unaligned_SSSE3(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* dst_rgba, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // Y - mov esi, [esp + 8 + 8] // U - mov edi, [esp + 8 + 12] // V - mov edx, [esp + 8 + 16] // rgba - mov ecx, [esp + 8 + 20] // width - sub edi, esi - pxor xmm4, xmm4 - - align 4 - convertloop: - READYUV422 - YUVTORGB - - // Step 3: Weave into RGBA - pcmpeqb xmm5, xmm5 // generate 0xffffffff for alpha - punpcklbw xmm1, xmm2 // GR - punpcklbw xmm5, xmm0 // AB - movdqa xmm0, xmm5 - punpcklwd xmm5, xmm1 // RGBA first 4 pixels - punpckhwd xmm0, xmm1 // RGBA next 4 pixels - movdqu [edx], xmm5 - movdqu [edx + 16], xmm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop edi - pop esi - ret - } -} - -#endif // HAS_I422TOARGBROW_SSSE3 - -#ifdef HAS_YTOARGBROW_SSE2 -__declspec(naked) __declspec(align(16)) -void YToARGBRow_SSE2(const uint8* y_buf, - uint8* rgb_buf, - int width) { - __asm { - pxor xmm5, xmm5 - pcmpeqb xmm4, xmm4 // generate mask 0xff000000 - pslld xmm4, 24 - mov eax, 0x00100010 - movd xmm3, eax - pshufd xmm3, xmm3, 0 - mov eax, 0x004a004a // 74 - movd xmm2, eax - pshufd xmm2, xmm2,0 - mov eax, [esp + 4] // Y - mov edx, [esp + 8] // rgb - mov ecx, [esp + 12] // width - - align 4 - convertloop: - // Step 1: Scale Y contribution to 8 G values. G = (y - 16) * 1.164 - movq xmm0, qword ptr [eax] - lea eax, [eax + 8] - punpcklbw xmm0, xmm5 // 0.Y - psubusw xmm0, xmm3 - pmullw xmm0, xmm2 - psrlw xmm0, 6 - packuswb xmm0, xmm0 // G - - // Step 2: Weave into ARGB - punpcklbw xmm0, xmm0 // GG - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm0 // BGRA first 4 pixels - punpckhwd xmm1, xmm1 // BGRA next 4 pixels - por xmm0, xmm4 - por xmm1, xmm4 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - ret - } -} -#endif // HAS_YTOARGBROW_SSE2 - -#ifdef HAS_MIRRORROW_SSSE3 -// Shuffle table for reversing the bytes. -static const uvec8 kShuffleMirror = { - 15u, 14u, 13u, 12u, 11u, 10u, 9u, 8u, 7u, 6u, 5u, 4u, 3u, 2u, 1u, 0u -}; - -__declspec(naked) __declspec(align(16)) -void MirrorRow_SSSE3(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // width - movdqa xmm5, kShuffleMirror - lea eax, [eax - 16] - - align 4 - convertloop: - movdqa xmm0, [eax + ecx] - pshufb xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} -#endif // HAS_MIRRORROW_SSSE3 - -#ifdef HAS_MIRRORROW_AVX2 -// Shuffle table for reversing the bytes. -static const ulvec8 kShuffleMirror_AVX2 = { - 15u, 14u, 13u, 12u, 11u, 10u, 9u, 8u, 7u, 6u, 5u, 4u, 3u, 2u, 1u, 0u, - 15u, 14u, 13u, 12u, 11u, 10u, 9u, 8u, 7u, 6u, 5u, 4u, 3u, 2u, 1u, 0u -}; - -__declspec(naked) __declspec(align(16)) -void MirrorRow_AVX2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // width - vmovdqa ymm5, kShuffleMirror_AVX2 - lea eax, [eax - 32] - - align 4 - convertloop: - vmovdqu ymm0, [eax + ecx] - vpshufb ymm0, ymm0, ymm5 - vpermq ymm0, ymm0, 0x4e // swap high and low halfs - sub ecx, 32 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - jg convertloop - vzeroupper - ret - } -} -#endif // HAS_MIRRORROW_AVX2 - -#ifdef HAS_MIRRORROW_SSE2 -// SSE2 version has movdqu so it can be used on unaligned buffers when SSSE3 -// version can not. -__declspec(naked) __declspec(align(16)) -void MirrorRow_SSE2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // width - lea eax, [eax - 16] - - align 4 - convertloop: - movdqu xmm0, [eax + ecx] - movdqa xmm1, xmm0 // swap bytes - psllw xmm0, 8 - psrlw xmm1, 8 - por xmm0, xmm1 - pshuflw xmm0, xmm0, 0x1b // swap words - pshufhw xmm0, xmm0, 0x1b - pshufd xmm0, xmm0, 0x4e // swap qwords - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} -#endif // HAS_MIRRORROW_SSE2 - -#ifdef HAS_MIRRORROW_UV_SSSE3 -// Shuffle table for reversing the bytes of UV channels. -static const uvec8 kShuffleMirrorUV = { - 14u, 12u, 10u, 8u, 6u, 4u, 2u, 0u, 15u, 13u, 11u, 9u, 7u, 5u, 3u, 1u -}; - -__declspec(naked) __declspec(align(16)) -void MirrorUVRow_SSSE3(const uint8* src, uint8* dst_u, uint8* dst_v, - int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // width - movdqa xmm1, kShuffleMirrorUV - lea eax, [eax + ecx * 2 - 16] - sub edi, edx - - align 4 - convertloop: - movdqa xmm0, [eax] - lea eax, [eax - 16] - pshufb xmm0, xmm1 - sub ecx, 8 - movlpd qword ptr [edx], xmm0 - movhpd qword ptr [edx + edi], xmm0 - lea edx, [edx + 8] - jg convertloop - - pop edi - ret - } -} -#endif // HAS_MIRRORROW_UV_SSSE3 - -#ifdef HAS_ARGBMIRRORROW_SSSE3 -// Shuffle table for reversing the bytes. -static const uvec8 kARGBShuffleMirror = { - 12u, 13u, 14u, 15u, 8u, 9u, 10u, 11u, 4u, 5u, 6u, 7u, 0u, 1u, 2u, 3u -}; - -__declspec(naked) __declspec(align(16)) -void ARGBMirrorRow_SSSE3(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // width - lea eax, [eax - 16 + ecx * 4] // last 4 pixels. - movdqa xmm5, kARGBShuffleMirror - - align 4 - convertloop: - movdqa xmm0, [eax] - lea eax, [eax - 16] - pshufb xmm0, xmm5 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} -#endif // HAS_ARGBMIRRORROW_SSSE3 - -#ifdef HAS_ARGBMIRRORROW_AVX2 -// Shuffle table for reversing the bytes. -static const ulvec32 kARGBShuffleMirror_AVX2 = { - 7u, 6u, 5u, 4u, 3u, 2u, 1u, 0u -}; - -__declspec(naked) __declspec(align(16)) -void ARGBMirrorRow_AVX2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // width - lea eax, [eax - 32] - vmovdqa ymm5, kARGBShuffleMirror_AVX2 - - align 4 - convertloop: - vpermd ymm0, ymm5, [eax + ecx * 4] // permute dword order - sub ecx, 8 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - jg convertloop - vzeroupper - ret - } -} -#endif // HAS_ARGBMIRRORROW_AVX2 - -#ifdef HAS_SPLITUVROW_SSE2 -__declspec(naked) __declspec(align(16)) -void SplitUVRow_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_uv - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - movdqa xmm2, xmm0 - movdqa xmm3, xmm1 - pand xmm0, xmm5 // even bytes - pand xmm1, xmm5 - packuswb xmm0, xmm1 - psrlw xmm2, 8 // odd bytes - psrlw xmm3, 8 - packuswb xmm2, xmm3 - movdqa [edx], xmm0 - movdqa [edx + edi], xmm2 - lea edx, [edx + 16] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void SplitUVRow_Unaligned_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, - int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_uv - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - movdqa xmm2, xmm0 - movdqa xmm3, xmm1 - pand xmm0, xmm5 // even bytes - pand xmm1, xmm5 - packuswb xmm0, xmm1 - psrlw xmm2, 8 // odd bytes - psrlw xmm3, 8 - packuswb xmm2, xmm3 - movdqu [edx], xmm0 - movdqu [edx + edi], xmm2 - lea edx, [edx + 16] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} -#endif // HAS_SPLITUVROW_SSE2 - -#ifdef HAS_SPLITUVROW_AVX2 -__declspec(naked) __declspec(align(16)) -void SplitUVRow_AVX2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_uv - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0x00ff00ff - vpsrlw ymm5, ymm5, 8 - sub edi, edx - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - lea eax, [eax + 64] - vpsrlw ymm2, ymm0, 8 // odd bytes - vpsrlw ymm3, ymm1, 8 - vpand ymm0, ymm0, ymm5 // even bytes - vpand ymm1, ymm1, ymm5 - vpackuswb ymm0, ymm0, ymm1 - vpackuswb ymm2, ymm2, ymm3 - vpermq ymm0, ymm0, 0xd8 - vpermq ymm2, ymm2, 0xd8 - vmovdqu [edx], ymm0 - vmovdqu [edx + edi], ymm2 - lea edx, [edx + 32] - sub ecx, 32 - jg convertloop - - pop edi - vzeroupper - ret - } -} -#endif // HAS_SPLITUVROW_AVX2 - -#ifdef HAS_MERGEUVROW_SSE2 -__declspec(naked) __declspec(align(16)) -void MergeUVRow_SSE2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_u - mov edx, [esp + 4 + 8] // src_v - mov edi, [esp + 4 + 12] // dst_uv - mov ecx, [esp + 4 + 16] // width - sub edx, eax - - align 4 - convertloop: - movdqa xmm0, [eax] // read 16 U's - movdqa xmm1, [eax + edx] // and 16 V's - lea eax, [eax + 16] - movdqa xmm2, xmm0 - punpcklbw xmm0, xmm1 // first 8 UV pairs - punpckhbw xmm2, xmm1 // next 8 UV pairs - movdqa [edi], xmm0 - movdqa [edi + 16], xmm2 - lea edi, [edi + 32] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void MergeUVRow_Unaligned_SSE2(const uint8* src_u, const uint8* src_v, - uint8* dst_uv, int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_u - mov edx, [esp + 4 + 8] // src_v - mov edi, [esp + 4 + 12] // dst_uv - mov ecx, [esp + 4 + 16] // width - sub edx, eax - - align 4 - convertloop: - movdqu xmm0, [eax] // read 16 U's - movdqu xmm1, [eax + edx] // and 16 V's - lea eax, [eax + 16] - movdqa xmm2, xmm0 - punpcklbw xmm0, xmm1 // first 8 UV pairs - punpckhbw xmm2, xmm1 // next 8 UV pairs - movdqu [edi], xmm0 - movdqu [edi + 16], xmm2 - lea edi, [edi + 32] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} -#endif // HAS_MERGEUVROW_SSE2 - -#ifdef HAS_MERGEUVROW_AVX2 -__declspec(naked) __declspec(align(16)) -void MergeUVRow_AVX2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, - int width) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_u - mov edx, [esp + 4 + 8] // src_v - mov edi, [esp + 4 + 12] // dst_uv - mov ecx, [esp + 4 + 16] // width - sub edx, eax - - align 4 - convertloop: - vmovdqu ymm0, [eax] // read 32 U's - vmovdqu ymm1, [eax + edx] // and 32 V's - lea eax, [eax + 32] - vpunpcklbw ymm2, ymm0, ymm1 // low 16 UV pairs. mutated qqword 0,2 - vpunpckhbw ymm0, ymm0, ymm1 // high 16 UV pairs. mutated qqword 1,3 - vperm2i128 ymm1, ymm2, ymm0, 0x20 // low 128 of ymm2 and low 128 of ymm0 - vperm2i128 ymm2, ymm2, ymm0, 0x31 // high 128 of ymm2 and high 128 of ymm0 - vmovdqu [edi], ymm1 - vmovdqu [edi + 32], ymm2 - lea edi, [edi + 64] - sub ecx, 32 - jg convertloop - - pop edi - vzeroupper - ret - } -} -#endif // HAS_MERGEUVROW_AVX2 - -#ifdef HAS_COPYROW_SSE2 -// CopyRow copys 'count' bytes using a 16 byte load/store, 32 bytes at time. -__declspec(naked) __declspec(align(16)) -void CopyRow_SSE2(const uint8* src, uint8* dst, int count) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // count - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - sub ecx, 32 - jg convertloop - ret - } -} -#endif // HAS_COPYROW_SSE2 - -// Unaligned Multiple of 1. -__declspec(naked) __declspec(align(16)) -void CopyRow_ERMS(const uint8* src, uint8* dst, int count) { - __asm { - mov eax, esi - mov edx, edi - mov esi, [esp + 4] // src - mov edi, [esp + 8] // dst - mov ecx, [esp + 12] // count - rep movsb - mov edi, edx - mov esi, eax - ret - } -} - -#ifdef HAS_COPYROW_X86 -__declspec(naked) __declspec(align(16)) -void CopyRow_X86(const uint8* src, uint8* dst, int count) { - __asm { - mov eax, esi - mov edx, edi - mov esi, [esp + 4] // src - mov edi, [esp + 8] // dst - mov ecx, [esp + 12] // count - shr ecx, 2 - rep movsd - mov edi, edx - mov esi, eax - ret - } -} -#endif // HAS_COPYROW_X86 - -#ifdef HAS_ARGBCOPYALPHAROW_SSE2 -// width in pixels -__declspec(naked) __declspec(align(16)) -void ARGBCopyAlphaRow_SSE2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // count - pcmpeqb xmm0, xmm0 // generate mask 0xff000000 - pslld xmm0, 24 - pcmpeqb xmm1, xmm1 // generate mask 0x00ffffff - psrld xmm1, 8 - - align 4 - convertloop: - movdqa xmm2, [eax] - movdqa xmm3, [eax + 16] - lea eax, [eax + 32] - movdqa xmm4, [edx] - movdqa xmm5, [edx + 16] - pand xmm2, xmm0 - pand xmm3, xmm0 - pand xmm4, xmm1 - pand xmm5, xmm1 - por xmm2, xmm4 - por xmm3, xmm5 - movdqa [edx], xmm2 - movdqa [edx + 16], xmm3 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - ret - } -} -#endif // HAS_ARGBCOPYALPHAROW_SSE2 - -#ifdef HAS_ARGBCOPYALPHAROW_AVX2 -// width in pixels -__declspec(naked) __declspec(align(16)) -void ARGBCopyAlphaRow_AVX2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // count - vpcmpeqb ymm0, ymm0, ymm0 - vpsrld ymm0, ymm0, 8 // generate mask 0x00ffffff - - align 4 - convertloop: - vmovdqu ymm1, [eax] - vmovdqu ymm2, [eax + 32] - lea eax, [eax + 64] - vpblendvb ymm1, ymm1, [edx], ymm0 - vpblendvb ymm2, ymm2, [edx + 32], ymm0 - vmovdqu [edx], ymm1 - vmovdqu [edx + 32], ymm2 - lea edx, [edx + 64] - sub ecx, 16 - jg convertloop - - vzeroupper - ret - } -} -#endif // HAS_ARGBCOPYALPHAROW_AVX2 - -#ifdef HAS_ARGBCOPYYTOALPHAROW_SSE2 -// width in pixels -__declspec(naked) __declspec(align(16)) -void ARGBCopyYToAlphaRow_SSE2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // count - pcmpeqb xmm0, xmm0 // generate mask 0xff000000 - pslld xmm0, 24 - pcmpeqb xmm1, xmm1 // generate mask 0x00ffffff - psrld xmm1, 8 - - align 4 - convertloop: - movq xmm2, qword ptr [eax] // 8 Y's - lea eax, [eax + 8] - punpcklbw xmm2, xmm2 - punpckhwd xmm3, xmm2 - punpcklwd xmm2, xmm2 - movdqa xmm4, [edx] - movdqa xmm5, [edx + 16] - pand xmm2, xmm0 - pand xmm3, xmm0 - pand xmm4, xmm1 - pand xmm5, xmm1 - por xmm2, xmm4 - por xmm3, xmm5 - movdqa [edx], xmm2 - movdqa [edx + 16], xmm3 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - ret - } -} -#endif // HAS_ARGBCOPYYTOALPHAROW_SSE2 - -#ifdef HAS_ARGBCOPYYTOALPHAROW_AVX2 -// width in pixels -__declspec(naked) __declspec(align(16)) -void ARGBCopyYToAlphaRow_AVX2(const uint8* src, uint8* dst, int width) { - __asm { - mov eax, [esp + 4] // src - mov edx, [esp + 8] // dst - mov ecx, [esp + 12] // count - vpcmpeqb ymm0, ymm0, ymm0 - vpsrld ymm0, ymm0, 8 // generate mask 0x00ffffff - - align 4 - convertloop: - vpmovzxbd ymm1, qword ptr [eax] - vpmovzxbd ymm2, qword ptr [eax + 8] - lea eax, [eax + 16] - vpslld ymm1, ymm1, 24 - vpslld ymm2, ymm2, 24 - vpblendvb ymm1, ymm1, [edx], ymm0 - vpblendvb ymm2, ymm2, [edx + 32], ymm0 - vmovdqu [edx], ymm1 - vmovdqu [edx + 32], ymm2 - lea edx, [edx + 64] - sub ecx, 16 - jg convertloop - - vzeroupper - ret - } -} -#endif // HAS_ARGBCOPYYTOALPHAROW_AVX2 - -#ifdef HAS_SETROW_X86 -// SetRow8 writes 'count' bytes using a 32 bit value repeated. -__declspec(naked) __declspec(align(16)) -void SetRow_X86(uint8* dst, uint32 v32, int count) { - __asm { - mov edx, edi - mov edi, [esp + 4] // dst - mov eax, [esp + 8] // v32 - mov ecx, [esp + 12] // count - shr ecx, 2 - rep stosd - mov edi, edx - ret - } -} - -// SetRow32 writes 'count' words using a 32 bit value repeated. -__declspec(naked) __declspec(align(16)) -void ARGBSetRows_X86(uint8* dst, uint32 v32, int width, - int dst_stride, int height) { - __asm { - push esi - push edi - push ebp - mov edi, [esp + 12 + 4] // dst - mov eax, [esp + 12 + 8] // v32 - mov ebp, [esp + 12 + 12] // width - mov edx, [esp + 12 + 16] // dst_stride - mov esi, [esp + 12 + 20] // height - lea ecx, [ebp * 4] - sub edx, ecx // stride - width * 4 - - align 4 - convertloop: - mov ecx, ebp - rep stosd - add edi, edx - sub esi, 1 - jg convertloop - - pop ebp - pop edi - pop esi - ret - } -} -#endif // HAS_SETROW_X86 - -#ifdef HAS_YUY2TOYROW_AVX2 -__declspec(naked) __declspec(align(16)) -void YUY2ToYRow_AVX2(const uint8* src_yuy2, - uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] // src_yuy2 - mov edx, [esp + 8] // dst_y - mov ecx, [esp + 12] // pix - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0x00ff00ff - vpsrlw ymm5, ymm5, 8 - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - lea eax, [eax + 64] - vpand ymm0, ymm0, ymm5 // even bytes are Y - vpand ymm1, ymm1, ymm5 - vpackuswb ymm0, ymm0, ymm1 // mutates. - vpermq ymm0, ymm0, 0xd8 - sub ecx, 32 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - jg convertloop - vzeroupper - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToUVRow_AVX2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_yuy2 - mov esi, [esp + 8 + 8] // stride_yuy2 - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0x00ff00ff - vpsrlw ymm5, ymm5, 8 - sub edi, edx - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - vpavgb ymm0, ymm0, [eax + esi] - vpavgb ymm1, ymm1, [eax + esi + 32] - lea eax, [eax + 64] - vpsrlw ymm0, ymm0, 8 // YUYV -> UVUV - vpsrlw ymm1, ymm1, 8 - vpackuswb ymm0, ymm0, ymm1 // mutates. - vpermq ymm0, ymm0, 0xd8 - vpand ymm1, ymm0, ymm5 // U - vpsrlw ymm0, ymm0, 8 // V - vpackuswb ymm1, ymm1, ymm1 // mutates. - vpackuswb ymm0, ymm0, ymm0 // mutates. - vpermq ymm1, ymm1, 0xd8 - vpermq ymm0, ymm0, 0xd8 - vextractf128 [edx], ymm1, 0 // U - vextractf128 [edx + edi], ymm0, 0 // V - lea edx, [edx + 16] - sub ecx, 32 - jg convertloop - - pop edi - pop esi - vzeroupper - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToUV422Row_AVX2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_yuy2 - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0x00ff00ff - vpsrlw ymm5, ymm5, 8 - sub edi, edx - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - lea eax, [eax + 64] - vpsrlw ymm0, ymm0, 8 // YUYV -> UVUV - vpsrlw ymm1, ymm1, 8 - vpackuswb ymm0, ymm0, ymm1 // mutates. - vpermq ymm0, ymm0, 0xd8 - vpand ymm1, ymm0, ymm5 // U - vpsrlw ymm0, ymm0, 8 // V - vpackuswb ymm1, ymm1, ymm1 // mutates. - vpackuswb ymm0, ymm0, ymm0 // mutates. - vpermq ymm1, ymm1, 0xd8 - vpermq ymm0, ymm0, 0xd8 - vextractf128 [edx], ymm1, 0 // U - vextractf128 [edx + edi], ymm0, 0 // V - lea edx, [edx + 16] - sub ecx, 32 - jg convertloop - - pop edi - vzeroupper - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToYRow_AVX2(const uint8* src_uyvy, - uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] // src_uyvy - mov edx, [esp + 8] // dst_y - mov ecx, [esp + 12] // pix - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - lea eax, [eax + 64] - vpsrlw ymm0, ymm0, 8 // odd bytes are Y - vpsrlw ymm1, ymm1, 8 - vpackuswb ymm0, ymm0, ymm1 // mutates. - vpermq ymm0, ymm0, 0xd8 - sub ecx, 32 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - jg convertloop - ret - vzeroupper - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToUVRow_AVX2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_yuy2 - mov esi, [esp + 8 + 8] // stride_yuy2 - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0x00ff00ff - vpsrlw ymm5, ymm5, 8 - sub edi, edx - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - vpavgb ymm0, ymm0, [eax + esi] - vpavgb ymm1, ymm1, [eax + esi + 32] - lea eax, [eax + 64] - vpand ymm0, ymm0, ymm5 // UYVY -> UVUV - vpand ymm1, ymm1, ymm5 - vpackuswb ymm0, ymm0, ymm1 // mutates. - vpermq ymm0, ymm0, 0xd8 - vpand ymm1, ymm0, ymm5 // U - vpsrlw ymm0, ymm0, 8 // V - vpackuswb ymm1, ymm1, ymm1 // mutates. - vpackuswb ymm0, ymm0, ymm0 // mutates. - vpermq ymm1, ymm1, 0xd8 - vpermq ymm0, ymm0, 0xd8 - vextractf128 [edx], ymm1, 0 // U - vextractf128 [edx + edi], ymm0, 0 // V - lea edx, [edx + 16] - sub ecx, 32 - jg convertloop - - pop edi - pop esi - vzeroupper - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToUV422Row_AVX2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_yuy2 - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0x00ff00ff - vpsrlw ymm5, ymm5, 8 - sub edi, edx - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - lea eax, [eax + 64] - vpand ymm0, ymm0, ymm5 // UYVY -> UVUV - vpand ymm1, ymm1, ymm5 - vpackuswb ymm0, ymm0, ymm1 // mutates. - vpermq ymm0, ymm0, 0xd8 - vpand ymm1, ymm0, ymm5 // U - vpsrlw ymm0, ymm0, 8 // V - vpackuswb ymm1, ymm1, ymm1 // mutates. - vpackuswb ymm0, ymm0, ymm0 // mutates. - vpermq ymm1, ymm1, 0xd8 - vpermq ymm0, ymm0, 0xd8 - vextractf128 [edx], ymm1, 0 // U - vextractf128 [edx + edi], ymm0, 0 // V - lea edx, [edx + 16] - sub ecx, 32 - jg convertloop - - pop edi - vzeroupper - ret - } -} -#endif // HAS_YUY2TOYROW_AVX2 - -#ifdef HAS_YUY2TOYROW_SSE2 -__declspec(naked) __declspec(align(16)) -void YUY2ToYRow_SSE2(const uint8* src_yuy2, - uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] // src_yuy2 - mov edx, [esp + 8] // dst_y - mov ecx, [esp + 12] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - pand xmm0, xmm5 // even bytes are Y - pand xmm1, xmm5 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToUVRow_SSE2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_yuy2 - mov esi, [esp + 8 + 8] // stride_yuy2 - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + esi] - movdqa xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 - pavgb xmm1, xmm3 - psrlw xmm0, 8 // YUYV -> UVUV - psrlw xmm1, 8 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToUV422Row_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_yuy2 - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - psrlw xmm0, 8 // YUYV -> UVUV - psrlw xmm1, 8 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToYRow_Unaligned_SSE2(const uint8* src_yuy2, - uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] // src_yuy2 - mov edx, [esp + 8] // dst_y - mov ecx, [esp + 12] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - pand xmm0, xmm5 // even bytes are Y - pand xmm1, xmm5 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToUVRow_Unaligned_SSE2(const uint8* src_yuy2, int stride_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_yuy2 - mov esi, [esp + 8 + 8] // stride_yuy2 - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + esi] - movdqu xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 - pavgb xmm1, xmm3 - psrlw xmm0, 8 // YUYV -> UVUV - psrlw xmm1, 8 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void YUY2ToUV422Row_Unaligned_SSE2(const uint8* src_yuy2, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_yuy2 - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - psrlw xmm0, 8 // YUYV -> UVUV - psrlw xmm1, 8 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToYRow_SSE2(const uint8* src_uyvy, - uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] // src_uyvy - mov edx, [esp + 8] // dst_y - mov ecx, [esp + 12] // pix - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - psrlw xmm0, 8 // odd bytes are Y - psrlw xmm1, 8 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToUVRow_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_yuy2 - mov esi, [esp + 8 + 8] // stride_yuy2 - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + esi] - movdqa xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 - pavgb xmm1, xmm3 - pand xmm0, xmm5 // UYVY -> UVUV - pand xmm1, xmm5 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToUV422Row_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_yuy2 - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - pand xmm0, xmm5 // UYVY -> UVUV - pand xmm1, xmm5 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToYRow_Unaligned_SSE2(const uint8* src_uyvy, - uint8* dst_y, int pix) { - __asm { - mov eax, [esp + 4] // src_uyvy - mov edx, [esp + 8] // dst_y - mov ecx, [esp + 12] // pix - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - psrlw xmm0, 8 // odd bytes are Y - psrlw xmm1, 8 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToUVRow_Unaligned_SSE2(const uint8* src_uyvy, int stride_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_yuy2 - mov esi, [esp + 8 + 8] // stride_yuy2 - mov edx, [esp + 8 + 12] // dst_u - mov edi, [esp + 8 + 16] // dst_v - mov ecx, [esp + 8 + 20] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + esi] - movdqu xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 - pavgb xmm1, xmm3 - pand xmm0, xmm5 // UYVY -> UVUV - pand xmm1, xmm5 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void UYVYToUV422Row_Unaligned_SSE2(const uint8* src_uyvy, - uint8* dst_u, uint8* dst_v, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_yuy2 - mov edx, [esp + 4 + 8] // dst_u - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - sub edi, edx - - align 4 - convertloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - pand xmm0, xmm5 // UYVY -> UVUV - pand xmm1, xmm5 - packuswb xmm0, xmm1 - movdqa xmm1, xmm0 - pand xmm0, xmm5 // U - packuswb xmm0, xmm0 - psrlw xmm1, 8 // V - packuswb xmm1, xmm1 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + edi], xmm1 - lea edx, [edx + 8] - sub ecx, 16 - jg convertloop - - pop edi - ret - } -} -#endif // HAS_YUY2TOYROW_SSE2 - -#ifdef HAS_ARGBBLENDROW_SSE2 -// Blend 8 pixels at a time. -__declspec(naked) __declspec(align(16)) -void ARGBBlendRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - pcmpeqb xmm7, xmm7 // generate constant 1 - psrlw xmm7, 15 - pcmpeqb xmm6, xmm6 // generate mask 0x00ff00ff - psrlw xmm6, 8 - pcmpeqb xmm5, xmm5 // generate mask 0xff00ff00 - psllw xmm5, 8 - pcmpeqb xmm4, xmm4 // generate mask 0xff000000 - pslld xmm4, 24 - - sub ecx, 1 - je convertloop1 // only 1 pixel? - jl convertloop1b - - // 1 pixel loop until destination pointer is aligned. - alignloop1: - test edx, 15 // aligned? - je alignloop1b - movd xmm3, [eax] - lea eax, [eax + 4] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movd xmm2, [esi] // _r_b - psrlw xmm3, 8 // alpha - pshufhw xmm3, xmm3, 0F5h // 8 alpha words - pshuflw xmm3, xmm3, 0F5h - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movd xmm1, [esi] // _a_g - lea esi, [esi + 4] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 1 - movd [edx], xmm0 - lea edx, [edx + 4] - jge alignloop1 - - alignloop1b: - add ecx, 1 - 4 - jl convertloop4b - - // 4 pixel loop. - convertloop4: - movdqu xmm3, [eax] // src argb - lea eax, [eax + 16] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movdqu xmm2, [esi] // _r_b - psrlw xmm3, 8 // alpha - pshufhw xmm3, xmm3, 0F5h // 8 alpha words - pshuflw xmm3, xmm3, 0F5h - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movdqu xmm1, [esi] // _a_g - lea esi, [esi + 16] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jge convertloop4 - - convertloop4b: - add ecx, 4 - 1 - jl convertloop1b - - // 1 pixel loop. - convertloop1: - movd xmm3, [eax] // src argb - lea eax, [eax + 4] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movd xmm2, [esi] // _r_b - psrlw xmm3, 8 // alpha - pshufhw xmm3, xmm3, 0F5h // 8 alpha words - pshuflw xmm3, xmm3, 0F5h - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movd xmm1, [esi] // _a_g - lea esi, [esi + 4] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 1 - movd [edx], xmm0 - lea edx, [edx + 4] - jge convertloop1 - - convertloop1b: - pop esi - ret - } -} -#endif // HAS_ARGBBLENDROW_SSE2 - -#ifdef HAS_ARGBBLENDROW_SSSE3 -// Shuffle table for isolating alpha. -static const uvec8 kShuffleAlpha = { - 3u, 0x80, 3u, 0x80, 7u, 0x80, 7u, 0x80, - 11u, 0x80, 11u, 0x80, 15u, 0x80, 15u, 0x80 -}; -// Same as SSE2, but replaces: -// psrlw xmm3, 8 // alpha -// pshufhw xmm3, xmm3, 0F5h // 8 alpha words -// pshuflw xmm3, xmm3, 0F5h -// with.. -// pshufb xmm3, kShuffleAlpha // alpha -// Blend 8 pixels at a time. - -__declspec(naked) __declspec(align(16)) -void ARGBBlendRow_SSSE3(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - pcmpeqb xmm7, xmm7 // generate constant 0x0001 - psrlw xmm7, 15 - pcmpeqb xmm6, xmm6 // generate mask 0x00ff00ff - psrlw xmm6, 8 - pcmpeqb xmm5, xmm5 // generate mask 0xff00ff00 - psllw xmm5, 8 - pcmpeqb xmm4, xmm4 // generate mask 0xff000000 - pslld xmm4, 24 - - sub ecx, 1 - je convertloop1 // only 1 pixel? - jl convertloop1b - - // 1 pixel loop until destination pointer is aligned. - alignloop1: - test edx, 15 // aligned? - je alignloop1b - movd xmm3, [eax] - lea eax, [eax + 4] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movd xmm2, [esi] // _r_b - pshufb xmm3, kShuffleAlpha // alpha - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movd xmm1, [esi] // _a_g - lea esi, [esi + 4] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 1 - movd [edx], xmm0 - lea edx, [edx + 4] - jge alignloop1 - - alignloop1b: - add ecx, 1 - 4 - jl convertloop4b - - test eax, 15 // unaligned? - jne convertuloop4 - test esi, 15 // unaligned? - jne convertuloop4 - - // 4 pixel loop. - convertloop4: - movdqa xmm3, [eax] // src argb - lea eax, [eax + 16] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movdqa xmm2, [esi] // _r_b - pshufb xmm3, kShuffleAlpha // alpha - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movdqa xmm1, [esi] // _a_g - lea esi, [esi + 16] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jge convertloop4 - jmp convertloop4b - - // 4 pixel unaligned loop. - convertuloop4: - movdqu xmm3, [eax] // src argb - lea eax, [eax + 16] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movdqu xmm2, [esi] // _r_b - pshufb xmm3, kShuffleAlpha // alpha - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movdqu xmm1, [esi] // _a_g - lea esi, [esi + 16] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jge convertuloop4 - - convertloop4b: - add ecx, 4 - 1 - jl convertloop1b - - // 1 pixel loop. - convertloop1: - movd xmm3, [eax] // src argb - lea eax, [eax + 4] - movdqa xmm0, xmm3 // src argb - pxor xmm3, xmm4 // ~alpha - movd xmm2, [esi] // _r_b - pshufb xmm3, kShuffleAlpha // alpha - pand xmm2, xmm6 // _r_b - paddw xmm3, xmm7 // 256 - alpha - pmullw xmm2, xmm3 // _r_b * alpha - movd xmm1, [esi] // _a_g - lea esi, [esi + 4] - psrlw xmm1, 8 // _a_g - por xmm0, xmm4 // set alpha to 255 - pmullw xmm1, xmm3 // _a_g * alpha - psrlw xmm2, 8 // _r_b convert to 8 bits again - paddusb xmm0, xmm2 // + src argb - pand xmm1, xmm5 // a_g_ convert to 8 bits again - paddusb xmm0, xmm1 // + src argb - sub ecx, 1 - movd [edx], xmm0 - lea edx, [edx + 4] - jge convertloop1 - - convertloop1b: - pop esi - ret - } -} -#endif // HAS_ARGBBLENDROW_SSSE3 - -#ifdef HAS_ARGBATTENUATEROW_SSE2 -// Attenuate 4 pixels at a time. -// Aligned to 16 bytes. -__declspec(naked) __declspec(align(16)) -void ARGBAttenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width) { - __asm { - mov eax, [esp + 4] // src_argb0 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // width - pcmpeqb xmm4, xmm4 // generate mask 0xff000000 - pslld xmm4, 24 - pcmpeqb xmm5, xmm5 // generate mask 0x00ffffff - psrld xmm5, 8 - - align 4 - convertloop: - movdqa xmm0, [eax] // read 4 pixels - punpcklbw xmm0, xmm0 // first 2 - pshufhw xmm2, xmm0, 0FFh // 8 alpha words - pshuflw xmm2, xmm2, 0FFh - pmulhuw xmm0, xmm2 // rgb * a - movdqa xmm1, [eax] // read 4 pixels - punpckhbw xmm1, xmm1 // next 2 pixels - pshufhw xmm2, xmm1, 0FFh // 8 alpha words - pshuflw xmm2, xmm2, 0FFh - pmulhuw xmm1, xmm2 // rgb * a - movdqa xmm2, [eax] // alphas - lea eax, [eax + 16] - psrlw xmm0, 8 - pand xmm2, xmm4 - psrlw xmm1, 8 - packuswb xmm0, xmm1 - pand xmm0, xmm5 // keep original alphas - por xmm0, xmm2 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - - ret - } -} -#endif // HAS_ARGBATTENUATEROW_SSE2 - -#ifdef HAS_ARGBATTENUATEROW_SSSE3 -// Shuffle table duplicating alpha. -static const uvec8 kShuffleAlpha0 = { - 3u, 3u, 3u, 3u, 3u, 3u, 128u, 128u, 7u, 7u, 7u, 7u, 7u, 7u, 128u, 128u, -}; -static const uvec8 kShuffleAlpha1 = { - 11u, 11u, 11u, 11u, 11u, 11u, 128u, 128u, - 15u, 15u, 15u, 15u, 15u, 15u, 128u, 128u, -}; -__declspec(naked) __declspec(align(16)) -void ARGBAttenuateRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width) { - __asm { - mov eax, [esp + 4] // src_argb0 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // width - pcmpeqb xmm3, xmm3 // generate mask 0xff000000 - pslld xmm3, 24 - movdqa xmm4, kShuffleAlpha0 - movdqa xmm5, kShuffleAlpha1 - - align 4 - convertloop: - movdqu xmm0, [eax] // read 4 pixels - pshufb xmm0, xmm4 // isolate first 2 alphas - movdqu xmm1, [eax] // read 4 pixels - punpcklbw xmm1, xmm1 // first 2 pixel rgbs - pmulhuw xmm0, xmm1 // rgb * a - movdqu xmm1, [eax] // read 4 pixels - pshufb xmm1, xmm5 // isolate next 2 alphas - movdqu xmm2, [eax] // read 4 pixels - punpckhbw xmm2, xmm2 // next 2 pixel rgbs - pmulhuw xmm1, xmm2 // rgb * a - movdqu xmm2, [eax] // mask original alpha - lea eax, [eax + 16] - pand xmm2, xmm3 - psrlw xmm0, 8 - psrlw xmm1, 8 - packuswb xmm0, xmm1 - por xmm0, xmm2 // copy original alpha - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - - ret - } -} -#endif // HAS_ARGBATTENUATEROW_SSSE3 - -#ifdef HAS_ARGBATTENUATEROW_AVX2 -// Shuffle table duplicating alpha. -static const ulvec8 kShuffleAlpha_AVX2 = { - 6u, 7u, 6u, 7u, 6u, 7u, 128u, 128u, - 14u, 15u, 14u, 15u, 14u, 15u, 128u, 128u, - 6u, 7u, 6u, 7u, 6u, 7u, 128u, 128u, - 14u, 15u, 14u, 15u, 14u, 15u, 128u, 128u, -}; -__declspec(naked) __declspec(align(16)) -void ARGBAttenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, int width) { - __asm { - mov eax, [esp + 4] // src_argb0 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // width - sub edx, eax - vmovdqa ymm4, kShuffleAlpha_AVX2 - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0xff000000 - vpslld ymm5, ymm5, 24 - - align 4 - convertloop: - vmovdqu ymm6, [eax] // read 8 pixels. - vpunpcklbw ymm0, ymm6, ymm6 // low 4 pixels. mutated. - vpunpckhbw ymm1, ymm6, ymm6 // high 4 pixels. mutated. - vpshufb ymm2, ymm0, ymm4 // low 4 alphas - vpshufb ymm3, ymm1, ymm4 // high 4 alphas - vpmulhuw ymm0, ymm0, ymm2 // rgb * a - vpmulhuw ymm1, ymm1, ymm3 // rgb * a - vpand ymm6, ymm6, ymm5 // isolate alpha - vpsrlw ymm0, ymm0, 8 - vpsrlw ymm1, ymm1, 8 - vpackuswb ymm0, ymm0, ymm1 // unmutated. - vpor ymm0, ymm0, ymm6 // copy original alpha - sub ecx, 8 - vmovdqu [eax + edx], ymm0 - lea eax, [eax + 32] - jg convertloop - - vzeroupper - ret - } -} -#endif // HAS_ARGBATTENUATEROW_AVX2 - -#ifdef HAS_ARGBUNATTENUATEROW_SSE2 -// Unattenuate 4 pixels at a time. -// Aligned to 16 bytes. -__declspec(naked) __declspec(align(16)) -void ARGBUnattenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, - int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_argb0 - mov edx, [esp + 8 + 8] // dst_argb - mov ecx, [esp + 8 + 12] // width - - align 4 - convertloop: - movdqu xmm0, [eax] // read 4 pixels - movzx esi, byte ptr [eax + 3] // first alpha - movzx edi, byte ptr [eax + 7] // second alpha - punpcklbw xmm0, xmm0 // first 2 - movd xmm2, dword ptr fixed_invtbl8[esi * 4] - movd xmm3, dword ptr fixed_invtbl8[edi * 4] - pshuflw xmm2, xmm2, 040h // first 4 inv_alpha words. 1, a, a, a - pshuflw xmm3, xmm3, 040h // next 4 inv_alpha words - movlhps xmm2, xmm3 - pmulhuw xmm0, xmm2 // rgb * a - - movdqu xmm1, [eax] // read 4 pixels - movzx esi, byte ptr [eax + 11] // third alpha - movzx edi, byte ptr [eax + 15] // forth alpha - punpckhbw xmm1, xmm1 // next 2 - movd xmm2, dword ptr fixed_invtbl8[esi * 4] - movd xmm3, dword ptr fixed_invtbl8[edi * 4] - pshuflw xmm2, xmm2, 040h // first 4 inv_alpha words - pshuflw xmm3, xmm3, 040h // next 4 inv_alpha words - movlhps xmm2, xmm3 - pmulhuw xmm1, xmm2 // rgb * a - lea eax, [eax + 16] - - packuswb xmm0, xmm1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - pop edi - pop esi - ret - } -} -#endif // HAS_ARGBUNATTENUATEROW_SSE2 - -#ifdef HAS_ARGBUNATTENUATEROW_AVX2 -// Shuffle table duplicating alpha. -static const ulvec8 kUnattenShuffleAlpha_AVX2 = { - 0u, 1u, 0u, 1u, 0u, 1u, 6u, 7u, 8u, 9u, 8u, 9u, 8u, 9u, 14u, 15, - 0u, 1u, 0u, 1u, 0u, 1u, 6u, 7u, 8u, 9u, 8u, 9u, 8u, 9u, 14u, 15, -}; -// TODO(fbarchard): Enable USE_GATHER for future hardware if faster. -// USE_GATHER is not on by default, due to being a slow instruction. -#ifdef USE_GATHER -__declspec(naked) __declspec(align(16)) -void ARGBUnattenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, - int width) { - __asm { - mov eax, [esp + 4] // src_argb0 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // width - sub edx, eax - vmovdqa ymm4, kUnattenShuffleAlpha_AVX2 - - align 4 - convertloop: - vmovdqu ymm6, [eax] // read 8 pixels. - vpcmpeqb ymm5, ymm5, ymm5 // generate mask 0xffffffff for gather. - vpsrld ymm2, ymm6, 24 // alpha in low 8 bits. - vpunpcklbw ymm0, ymm6, ymm6 // low 4 pixels. mutated. - vpunpckhbw ymm1, ymm6, ymm6 // high 4 pixels. mutated. - vpgatherdd ymm3, [ymm2 * 4 + fixed_invtbl8], ymm5 // ymm5 cleared. 1, a - vpunpcklwd ymm2, ymm3, ymm3 // low 4 inverted alphas. mutated. 1, 1, a, a - vpunpckhwd ymm3, ymm3, ymm3 // high 4 inverted alphas. mutated. - vpshufb ymm2, ymm2, ymm4 // replicate low 4 alphas. 1, a, a, a - vpshufb ymm3, ymm3, ymm4 // replicate high 4 alphas - vpmulhuw ymm0, ymm0, ymm2 // rgb * ia - vpmulhuw ymm1, ymm1, ymm3 // rgb * ia - vpackuswb ymm0, ymm0, ymm1 // unmutated. - sub ecx, 8 - vmovdqu [eax + edx], ymm0 - lea eax, [eax + 32] - jg convertloop - - vzeroupper - ret - } -} -#else // USE_GATHER -__declspec(naked) __declspec(align(16)) -void ARGBUnattenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, - int width) { - __asm { - - mov eax, [esp + 4] // src_argb0 - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // width - sub edx, eax - vmovdqa ymm5, kUnattenShuffleAlpha_AVX2 - - push esi - push edi - - align 4 - convertloop: - // replace VPGATHER - movzx esi, byte ptr [eax + 3] // alpha0 - movzx edi, byte ptr [eax + 7] // alpha1 - vmovd xmm0, dword ptr fixed_invtbl8[esi * 4] // [1,a0] - vmovd xmm1, dword ptr fixed_invtbl8[edi * 4] // [1,a1] - movzx esi, byte ptr [eax + 11] // alpha2 - movzx edi, byte ptr [eax + 15] // alpha3 - vpunpckldq xmm6, xmm0, xmm1 // [1,a1,1,a0] - vmovd xmm2, dword ptr fixed_invtbl8[esi * 4] // [1,a2] - vmovd xmm3, dword ptr fixed_invtbl8[edi * 4] // [1,a3] - movzx esi, byte ptr [eax + 19] // alpha4 - movzx edi, byte ptr [eax + 23] // alpha5 - vpunpckldq xmm7, xmm2, xmm3 // [1,a3,1,a2] - vmovd xmm0, dword ptr fixed_invtbl8[esi * 4] // [1,a4] - vmovd xmm1, dword ptr fixed_invtbl8[edi * 4] // [1,a5] - movzx esi, byte ptr [eax + 27] // alpha6 - movzx edi, byte ptr [eax + 31] // alpha7 - vpunpckldq xmm0, xmm0, xmm1 // [1,a5,1,a4] - vmovd xmm2, dword ptr fixed_invtbl8[esi * 4] // [1,a6] - vmovd xmm3, dword ptr fixed_invtbl8[edi * 4] // [1,a7] - vpunpckldq xmm2, xmm2, xmm3 // [1,a7,1,a6] - vpunpcklqdq xmm3, xmm6, xmm7 // [1,a3,1,a2,1,a1,1,a0] - vpunpcklqdq xmm0, xmm0, xmm2 // [1,a7,1,a6,1,a5,1,a4] - vinserti128 ymm3, ymm3, xmm0, 1 // [1,a7,1,a6,1,a5,1,a4,1,a3,1,a2,1,a1,1,a0] - // end of VPGATHER - - vmovdqu ymm6, [eax] // read 8 pixels. - vpunpcklbw ymm0, ymm6, ymm6 // low 4 pixels. mutated. - vpunpckhbw ymm1, ymm6, ymm6 // high 4 pixels. mutated. - vpunpcklwd ymm2, ymm3, ymm3 // low 4 inverted alphas. mutated. 1, 1, a, a - vpunpckhwd ymm3, ymm3, ymm3 // high 4 inverted alphas. mutated. - vpshufb ymm2, ymm2, ymm5 // replicate low 4 alphas. 1, a, a, a - vpshufb ymm3, ymm3, ymm5 // replicate high 4 alphas - vpmulhuw ymm0, ymm0, ymm2 // rgb * ia - vpmulhuw ymm1, ymm1, ymm3 // rgb * ia - vpackuswb ymm0, ymm0, ymm1 // unmutated. - sub ecx, 8 - vmovdqu [eax + edx], ymm0 - lea eax, [eax + 32] - jg convertloop - - pop edi - pop esi - vzeroupper - ret - } -} -#endif // USE_GATHER -#endif // HAS_ARGBATTENUATEROW_AVX2 - -#ifdef HAS_ARGBGRAYROW_SSSE3 -// Convert 8 ARGB pixels (64 bytes) to 8 Gray ARGB pixels. -__declspec(naked) __declspec(align(16)) -void ARGBGrayRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_argb */ - mov ecx, [esp + 12] /* width */ - movdqa xmm4, kARGBToYJ - movdqa xmm5, kAddYJ64 - - align 4 - convertloop: - movdqa xmm0, [eax] // G - movdqa xmm1, [eax + 16] - pmaddubsw xmm0, xmm4 - pmaddubsw xmm1, xmm4 - phaddw xmm0, xmm1 - paddw xmm0, xmm5 // Add .5 for rounding. - psrlw xmm0, 7 - packuswb xmm0, xmm0 // 8 G bytes - movdqa xmm2, [eax] // A - movdqa xmm3, [eax + 16] - lea eax, [eax + 32] - psrld xmm2, 24 - psrld xmm3, 24 - packuswb xmm2, xmm3 - packuswb xmm2, xmm2 // 8 A bytes - movdqa xmm3, xmm0 // Weave into GG, GA, then GGGA - punpcklbw xmm0, xmm0 // 8 GG words - punpcklbw xmm3, xmm2 // 8 GA words - movdqa xmm1, xmm0 - punpcklwd xmm0, xmm3 // GGGA first 4 - punpckhwd xmm1, xmm3 // GGGA next 4 - sub ecx, 8 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - jg convertloop - ret - } -} -#endif // HAS_ARGBGRAYROW_SSSE3 - -#ifdef HAS_ARGBSEPIAROW_SSSE3 -// b = (r * 35 + g * 68 + b * 17) >> 7 -// g = (r * 45 + g * 88 + b * 22) >> 7 -// r = (r * 50 + g * 98 + b * 24) >> 7 -// Constant for ARGB color to sepia tone. -static const vec8 kARGBToSepiaB = { - 17, 68, 35, 0, 17, 68, 35, 0, 17, 68, 35, 0, 17, 68, 35, 0 -}; - -static const vec8 kARGBToSepiaG = { - 22, 88, 45, 0, 22, 88, 45, 0, 22, 88, 45, 0, 22, 88, 45, 0 -}; - -static const vec8 kARGBToSepiaR = { - 24, 98, 50, 0, 24, 98, 50, 0, 24, 98, 50, 0, 24, 98, 50, 0 -}; - -// Convert 8 ARGB pixels (32 bytes) to 8 Sepia ARGB pixels. -__declspec(naked) __declspec(align(16)) -void ARGBSepiaRow_SSSE3(uint8* dst_argb, int width) { - __asm { - mov eax, [esp + 4] /* dst_argb */ - mov ecx, [esp + 8] /* width */ - movdqa xmm2, kARGBToSepiaB - movdqa xmm3, kARGBToSepiaG - movdqa xmm4, kARGBToSepiaR - - align 4 - convertloop: - movdqa xmm0, [eax] // B - movdqa xmm6, [eax + 16] - pmaddubsw xmm0, xmm2 - pmaddubsw xmm6, xmm2 - phaddw xmm0, xmm6 - psrlw xmm0, 7 - packuswb xmm0, xmm0 // 8 B values - movdqa xmm5, [eax] // G - movdqa xmm1, [eax + 16] - pmaddubsw xmm5, xmm3 - pmaddubsw xmm1, xmm3 - phaddw xmm5, xmm1 - psrlw xmm5, 7 - packuswb xmm5, xmm5 // 8 G values - punpcklbw xmm0, xmm5 // 8 BG values - movdqa xmm5, [eax] // R - movdqa xmm1, [eax + 16] - pmaddubsw xmm5, xmm4 - pmaddubsw xmm1, xmm4 - phaddw xmm5, xmm1 - psrlw xmm5, 7 - packuswb xmm5, xmm5 // 8 R values - movdqa xmm6, [eax] // A - movdqa xmm1, [eax + 16] - psrld xmm6, 24 - psrld xmm1, 24 - packuswb xmm6, xmm1 - packuswb xmm6, xmm6 // 8 A values - punpcklbw xmm5, xmm6 // 8 RA values - movdqa xmm1, xmm0 // Weave BG, RA together - punpcklwd xmm0, xmm5 // BGRA first 4 - punpckhwd xmm1, xmm5 // BGRA next 4 - sub ecx, 8 - movdqa [eax], xmm0 - movdqa [eax + 16], xmm1 - lea eax, [eax + 32] - jg convertloop - ret - } -} -#endif // HAS_ARGBSEPIAROW_SSSE3 - -#ifdef HAS_ARGBCOLORMATRIXROW_SSSE3 -// Tranform 8 ARGB pixels (32 bytes) with color matrix. -// Same as Sepia except matrix is provided. -// TODO(fbarchard): packuswbs only use half of the reg. To make RGBA, combine R -// and B into a high and low, then G/A, unpackl/hbw and then unpckl/hwd. -__declspec(naked) __declspec(align(16)) -void ARGBColorMatrixRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - const int8* matrix_argb, int width) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_argb */ - mov ecx, [esp + 12] /* matrix_argb */ - movdqu xmm5, [ecx] - pshufd xmm2, xmm5, 0x00 - pshufd xmm3, xmm5, 0x55 - pshufd xmm4, xmm5, 0xaa - pshufd xmm5, xmm5, 0xff - mov ecx, [esp + 16] /* width */ - - align 4 - convertloop: - movdqa xmm0, [eax] // B - movdqa xmm7, [eax + 16] - pmaddubsw xmm0, xmm2 - pmaddubsw xmm7, xmm2 - movdqa xmm6, [eax] // G - movdqa xmm1, [eax + 16] - pmaddubsw xmm6, xmm3 - pmaddubsw xmm1, xmm3 - phaddsw xmm0, xmm7 // B - phaddsw xmm6, xmm1 // G - psraw xmm0, 6 // B - psraw xmm6, 6 // G - packuswb xmm0, xmm0 // 8 B values - packuswb xmm6, xmm6 // 8 G values - punpcklbw xmm0, xmm6 // 8 BG values - movdqa xmm1, [eax] // R - movdqa xmm7, [eax + 16] - pmaddubsw xmm1, xmm4 - pmaddubsw xmm7, xmm4 - phaddsw xmm1, xmm7 // R - movdqa xmm6, [eax] // A - movdqa xmm7, [eax + 16] - pmaddubsw xmm6, xmm5 - pmaddubsw xmm7, xmm5 - phaddsw xmm6, xmm7 // A - psraw xmm1, 6 // R - psraw xmm6, 6 // A - packuswb xmm1, xmm1 // 8 R values - packuswb xmm6, xmm6 // 8 A values - punpcklbw xmm1, xmm6 // 8 RA values - movdqa xmm6, xmm0 // Weave BG, RA together - punpcklwd xmm0, xmm1 // BGRA first 4 - punpckhwd xmm6, xmm1 // BGRA next 4 - sub ecx, 8 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm6 - lea eax, [eax + 32] - lea edx, [edx + 32] - jg convertloop - ret - } -} -#endif // HAS_ARGBCOLORMATRIXROW_SSSE3 - -#ifdef HAS_ARGBQUANTIZEROW_SSE2 -// Quantize 4 ARGB pixels (16 bytes). -// Aligned to 16 bytes. -__declspec(naked) __declspec(align(16)) -void ARGBQuantizeRow_SSE2(uint8* dst_argb, int scale, int interval_size, - int interval_offset, int width) { - __asm { - mov eax, [esp + 4] /* dst_argb */ - movd xmm2, [esp + 8] /* scale */ - movd xmm3, [esp + 12] /* interval_size */ - movd xmm4, [esp + 16] /* interval_offset */ - mov ecx, [esp + 20] /* width */ - pshuflw xmm2, xmm2, 040h - pshufd xmm2, xmm2, 044h - pshuflw xmm3, xmm3, 040h - pshufd xmm3, xmm3, 044h - pshuflw xmm4, xmm4, 040h - pshufd xmm4, xmm4, 044h - pxor xmm5, xmm5 // constant 0 - pcmpeqb xmm6, xmm6 // generate mask 0xff000000 - pslld xmm6, 24 - - align 4 - convertloop: - movdqa xmm0, [eax] // read 4 pixels - punpcklbw xmm0, xmm5 // first 2 pixels - pmulhuw xmm0, xmm2 // pixel * scale >> 16 - movdqa xmm1, [eax] // read 4 pixels - punpckhbw xmm1, xmm5 // next 2 pixels - pmulhuw xmm1, xmm2 - pmullw xmm0, xmm3 // * interval_size - movdqa xmm7, [eax] // read 4 pixels - pmullw xmm1, xmm3 - pand xmm7, xmm6 // mask alpha - paddw xmm0, xmm4 // + interval_size / 2 - paddw xmm1, xmm4 - packuswb xmm0, xmm1 - por xmm0, xmm7 - sub ecx, 4 - movdqa [eax], xmm0 - lea eax, [eax + 16] - jg convertloop - ret - } -} -#endif // HAS_ARGBQUANTIZEROW_SSE2 - -#ifdef HAS_ARGBSHADEROW_SSE2 -// Shade 4 pixels at a time by specified value. -// Aligned to 16 bytes. -__declspec(naked) __declspec(align(16)) -void ARGBShadeRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width, - uint32 value) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // width - movd xmm2, [esp + 16] // value - punpcklbw xmm2, xmm2 - punpcklqdq xmm2, xmm2 - - align 4 - convertloop: - movdqa xmm0, [eax] // read 4 pixels - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm0 // first 2 - punpckhbw xmm1, xmm1 // next 2 - pmulhuw xmm0, xmm2 // argb * value - pmulhuw xmm1, xmm2 // argb * value - psrlw xmm0, 8 - psrlw xmm1, 8 - packuswb xmm0, xmm1 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - - ret - } -} -#endif // HAS_ARGBSHADEROW_SSE2 - -#ifdef HAS_ARGBMULTIPLYROW_SSE2 -// Multiply 2 rows of ARGB pixels together, 4 pixels at a time. -__declspec(naked) __declspec(align(16)) -void ARGBMultiplyRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - pxor xmm5, xmm5 // constant 0 - - align 4 - convertloop: - movdqu xmm0, [eax] // read 4 pixels from src_argb0 - movdqu xmm2, [esi] // read 4 pixels from src_argb1 - movdqu xmm1, xmm0 - movdqu xmm3, xmm2 - punpcklbw xmm0, xmm0 // first 2 - punpckhbw xmm1, xmm1 // next 2 - punpcklbw xmm2, xmm5 // first 2 - punpckhbw xmm3, xmm5 // next 2 - pmulhuw xmm0, xmm2 // src_argb0 * src_argb1 first 2 - pmulhuw xmm1, xmm3 // src_argb0 * src_argb1 next 2 - lea eax, [eax + 16] - lea esi, [esi + 16] - packuswb xmm0, xmm1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - - pop esi - ret - } -} -#endif // HAS_ARGBMULTIPLYROW_SSE2 - -#ifdef HAS_ARGBADDROW_SSE2 -// Add 2 rows of ARGB pixels together, 4 pixels at a time. -// TODO(fbarchard): Port this to posix, neon and other math functions. -__declspec(naked) __declspec(align(16)) -void ARGBAddRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - - sub ecx, 4 - jl convertloop49 - - align 4 - convertloop4: - movdqu xmm0, [eax] // read 4 pixels from src_argb0 - lea eax, [eax + 16] - movdqu xmm1, [esi] // read 4 pixels from src_argb1 - lea esi, [esi + 16] - paddusb xmm0, xmm1 // src_argb0 + src_argb1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jge convertloop4 - - convertloop49: - add ecx, 4 - 1 - jl convertloop19 - - convertloop1: - movd xmm0, [eax] // read 1 pixels from src_argb0 - lea eax, [eax + 4] - movd xmm1, [esi] // read 1 pixels from src_argb1 - lea esi, [esi + 4] - paddusb xmm0, xmm1 // src_argb0 + src_argb1 - sub ecx, 1 - movd [edx], xmm0 - lea edx, [edx + 4] - jge convertloop1 - - convertloop19: - pop esi - ret - } -} -#endif // HAS_ARGBADDROW_SSE2 - -#ifdef HAS_ARGBSUBTRACTROW_SSE2 -// Subtract 2 rows of ARGB pixels together, 4 pixels at a time. -__declspec(naked) __declspec(align(16)) -void ARGBSubtractRow_SSE2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - - align 4 - convertloop: - movdqu xmm0, [eax] // read 4 pixels from src_argb0 - lea eax, [eax + 16] - movdqu xmm1, [esi] // read 4 pixels from src_argb1 - lea esi, [esi + 16] - psubusb xmm0, xmm1 // src_argb0 - src_argb1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - - pop esi - ret - } -} -#endif // HAS_ARGBSUBTRACTROW_SSE2 - -#ifdef HAS_ARGBMULTIPLYROW_AVX2 -// Multiply 2 rows of ARGB pixels together, 8 pixels at a time. -__declspec(naked) __declspec(align(16)) -void ARGBMultiplyRow_AVX2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - vpxor ymm5, ymm5, ymm5 // constant 0 - - align 4 - convertloop: - vmovdqu ymm1, [eax] // read 8 pixels from src_argb0 - lea eax, [eax + 32] - vmovdqu ymm3, [esi] // read 8 pixels from src_argb1 - lea esi, [esi + 32] - vpunpcklbw ymm0, ymm1, ymm1 // low 4 - vpunpckhbw ymm1, ymm1, ymm1 // high 4 - vpunpcklbw ymm2, ymm3, ymm5 // low 4 - vpunpckhbw ymm3, ymm3, ymm5 // high 4 - vpmulhuw ymm0, ymm0, ymm2 // src_argb0 * src_argb1 low 4 - vpmulhuw ymm1, ymm1, ymm3 // src_argb0 * src_argb1 high 4 - vpackuswb ymm0, ymm0, ymm1 - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - vzeroupper - ret - } -} -#endif // HAS_ARGBMULTIPLYROW_AVX2 - -#ifdef HAS_ARGBADDROW_AVX2 -// Add 2 rows of ARGB pixels together, 8 pixels at a time. -__declspec(naked) __declspec(align(16)) -void ARGBAddRow_AVX2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - - align 4 - convertloop: - vmovdqu ymm0, [eax] // read 8 pixels from src_argb0 - lea eax, [eax + 32] - vpaddusb ymm0, ymm0, [esi] // add 8 pixels from src_argb1 - lea esi, [esi + 32] - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - vzeroupper - ret - } -} -#endif // HAS_ARGBADDROW_AVX2 - -#ifdef HAS_ARGBSUBTRACTROW_AVX2 -// Subtract 2 rows of ARGB pixels together, 8 pixels at a time. -__declspec(naked) __declspec(align(16)) -void ARGBSubtractRow_AVX2(const uint8* src_argb0, const uint8* src_argb1, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb0 - mov esi, [esp + 4 + 8] // src_argb1 - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - - align 4 - convertloop: - vmovdqu ymm0, [eax] // read 8 pixels from src_argb0 - lea eax, [eax + 32] - vpsubusb ymm0, ymm0, [esi] // src_argb0 - src_argb1 - lea esi, [esi + 32] - vmovdqu [edx], ymm0 - lea edx, [edx + 32] - sub ecx, 8 - jg convertloop - - pop esi - vzeroupper - ret - } -} -#endif // HAS_ARGBSUBTRACTROW_AVX2 - -#ifdef HAS_SOBELXROW_SSE2 -// SobelX as a matrix is -// -1 0 1 -// -2 0 2 -// -1 0 1 -__declspec(naked) __declspec(align(16)) -void SobelXRow_SSE2(const uint8* src_y0, const uint8* src_y1, - const uint8* src_y2, uint8* dst_sobelx, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_y0 - mov esi, [esp + 8 + 8] // src_y1 - mov edi, [esp + 8 + 12] // src_y2 - mov edx, [esp + 8 + 16] // dst_sobelx - mov ecx, [esp + 8 + 20] // width - sub esi, eax - sub edi, eax - sub edx, eax - pxor xmm5, xmm5 // constant 0 - - align 4 - convertloop: - movq xmm0, qword ptr [eax] // read 8 pixels from src_y0[0] - movq xmm1, qword ptr [eax + 2] // read 8 pixels from src_y0[2] - punpcklbw xmm0, xmm5 - punpcklbw xmm1, xmm5 - psubw xmm0, xmm1 - movq xmm1, qword ptr [eax + esi] // read 8 pixels from src_y1[0] - movq xmm2, qword ptr [eax + esi + 2] // read 8 pixels from src_y1[2] - punpcklbw xmm1, xmm5 - punpcklbw xmm2, xmm5 - psubw xmm1, xmm2 - movq xmm2, qword ptr [eax + edi] // read 8 pixels from src_y2[0] - movq xmm3, qword ptr [eax + edi + 2] // read 8 pixels from src_y2[2] - punpcklbw xmm2, xmm5 - punpcklbw xmm3, xmm5 - psubw xmm2, xmm3 - paddw xmm0, xmm2 - paddw xmm0, xmm1 - paddw xmm0, xmm1 - pxor xmm1, xmm1 // abs = max(xmm0, -xmm0). SSSE3 could use pabsw - psubw xmm1, xmm0 - pmaxsw xmm0, xmm1 - packuswb xmm0, xmm0 - sub ecx, 8 - movq qword ptr [eax + edx], xmm0 - lea eax, [eax + 8] - jg convertloop - - pop edi - pop esi - ret - } -} -#endif // HAS_SOBELXROW_SSE2 - -#ifdef HAS_SOBELYROW_SSE2 -// SobelY as a matrix is -// -1 -2 -1 -// 0 0 0 -// 1 2 1 -__declspec(naked) __declspec(align(16)) -void SobelYRow_SSE2(const uint8* src_y0, const uint8* src_y1, - uint8* dst_sobely, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_y0 - mov esi, [esp + 4 + 8] // src_y1 - mov edx, [esp + 4 + 12] // dst_sobely - mov ecx, [esp + 4 + 16] // width - sub esi, eax - sub edx, eax - pxor xmm5, xmm5 // constant 0 - - align 4 - convertloop: - movq xmm0, qword ptr [eax] // read 8 pixels from src_y0[0] - movq xmm1, qword ptr [eax + esi] // read 8 pixels from src_y1[0] - punpcklbw xmm0, xmm5 - punpcklbw xmm1, xmm5 - psubw xmm0, xmm1 - movq xmm1, qword ptr [eax + 1] // read 8 pixels from src_y0[1] - movq xmm2, qword ptr [eax + esi + 1] // read 8 pixels from src_y1[1] - punpcklbw xmm1, xmm5 - punpcklbw xmm2, xmm5 - psubw xmm1, xmm2 - movq xmm2, qword ptr [eax + 2] // read 8 pixels from src_y0[2] - movq xmm3, qword ptr [eax + esi + 2] // read 8 pixels from src_y1[2] - punpcklbw xmm2, xmm5 - punpcklbw xmm3, xmm5 - psubw xmm2, xmm3 - paddw xmm0, xmm2 - paddw xmm0, xmm1 - paddw xmm0, xmm1 - pxor xmm1, xmm1 // abs = max(xmm0, -xmm0). SSSE3 could use pabsw - psubw xmm1, xmm0 - pmaxsw xmm0, xmm1 - packuswb xmm0, xmm0 - sub ecx, 8 - movq qword ptr [eax + edx], xmm0 - lea eax, [eax + 8] - jg convertloop - - pop esi - ret - } -} -#endif // HAS_SOBELYROW_SSE2 - -#ifdef HAS_SOBELROW_SSE2 -// Adds Sobel X and Sobel Y and stores Sobel into ARGB. -// A = 255 -// R = Sobel -// G = Sobel -// B = Sobel -__declspec(naked) __declspec(align(16)) -void SobelRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_sobelx - mov esi, [esp + 4 + 8] // src_sobely - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - sub esi, eax - pcmpeqb xmm5, xmm5 // alpha 255 - pslld xmm5, 24 // 0xff000000 - - align 4 - convertloop: - movdqa xmm0, [eax] // read 16 pixels src_sobelx - movdqa xmm1, [eax + esi] // read 16 pixels src_sobely - lea eax, [eax + 16] - paddusb xmm0, xmm1 // sobel = sobelx + sobely - movdqa xmm2, xmm0 // GG - punpcklbw xmm2, xmm0 // First 8 - punpckhbw xmm0, xmm0 // Next 8 - movdqa xmm1, xmm2 // GGGG - punpcklwd xmm1, xmm2 // First 4 - punpckhwd xmm2, xmm2 // Next 4 - por xmm1, xmm5 // GGGA - por xmm2, xmm5 - movdqa xmm3, xmm0 // GGGG - punpcklwd xmm3, xmm0 // Next 4 - punpckhwd xmm0, xmm0 // Last 4 - por xmm3, xmm5 // GGGA - por xmm0, xmm5 - sub ecx, 16 - movdqa [edx], xmm1 - movdqa [edx + 16], xmm2 - movdqa [edx + 32], xmm3 - movdqa [edx + 48], xmm0 - lea edx, [edx + 64] - jg convertloop - - pop esi - ret - } -} -#endif // HAS_SOBELROW_SSE2 - -#ifdef HAS_SOBELTOPLANEROW_SSE2 -// Adds Sobel X and Sobel Y and stores Sobel into a plane. -__declspec(naked) __declspec(align(16)) -void SobelToPlaneRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_y, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_sobelx - mov esi, [esp + 4 + 8] // src_sobely - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - sub esi, eax - - align 4 - convertloop: - movdqa xmm0, [eax] // read 16 pixels src_sobelx - movdqa xmm1, [eax + esi] // read 16 pixels src_sobely - lea eax, [eax + 16] - paddusb xmm0, xmm1 // sobel = sobelx + sobely - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg convertloop - - pop esi - ret - } -} -#endif // HAS_SOBELTOPLANEROW_SSE2 - -#ifdef HAS_SOBELXYROW_SSE2 -// Mixes Sobel X, Sobel Y and Sobel into ARGB. -// A = 255 -// R = Sobel X -// G = Sobel -// B = Sobel Y -__declspec(naked) __declspec(align(16)) -void SobelXYRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely, - uint8* dst_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_sobelx - mov esi, [esp + 4 + 8] // src_sobely - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // width - sub esi, eax - pcmpeqb xmm5, xmm5 // alpha 255 - - align 4 - convertloop: - movdqa xmm0, [eax] // read 16 pixels src_sobelx - movdqa xmm1, [eax + esi] // read 16 pixels src_sobely - lea eax, [eax + 16] - movdqa xmm2, xmm0 - paddusb xmm2, xmm1 // sobel = sobelx + sobely - movdqa xmm3, xmm0 // XA - punpcklbw xmm3, xmm5 - punpckhbw xmm0, xmm5 - movdqa xmm4, xmm1 // YS - punpcklbw xmm4, xmm2 - punpckhbw xmm1, xmm2 - movdqa xmm6, xmm4 // YSXA - punpcklwd xmm6, xmm3 // First 4 - punpckhwd xmm4, xmm3 // Next 4 - movdqa xmm7, xmm1 // YSXA - punpcklwd xmm7, xmm0 // Next 4 - punpckhwd xmm1, xmm0 // Last 4 - sub ecx, 16 - movdqa [edx], xmm6 - movdqa [edx + 16], xmm4 - movdqa [edx + 32], xmm7 - movdqa [edx + 48], xmm1 - lea edx, [edx + 64] - jg convertloop - - pop esi - ret - } -} -#endif // HAS_SOBELXYROW_SSE2 - -#ifdef HAS_CUMULATIVESUMTOAVERAGEROW_SSE2 -// Consider float CumulativeSum. -// Consider calling CumulativeSum one row at time as needed. -// Consider circular CumulativeSum buffer of radius * 2 + 1 height. -// Convert cumulative sum for an area to an average for 1 pixel. -// topleft is pointer to top left of CumulativeSum buffer for area. -// botleft is pointer to bottom left of CumulativeSum buffer. -// width is offset from left to right of area in CumulativeSum buffer measured -// in number of ints. -// area is the number of pixels in the area being averaged. -// dst points to pixel to store result to. -// count is number of averaged pixels to produce. -// Does 4 pixels at a time, requires CumulativeSum pointers to be 16 byte -// aligned. -void CumulativeSumToAverageRow_SSE2(const int32* topleft, const int32* botleft, - int width, int area, uint8* dst, - int count) { - __asm { - mov eax, topleft // eax topleft - mov esi, botleft // esi botleft - mov edx, width - movd xmm5, area - mov edi, dst - mov ecx, count - cvtdq2ps xmm5, xmm5 - rcpss xmm4, xmm5 // 1.0f / area - pshufd xmm4, xmm4, 0 - sub ecx, 4 - jl l4b - - cmp area, 128 // 128 pixels will not overflow 15 bits. - ja l4 - - pshufd xmm5, xmm5, 0 // area - pcmpeqb xmm6, xmm6 // constant of 65536.0 - 1 = 65535.0 - psrld xmm6, 16 - cvtdq2ps xmm6, xmm6 - addps xmm5, xmm6 // (65536.0 + area - 1) - mulps xmm5, xmm4 // (65536.0 + area - 1) * 1 / area - cvtps2dq xmm5, xmm5 // 0.16 fixed point - packssdw xmm5, xmm5 // 16 bit shorts - - // 4 pixel loop small blocks. - align 4 - s4: - // top left - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - - // - top right - psubd xmm0, [eax + edx * 4] - psubd xmm1, [eax + edx * 4 + 16] - psubd xmm2, [eax + edx * 4 + 32] - psubd xmm3, [eax + edx * 4 + 48] - lea eax, [eax + 64] - - // - bottom left - psubd xmm0, [esi] - psubd xmm1, [esi + 16] - psubd xmm2, [esi + 32] - psubd xmm3, [esi + 48] - - // + bottom right - paddd xmm0, [esi + edx * 4] - paddd xmm1, [esi + edx * 4 + 16] - paddd xmm2, [esi + edx * 4 + 32] - paddd xmm3, [esi + edx * 4 + 48] - lea esi, [esi + 64] - - packssdw xmm0, xmm1 // pack 4 pixels into 2 registers - packssdw xmm2, xmm3 - - pmulhuw xmm0, xmm5 - pmulhuw xmm2, xmm5 - - packuswb xmm0, xmm2 - movdqu [edi], xmm0 - lea edi, [edi + 16] - sub ecx, 4 - jge s4 - - jmp l4b - - // 4 pixel loop - align 4 - l4: - // top left - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + 32] - movdqa xmm3, [eax + 48] - - // - top right - psubd xmm0, [eax + edx * 4] - psubd xmm1, [eax + edx * 4 + 16] - psubd xmm2, [eax + edx * 4 + 32] - psubd xmm3, [eax + edx * 4 + 48] - lea eax, [eax + 64] - - // - bottom left - psubd xmm0, [esi] - psubd xmm1, [esi + 16] - psubd xmm2, [esi + 32] - psubd xmm3, [esi + 48] - - // + bottom right - paddd xmm0, [esi + edx * 4] - paddd xmm1, [esi + edx * 4 + 16] - paddd xmm2, [esi + edx * 4 + 32] - paddd xmm3, [esi + edx * 4 + 48] - lea esi, [esi + 64] - - cvtdq2ps xmm0, xmm0 // Average = Sum * 1 / Area - cvtdq2ps xmm1, xmm1 - mulps xmm0, xmm4 - mulps xmm1, xmm4 - cvtdq2ps xmm2, xmm2 - cvtdq2ps xmm3, xmm3 - mulps xmm2, xmm4 - mulps xmm3, xmm4 - cvtps2dq xmm0, xmm0 - cvtps2dq xmm1, xmm1 - cvtps2dq xmm2, xmm2 - cvtps2dq xmm3, xmm3 - packssdw xmm0, xmm1 - packssdw xmm2, xmm3 - packuswb xmm0, xmm2 - movdqu [edi], xmm0 - lea edi, [edi + 16] - sub ecx, 4 - jge l4 - - l4b: - add ecx, 4 - 1 - jl l1b - - // 1 pixel loop - align 4 - l1: - movdqa xmm0, [eax] - psubd xmm0, [eax + edx * 4] - lea eax, [eax + 16] - psubd xmm0, [esi] - paddd xmm0, [esi + edx * 4] - lea esi, [esi + 16] - cvtdq2ps xmm0, xmm0 - mulps xmm0, xmm4 - cvtps2dq xmm0, xmm0 - packssdw xmm0, xmm0 - packuswb xmm0, xmm0 - movd dword ptr [edi], xmm0 - lea edi, [edi + 4] - sub ecx, 1 - jge l1 - l1b: - } -} -#endif // HAS_CUMULATIVESUMTOAVERAGEROW_SSE2 - -#ifdef HAS_COMPUTECUMULATIVESUMROW_SSE2 -// Creates a table of cumulative sums where each value is a sum of all values -// above and to the left of the value. -void ComputeCumulativeSumRow_SSE2(const uint8* row, int32* cumsum, - const int32* previous_cumsum, int width) { - __asm { - mov eax, row - mov edx, cumsum - mov esi, previous_cumsum - mov ecx, width - pxor xmm0, xmm0 - pxor xmm1, xmm1 - - sub ecx, 4 - jl l4b - test edx, 15 - jne l4b - - // 4 pixel loop - align 4 - l4: - movdqu xmm2, [eax] // 4 argb pixels 16 bytes. - lea eax, [eax + 16] - movdqa xmm4, xmm2 - - punpcklbw xmm2, xmm1 - movdqa xmm3, xmm2 - punpcklwd xmm2, xmm1 - punpckhwd xmm3, xmm1 - - punpckhbw xmm4, xmm1 - movdqa xmm5, xmm4 - punpcklwd xmm4, xmm1 - punpckhwd xmm5, xmm1 - - paddd xmm0, xmm2 - movdqa xmm2, [esi] // previous row above. - paddd xmm2, xmm0 - - paddd xmm0, xmm3 - movdqa xmm3, [esi + 16] - paddd xmm3, xmm0 - - paddd xmm0, xmm4 - movdqa xmm4, [esi + 32] - paddd xmm4, xmm0 - - paddd xmm0, xmm5 - movdqa xmm5, [esi + 48] - lea esi, [esi + 64] - paddd xmm5, xmm0 - - movdqa [edx], xmm2 - movdqa [edx + 16], xmm3 - movdqa [edx + 32], xmm4 - movdqa [edx + 48], xmm5 - - lea edx, [edx + 64] - sub ecx, 4 - jge l4 - - l4b: - add ecx, 4 - 1 - jl l1b - - // 1 pixel loop - align 4 - l1: - movd xmm2, dword ptr [eax] // 1 argb pixel 4 bytes. - lea eax, [eax + 4] - punpcklbw xmm2, xmm1 - punpcklwd xmm2, xmm1 - paddd xmm0, xmm2 - movdqu xmm2, [esi] - lea esi, [esi + 16] - paddd xmm2, xmm0 - movdqu [edx], xmm2 - lea edx, [edx + 16] - sub ecx, 1 - jge l1 - - l1b: - } -} -#endif // HAS_COMPUTECUMULATIVESUMROW_SSE2 - -#ifdef HAS_ARGBAFFINEROW_SSE2 -// Copy ARGB pixels from source image with slope to a row of destination. -__declspec(naked) __declspec(align(16)) -LIBYUV_API -void ARGBAffineRow_SSE2(const uint8* src_argb, int src_argb_stride, - uint8* dst_argb, const float* uv_dudv, int width) { - __asm { - push esi - push edi - mov eax, [esp + 12] // src_argb - mov esi, [esp + 16] // stride - mov edx, [esp + 20] // dst_argb - mov ecx, [esp + 24] // pointer to uv_dudv - movq xmm2, qword ptr [ecx] // uv - movq xmm7, qword ptr [ecx + 8] // dudv - mov ecx, [esp + 28] // width - shl esi, 16 // 4, stride - add esi, 4 - movd xmm5, esi - sub ecx, 4 - jl l4b - - // setup for 4 pixel loop - pshufd xmm7, xmm7, 0x44 // dup dudv - pshufd xmm5, xmm5, 0 // dup 4, stride - movdqa xmm0, xmm2 // x0, y0, x1, y1 - addps xmm0, xmm7 - movlhps xmm2, xmm0 - movdqa xmm4, xmm7 - addps xmm4, xmm4 // dudv *= 2 - movdqa xmm3, xmm2 // x2, y2, x3, y3 - addps xmm3, xmm4 - addps xmm4, xmm4 // dudv *= 4 - - // 4 pixel loop - align 4 - l4: - cvttps2dq xmm0, xmm2 // x, y float to int first 2 - cvttps2dq xmm1, xmm3 // x, y float to int next 2 - packssdw xmm0, xmm1 // x, y as 8 shorts - pmaddwd xmm0, xmm5 // offsets = x * 4 + y * stride. - movd esi, xmm0 - pshufd xmm0, xmm0, 0x39 // shift right - movd edi, xmm0 - pshufd xmm0, xmm0, 0x39 // shift right - movd xmm1, [eax + esi] // read pixel 0 - movd xmm6, [eax + edi] // read pixel 1 - punpckldq xmm1, xmm6 // combine pixel 0 and 1 - addps xmm2, xmm4 // x, y += dx, dy first 2 - movq qword ptr [edx], xmm1 - movd esi, xmm0 - pshufd xmm0, xmm0, 0x39 // shift right - movd edi, xmm0 - movd xmm6, [eax + esi] // read pixel 2 - movd xmm0, [eax + edi] // read pixel 3 - punpckldq xmm6, xmm0 // combine pixel 2 and 3 - addps xmm3, xmm4 // x, y += dx, dy next 2 - sub ecx, 4 - movq qword ptr 8[edx], xmm6 - lea edx, [edx + 16] - jge l4 - - l4b: - add ecx, 4 - 1 - jl l1b - - // 1 pixel loop - align 4 - l1: - cvttps2dq xmm0, xmm2 // x, y float to int - packssdw xmm0, xmm0 // x, y as shorts - pmaddwd xmm0, xmm5 // offset = x * 4 + y * stride - addps xmm2, xmm7 // x, y += dx, dy - movd esi, xmm0 - movd xmm0, [eax + esi] // copy a pixel - sub ecx, 1 - movd [edx], xmm0 - lea edx, [edx + 4] - jge l1 - l1b: - pop edi - pop esi - ret - } -} -#endif // HAS_ARGBAFFINEROW_SSE2 - -#ifdef HAS_INTERPOLATEROW_AVX2 -// Bilinear filter 16x2 -> 16x1 -__declspec(naked) __declspec(align(16)) -void InterpolateRow_AVX2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - __asm { - push esi - push edi - mov edi, [esp + 8 + 4] // dst_ptr - mov esi, [esp + 8 + 8] // src_ptr - mov edx, [esp + 8 + 12] // src_stride - mov ecx, [esp + 8 + 16] // dst_width - mov eax, [esp + 8 + 20] // source_y_fraction (0..255) - shr eax, 1 - // Dispatch to specialized filters if applicable. - cmp eax, 0 - je xloop100 // 0 / 128. Blend 100 / 0. - sub edi, esi - cmp eax, 32 - je xloop75 // 32 / 128 is 0.25. Blend 75 / 25. - cmp eax, 64 - je xloop50 // 64 / 128 is 0.50. Blend 50 / 50. - cmp eax, 96 - je xloop25 // 96 / 128 is 0.75. Blend 25 / 75. - - vmovd xmm0, eax // high fraction 0..127 - neg eax - add eax, 128 - vmovd xmm5, eax // low fraction 128..1 - vpunpcklbw xmm5, xmm5, xmm0 - vpunpcklwd xmm5, xmm5, xmm5 - vpxor ymm0, ymm0, ymm0 - vpermd ymm5, ymm0, ymm5 - - align 4 - xloop: - vmovdqu ymm0, [esi] - vmovdqu ymm2, [esi + edx] - vpunpckhbw ymm1, ymm0, ymm2 // mutates - vpunpcklbw ymm0, ymm0, ymm2 // mutates - vpmaddubsw ymm0, ymm0, ymm5 - vpmaddubsw ymm1, ymm1, ymm5 - vpsrlw ymm0, ymm0, 7 - vpsrlw ymm1, ymm1, 7 - vpackuswb ymm0, ymm0, ymm1 // unmutates - sub ecx, 32 - vmovdqu [esi + edi], ymm0 - lea esi, [esi + 32] - jg xloop - jmp xloop99 - - // Blend 25 / 75. - align 4 - xloop25: - vmovdqu ymm0, [esi] - vpavgb ymm0, ymm0, [esi + edx] - vpavgb ymm0, ymm0, [esi + edx] - sub ecx, 32 - vmovdqu [esi + edi], ymm0 - lea esi, [esi + 32] - jg xloop25 - jmp xloop99 - - // Blend 50 / 50. - align 4 - xloop50: - vmovdqu ymm0, [esi] - vpavgb ymm0, ymm0, [esi + edx] - sub ecx, 32 - vmovdqu [esi + edi], ymm0 - lea esi, [esi + 32] - jg xloop50 - jmp xloop99 - - // Blend 75 / 25. - align 4 - xloop75: - vmovdqu ymm0, [esi + edx] - vpavgb ymm0, ymm0, [esi] - vpavgb ymm0, ymm0, [esi] - sub ecx, 32 - vmovdqu [esi + edi], ymm0 - lea esi, [esi + 32] - jg xloop75 - jmp xloop99 - - // Blend 100 / 0 - Copy row unchanged. - align 4 - xloop100: - rep movsb - - xloop99: - pop edi - pop esi - vzeroupper - ret - } -} -#endif // HAS_INTERPOLATEROW_AVX2 - -#ifdef HAS_INTERPOLATEROW_SSSE3 -// Bilinear filter 16x2 -> 16x1 -__declspec(naked) __declspec(align(16)) -void InterpolateRow_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - __asm { - push esi - push edi - mov edi, [esp + 8 + 4] // dst_ptr - mov esi, [esp + 8 + 8] // src_ptr - mov edx, [esp + 8 + 12] // src_stride - mov ecx, [esp + 8 + 16] // dst_width - mov eax, [esp + 8 + 20] // source_y_fraction (0..255) - sub edi, esi - shr eax, 1 - // Dispatch to specialized filters if applicable. - cmp eax, 0 - je xloop100 // 0 / 128. Blend 100 / 0. - cmp eax, 32 - je xloop75 // 32 / 128 is 0.25. Blend 75 / 25. - cmp eax, 64 - je xloop50 // 64 / 128 is 0.50. Blend 50 / 50. - cmp eax, 96 - je xloop25 // 96 / 128 is 0.75. Blend 25 / 75. - - movd xmm0, eax // high fraction 0..127 - neg eax - add eax, 128 - movd xmm5, eax // low fraction 128..1 - punpcklbw xmm5, xmm0 - punpcklwd xmm5, xmm5 - pshufd xmm5, xmm5, 0 - - align 4 - xloop: - movdqa xmm0, [esi] - movdqa xmm2, [esi + edx] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm2 - punpckhbw xmm1, xmm2 - pmaddubsw xmm0, xmm5 - pmaddubsw xmm1, xmm5 - psrlw xmm0, 7 - psrlw xmm1, 7 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop - jmp xloop99 - - // Blend 25 / 75. - align 4 - xloop25: - movdqa xmm0, [esi] - movdqa xmm1, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop25 - jmp xloop99 - - // Blend 50 / 50. - align 4 - xloop50: - movdqa xmm0, [esi] - movdqa xmm1, [esi + edx] - pavgb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop50 - jmp xloop99 - - // Blend 75 / 25. - align 4 - xloop75: - movdqa xmm1, [esi] - movdqa xmm0, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop75 - jmp xloop99 - - // Blend 100 / 0 - Copy row unchanged. - align 4 - xloop100: - movdqa xmm0, [esi] - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop100 - - xloop99: - pop edi - pop esi - ret - } -} -#endif // HAS_INTERPOLATEROW_SSSE3 - -#ifdef HAS_INTERPOLATEROW_SSE2 -// Bilinear filter 16x2 -> 16x1 -__declspec(naked) __declspec(align(16)) -void InterpolateRow_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - __asm { - push esi - push edi - mov edi, [esp + 8 + 4] // dst_ptr - mov esi, [esp + 8 + 8] // src_ptr - mov edx, [esp + 8 + 12] // src_stride - mov ecx, [esp + 8 + 16] // dst_width - mov eax, [esp + 8 + 20] // source_y_fraction (0..255) - sub edi, esi - // Dispatch to specialized filters if applicable. - cmp eax, 0 - je xloop100 // 0 / 256. Blend 100 / 0. - cmp eax, 64 - je xloop75 // 64 / 256 is 0.25. Blend 75 / 25. - cmp eax, 128 - je xloop50 // 128 / 256 is 0.50. Blend 50 / 50. - cmp eax, 192 - je xloop25 // 192 / 256 is 0.75. Blend 25 / 75. - - movd xmm5, eax // xmm5 = y fraction - punpcklbw xmm5, xmm5 - psrlw xmm5, 1 - punpcklwd xmm5, xmm5 - punpckldq xmm5, xmm5 - punpcklqdq xmm5, xmm5 - pxor xmm4, xmm4 - - align 4 - xloop: - movdqa xmm0, [esi] // row0 - movdqa xmm2, [esi + edx] // row1 - movdqa xmm1, xmm0 - movdqa xmm3, xmm2 - punpcklbw xmm2, xmm4 - punpckhbw xmm3, xmm4 - punpcklbw xmm0, xmm4 - punpckhbw xmm1, xmm4 - psubw xmm2, xmm0 // row1 - row0 - psubw xmm3, xmm1 - paddw xmm2, xmm2 // 9 bits * 15 bits = 8.16 - paddw xmm3, xmm3 - pmulhw xmm2, xmm5 // scale diff - pmulhw xmm3, xmm5 - paddw xmm0, xmm2 // sum rows - paddw xmm1, xmm3 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop - jmp xloop99 - - // Blend 25 / 75. - align 4 - xloop25: - movdqa xmm0, [esi] - movdqa xmm1, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop25 - jmp xloop99 - - // Blend 50 / 50. - align 4 - xloop50: - movdqa xmm0, [esi] - movdqa xmm1, [esi + edx] - pavgb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop50 - jmp xloop99 - - // Blend 75 / 25. - align 4 - xloop75: - movdqa xmm1, [esi] - movdqa xmm0, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop75 - jmp xloop99 - - // Blend 100 / 0 - Copy row unchanged. - align 4 - xloop100: - movdqa xmm0, [esi] - sub ecx, 16 - movdqa [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop100 - - xloop99: - pop edi - pop esi - ret - } -} -#endif // HAS_INTERPOLATEROW_SSE2 - -// Bilinear filter 16x2 -> 16x1 -__declspec(naked) __declspec(align(16)) -void InterpolateRow_Unaligned_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - __asm { - push esi - push edi - mov edi, [esp + 8 + 4] // dst_ptr - mov esi, [esp + 8 + 8] // src_ptr - mov edx, [esp + 8 + 12] // src_stride - mov ecx, [esp + 8 + 16] // dst_width - mov eax, [esp + 8 + 20] // source_y_fraction (0..255) - sub edi, esi - shr eax, 1 - // Dispatch to specialized filters if applicable. - cmp eax, 0 - je xloop100 // 0 / 128. Blend 100 / 0. - cmp eax, 32 - je xloop75 // 32 / 128 is 0.25. Blend 75 / 25. - cmp eax, 64 - je xloop50 // 64 / 128 is 0.50. Blend 50 / 50. - cmp eax, 96 - je xloop25 // 96 / 128 is 0.75. Blend 25 / 75. - - movd xmm0, eax // high fraction 0..127 - neg eax - add eax, 128 - movd xmm5, eax // low fraction 128..1 - punpcklbw xmm5, xmm0 - punpcklwd xmm5, xmm5 - pshufd xmm5, xmm5, 0 - - align 4 - xloop: - movdqu xmm0, [esi] - movdqu xmm2, [esi + edx] - movdqu xmm1, xmm0 - punpcklbw xmm0, xmm2 - punpckhbw xmm1, xmm2 - pmaddubsw xmm0, xmm5 - pmaddubsw xmm1, xmm5 - psrlw xmm0, 7 - psrlw xmm1, 7 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop - jmp xloop99 - - // Blend 25 / 75. - align 4 - xloop25: - movdqu xmm0, [esi] - movdqu xmm1, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop25 - jmp xloop99 - - // Blend 50 / 50. - align 4 - xloop50: - movdqu xmm0, [esi] - movdqu xmm1, [esi + edx] - pavgb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop50 - jmp xloop99 - - // Blend 75 / 25. - align 4 - xloop75: - movdqu xmm1, [esi] - movdqu xmm0, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop75 - jmp xloop99 - - // Blend 100 / 0 - Copy row unchanged. - align 4 - xloop100: - movdqu xmm0, [esi] - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop100 - - xloop99: - pop edi - pop esi - ret - } -} - -#ifdef HAS_INTERPOLATEROW_SSE2 -// Bilinear filter 16x2 -> 16x1 -__declspec(naked) __declspec(align(16)) -void InterpolateRow_Unaligned_SSE2(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, - int source_y_fraction) { - __asm { - push esi - push edi - mov edi, [esp + 8 + 4] // dst_ptr - mov esi, [esp + 8 + 8] // src_ptr - mov edx, [esp + 8 + 12] // src_stride - mov ecx, [esp + 8 + 16] // dst_width - mov eax, [esp + 8 + 20] // source_y_fraction (0..255) - sub edi, esi - // Dispatch to specialized filters if applicable. - cmp eax, 0 - je xloop100 // 0 / 256. Blend 100 / 0. - cmp eax, 64 - je xloop75 // 64 / 256 is 0.25. Blend 75 / 25. - cmp eax, 128 - je xloop50 // 128 / 256 is 0.50. Blend 50 / 50. - cmp eax, 192 - je xloop25 // 192 / 256 is 0.75. Blend 25 / 75. - - movd xmm5, eax // xmm5 = y fraction - punpcklbw xmm5, xmm5 - psrlw xmm5, 1 - punpcklwd xmm5, xmm5 - punpckldq xmm5, xmm5 - punpcklqdq xmm5, xmm5 - pxor xmm4, xmm4 - - align 4 - xloop: - movdqu xmm0, [esi] // row0 - movdqu xmm2, [esi + edx] // row1 - movdqu xmm1, xmm0 - movdqu xmm3, xmm2 - punpcklbw xmm2, xmm4 - punpckhbw xmm3, xmm4 - punpcklbw xmm0, xmm4 - punpckhbw xmm1, xmm4 - psubw xmm2, xmm0 // row1 - row0 - psubw xmm3, xmm1 - paddw xmm2, xmm2 // 9 bits * 15 bits = 8.16 - paddw xmm3, xmm3 - pmulhw xmm2, xmm5 // scale diff - pmulhw xmm3, xmm5 - paddw xmm0, xmm2 // sum rows - paddw xmm1, xmm3 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop - jmp xloop99 - - // Blend 25 / 75. - align 4 - xloop25: - movdqu xmm0, [esi] - movdqu xmm1, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop25 - jmp xloop99 - - // Blend 50 / 50. - align 4 - xloop50: - movdqu xmm0, [esi] - movdqu xmm1, [esi + edx] - pavgb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop50 - jmp xloop99 - - // Blend 75 / 25. - align 4 - xloop75: - movdqu xmm1, [esi] - movdqu xmm0, [esi + edx] - pavgb xmm0, xmm1 - pavgb xmm0, xmm1 - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop75 - jmp xloop99 - - // Blend 100 / 0 - Copy row unchanged. - align 4 - xloop100: - movdqu xmm0, [esi] - sub ecx, 16 - movdqu [esi + edi], xmm0 - lea esi, [esi + 16] - jg xloop100 - - xloop99: - pop edi - pop esi - ret - } -} -#endif // HAS_INTERPOLATEROW_SSE2 - -__declspec(naked) __declspec(align(16)) -void HalfRow_SSE2(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_uv - mov edx, [esp + 4 + 8] // src_uv_stride - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - sub edi, eax - - align 4 - convertloop: - movdqa xmm0, [eax] - pavgb xmm0, [eax + edx] - sub ecx, 16 - movdqa [eax + edi], xmm0 - lea eax, [eax + 16] - jg convertloop - pop edi - ret - } -} - -#ifdef HAS_HALFROW_AVX2 -__declspec(naked) __declspec(align(16)) -void HalfRow_AVX2(const uint8* src_uv, int src_uv_stride, - uint8* dst_uv, int pix) { - __asm { - push edi - mov eax, [esp + 4 + 4] // src_uv - mov edx, [esp + 4 + 8] // src_uv_stride - mov edi, [esp + 4 + 12] // dst_v - mov ecx, [esp + 4 + 16] // pix - sub edi, eax - - align 4 - convertloop: - vmovdqu ymm0, [eax] - vpavgb ymm0, ymm0, [eax + edx] - sub ecx, 32 - vmovdqu [eax + edi], ymm0 - lea eax, [eax + 32] - jg convertloop - - pop edi - vzeroupper - ret - } -} -#endif // HAS_HALFROW_AVX2 - -__declspec(naked) __declspec(align(16)) -void ARGBToBayerRow_SSSE3(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_bayer - movd xmm5, [esp + 12] // selector - mov ecx, [esp + 16] // pix - pshufd xmm5, xmm5, 0 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - pshufb xmm0, xmm5 - pshufb xmm1, xmm5 - punpckldq xmm0, xmm1 - sub ecx, 8 - movq qword ptr [edx], xmm0 - lea edx, [edx + 8] - jg wloop - ret - } -} - -// Specialized ARGB to Bayer that just isolates G channel. -__declspec(naked) __declspec(align(16)) -void ARGBToBayerGGRow_SSE2(const uint8* src_argb, uint8* dst_bayer, - uint32 selector, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_bayer - // selector - mov ecx, [esp + 16] // pix - pcmpeqb xmm5, xmm5 // generate mask 0x000000ff - psrld xmm5, 24 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - psrld xmm0, 8 // Move green to bottom. - psrld xmm1, 8 - pand xmm0, xmm5 - pand xmm1, xmm5 - packssdw xmm0, xmm1 - packuswb xmm0, xmm1 - sub ecx, 8 - movq qword ptr [edx], xmm0 - lea edx, [edx + 8] - jg wloop - ret - } -} - -// For BGRAToARGB, ABGRToARGB, RGBAToARGB, and ARGBToRGBA. -__declspec(naked) __declspec(align(16)) -void ARGBShuffleRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // shuffler - movdqa xmm5, [ecx] - mov ecx, [esp + 16] // pix - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - pshufb xmm0, xmm5 - pshufb xmm1, xmm5 - sub ecx, 8 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - jg wloop - ret - } -} - -__declspec(naked) __declspec(align(16)) -void ARGBShuffleRow_Unaligned_SSSE3(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // shuffler - movdqa xmm5, [ecx] - mov ecx, [esp + 16] // pix - - align 4 - wloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - pshufb xmm0, xmm5 - pshufb xmm1, xmm5 - sub ecx, 8 - movdqu [edx], xmm0 - movdqu [edx + 16], xmm1 - lea edx, [edx + 32] - jg wloop - ret - } -} - -#ifdef HAS_ARGBSHUFFLEROW_AVX2 -__declspec(naked) __declspec(align(16)) -void ARGBShuffleRow_AVX2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - __asm { - mov eax, [esp + 4] // src_argb - mov edx, [esp + 8] // dst_argb - mov ecx, [esp + 12] // shuffler - vbroadcastf128 ymm5, [ecx] // same shuffle in high as low. - mov ecx, [esp + 16] // pix - - align 4 - wloop: - vmovdqu ymm0, [eax] - vmovdqu ymm1, [eax + 32] - lea eax, [eax + 64] - vpshufb ymm0, ymm0, ymm5 - vpshufb ymm1, ymm1, ymm5 - sub ecx, 16 - vmovdqu [edx], ymm0 - vmovdqu [edx + 32], ymm1 - lea edx, [edx + 64] - jg wloop - - vzeroupper - ret - } -} -#endif // HAS_ARGBSHUFFLEROW_AVX2 - -__declspec(naked) __declspec(align(16)) -void ARGBShuffleRow_SSE2(const uint8* src_argb, uint8* dst_argb, - const uint8* shuffler, int pix) { - __asm { - push ebx - push esi - mov eax, [esp + 8 + 4] // src_argb - mov edx, [esp + 8 + 8] // dst_argb - mov esi, [esp + 8 + 12] // shuffler - mov ecx, [esp + 8 + 16] // pix - pxor xmm5, xmm5 - - mov ebx, [esi] // shuffler - cmp ebx, 0x03000102 - je shuf_3012 - cmp ebx, 0x00010203 - je shuf_0123 - cmp ebx, 0x00030201 - je shuf_0321 - cmp ebx, 0x02010003 - je shuf_2103 - - // TODO(fbarchard): Use one source pointer and 3 offsets. - shuf_any1: - movzx ebx, byte ptr [esi] - movzx ebx, byte ptr [eax + ebx] - mov [edx], bl - movzx ebx, byte ptr [esi + 1] - movzx ebx, byte ptr [eax + ebx] - mov [edx + 1], bl - movzx ebx, byte ptr [esi + 2] - movzx ebx, byte ptr [eax + ebx] - mov [edx + 2], bl - movzx ebx, byte ptr [esi + 3] - movzx ebx, byte ptr [eax + ebx] - mov [edx + 3], bl - lea eax, [eax + 4] - lea edx, [edx + 4] - sub ecx, 1 - jg shuf_any1 - jmp shuf99 - - align 4 - shuf_0123: - movdqu xmm0, [eax] - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm5 - punpckhbw xmm1, xmm5 - pshufhw xmm0, xmm0, 01Bh // 1B = 00011011 = 0x0123 = BGRAToARGB - pshuflw xmm0, xmm0, 01Bh - pshufhw xmm1, xmm1, 01Bh - pshuflw xmm1, xmm1, 01Bh - packuswb xmm0, xmm1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg shuf_0123 - jmp shuf99 - - align 4 - shuf_0321: - movdqu xmm0, [eax] - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm5 - punpckhbw xmm1, xmm5 - pshufhw xmm0, xmm0, 039h // 39 = 00111001 = 0x0321 = RGBAToARGB - pshuflw xmm0, xmm0, 039h - pshufhw xmm1, xmm1, 039h - pshuflw xmm1, xmm1, 039h - packuswb xmm0, xmm1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg shuf_0321 - jmp shuf99 - - align 4 - shuf_2103: - movdqu xmm0, [eax] - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm5 - punpckhbw xmm1, xmm5 - pshufhw xmm0, xmm0, 093h // 93 = 10010011 = 0x2103 = ARGBToRGBA - pshuflw xmm0, xmm0, 093h - pshufhw xmm1, xmm1, 093h - pshuflw xmm1, xmm1, 093h - packuswb xmm0, xmm1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg shuf_2103 - jmp shuf99 - - align 4 - shuf_3012: - movdqu xmm0, [eax] - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm5 - punpckhbw xmm1, xmm5 - pshufhw xmm0, xmm0, 0C6h // C6 = 11000110 = 0x3012 = ABGRToARGB - pshuflw xmm0, xmm0, 0C6h - pshufhw xmm1, xmm1, 0C6h - pshuflw xmm1, xmm1, 0C6h - packuswb xmm0, xmm1 - sub ecx, 4 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg shuf_3012 - - shuf99: - pop esi - pop ebx - ret - } -} - -// YUY2 - Macro-pixel = 2 image pixels -// Y0U0Y1V0....Y2U2Y3V2...Y4U4Y5V4.... - -// UYVY - Macro-pixel = 2 image pixels -// U0Y0V0Y1 - -__declspec(naked) __declspec(align(16)) -void I422ToYUY2Row_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_frame, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_y - mov esi, [esp + 8 + 8] // src_u - mov edx, [esp + 8 + 12] // src_v - mov edi, [esp + 8 + 16] // dst_frame - mov ecx, [esp + 8 + 20] // width - sub edx, esi - - align 4 - convertloop: - movq xmm2, qword ptr [esi] // U - movq xmm3, qword ptr [esi + edx] // V - lea esi, [esi + 8] - punpcklbw xmm2, xmm3 // UV - movdqu xmm0, [eax] // Y - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm2 // YUYV - punpckhbw xmm1, xmm2 - movdqu [edi], xmm0 - movdqu [edi + 16], xmm1 - lea edi, [edi + 32] - sub ecx, 16 - jg convertloop - - pop edi - pop esi - ret - } -} - -__declspec(naked) __declspec(align(16)) -void I422ToUYVYRow_SSE2(const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_frame, int width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_y - mov esi, [esp + 8 + 8] // src_u - mov edx, [esp + 8 + 12] // src_v - mov edi, [esp + 8 + 16] // dst_frame - mov ecx, [esp + 8 + 20] // width - sub edx, esi - - align 4 - convertloop: - movq xmm2, qword ptr [esi] // U - movq xmm3, qword ptr [esi + edx] // V - lea esi, [esi + 8] - punpcklbw xmm2, xmm3 // UV - movdqu xmm0, [eax] // Y - movdqa xmm1, xmm2 - lea eax, [eax + 16] - punpcklbw xmm1, xmm0 // UYVY - punpckhbw xmm2, xmm0 - movdqu [edi], xmm1 - movdqu [edi + 16], xmm2 - lea edi, [edi + 32] - sub ecx, 16 - jg convertloop - - pop edi - pop esi - ret - } -} - -#ifdef HAS_ARGBPOLYNOMIALROW_SSE2 -__declspec(naked) __declspec(align(16)) -void ARGBPolynomialRow_SSE2(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] /* src_argb */ - mov edx, [esp + 4 + 8] /* dst_argb */ - mov esi, [esp + 4 + 12] /* poly */ - mov ecx, [esp + 4 + 16] /* width */ - pxor xmm3, xmm3 // 0 constant for zero extending bytes to ints. - - // 2 pixel loop. - align 4 - convertloop: -// pmovzxbd xmm0, dword ptr [eax] // BGRA pixel -// pmovzxbd xmm4, dword ptr [eax + 4] // BGRA pixel - movq xmm0, qword ptr [eax] // BGRABGRA - lea eax, [eax + 8] - punpcklbw xmm0, xmm3 - movdqa xmm4, xmm0 - punpcklwd xmm0, xmm3 // pixel 0 - punpckhwd xmm4, xmm3 // pixel 1 - cvtdq2ps xmm0, xmm0 // 4 floats - cvtdq2ps xmm4, xmm4 - movdqa xmm1, xmm0 // X - movdqa xmm5, xmm4 - mulps xmm0, [esi + 16] // C1 * X - mulps xmm4, [esi + 16] - addps xmm0, [esi] // result = C0 + C1 * X - addps xmm4, [esi] - movdqa xmm2, xmm1 - movdqa xmm6, xmm5 - mulps xmm2, xmm1 // X * X - mulps xmm6, xmm5 - mulps xmm1, xmm2 // X * X * X - mulps xmm5, xmm6 - mulps xmm2, [esi + 32] // C2 * X * X - mulps xmm6, [esi + 32] - mulps xmm1, [esi + 48] // C3 * X * X * X - mulps xmm5, [esi + 48] - addps xmm0, xmm2 // result += C2 * X * X - addps xmm4, xmm6 - addps xmm0, xmm1 // result += C3 * X * X * X - addps xmm4, xmm5 - cvttps2dq xmm0, xmm0 - cvttps2dq xmm4, xmm4 - packuswb xmm0, xmm4 - packuswb xmm0, xmm0 - sub ecx, 2 - movq qword ptr [edx], xmm0 - lea edx, [edx + 8] - jg convertloop - pop esi - ret - } -} -#endif // HAS_ARGBPOLYNOMIALROW_SSE2 - -#ifdef HAS_ARGBPOLYNOMIALROW_AVX2 -__declspec(naked) __declspec(align(16)) -void ARGBPolynomialRow_AVX2(const uint8* src_argb, - uint8* dst_argb, const float* poly, - int width) { - __asm { - mov eax, [esp + 4] /* src_argb */ - mov edx, [esp + 8] /* dst_argb */ - mov ecx, [esp + 12] /* poly */ - vbroadcastf128 ymm4, [ecx] // C0 - vbroadcastf128 ymm5, [ecx + 16] // C1 - vbroadcastf128 ymm6, [ecx + 32] // C2 - vbroadcastf128 ymm7, [ecx + 48] // C3 - mov ecx, [esp + 16] /* width */ - - // 2 pixel loop. - align 4 - convertloop: - vpmovzxbd ymm0, qword ptr [eax] // 2 BGRA pixels - lea eax, [eax + 8] - vcvtdq2ps ymm0, ymm0 // X 8 floats - vmulps ymm2, ymm0, ymm0 // X * X - vmulps ymm3, ymm0, ymm7 // C3 * X - vfmadd132ps ymm0, ymm4, ymm5 // result = C0 + C1 * X - vfmadd231ps ymm0, ymm2, ymm6 // result += C2 * X * X - vfmadd231ps ymm0, ymm2, ymm3 // result += C3 * X * X * X - vcvttps2dq ymm0, ymm0 - vpackusdw ymm0, ymm0, ymm0 // b0g0r0a0_00000000_b0g0r0a0_00000000 - vpermq ymm0, ymm0, 0xd8 // b0g0r0a0_b0g0r0a0_00000000_00000000 - vpackuswb xmm0, xmm0, xmm0 // bgrabgra_00000000_00000000_00000000 - sub ecx, 2 - vmovq qword ptr [edx], xmm0 - lea edx, [edx + 8] - jg convertloop - vzeroupper - ret - } -} -#endif // HAS_ARGBPOLYNOMIALROW_AVX2 - -#ifdef HAS_ARGBCOLORTABLEROW_X86 -// Tranform ARGB pixels with color table. -__declspec(naked) __declspec(align(16)) -void ARGBColorTableRow_X86(uint8* dst_argb, const uint8* table_argb, - int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] /* dst_argb */ - mov esi, [esp + 4 + 8] /* table_argb */ - mov ecx, [esp + 4 + 12] /* width */ - - // 1 pixel loop. - align 4 - convertloop: - movzx edx, byte ptr [eax] - lea eax, [eax + 4] - movzx edx, byte ptr [esi + edx * 4] - mov byte ptr [eax - 4], dl - movzx edx, byte ptr [eax - 4 + 1] - movzx edx, byte ptr [esi + edx * 4 + 1] - mov byte ptr [eax - 4 + 1], dl - movzx edx, byte ptr [eax - 4 + 2] - movzx edx, byte ptr [esi + edx * 4 + 2] - mov byte ptr [eax - 4 + 2], dl - movzx edx, byte ptr [eax - 4 + 3] - movzx edx, byte ptr [esi + edx * 4 + 3] - mov byte ptr [eax - 4 + 3], dl - dec ecx - jg convertloop - pop esi - ret - } -} -#endif // HAS_ARGBCOLORTABLEROW_X86 - -#ifdef HAS_RGBCOLORTABLEROW_X86 -// Tranform RGB pixels with color table. -__declspec(naked) __declspec(align(16)) -void RGBColorTableRow_X86(uint8* dst_argb, const uint8* table_argb, int width) { - __asm { - push esi - mov eax, [esp + 4 + 4] /* dst_argb */ - mov esi, [esp + 4 + 8] /* table_argb */ - mov ecx, [esp + 4 + 12] /* width */ - - // 1 pixel loop. - align 4 - convertloop: - movzx edx, byte ptr [eax] - lea eax, [eax + 4] - movzx edx, byte ptr [esi + edx * 4] - mov byte ptr [eax - 4], dl - movzx edx, byte ptr [eax - 4 + 1] - movzx edx, byte ptr [esi + edx * 4 + 1] - mov byte ptr [eax - 4 + 1], dl - movzx edx, byte ptr [eax - 4 + 2] - movzx edx, byte ptr [esi + edx * 4 + 2] - mov byte ptr [eax - 4 + 2], dl - dec ecx - jg convertloop - - pop esi - ret - } -} -#endif // HAS_RGBCOLORTABLEROW_X86 - -#ifdef HAS_ARGBLUMACOLORTABLEROW_SSSE3 -// Tranform RGB pixels with luma table. -__declspec(naked) __declspec(align(16)) -void ARGBLumaColorTableRow_SSSE3(const uint8* src_argb, uint8* dst_argb, - int width, - const uint8* luma, uint32 lumacoeff) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] /* src_argb */ - mov edi, [esp + 8 + 8] /* dst_argb */ - mov ecx, [esp + 8 + 12] /* width */ - movd xmm2, dword ptr [esp + 8 + 16] // luma table - movd xmm3, dword ptr [esp + 8 + 20] // lumacoeff - pshufd xmm2, xmm2, 0 - pshufd xmm3, xmm3, 0 - pcmpeqb xmm4, xmm4 // generate mask 0xff00ff00 - psllw xmm4, 8 - pxor xmm5, xmm5 - - // 4 pixel loop. - align 4 - convertloop: - movdqu xmm0, qword ptr [eax] // generate luma ptr - pmaddubsw xmm0, xmm3 - phaddw xmm0, xmm0 - pand xmm0, xmm4 // mask out low bits - punpcklwd xmm0, xmm5 - paddd xmm0, xmm2 // add table base - movd esi, xmm0 - pshufd xmm0, xmm0, 0x39 // 00111001 to rotate right 32 - - movzx edx, byte ptr [eax] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi], dl - movzx edx, byte ptr [eax + 1] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 1], dl - movzx edx, byte ptr [eax + 2] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 2], dl - movzx edx, byte ptr [eax + 3] // copy alpha. - mov byte ptr [edi + 3], dl - - movd esi, xmm0 - pshufd xmm0, xmm0, 0x39 // 00111001 to rotate right 32 - - movzx edx, byte ptr [eax + 4] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 4], dl - movzx edx, byte ptr [eax + 5] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 5], dl - movzx edx, byte ptr [eax + 6] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 6], dl - movzx edx, byte ptr [eax + 7] // copy alpha. - mov byte ptr [edi + 7], dl - - movd esi, xmm0 - pshufd xmm0, xmm0, 0x39 // 00111001 to rotate right 32 - - movzx edx, byte ptr [eax + 8] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 8], dl - movzx edx, byte ptr [eax + 9] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 9], dl - movzx edx, byte ptr [eax + 10] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 10], dl - movzx edx, byte ptr [eax + 11] // copy alpha. - mov byte ptr [edi + 11], dl - - movd esi, xmm0 - - movzx edx, byte ptr [eax + 12] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 12], dl - movzx edx, byte ptr [eax + 13] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 13], dl - movzx edx, byte ptr [eax + 14] - movzx edx, byte ptr [esi + edx] - mov byte ptr [edi + 14], dl - movzx edx, byte ptr [eax + 15] // copy alpha. - mov byte ptr [edi + 15], dl - - sub ecx, 4 - lea eax, [eax + 16] - lea edi, [edi + 16] - jg convertloop - - pop edi - pop esi - ret - } -} -#endif // HAS_ARGBLUMACOLORTABLEROW_SSSE3 - -#endif // defined(_M_X64) -#endif // !defined(LIBYUV_DISABLE_X86) && defined(_MSC_VER) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/row_x86.asm b/src/main/jni/libyuv/source/row_x86.asm deleted file mode 100644 index 0cb326f8e..000000000 --- a/src/main/jni/libyuv/source/row_x86.asm +++ /dev/null @@ -1,146 +0,0 @@ -; -; Copyright 2012 The LibYuv Project Authors. All rights reserved. -; -; Use of this source code is governed by a BSD-style license -; that can be found in the LICENSE file in the root of the source -; tree. An additional intellectual property rights grant can be found -; in the file PATENTS. All contributing project authors may -; be found in the AUTHORS file in the root of the source tree. -; - -%ifdef __YASM_VERSION_ID__ -%if __YASM_VERSION_ID__ < 01020000h -%error AVX2 is supported only by yasm 1.2.0 or later. -%endif -%endif -%include "x86inc.asm" - -SECTION .text - -; cglobal numeric constants are parameters, gpr regs, mm regs - -; void YUY2ToYRow_SSE2(const uint8* src_yuy2, uint8* dst_y, int pix) - -%macro YUY2TOYROW 2-3 -cglobal %1ToYRow%3, 3, 3, 3, src_yuy2, dst_y, pix -%ifidn %1,YUY2 - pcmpeqb m2, m2, m2 ; generate mask 0x00ff00ff - psrlw m2, m2, 8 -%endif - - ALIGN 4 -.convertloop: - mov%2 m0, [src_yuy2q] - mov%2 m1, [src_yuy2q + mmsize] - lea src_yuy2q, [src_yuy2q + mmsize * 2] -%ifidn %1,YUY2 - pand m0, m0, m2 ; YUY2 even bytes are Y - pand m1, m1, m2 -%else - psrlw m0, m0, 8 ; UYVY odd bytes are Y - psrlw m1, m1, 8 -%endif - packuswb m0, m0, m1 -%if cpuflag(AVX2) - vpermq m0, m0, 0xd8 -%endif - sub pixd, mmsize - mov%2 [dst_yq], m0 - lea dst_yq, [dst_yq + mmsize] - jg .convertloop - REP_RET -%endmacro - -; TODO(fbarchard): Remove MMX. Add SSSE3 pshufb version. -INIT_MMX MMX -YUY2TOYROW YUY2,a, -YUY2TOYROW YUY2,u,_Unaligned -YUY2TOYROW UYVY,a, -YUY2TOYROW UYVY,u,_Unaligned -INIT_XMM SSE2 -YUY2TOYROW YUY2,a, -YUY2TOYROW YUY2,u,_Unaligned -YUY2TOYROW UYVY,a, -YUY2TOYROW UYVY,u,_Unaligned -INIT_YMM AVX2 -YUY2TOYROW YUY2,a, -YUY2TOYROW UYVY,a, - -; void SplitUVRow_SSE2(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix) - -%macro SplitUVRow 1-2 -cglobal SplitUVRow%2, 4, 4, 5, src_uv, dst_u, dst_v, pix - pcmpeqb m4, m4, m4 ; generate mask 0x00ff00ff - psrlw m4, m4, 8 - sub dst_vq, dst_uq - - ALIGN 4 -.convertloop: - mov%1 m0, [src_uvq] - mov%1 m1, [src_uvq + mmsize] - lea src_uvq, [src_uvq + mmsize * 2] - psrlw m2, m0, 8 ; odd bytes - psrlw m3, m1, 8 - pand m0, m0, m4 ; even bytes - pand m1, m1, m4 - packuswb m0, m0, m1 - packuswb m2, m2, m3 -%if cpuflag(AVX2) - vpermq m0, m0, 0xd8 - vpermq m2, m2, 0xd8 -%endif - mov%1 [dst_uq], m0 - mov%1 [dst_uq + dst_vq], m2 - lea dst_uq, [dst_uq + mmsize] - sub pixd, mmsize - jg .convertloop - REP_RET -%endmacro - -INIT_MMX MMX -SplitUVRow a, -SplitUVRow u,_Unaligned -INIT_XMM SSE2 -SplitUVRow a, -SplitUVRow u,_Unaligned -INIT_YMM AVX2 -SplitUVRow a, - -; void MergeUVRow_SSE2(const uint8* src_u, const uint8* src_v, uint8* dst_uv, -; int width); - -%macro MergeUVRow_ 1-2 -cglobal MergeUVRow_%2, 4, 4, 3, src_u, src_v, dst_uv, pix - sub src_vq, src_uq - - ALIGN 4 -.convertloop: - mov%1 m0, [src_uq] - mov%1 m1, [src_vq] - lea src_uq, [src_uq + mmsize] - punpcklbw m2, m0, m1 // first 8 UV pairs - punpckhbw m0, m0, m1 // next 8 UV pairs -%if cpuflag(AVX2) - vperm2i128 m1, m2, m0, 0x20 // low 128 of ymm2 and low 128 of ymm0 - vperm2i128 m2, m2, m0, 0x31 // high 128 of ymm2 and high 128 of ymm0 - mov%1 [dst_uvq], m1 - mov%1 [dst_uvq + mmsize], m2 -%else - mov%1 [dst_uvq], m2 - mov%1 [dst_uvq + mmsize], m0 -%endif - lea dst_uvq, [dst_uvq + mmsize * 2] - sub pixd, mmsize - jg .convertloop - REP_RET -%endmacro - -INIT_MMX MMX -MergeUVRow_ a, -MergeUVRow_ u,_Unaligned -INIT_XMM SSE2 -MergeUVRow_ a, -MergeUVRow_ u,_Unaligned -INIT_YMM AVX2 -MergeUVRow_ a, - diff --git a/src/main/jni/libyuv/source/scale.cc b/src/main/jni/libyuv/source/scale.cc deleted file mode 100644 index 5b33b5f04..000000000 --- a/src/main/jni/libyuv/source/scale.cc +++ /dev/null @@ -1,1716 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/scale.h" - -#include <assert.h> -#include <string.h> - -#include "libyuv/cpu_id.h" -#include "libyuv/planar_functions.h" // For CopyPlane -#include "libyuv/row.h" -#include "libyuv/scale_row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// Remove this macro if OVERREAD is safe. -#define AVOID_OVERREAD 1 - -static __inline int Abs(int v) { - return v >= 0 ? v : -v; -} - -#define SUBSAMPLE(v, a, s) (v < 0) ? (-((-v + a) >> s)) : ((v + a) >> s) - -// Scale plane, 1/2 -// This is an optimized version for scaling down a plane to 1/2 of -// its original size. - -static void ScalePlaneDown2(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown2)(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) = - filtering == kFilterNone ? ScaleRowDown2_C : - (filtering == kFilterLinear ? ScaleRowDown2Linear_C : - ScaleRowDown2Box_C); - int row_stride = src_stride << 1; - if (!filtering) { - src_ptr += src_stride; // Point to odd rows. - src_stride = 0; - } - -#if defined(HAS_SCALEROWDOWN2_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 16)) { - ScaleRowDown2 = filtering ? ScaleRowDown2Box_NEON : ScaleRowDown2_NEON; - } -#elif defined(HAS_SCALEROWDOWN2_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 16)) { - ScaleRowDown2 = filtering == kFilterNone ? ScaleRowDown2_Unaligned_SSE2 : - (filtering == kFilterLinear ? ScaleRowDown2Linear_Unaligned_SSE2 : - ScaleRowDown2Box_Unaligned_SSE2); - if (IS_ALIGNED(src_ptr, 16) && - IS_ALIGNED(src_stride, 16) && IS_ALIGNED(row_stride, 16) && - IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleRowDown2 = filtering == kFilterNone ? ScaleRowDown2_SSE2 : - (filtering == kFilterLinear ? ScaleRowDown2Linear_SSE2 : - ScaleRowDown2Box_SSE2); - } - } -#elif defined(HAS_SCALEROWDOWN2_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(src_ptr, 4) && - IS_ALIGNED(src_stride, 4) && IS_ALIGNED(row_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - ScaleRowDown2 = filtering ? - ScaleRowDown2Box_MIPS_DSPR2 : ScaleRowDown2_MIPS_DSPR2; - } -#endif - - if (filtering == kFilterLinear) { - src_stride = 0; - } - // TODO(fbarchard): Loop through source height to allow odd height. - for (y = 0; y < dst_height; ++y) { - ScaleRowDown2(src_ptr, src_stride, dst_ptr, dst_width); - src_ptr += row_stride; - dst_ptr += dst_stride; - } -} - -static void ScalePlaneDown2_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown2)(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width) = - filtering == kFilterNone ? ScaleRowDown2_16_C : - (filtering == kFilterLinear ? ScaleRowDown2Linear_16_C : - ScaleRowDown2Box_16_C); - int row_stride = src_stride << 1; - if (!filtering) { - src_ptr += src_stride; // Point to odd rows. - src_stride = 0; - } - -#if defined(HAS_SCALEROWDOWN2_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 16)) { - ScaleRowDown2 = filtering ? ScaleRowDown2Box_16_NEON : - ScaleRowDown2_16_NEON; - } -#elif defined(HAS_SCALEROWDOWN2_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 16)) { - ScaleRowDown2 = filtering == kFilterNone ? - ScaleRowDown2_Unaligned_16_SSE2 : - (filtering == kFilterLinear ? ScaleRowDown2Linear_Unaligned_16_SSE2 : - ScaleRowDown2Box_Unaligned_16_SSE2); - if (IS_ALIGNED(src_ptr, 16) && - IS_ALIGNED(src_stride, 16) && IS_ALIGNED(row_stride, 16) && - IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleRowDown2 = filtering == kFilterNone ? ScaleRowDown2_16_SSE2 : - (filtering == kFilterLinear ? ScaleRowDown2Linear_16_SSE2 : - ScaleRowDown2Box_16_SSE2); - } - } -#elif defined(HAS_SCALEROWDOWN2_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(src_ptr, 4) && - IS_ALIGNED(src_stride, 4) && IS_ALIGNED(row_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - ScaleRowDown2 = filtering ? - ScaleRowDown2Box_16_MIPS_DSPR2 : ScaleRowDown2_16_MIPS_DSPR2; - } -#endif - - if (filtering == kFilterLinear) { - src_stride = 0; - } - // TODO(fbarchard): Loop through source height to allow odd height. - for (y = 0; y < dst_height; ++y) { - ScaleRowDown2(src_ptr, src_stride, dst_ptr, dst_width); - src_ptr += row_stride; - dst_ptr += dst_stride; - } -} - -// Scale plane, 1/4 -// This is an optimized version for scaling down a plane to 1/4 of -// its original size. - -static void ScalePlaneDown4(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown4)(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) = - filtering ? ScaleRowDown4Box_C : ScaleRowDown4_C; - int row_stride = src_stride << 2; - if (!filtering) { - src_ptr += src_stride * 2; // Point to row 2. - src_stride = 0; - } -#if defined(HAS_SCALEROWDOWN4_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 8)) { - ScaleRowDown4 = filtering ? ScaleRowDown4Box_NEON : ScaleRowDown4_NEON; - } -#elif defined(HAS_SCALEROWDOWN4_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && - IS_ALIGNED(dst_width, 8) && IS_ALIGNED(row_stride, 16) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - ScaleRowDown4 = filtering ? ScaleRowDown4Box_SSE2 : ScaleRowDown4_SSE2; - } -#elif defined(HAS_SCALEROWDOWN4_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(row_stride, 4) && - IS_ALIGNED(src_ptr, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - ScaleRowDown4 = filtering ? - ScaleRowDown4Box_MIPS_DSPR2 : ScaleRowDown4_MIPS_DSPR2; - } -#endif - - if (filtering == kFilterLinear) { - src_stride = 0; - } - for (y = 0; y < dst_height; ++y) { - ScaleRowDown4(src_ptr, src_stride, dst_ptr, dst_width); - src_ptr += row_stride; - dst_ptr += dst_stride; - } -} - -static void ScalePlaneDown4_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown4)(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width) = - filtering ? ScaleRowDown4Box_16_C : ScaleRowDown4_16_C; - int row_stride = src_stride << 2; - if (!filtering) { - src_ptr += src_stride * 2; // Point to row 2. - src_stride = 0; - } -#if defined(HAS_SCALEROWDOWN4_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 8)) { - ScaleRowDown4 = filtering ? ScaleRowDown4Box_16_NEON : - ScaleRowDown4_16_NEON; - } -#elif defined(HAS_SCALEROWDOWN4_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && - IS_ALIGNED(dst_width, 8) && IS_ALIGNED(row_stride, 16) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - ScaleRowDown4 = filtering ? ScaleRowDown4Box_16_SSE2 : - ScaleRowDown4_16_SSE2; - } -#elif defined(HAS_SCALEROWDOWN4_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(row_stride, 4) && - IS_ALIGNED(src_ptr, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - ScaleRowDown4 = filtering ? - ScaleRowDown4Box_16_MIPS_DSPR2 : ScaleRowDown4_16_MIPS_DSPR2; - } -#endif - - if (filtering == kFilterLinear) { - src_stride = 0; - } - for (y = 0; y < dst_height; ++y) { - ScaleRowDown4(src_ptr, src_stride, dst_ptr, dst_width); - src_ptr += row_stride; - dst_ptr += dst_stride; - } -} - -// Scale plane down, 3/4 - -static void ScalePlaneDown34(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown34_0)(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - void (*ScaleRowDown34_1)(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - const int filter_stride = (filtering == kFilterLinear) ? 0 : src_stride; - assert(dst_width % 3 == 0); - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_C; - ScaleRowDown34_1 = ScaleRowDown34_C; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_C; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_C; - } -#if defined(HAS_SCALEROWDOWN34_NEON) - if (TestCpuFlag(kCpuHasNEON) && (dst_width % 24 == 0)) { - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_NEON; - ScaleRowDown34_1 = ScaleRowDown34_NEON; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_NEON; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_NEON; - } - } -#endif -#if defined(HAS_SCALEROWDOWN34_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && (dst_width % 24 == 0) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_SSSE3; - ScaleRowDown34_1 = ScaleRowDown34_SSSE3; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_SSSE3; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_SSSE3; - } - } -#endif -#if defined(HAS_SCALEROWDOWN34_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && (dst_width % 24 == 0) && - IS_ALIGNED(src_ptr, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_MIPS_DSPR2; - ScaleRowDown34_1 = ScaleRowDown34_MIPS_DSPR2; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_MIPS_DSPR2; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_MIPS_DSPR2; - } - } -#endif - - for (y = 0; y < dst_height - 2; y += 3) { - ScaleRowDown34_0(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride; - dst_ptr += dst_stride; - ScaleRowDown34_1(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride; - dst_ptr += dst_stride; - ScaleRowDown34_0(src_ptr + src_stride, -filter_stride, - dst_ptr, dst_width); - src_ptr += src_stride * 2; - dst_ptr += dst_stride; - } - - // Remainder 1 or 2 rows with last row vertically unfiltered - if ((dst_height % 3) == 2) { - ScaleRowDown34_0(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride; - dst_ptr += dst_stride; - ScaleRowDown34_1(src_ptr, 0, dst_ptr, dst_width); - } else if ((dst_height % 3) == 1) { - ScaleRowDown34_0(src_ptr, 0, dst_ptr, dst_width); - } -} - -static void ScalePlaneDown34_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown34_0)(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width); - void (*ScaleRowDown34_1)(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width); - const int filter_stride = (filtering == kFilterLinear) ? 0 : src_stride; - assert(dst_width % 3 == 0); - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_16_C; - ScaleRowDown34_1 = ScaleRowDown34_16_C; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_16_C; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_16_C; - } -#if defined(HAS_SCALEROWDOWN34_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && (dst_width % 24 == 0)) { - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_16_NEON; - ScaleRowDown34_1 = ScaleRowDown34_16_NEON; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_16_NEON; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_16_NEON; - } - } -#endif -#if defined(HAS_SCALEROWDOWN34_16_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && (dst_width % 24 == 0) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_16_SSSE3; - ScaleRowDown34_1 = ScaleRowDown34_16_SSSE3; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_16_SSSE3; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_16_SSSE3; - } - } -#endif -#if defined(HAS_SCALEROWDOWN34_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && (dst_width % 24 == 0) && - IS_ALIGNED(src_ptr, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - if (!filtering) { - ScaleRowDown34_0 = ScaleRowDown34_16_MIPS_DSPR2; - ScaleRowDown34_1 = ScaleRowDown34_16_MIPS_DSPR2; - } else { - ScaleRowDown34_0 = ScaleRowDown34_0_Box_16_MIPS_DSPR2; - ScaleRowDown34_1 = ScaleRowDown34_1_Box_16_MIPS_DSPR2; - } - } -#endif - - for (y = 0; y < dst_height - 2; y += 3) { - ScaleRowDown34_0(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride; - dst_ptr += dst_stride; - ScaleRowDown34_1(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride; - dst_ptr += dst_stride; - ScaleRowDown34_0(src_ptr + src_stride, -filter_stride, - dst_ptr, dst_width); - src_ptr += src_stride * 2; - dst_ptr += dst_stride; - } - - // Remainder 1 or 2 rows with last row vertically unfiltered - if ((dst_height % 3) == 2) { - ScaleRowDown34_0(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride; - dst_ptr += dst_stride; - ScaleRowDown34_1(src_ptr, 0, dst_ptr, dst_width); - } else if ((dst_height % 3) == 1) { - ScaleRowDown34_0(src_ptr, 0, dst_ptr, dst_width); - } -} - - -// Scale plane, 3/8 -// This is an optimized version for scaling down a plane to 3/8 -// of its original size. -// -// Uses box filter arranges like this -// aaabbbcc -> abc -// aaabbbcc def -// aaabbbcc ghi -// dddeeeff -// dddeeeff -// dddeeeff -// ggghhhii -// ggghhhii -// Boxes are 3x3, 2x3, 3x2 and 2x2 - -static void ScalePlaneDown38(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown38_3)(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - void (*ScaleRowDown38_2)(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width); - const int filter_stride = (filtering == kFilterLinear) ? 0 : src_stride; - assert(dst_width % 3 == 0); - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_C; - ScaleRowDown38_2 = ScaleRowDown38_C; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_C; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_C; - } -#if defined(HAS_SCALEROWDOWN38_NEON) - if (TestCpuFlag(kCpuHasNEON) && (dst_width % 12 == 0)) { - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_NEON; - ScaleRowDown38_2 = ScaleRowDown38_NEON; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_NEON; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_NEON; - } - } -#elif defined(HAS_SCALEROWDOWN38_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && (dst_width % 24 == 0) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_SSSE3; - ScaleRowDown38_2 = ScaleRowDown38_SSSE3; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_SSSE3; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_SSSE3; - } - } -#elif defined(HAS_SCALEROWDOWN38_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && (dst_width % 12 == 0) && - IS_ALIGNED(src_ptr, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_MIPS_DSPR2; - ScaleRowDown38_2 = ScaleRowDown38_MIPS_DSPR2; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_MIPS_DSPR2; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_MIPS_DSPR2; - } - } -#endif - - for (y = 0; y < dst_height - 2; y += 3) { - ScaleRowDown38_3(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 3; - dst_ptr += dst_stride; - ScaleRowDown38_3(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 3; - dst_ptr += dst_stride; - ScaleRowDown38_2(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 2; - dst_ptr += dst_stride; - } - - // Remainder 1 or 2 rows with last row vertically unfiltered - if ((dst_height % 3) == 2) { - ScaleRowDown38_3(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 3; - dst_ptr += dst_stride; - ScaleRowDown38_3(src_ptr, 0, dst_ptr, dst_width); - } else if ((dst_height % 3) == 1) { - ScaleRowDown38_3(src_ptr, 0, dst_ptr, dst_width); - } -} - -static void ScalePlaneDown38_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr, - enum FilterMode filtering) { - int y; - void (*ScaleRowDown38_3)(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width); - void (*ScaleRowDown38_2)(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width); - const int filter_stride = (filtering == kFilterLinear) ? 0 : src_stride; - assert(dst_width % 3 == 0); - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_16_C; - ScaleRowDown38_2 = ScaleRowDown38_16_C; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_16_C; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_16_C; - } -#if defined(HAS_SCALEROWDOWN38_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && (dst_width % 12 == 0)) { - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_16_NEON; - ScaleRowDown38_2 = ScaleRowDown38_16_NEON; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_16_NEON; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_16_NEON; - } - } -#elif defined(HAS_SCALEROWDOWN38_16_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && (dst_width % 24 == 0) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_16_SSSE3; - ScaleRowDown38_2 = ScaleRowDown38_16_SSSE3; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_16_SSSE3; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_16_SSSE3; - } - } -#elif defined(HAS_SCALEROWDOWN38_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && (dst_width % 12 == 0) && - IS_ALIGNED(src_ptr, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_ptr, 4) && IS_ALIGNED(dst_stride, 4)) { - if (!filtering) { - ScaleRowDown38_3 = ScaleRowDown38_16_MIPS_DSPR2; - ScaleRowDown38_2 = ScaleRowDown38_16_MIPS_DSPR2; - } else { - ScaleRowDown38_3 = ScaleRowDown38_3_Box_16_MIPS_DSPR2; - ScaleRowDown38_2 = ScaleRowDown38_2_Box_16_MIPS_DSPR2; - } - } -#endif - - for (y = 0; y < dst_height - 2; y += 3) { - ScaleRowDown38_3(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 3; - dst_ptr += dst_stride; - ScaleRowDown38_3(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 3; - dst_ptr += dst_stride; - ScaleRowDown38_2(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 2; - dst_ptr += dst_stride; - } - - // Remainder 1 or 2 rows with last row vertically unfiltered - if ((dst_height % 3) == 2) { - ScaleRowDown38_3(src_ptr, filter_stride, dst_ptr, dst_width); - src_ptr += src_stride * 3; - dst_ptr += dst_stride; - ScaleRowDown38_3(src_ptr, 0, dst_ptr, dst_width); - } else if ((dst_height % 3) == 1) { - ScaleRowDown38_3(src_ptr, 0, dst_ptr, dst_width); - } -} - -static __inline uint32 SumBox(int iboxwidth, int iboxheight, - ptrdiff_t src_stride, const uint8* src_ptr) { - uint32 sum = 0u; - int y; - assert(iboxwidth > 0); - assert(iboxheight > 0); - for (y = 0; y < iboxheight; ++y) { - int x; - for (x = 0; x < iboxwidth; ++x) { - sum += src_ptr[x]; - } - src_ptr += src_stride; - } - return sum; -} - -static __inline uint32 SumBox_16(int iboxwidth, int iboxheight, - ptrdiff_t src_stride, const uint16* src_ptr) { - uint32 sum = 0u; - int y; - assert(iboxwidth > 0); - assert(iboxheight > 0); - for (y = 0; y < iboxheight; ++y) { - int x; - for (x = 0; x < iboxwidth; ++x) { - sum += src_ptr[x]; - } - src_ptr += src_stride; - } - return sum; -} - -static void ScalePlaneBoxRow_C(int dst_width, int boxheight, - int x, int dx, ptrdiff_t src_stride, - const uint8* src_ptr, uint8* dst_ptr) { - int i; - int boxwidth; - for (i = 0; i < dst_width; ++i) { - int ix = x >> 16; - x += dx; - boxwidth = (x >> 16) - ix; - *dst_ptr++ = SumBox(boxwidth, boxheight, src_stride, src_ptr + ix) / - (boxwidth * boxheight); - } -} - -static void ScalePlaneBoxRow_16_C(int dst_width, int boxheight, - int x, int dx, ptrdiff_t src_stride, - const uint16* src_ptr, uint16* dst_ptr) { - int i; - int boxwidth; - for (i = 0; i < dst_width; ++i) { - int ix = x >> 16; - x += dx; - boxwidth = (x >> 16) - ix; - *dst_ptr++ = SumBox_16(boxwidth, boxheight, src_stride, src_ptr + ix) / - (boxwidth * boxheight); - } -} - -static __inline uint32 SumPixels(int iboxwidth, const uint16* src_ptr) { - uint32 sum = 0u; - int x; - assert(iboxwidth > 0); - for (x = 0; x < iboxwidth; ++x) { - sum += src_ptr[x]; - } - return sum; -} - -static __inline uint32 SumPixels_16(int iboxwidth, const uint32* src_ptr) { - uint32 sum = 0u; - int x; - assert(iboxwidth > 0); - for (x = 0; x < iboxwidth; ++x) { - sum += src_ptr[x]; - } - return sum; -} - -static void ScaleAddCols2_C(int dst_width, int boxheight, int x, int dx, - const uint16* src_ptr, uint8* dst_ptr) { - int i; - int scaletbl[2]; - int minboxwidth = (dx >> 16); - int* scaleptr = scaletbl - minboxwidth; - int boxwidth; - scaletbl[0] = 65536 / (minboxwidth * boxheight); - scaletbl[1] = 65536 / ((minboxwidth + 1) * boxheight); - for (i = 0; i < dst_width; ++i) { - int ix = x >> 16; - x += dx; - boxwidth = (x >> 16) - ix; - *dst_ptr++ = SumPixels(boxwidth, src_ptr + ix) * scaleptr[boxwidth] >> 16; - } -} - -static void ScaleAddCols2_16_C(int dst_width, int boxheight, int x, int dx, - const uint32* src_ptr, uint16* dst_ptr) { - int i; - int scaletbl[2]; - int minboxwidth = (dx >> 16); - int* scaleptr = scaletbl - minboxwidth; - int boxwidth; - scaletbl[0] = 65536 / (minboxwidth * boxheight); - scaletbl[1] = 65536 / ((minboxwidth + 1) * boxheight); - for (i = 0; i < dst_width; ++i) { - int ix = x >> 16; - x += dx; - boxwidth = (x >> 16) - ix; - *dst_ptr++ = SumPixels_16(boxwidth, src_ptr + ix) * - scaleptr[boxwidth] >> 16; - } -} - -static void ScaleAddCols1_C(int dst_width, int boxheight, int x, int dx, - const uint16* src_ptr, uint8* dst_ptr) { - int boxwidth = (dx >> 16); - int scaleval = 65536 / (boxwidth * boxheight); - int i; - for (i = 0; i < dst_width; ++i) { - *dst_ptr++ = SumPixels(boxwidth, src_ptr + x) * scaleval >> 16; - x += boxwidth; - } -} - -static void ScaleAddCols1_16_C(int dst_width, int boxheight, int x, int dx, - const uint32* src_ptr, uint16* dst_ptr) { - int boxwidth = (dx >> 16); - int scaleval = 65536 / (boxwidth * boxheight); - int i; - for (i = 0; i < dst_width; ++i) { - *dst_ptr++ = SumPixels_16(boxwidth, src_ptr + x) * scaleval >> 16; - x += boxwidth; - } -} - -// Scale plane down to any dimensions, with interpolation. -// (boxfilter). -// -// Same method as SimpleScale, which is fixed point, outputting -// one pixel of destination using fixed point (16.16) to step -// through source, sampling a box of pixel with simple -// averaging. -static void ScalePlaneBox(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr) { - int j; - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - const int max_y = (src_height << 16); - ScaleSlope(src_width, src_height, dst_width, dst_height, kFilterBox, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - // TODO(fbarchard): Remove this and make AddRows handle boxheight 1. - if (!IS_ALIGNED(src_width, 16) || dst_height * 2 > src_height) { - uint8* dst = dst_ptr; - int j; - for (j = 0; j < dst_height; ++j) { - int boxheight; - int iy = y >> 16; - const uint8* src = src_ptr + iy * src_stride; - y += dy; - if (y > max_y) { - y = max_y; - } - boxheight = (y >> 16) - iy; - ScalePlaneBoxRow_C(dst_width, boxheight, - x, dx, src_stride, - src, dst); - dst += dst_stride; - } - return; - } - { - // Allocate a row buffer of uint16. - align_buffer_64(row16, src_width * 2); - void (*ScaleAddCols)(int dst_width, int boxheight, int x, int dx, - const uint16* src_ptr, uint8* dst_ptr) = - (dx & 0xffff) ? ScaleAddCols2_C: ScaleAddCols1_C; - void (*ScaleAddRows)(const uint8* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int src_width, int src_height) = ScaleAddRows_C; - -#if defined(HAS_SCALEADDROWS_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && -#ifdef AVOID_OVERREAD - IS_ALIGNED(src_width, 16) && -#endif - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - ScaleAddRows = ScaleAddRows_SSE2; - } -#endif - - for (j = 0; j < dst_height; ++j) { - int boxheight; - int iy = y >> 16; - const uint8* src = src_ptr + iy * src_stride; - y += dy; - if (y > (src_height << 16)) { - y = (src_height << 16); - } - boxheight = (y >> 16) - iy; - ScaleAddRows(src, src_stride, (uint16*)(row16), - src_width, boxheight); - ScaleAddCols(dst_width, boxheight, x, dx, (uint16*)(row16), - dst_ptr); - dst_ptr += dst_stride; - } - free_aligned_buffer_64(row16); - } -} - -static void ScalePlaneBox_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr) { - int j; - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - const int max_y = (src_height << 16); - ScaleSlope(src_width, src_height, dst_width, dst_height, kFilterBox, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - // TODO(fbarchard): Remove this and make AddRows handle boxheight 1. - if (!IS_ALIGNED(src_width, 16) || dst_height * 2 > src_height) { - uint16* dst = dst_ptr; - int j; - for (j = 0; j < dst_height; ++j) { - int boxheight; - int iy = y >> 16; - const uint16* src = src_ptr + iy * src_stride; - y += dy; - if (y > max_y) { - y = max_y; - } - boxheight = (y >> 16) - iy; - ScalePlaneBoxRow_16_C(dst_width, boxheight, - x, dx, src_stride, - src, dst); - dst += dst_stride; - } - return; - } - { - // Allocate a row buffer of uint32. - align_buffer_64(row32, src_width * 4); - void (*ScaleAddCols)(int dst_width, int boxheight, int x, int dx, - const uint32* src_ptr, uint16* dst_ptr) = - (dx & 0xffff) ? ScaleAddCols2_16_C: ScaleAddCols1_16_C; - void (*ScaleAddRows)(const uint16* src_ptr, ptrdiff_t src_stride, - uint32* dst_ptr, int src_width, int src_height) = ScaleAddRows_16_C; - -#if defined(HAS_SCALEADDROWS_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && -#ifdef AVOID_OVERREAD - IS_ALIGNED(src_width, 16) && -#endif - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - ScaleAddRows = ScaleAddRows_16_SSE2; - } -#endif - - for (j = 0; j < dst_height; ++j) { - int boxheight; - int iy = y >> 16; - const uint16* src = src_ptr + iy * src_stride; - y += dy; - if (y > (src_height << 16)) { - y = (src_height << 16); - } - boxheight = (y >> 16) - iy; - ScaleAddRows(src, src_stride, (uint32*)(row32), - src_width, boxheight); - ScaleAddCols(dst_width, boxheight, x, dx, (uint32*)(row32), - dst_ptr); - dst_ptr += dst_stride; - } - free_aligned_buffer_64(row32); - } -} - -// Scale plane down with bilinear interpolation. -void ScalePlaneBilinearDown(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr, - enum FilterMode filtering) { - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - // TODO(fbarchard): Consider not allocating row buffer for kFilterLinear. - // Allocate a row buffer. - align_buffer_64(row, src_width); - - const int max_y = (src_height - 1) << 16; - int j; - void (*ScaleFilterCols)(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) = - (src_width >= 32768) ? ScaleFilterCols64_C : ScaleFilterCols_C; - void (*InterpolateRow)(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_C; - ScaleSlope(src_width, src_height, dst_width, dst_height, filtering, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && src_width >= 16) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(src_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && src_width >= 16) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(src_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && src_width >= 32) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(src_width, 32)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && src_width >= 16) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(src_width, 16)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && src_width >= 4) { - InterpolateRow = InterpolateRow_Any_MIPS_DSPR2; - if (IS_ALIGNED(src_width, 4)) { - InterpolateRow = InterpolateRow_MIPS_DSPR2; - } - } -#endif - - -#if defined(HAS_SCALEFILTERCOLS_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleFilterCols = ScaleFilterCols_SSSE3; - } -#endif - if (y > max_y) { - y = max_y; - } - - for (j = 0; j < dst_height; ++j) { - int yi = y >> 16; - const uint8* src = src_ptr + yi * src_stride; - if (filtering == kFilterLinear) { - ScaleFilterCols(dst_ptr, src, dst_width, x, dx); - } else { - int yf = (y >> 8) & 255; - InterpolateRow(row, src, src_stride, src_width, yf); - ScaleFilterCols(dst_ptr, row, dst_width, x, dx); - } - dst_ptr += dst_stride; - y += dy; - if (y > max_y) { - y = max_y; - } - } - free_aligned_buffer_64(row); -} - -void ScalePlaneBilinearDown_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr, - enum FilterMode filtering) { - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - // TODO(fbarchard): Consider not allocating row buffer for kFilterLinear. - // Allocate a row buffer. - align_buffer_64(row, src_width * 2); - - const int max_y = (src_height - 1) << 16; - int j; - void (*ScaleFilterCols)(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx) = - (src_width >= 32768) ? ScaleFilterCols64_16_C : ScaleFilterCols_16_C; - void (*InterpolateRow)(uint16* dst_ptr, const uint16* src_ptr, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_16_C; - ScaleSlope(src_width, src_height, dst_width, dst_height, filtering, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - -#if defined(HAS_INTERPOLATEROW_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && src_width >= 16) { - InterpolateRow = InterpolateRow_Any_16_SSE2; - if (IS_ALIGNED(src_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_16_SSE2; - if (IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - InterpolateRow = InterpolateRow_16_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && src_width >= 16) { - InterpolateRow = InterpolateRow_Any_16_SSSE3; - if (IS_ALIGNED(src_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_16_SSSE3; - if (IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16)) { - InterpolateRow = InterpolateRow_16_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && src_width >= 32) { - InterpolateRow = InterpolateRow_Any_16_AVX2; - if (IS_ALIGNED(src_width, 32)) { - InterpolateRow = InterpolateRow_16_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && src_width >= 16) { - InterpolateRow = InterpolateRow_Any_16_NEON; - if (IS_ALIGNED(src_width, 16)) { - InterpolateRow = InterpolateRow_16_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && src_width >= 4) { - InterpolateRow = InterpolateRow_Any_16_MIPS_DSPR2; - if (IS_ALIGNED(src_width, 4)) { - InterpolateRow = InterpolateRow_16_MIPS_DSPR2; - } - } -#endif - - -#if defined(HAS_SCALEFILTERCOLS_16_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleFilterCols = ScaleFilterCols_16_SSSE3; - } -#endif - if (y > max_y) { - y = max_y; - } - - for (j = 0; j < dst_height; ++j) { - int yi = y >> 16; - const uint16* src = src_ptr + yi * src_stride; - if (filtering == kFilterLinear) { - ScaleFilterCols(dst_ptr, src, dst_width, x, dx); - } else { - int yf = (y >> 8) & 255; - InterpolateRow((uint16*)row, src, src_stride, src_width, yf); - ScaleFilterCols(dst_ptr, (uint16*)row, dst_width, x, dx); - } - dst_ptr += dst_stride; - y += dy; - if (y > max_y) { - y = max_y; - } - } - free_aligned_buffer_64(row); -} - -// Scale up down with bilinear interpolation. -void ScalePlaneBilinearUp(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr, - enum FilterMode filtering) { - int j; - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - const int max_y = (src_height - 1) << 16; - void (*InterpolateRow)(uint8* dst_ptr, const uint8* src_ptr, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_C; - void (*ScaleFilterCols)(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) = - filtering ? ScaleFilterCols_C : ScaleCols_C; - ScaleSlope(src_width, src_height, dst_width, dst_height, filtering, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && dst_width >= 16) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(dst_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && dst_width >= 16) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(dst_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && dst_width >= 32) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(dst_width, 32)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && dst_width >= 16) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(dst_width, 16)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_MIPS_DSPR2; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_MIPS_DSPR2; - } - } -#endif - - if (filtering && src_width >= 32768) { - ScaleFilterCols = ScaleFilterCols64_C; - } -#if defined(HAS_SCALEFILTERCOLS_SSSE3) - if (filtering && TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleFilterCols = ScaleFilterCols_SSSE3; - } -#endif - if (!filtering && src_width * 2 == dst_width && x < 0x8000) { - ScaleFilterCols = ScaleColsUp2_C; -#if defined(HAS_SCALECOLS_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleFilterCols = ScaleColsUp2_SSE2; - } -#endif - } - - if (y > max_y) { - y = max_y; - } - { - int yi = y >> 16; - const uint8* src = src_ptr + yi * src_stride; - - // Allocate 2 row buffers. - const int kRowSize = (dst_width + 15) & ~15; - align_buffer_64(row, kRowSize * 2); - - uint8* rowptr = row; - int rowstride = kRowSize; - int lasty = yi; - - ScaleFilterCols(rowptr, src, dst_width, x, dx); - if (src_height > 1) { - src += src_stride; - } - ScaleFilterCols(rowptr + rowstride, src, dst_width, x, dx); - src += src_stride; - - for (j = 0; j < dst_height; ++j) { - yi = y >> 16; - if (yi != lasty) { - if (y > max_y) { - y = max_y; - yi = y >> 16; - src = src_ptr + yi * src_stride; - } - if (yi != lasty) { - ScaleFilterCols(rowptr, src, dst_width, x, dx); - rowptr += rowstride; - rowstride = -rowstride; - lasty = yi; - src += src_stride; - } - } - if (filtering == kFilterLinear) { - InterpolateRow(dst_ptr, rowptr, 0, dst_width, 0); - } else { - int yf = (y >> 8) & 255; - InterpolateRow(dst_ptr, rowptr, rowstride, dst_width, yf); - } - dst_ptr += dst_stride; - y += dy; - } - free_aligned_buffer_64(row); - } -} - -void ScalePlaneBilinearUp_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr, - enum FilterMode filtering) { - int j; - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - const int max_y = (src_height - 1) << 16; - void (*InterpolateRow)(uint16* dst_ptr, const uint16* src_ptr, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_16_C; - void (*ScaleFilterCols)(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx) = - filtering ? ScaleFilterCols_16_C : ScaleCols_16_C; - ScaleSlope(src_width, src_height, dst_width, dst_height, filtering, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - -#if defined(HAS_INTERPOLATEROW_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && dst_width >= 16) { - InterpolateRow = InterpolateRow_Any_16_SSE2; - if (IS_ALIGNED(dst_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_16_SSE2; - if (IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_16_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && dst_width >= 16) { - InterpolateRow = InterpolateRow_Any_16_SSSE3; - if (IS_ALIGNED(dst_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_16_SSSE3; - if (IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_16_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && dst_width >= 32) { - InterpolateRow = InterpolateRow_Any_16_AVX2; - if (IS_ALIGNED(dst_width, 32)) { - InterpolateRow = InterpolateRow_16_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && dst_width >= 16) { - InterpolateRow = InterpolateRow_Any_16_NEON; - if (IS_ALIGNED(dst_width, 16)) { - InterpolateRow = InterpolateRow_16_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_16_MIPS_DSPR2; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_16_MIPS_DSPR2; - } - } -#endif - - if (filtering && src_width >= 32768) { - ScaleFilterCols = ScaleFilterCols64_16_C; - } -#if defined(HAS_SCALEFILTERCOLS_16_SSSE3) - if (filtering && TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleFilterCols = ScaleFilterCols_16_SSSE3; - } -#endif - if (!filtering && src_width * 2 == dst_width && x < 0x8000) { - ScaleFilterCols = ScaleColsUp2_16_C; -#if defined(HAS_SCALECOLS_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleFilterCols = ScaleColsUp2_16_SSE2; - } -#endif - } - - if (y > max_y) { - y = max_y; - } - { - int yi = y >> 16; - const uint16* src = src_ptr + yi * src_stride; - - // Allocate 2 row buffers. - const int kRowSize = (dst_width + 15) & ~15; - align_buffer_64(row, kRowSize * 4); - - uint16* rowptr = (uint16*)row; - int rowstride = kRowSize; - int lasty = yi; - - ScaleFilterCols(rowptr, src, dst_width, x, dx); - if (src_height > 1) { - src += src_stride; - } - ScaleFilterCols(rowptr + rowstride, src, dst_width, x, dx); - src += src_stride; - - for (j = 0; j < dst_height; ++j) { - yi = y >> 16; - if (yi != lasty) { - if (y > max_y) { - y = max_y; - yi = y >> 16; - src = src_ptr + yi * src_stride; - } - if (yi != lasty) { - ScaleFilterCols(rowptr, src, dst_width, x, dx); - rowptr += rowstride; - rowstride = -rowstride; - lasty = yi; - src += src_stride; - } - } - if (filtering == kFilterLinear) { - InterpolateRow(dst_ptr, rowptr, 0, dst_width, 0); - } else { - int yf = (y >> 8) & 255; - InterpolateRow(dst_ptr, rowptr, rowstride, dst_width, yf); - } - dst_ptr += dst_stride; - y += dy; - } - free_aligned_buffer_64(row); - } -} - -// Scale Plane to/from any dimensions, without interpolation. -// Fixed point math is used for performance: The upper 16 bits -// of x and dx is the integer part of the source position and -// the lower 16 bits are the fixed decimal part. - -static void ScalePlaneSimple(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_ptr, uint8* dst_ptr) { - int i; - void (*ScaleCols)(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) = ScaleCols_C; - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - ScaleSlope(src_width, src_height, dst_width, dst_height, kFilterNone, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - - if (src_width * 2 == dst_width && x < 0x8000) { - ScaleCols = ScaleColsUp2_C; -#if defined(HAS_SCALECOLS_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleCols = ScaleColsUp2_SSE2; - } -#endif - } - - for (i = 0; i < dst_height; ++i) { - ScaleCols(dst_ptr, src_ptr + (y >> 16) * src_stride, - dst_width, x, dx); - dst_ptr += dst_stride; - y += dy; - } -} - -static void ScalePlaneSimple_16(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_ptr, uint16* dst_ptr) { - int i; - void (*ScaleCols)(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx) = ScaleCols_16_C; - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - ScaleSlope(src_width, src_height, dst_width, dst_height, kFilterNone, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - - if (src_width * 2 == dst_width && x < 0x8000) { - ScaleCols = ScaleColsUp2_16_C; -#if defined(HAS_SCALECOLS_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_ptr, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_ptr, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleCols = ScaleColsUp2_16_SSE2; - } -#endif - } - - for (i = 0; i < dst_height; ++i) { - ScaleCols(dst_ptr, src_ptr + (y >> 16) * src_stride, - dst_width, x, dx); - dst_ptr += dst_stride; - y += dy; - } -} - -// Scale a plane. -// This function dispatches to a specialized scaler based on scale factor. - -LIBYUV_API -void ScalePlane(const uint8* src, int src_stride, - int src_width, int src_height, - uint8* dst, int dst_stride, - int dst_width, int dst_height, - enum FilterMode filtering) { - // Simplify filtering when possible. - filtering = ScaleFilterReduce(src_width, src_height, - dst_width, dst_height, - filtering); - - // Negative height means invert the image. - if (src_height < 0) { - src_height = -src_height; - src = src + (src_height - 1) * src_stride; - src_stride = -src_stride; - } - - // Use specialized scales to improve performance for common resolutions. - // For example, all the 1/2 scalings will use ScalePlaneDown2() - if (dst_width == src_width && dst_height == src_height) { - // Straight copy. - CopyPlane(src, src_stride, dst, dst_stride, dst_width, dst_height); - return; - } - if (dst_width == src_width) { - int dy = FixedDiv(src_height, dst_height); - // Arbitrary scale vertically, but unscaled vertically. - ScalePlaneVertical(src_height, - dst_width, dst_height, - src_stride, dst_stride, src, dst, - 0, 0, dy, 1, filtering); - return; - } - if (dst_width <= Abs(src_width) && dst_height <= src_height) { - // Scale down. - if (4 * dst_width == 3 * src_width && - 4 * dst_height == 3 * src_height) { - // optimized, 3/4 - ScalePlaneDown34(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - if (2 * dst_width == src_width && 2 * dst_height == src_height) { - // optimized, 1/2 - ScalePlaneDown2(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - // 3/8 rounded up for odd sized chroma height. - if (8 * dst_width == 3 * src_width && - dst_height == ((src_height * 3 + 7) / 8)) { - // optimized, 3/8 - ScalePlaneDown38(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - if (4 * dst_width == src_width && 4 * dst_height == src_height && - filtering != kFilterBilinear) { - // optimized, 1/4 - ScalePlaneDown4(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - } - if (filtering == kFilterBox && dst_height * 2 < src_height) { - ScalePlaneBox(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst); - return; - } - if (filtering && dst_height > src_height) { - ScalePlaneBilinearUp(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - if (filtering) { - ScalePlaneBilinearDown(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - ScalePlaneSimple(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst); -} - -LIBYUV_API -void ScalePlane_16(const uint16* src, int src_stride, - int src_width, int src_height, - uint16* dst, int dst_stride, - int dst_width, int dst_height, - enum FilterMode filtering) { - // Simplify filtering when possible. - filtering = ScaleFilterReduce(src_width, src_height, - dst_width, dst_height, - filtering); - - // Negative height means invert the image. - if (src_height < 0) { - src_height = -src_height; - src = src + (src_height - 1) * src_stride; - src_stride = -src_stride; - } - - // Use specialized scales to improve performance for common resolutions. - // For example, all the 1/2 scalings will use ScalePlaneDown2() - if (dst_width == src_width && dst_height == src_height) { - // Straight copy. - CopyPlane_16(src, src_stride, dst, dst_stride, dst_width, dst_height); - return; - } - if (dst_width == src_width) { - int dy = FixedDiv(src_height, dst_height); - // Arbitrary scale vertically, but unscaled vertically. - ScalePlaneVertical_16(src_height, - dst_width, dst_height, - src_stride, dst_stride, src, dst, - 0, 0, dy, 1, filtering); - return; - } - if (dst_width <= Abs(src_width) && dst_height <= src_height) { - // Scale down. - if (4 * dst_width == 3 * src_width && - 4 * dst_height == 3 * src_height) { - // optimized, 3/4 - ScalePlaneDown34_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - if (2 * dst_width == src_width && 2 * dst_height == src_height) { - // optimized, 1/2 - ScalePlaneDown2_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - // 3/8 rounded up for odd sized chroma height. - if (8 * dst_width == 3 * src_width && - dst_height == ((src_height * 3 + 7) / 8)) { - // optimized, 3/8 - ScalePlaneDown38_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - if (4 * dst_width == src_width && 4 * dst_height == src_height && - filtering != kFilterBilinear) { - // optimized, 1/4 - ScalePlaneDown4_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - } - if (filtering == kFilterBox && dst_height * 2 < src_height) { - ScalePlaneBox_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst); - return; - } - if (filtering && dst_height > src_height) { - ScalePlaneBilinearUp_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - if (filtering) { - ScalePlaneBilinearDown_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst, filtering); - return; - } - ScalePlaneSimple_16(src_width, src_height, dst_width, dst_height, - src_stride, dst_stride, src, dst); -} - -// Scale an I420 image. -// This function in turn calls a scaling function for each plane. - -LIBYUV_API -int I420Scale(const uint8* src_y, int src_stride_y, - const uint8* src_u, int src_stride_u, - const uint8* src_v, int src_stride_v, - int src_width, int src_height, - uint8* dst_y, int dst_stride_y, - uint8* dst_u, int dst_stride_u, - uint8* dst_v, int dst_stride_v, - int dst_width, int dst_height, - enum FilterMode filtering) { - int src_halfwidth = SUBSAMPLE(src_width, 1, 1); - int src_halfheight = SUBSAMPLE(src_height, 1, 1); - int dst_halfwidth = SUBSAMPLE(dst_width, 1, 1); - int dst_halfheight = SUBSAMPLE(dst_height, 1, 1); - if (!src_y || !src_u || !src_v || src_width == 0 || src_height == 0 || - !dst_y || !dst_u || !dst_v || dst_width <= 0 || dst_height <= 0) { - return -1; - } - - ScalePlane(src_y, src_stride_y, src_width, src_height, - dst_y, dst_stride_y, dst_width, dst_height, - filtering); - ScalePlane(src_u, src_stride_u, src_halfwidth, src_halfheight, - dst_u, dst_stride_u, dst_halfwidth, dst_halfheight, - filtering); - ScalePlane(src_v, src_stride_v, src_halfwidth, src_halfheight, - dst_v, dst_stride_v, dst_halfwidth, dst_halfheight, - filtering); - return 0; -} - -LIBYUV_API -int I420Scale_16(const uint16* src_y, int src_stride_y, - const uint16* src_u, int src_stride_u, - const uint16* src_v, int src_stride_v, - int src_width, int src_height, - uint16* dst_y, int dst_stride_y, - uint16* dst_u, int dst_stride_u, - uint16* dst_v, int dst_stride_v, - int dst_width, int dst_height, - enum FilterMode filtering) { - int src_halfwidth = SUBSAMPLE(src_width, 1, 1); - int src_halfheight = SUBSAMPLE(src_height, 1, 1); - int dst_halfwidth = SUBSAMPLE(dst_width, 1, 1); - int dst_halfheight = SUBSAMPLE(dst_height, 1, 1); - if (!src_y || !src_u || !src_v || src_width == 0 || src_height == 0 || - !dst_y || !dst_u || !dst_v || dst_width <= 0 || dst_height <= 0) { - return -1; - } - - ScalePlane_16(src_y, src_stride_y, src_width, src_height, - dst_y, dst_stride_y, dst_width, dst_height, - filtering); - ScalePlane_16(src_u, src_stride_u, src_halfwidth, src_halfheight, - dst_u, dst_stride_u, dst_halfwidth, dst_halfheight, - filtering); - ScalePlane_16(src_v, src_stride_v, src_halfwidth, src_halfheight, - dst_v, dst_stride_v, dst_halfwidth, dst_halfheight, - filtering); - return 0; -} - -// Deprecated api -LIBYUV_API -int Scale(const uint8* src_y, const uint8* src_u, const uint8* src_v, - int src_stride_y, int src_stride_u, int src_stride_v, - int src_width, int src_height, - uint8* dst_y, uint8* dst_u, uint8* dst_v, - int dst_stride_y, int dst_stride_u, int dst_stride_v, - int dst_width, int dst_height, - LIBYUV_BOOL interpolate) { - return I420Scale(src_y, src_stride_y, - src_u, src_stride_u, - src_v, src_stride_v, - src_width, src_height, - dst_y, dst_stride_y, - dst_u, dst_stride_u, - dst_v, dst_stride_v, - dst_width, dst_height, - interpolate ? kFilterBox : kFilterNone); -} - -// Deprecated api -LIBYUV_API -int ScaleOffset(const uint8* src, int src_width, int src_height, - uint8* dst, int dst_width, int dst_height, int dst_yoffset, - LIBYUV_BOOL interpolate) { - // Chroma requires offset to multiple of 2. - int dst_yoffset_even = dst_yoffset & ~1; - int src_halfwidth = SUBSAMPLE(src_width, 1, 1); - int src_halfheight = SUBSAMPLE(src_height, 1, 1); - int dst_halfwidth = SUBSAMPLE(dst_width, 1, 1); - int dst_halfheight = SUBSAMPLE(dst_height, 1, 1); - int aheight = dst_height - dst_yoffset_even * 2; // actual output height - const uint8* src_y = src; - const uint8* src_u = src + src_width * src_height; - const uint8* src_v = src + src_width * src_height + - src_halfwidth * src_halfheight; - uint8* dst_y = dst + dst_yoffset_even * dst_width; - uint8* dst_u = dst + dst_width * dst_height + - (dst_yoffset_even >> 1) * dst_halfwidth; - uint8* dst_v = dst + dst_width * dst_height + dst_halfwidth * dst_halfheight + - (dst_yoffset_even >> 1) * dst_halfwidth; - if (!src || src_width <= 0 || src_height <= 0 || - !dst || dst_width <= 0 || dst_height <= 0 || dst_yoffset_even < 0 || - dst_yoffset_even >= dst_height) { - return -1; - } - return I420Scale(src_y, src_width, - src_u, src_halfwidth, - src_v, src_halfwidth, - src_width, src_height, - dst_y, dst_width, - dst_u, dst_halfwidth, - dst_v, dst_halfwidth, - dst_width, aheight, - interpolate ? kFilterBox : kFilterNone); -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/scale_argb.cc b/src/main/jni/libyuv/source/scale_argb.cc deleted file mode 100644 index e339cd7c7..000000000 --- a/src/main/jni/libyuv/source/scale_argb.cc +++ /dev/null @@ -1,809 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/scale.h" - -#include <assert.h> -#include <string.h> - -#include "libyuv/cpu_id.h" -#include "libyuv/planar_functions.h" // For CopyARGB -#include "libyuv/row.h" -#include "libyuv/scale_row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -static __inline int Abs(int v) { - return v >= 0 ? v : -v; -} - -// ScaleARGB ARGB, 1/2 -// This is an optimized version for scaling down a ARGB to 1/2 of -// its original size. -static void ScaleARGBDown2(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int dx, int y, int dy, - enum FilterMode filtering) { - int j; - int row_stride = src_stride * (dy >> 16); - void (*ScaleARGBRowDown2)(const uint8* src_argb, ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) = - filtering == kFilterNone ? ScaleARGBRowDown2_C : - (filtering == kFilterLinear ? ScaleARGBRowDown2Linear_C : - ScaleARGBRowDown2Box_C); - assert(dx == 65536 * 2); // Test scale factor of 2. - assert((dy & 0x1ffff) == 0); // Test vertical scale is multiple of 2. - // Advance to odd row, even column. - if (filtering == kFilterBilinear) { - src_argb += (y >> 16) * src_stride + (x >> 16) * 4; - } else { - src_argb += (y >> 16) * src_stride + ((x >> 16) - 1) * 4; - } - -#if defined(HAS_SCALEARGBROWDOWN2_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 4) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(row_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBRowDown2 = filtering == kFilterNone ? ScaleARGBRowDown2_SSE2 : - (filtering == kFilterLinear ? ScaleARGBRowDown2Linear_SSE2 : - ScaleARGBRowDown2Box_SSE2); - } -#elif defined(HAS_SCALEARGBROWDOWN2_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_argb, 4) && IS_ALIGNED(row_stride, 4)) { - ScaleARGBRowDown2 = filtering ? ScaleARGBRowDown2Box_NEON : - ScaleARGBRowDown2_NEON; - } -#endif - - if (filtering == kFilterLinear) { - src_stride = 0; - } - for (j = 0; j < dst_height; ++j) { - ScaleARGBRowDown2(src_argb, src_stride, dst_argb, dst_width); - src_argb += row_stride; - dst_argb += dst_stride; - } -} - -// ScaleARGB ARGB, 1/4 -// This is an optimized version for scaling down a ARGB to 1/4 of -// its original size. -static void ScaleARGBDown4Box(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int dx, int y, int dy) { - int j; - // Allocate 2 rows of ARGB. - const int kRowSize = (dst_width * 2 * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); - int row_stride = src_stride * (dy >> 16); - void (*ScaleARGBRowDown2)(const uint8* src_argb, ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) = ScaleARGBRowDown2Box_C; - // Advance to odd row, even column. - src_argb += (y >> 16) * src_stride + (x >> 16) * 4; - assert(dx == 65536 * 4); // Test scale factor of 4. - assert((dy & 0x3ffff) == 0); // Test vertical scale is multiple of 4. -#if defined(HAS_SCALEARGBROWDOWN2_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 4) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(row_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBRowDown2 = ScaleARGBRowDown2Box_SSE2; - } -#elif defined(HAS_SCALEARGBROWDOWN2_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_argb, 4) && IS_ALIGNED(row_stride, 4)) { - ScaleARGBRowDown2 = ScaleARGBRowDown2Box_NEON; - } -#endif - for (j = 0; j < dst_height; ++j) { - ScaleARGBRowDown2(src_argb, src_stride, row, dst_width * 2); - ScaleARGBRowDown2(src_argb + src_stride * 2, src_stride, - row + kRowSize, dst_width * 2); - ScaleARGBRowDown2(row, kRowSize, dst_argb, dst_width); - src_argb += row_stride; - dst_argb += dst_stride; - } - free_aligned_buffer_64(row); -} - -// ScaleARGB ARGB Even -// This is an optimized version for scaling down a ARGB to even -// multiple of its original size. -static void ScaleARGBDownEven(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int dx, int y, int dy, - enum FilterMode filtering) { - int j; - int col_step = dx >> 16; - int row_stride = (dy >> 16) * src_stride; - void (*ScaleARGBRowDownEven)(const uint8* src_argb, ptrdiff_t src_stride, - int src_step, uint8* dst_argb, int dst_width) = - filtering ? ScaleARGBRowDownEvenBox_C : ScaleARGBRowDownEven_C; - assert(IS_ALIGNED(src_width, 2)); - assert(IS_ALIGNED(src_height, 2)); - src_argb += (y >> 16) * src_stride + (x >> 16) * 4; -#if defined(HAS_SCALEARGBROWDOWNEVEN_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 4) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBRowDownEven = filtering ? ScaleARGBRowDownEvenBox_SSE2 : - ScaleARGBRowDownEven_SSE2; - } -#elif defined(HAS_SCALEARGBROWDOWNEVEN_NEON) - if (TestCpuFlag(kCpuHasNEON) && IS_ALIGNED(dst_width, 4) && - IS_ALIGNED(src_argb, 4)) { - ScaleARGBRowDownEven = filtering ? ScaleARGBRowDownEvenBox_NEON : - ScaleARGBRowDownEven_NEON; - } -#endif - - if (filtering == kFilterLinear) { - src_stride = 0; - } - for (j = 0; j < dst_height; ++j) { - ScaleARGBRowDownEven(src_argb, src_stride, col_step, dst_argb, dst_width); - src_argb += row_stride; - dst_argb += dst_stride; - } -} - -// Scale ARGB down with bilinear interpolation. -static void ScaleARGBBilinearDown(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int dx, int y, int dy, - enum FilterMode filtering) { - int j; - void (*InterpolateRow)(uint8* dst_argb, const uint8* src_argb, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_C; - void (*ScaleARGBFilterCols)(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) = - (src_width >= 32768) ? ScaleARGBFilterCols64_C : ScaleARGBFilterCols_C; - int64 xlast = x + (int64)(dst_width - 1) * dx; - int64 xl = (dx >= 0) ? x : xlast; - int64 xr = (dx >= 0) ? xlast : x; - int clip_src_width; - xl = (xl >> 16) & ~3; // Left edge aligned. - xr = (xr >> 16) + 1; // Right most pixel used. Bilinear uses 2 pixels. - xr = (xr + 1 + 3) & ~3; // 1 beyond 4 pixel aligned right most pixel. - if (xr > src_width) { - xr = src_width; - } - clip_src_width = (int)(xr - xl) * 4; // Width aligned to 4. - src_argb += xl * 4; - x -= (int)(xl << 16); -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && clip_src_width >= 16) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(clip_src_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && clip_src_width >= 16) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(clip_src_width, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && clip_src_width >= 32) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(clip_src_width, 32)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && clip_src_width >= 16) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(clip_src_width, 16)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROWS_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && clip_src_width >= 4 && - IS_ALIGNED(src_argb, 4) && IS_ALIGNED(src_stride, 4)) { - InterpolateRow = InterpolateRow_Any_MIPS_DSPR2; - if (IS_ALIGNED(clip_src_width, 4)) { - InterpolateRow = InterpolateRow_MIPS_DSPR2; - } - } -#endif -#if defined(HAS_SCALEARGBFILTERCOLS_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleARGBFilterCols = ScaleARGBFilterCols_SSSE3; - } -#endif - // TODO(fbarchard): Consider not allocating row buffer for kFilterLinear. - // Allocate a row of ARGB. - { - align_buffer_64(row, clip_src_width * 4); - - const int max_y = (src_height - 1) << 16; - if (y > max_y) { - y = max_y; - } - for (j = 0; j < dst_height; ++j) { - int yi = y >> 16; - const uint8* src = src_argb + yi * src_stride; - if (filtering == kFilterLinear) { - ScaleARGBFilterCols(dst_argb, src, dst_width, x, dx); - } else { - int yf = (y >> 8) & 255; - InterpolateRow(row, src, src_stride, clip_src_width, yf); - ScaleARGBFilterCols(dst_argb, row, dst_width, x, dx); - } - dst_argb += dst_stride; - y += dy; - if (y > max_y) { - y = max_y; - } - } - free_aligned_buffer_64(row); - } -} - -// Scale ARGB up with bilinear interpolation. -static void ScaleARGBBilinearUp(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int dx, int y, int dy, - enum FilterMode filtering) { - int j; - void (*InterpolateRow)(uint8* dst_argb, const uint8* src_argb, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_C; - void (*ScaleARGBFilterCols)(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) = - filtering ? ScaleARGBFilterCols_C : ScaleARGBCols_C; - const int max_y = (src_height - 1) << 16; -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && dst_width >= 8) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(dst_width, 8)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROWS_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && dst_width >= 1 && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride, 4)) { - InterpolateRow = InterpolateRow_MIPS_DSPR2; - } -#endif - if (src_width >= 32768) { - ScaleARGBFilterCols = filtering ? - ScaleARGBFilterCols64_C : ScaleARGBCols64_C; - } -#if defined(HAS_SCALEARGBFILTERCOLS_SSSE3) - if (filtering && TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleARGBFilterCols = ScaleARGBFilterCols_SSSE3; - } -#endif -#if defined(HAS_SCALEARGBCOLS_SSE2) - if (!filtering && TestCpuFlag(kCpuHasSSE2) && src_width < 32768) { - ScaleARGBFilterCols = ScaleARGBCols_SSE2; - } -#endif - if (!filtering && src_width * 2 == dst_width && x < 0x8000) { - ScaleARGBFilterCols = ScaleARGBColsUp2_C; -#if defined(HAS_SCALEARGBCOLSUP2_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBFilterCols = ScaleARGBColsUp2_SSE2; - } -#endif - } - - if (y > max_y) { - y = max_y; - } - - { - int yi = y >> 16; - const uint8* src = src_argb + yi * src_stride; - - // Allocate 2 rows of ARGB. - const int kRowSize = (dst_width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); - - uint8* rowptr = row; - int rowstride = kRowSize; - int lasty = yi; - - ScaleARGBFilterCols(rowptr, src, dst_width, x, dx); - if (src_height > 1) { - src += src_stride; - } - ScaleARGBFilterCols(rowptr + rowstride, src, dst_width, x, dx); - src += src_stride; - - for (j = 0; j < dst_height; ++j) { - yi = y >> 16; - if (yi != lasty) { - if (y > max_y) { - y = max_y; - yi = y >> 16; - src = src_argb + yi * src_stride; - } - if (yi != lasty) { - ScaleARGBFilterCols(rowptr, src, dst_width, x, dx); - rowptr += rowstride; - rowstride = -rowstride; - lasty = yi; - src += src_stride; - } - } - if (filtering == kFilterLinear) { - InterpolateRow(dst_argb, rowptr, 0, dst_width * 4, 0); - } else { - int yf = (y >> 8) & 255; - InterpolateRow(dst_argb, rowptr, rowstride, dst_width * 4, yf); - } - dst_argb += dst_stride; - y += dy; - } - free_aligned_buffer_64(row); - } -} - -#ifdef YUVSCALEUP -// Scale YUV to ARGB up with bilinear interpolation. -static void ScaleYUVToARGBBilinearUp(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride_y, - int src_stride_u, - int src_stride_v, - int dst_stride_argb, - const uint8* src_y, - const uint8* src_u, - const uint8* src_v, - uint8* dst_argb, - int x, int dx, int y, int dy, - enum FilterMode filtering) { - int j; - void (*I422ToARGBRow)(const uint8* y_buf, - const uint8* u_buf, - const uint8* v_buf, - uint8* rgb_buf, - int width) = I422ToARGBRow_C; -#if defined(HAS_I422TOARGBROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && src_width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_SSSE3; - if (IS_ALIGNED(src_width, 8)) { - I422ToARGBRow = I422ToARGBRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - I422ToARGBRow = I422ToARGBRow_SSSE3; - } - } - } -#endif -#if defined(HAS_I422TOARGBROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && src_width >= 16) { - I422ToARGBRow = I422ToARGBRow_Any_AVX2; - if (IS_ALIGNED(src_width, 16)) { - I422ToARGBRow = I422ToARGBRow_AVX2; - } - } -#endif -#if defined(HAS_I422TOARGBROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && src_width >= 8) { - I422ToARGBRow = I422ToARGBRow_Any_NEON; - if (IS_ALIGNED(src_width, 8)) { - I422ToARGBRow = I422ToARGBRow_NEON; - } - } -#endif -#if defined(HAS_I422TOARGBROW_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && IS_ALIGNED(src_width, 4) && - IS_ALIGNED(src_y, 4) && IS_ALIGNED(src_stride_y, 4) && - IS_ALIGNED(src_u, 2) && IS_ALIGNED(src_stride_u, 2) && - IS_ALIGNED(src_v, 2) && IS_ALIGNED(src_stride_v, 2) && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride_argb, 4)) { - I422ToARGBRow = I422ToARGBRow_MIPS_DSPR2; - } -#endif - - void (*InterpolateRow)(uint8* dst_argb, const uint8* src_argb, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_C; -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride_argb, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && dst_width >= 8) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(dst_width, 8)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && dst_width >= 4) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(dst_width, 4)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROWS_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && dst_width >= 1 && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride_argb, 4)) { - InterpolateRow = InterpolateRow_MIPS_DSPR2; - } -#endif - - void (*ScaleARGBFilterCols)(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) = - filtering ? ScaleARGBFilterCols_C : ScaleARGBCols_C; - if (src_width >= 32768) { - ScaleARGBFilterCols = filtering ? - ScaleARGBFilterCols64_C : ScaleARGBCols64_C; - } -#if defined(HAS_SCALEARGBFILTERCOLS_SSSE3) - if (filtering && TestCpuFlag(kCpuHasSSSE3) && src_width < 32768) { - ScaleARGBFilterCols = ScaleARGBFilterCols_SSSE3; - } -#endif -#if defined(HAS_SCALEARGBCOLS_SSE2) - if (!filtering && TestCpuFlag(kCpuHasSSE2) && src_width < 32768) { - ScaleARGBFilterCols = ScaleARGBCols_SSE2; - } -#endif - if (!filtering && src_width * 2 == dst_width && x < 0x8000) { - ScaleARGBFilterCols = ScaleARGBColsUp2_C; -#if defined(HAS_SCALEARGBCOLSUP2_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBFilterCols = ScaleARGBColsUp2_SSE2; - } -#endif - } - - const int max_y = (src_height - 1) << 16; - if (y > max_y) { - y = max_y; - } - const int kYShift = 1; // Shift Y by 1 to convert Y plane to UV coordinate. - int yi = y >> 16; - int uv_yi = yi >> kYShift; - const uint8* src_row_y = src_y + yi * src_stride_y; - const uint8* src_row_u = src_u + uv_yi * src_stride_u; - const uint8* src_row_v = src_v + uv_yi * src_stride_v; - - // Allocate 2 rows of ARGB. - const int kRowSize = (dst_width * 4 + 15) & ~15; - align_buffer_64(row, kRowSize * 2); - - // Allocate 1 row of ARGB for source conversion. - align_buffer_64(argb_row, src_width * 4); - - uint8* rowptr = row; - int rowstride = kRowSize; - int lasty = yi; - - // TODO(fbarchard): Convert first 2 rows of YUV to ARGB. - ScaleARGBFilterCols(rowptr, src_row_y, dst_width, x, dx); - if (src_height > 1) { - src_row_y += src_stride_y; - if (yi & 1) { - src_row_u += src_stride_u; - src_row_v += src_stride_v; - } - } - ScaleARGBFilterCols(rowptr + rowstride, src_row_y, dst_width, x, dx); - if (src_height > 2) { - src_row_y += src_stride_y; - if (!(yi & 1)) { - src_row_u += src_stride_u; - src_row_v += src_stride_v; - } - } - - for (j = 0; j < dst_height; ++j) { - yi = y >> 16; - if (yi != lasty) { - if (y > max_y) { - y = max_y; - yi = y >> 16; - uv_yi = yi >> kYShift; - src_row_y = src_y + yi * src_stride_y; - src_row_u = src_u + uv_yi * src_stride_u; - src_row_v = src_v + uv_yi * src_stride_v; - } - if (yi != lasty) { - // TODO(fbarchard): Convert the clipped region of row. - I422ToARGBRow(src_row_y, src_row_u, src_row_v, argb_row, src_width); - ScaleARGBFilterCols(rowptr, argb_row, dst_width, x, dx); - rowptr += rowstride; - rowstride = -rowstride; - lasty = yi; - src_row_y += src_stride_y; - if (yi & 1) { - src_row_u += src_stride_u; - src_row_v += src_stride_v; - } - } - } - if (filtering == kFilterLinear) { - InterpolateRow(dst_argb, rowptr, 0, dst_width * 4, 0); - } else { - int yf = (y >> 8) & 255; - InterpolateRow(dst_argb, rowptr, rowstride, dst_width * 4, yf); - } - dst_argb += dst_stride_argb; - y += dy; - } - free_aligned_buffer_64(row); - free_aligned_buffer_64(row_argb); -} -#endif - -// Scale ARGB to/from any dimensions, without interpolation. -// Fixed point math is used for performance: The upper 16 bits -// of x and dx is the integer part of the source position and -// the lower 16 bits are the fixed decimal part. - -static void ScaleARGBSimple(int src_width, int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int dx, int y, int dy) { - int j; - void (*ScaleARGBCols)(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) = - (src_width >= 32768) ? ScaleARGBCols64_C : ScaleARGBCols_C; -#if defined(HAS_SCALEARGBCOLS_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && src_width < 32768) { - ScaleARGBCols = ScaleARGBCols_SSE2; - } -#endif - if (src_width * 2 == dst_width && x < 0x8000) { - ScaleARGBCols = ScaleARGBColsUp2_C; -#if defined(HAS_SCALEARGBCOLSUP2_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && IS_ALIGNED(dst_width, 8) && - IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - ScaleARGBCols = ScaleARGBColsUp2_SSE2; - } -#endif - } - - for (j = 0; j < dst_height; ++j) { - ScaleARGBCols(dst_argb, src_argb + (y >> 16) * src_stride, - dst_width, x, dx); - dst_argb += dst_stride; - y += dy; - } -} - -// ScaleARGB a ARGB. -// This function in turn calls a scaling function -// suitable for handling the desired resolutions. -static void ScaleARGB(const uint8* src, int src_stride, - int src_width, int src_height, - uint8* dst, int dst_stride, - int dst_width, int dst_height, - int clip_x, int clip_y, int clip_width, int clip_height, - enum FilterMode filtering) { - // Initial source x/y coordinate and step values as 16.16 fixed point. - int x = 0; - int y = 0; - int dx = 0; - int dy = 0; - // ARGB does not support box filter yet, but allow the user to pass it. - // Simplify filtering when possible. - filtering = ScaleFilterReduce(src_width, src_height, - dst_width, dst_height, - filtering); - - // Negative src_height means invert the image. - if (src_height < 0) { - src_height = -src_height; - src = src + (src_height - 1) * src_stride; - src_stride = -src_stride; - } - ScaleSlope(src_width, src_height, dst_width, dst_height, filtering, - &x, &y, &dx, &dy); - src_width = Abs(src_width); - if (clip_x) { - int64 clipf = (int64)(clip_x) * dx; - x += (clipf & 0xffff); - src += (clipf >> 16) * 4; - dst += clip_x * 4; - } - if (clip_y) { - int64 clipf = (int64)(clip_y) * dy; - y += (clipf & 0xffff); - src += (clipf >> 16) * src_stride; - dst += clip_y * dst_stride; - } - - // Special case for integer step values. - if (((dx | dy) & 0xffff) == 0) { - if (!dx || !dy) { // 1 pixel wide and/or tall. - filtering = kFilterNone; - } else { - // Optimized even scale down. ie 2, 4, 6, 8, 10x. - if (!(dx & 0x10000) && !(dy & 0x10000)) { - if (dx == 0x20000) { - // Optimized 1/2 downsample. - ScaleARGBDown2(src_width, src_height, - clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, dx, y, dy, filtering); - return; - } - if (dx == 0x40000 && filtering == kFilterBox) { - // Optimized 1/4 box downsample. - ScaleARGBDown4Box(src_width, src_height, - clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, dx, y, dy); - return; - } - ScaleARGBDownEven(src_width, src_height, - clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, dx, y, dy, filtering); - return; - } - // Optimized odd scale down. ie 3, 5, 7, 9x. - if ((dx & 0x10000) && (dy & 0x10000)) { - filtering = kFilterNone; - if (dx == 0x10000 && dy == 0x10000) { - // Straight copy. - ARGBCopy(src + (y >> 16) * src_stride + (x >> 16) * 4, src_stride, - dst, dst_stride, clip_width, clip_height); - return; - } - } - } - } - if (dx == 0x10000 && (x & 0xffff) == 0) { - // Arbitrary scale vertically, but unscaled vertically. - ScalePlaneVertical(src_height, - clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, y, dy, 4, filtering); - return; - } - if (filtering && dy < 65536) { - ScaleARGBBilinearUp(src_width, src_height, - clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, dx, y, dy, filtering); - return; - } - if (filtering) { - ScaleARGBBilinearDown(src_width, src_height, - clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, dx, y, dy, filtering); - return; - } - ScaleARGBSimple(src_width, src_height, clip_width, clip_height, - src_stride, dst_stride, src, dst, - x, dx, y, dy); -} - -LIBYUV_API -int ARGBScaleClip(const uint8* src_argb, int src_stride_argb, - int src_width, int src_height, - uint8* dst_argb, int dst_stride_argb, - int dst_width, int dst_height, - int clip_x, int clip_y, int clip_width, int clip_height, - enum FilterMode filtering) { - if (!src_argb || src_width == 0 || src_height == 0 || - !dst_argb || dst_width <= 0 || dst_height <= 0 || - clip_x < 0 || clip_y < 0 || - (clip_x + clip_width) > dst_width || - (clip_y + clip_height) > dst_height) { - return -1; - } - ScaleARGB(src_argb, src_stride_argb, src_width, src_height, - dst_argb, dst_stride_argb, dst_width, dst_height, - clip_x, clip_y, clip_width, clip_height, filtering); - return 0; -} - -// Scale an ARGB image. -LIBYUV_API -int ARGBScale(const uint8* src_argb, int src_stride_argb, - int src_width, int src_height, - uint8* dst_argb, int dst_stride_argb, - int dst_width, int dst_height, - enum FilterMode filtering) { - if (!src_argb || src_width == 0 || src_height == 0 || - !dst_argb || dst_width <= 0 || dst_height <= 0) { - return -1; - } - ScaleARGB(src_argb, src_stride_argb, src_width, src_height, - dst_argb, dst_stride_argb, dst_width, dst_height, - 0, 0, dst_width, dst_height, filtering); - return 0; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/scale_common.cc b/src/main/jni/libyuv/source/scale_common.cc deleted file mode 100644 index e4b2acc41..000000000 --- a/src/main/jni/libyuv/source/scale_common.cc +++ /dev/null @@ -1,1165 +0,0 @@ -/* - * Copyright 2013 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/scale.h" - -#include <assert.h> -#include <string.h> - -#include "libyuv/cpu_id.h" -#include "libyuv/planar_functions.h" // For CopyARGB -#include "libyuv/row.h" -#include "libyuv/scale_row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -static __inline int Abs(int v) { - return v >= 0 ? v : -v; -} - -// CPU agnostic row functions -void ScaleRowDown2_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = src_ptr[1]; - dst[1] = src_ptr[3]; - dst += 2; - src_ptr += 4; - } - if (dst_width & 1) { - dst[0] = src_ptr[1]; - } -} - -void ScaleRowDown2_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = src_ptr[1]; - dst[1] = src_ptr[3]; - dst += 2; - src_ptr += 4; - } - if (dst_width & 1) { - dst[0] = src_ptr[1]; - } -} - -void ScaleRowDown2Linear_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - const uint8* s = src_ptr; - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = (s[0] + s[1] + 1) >> 1; - dst[1] = (s[2] + s[3] + 1) >> 1; - dst += 2; - s += 4; - } - if (dst_width & 1) { - dst[0] = (s[0] + s[1] + 1) >> 1; - } -} - -void ScaleRowDown2Linear_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - const uint16* s = src_ptr; - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = (s[0] + s[1] + 1) >> 1; - dst[1] = (s[2] + s[3] + 1) >> 1; - dst += 2; - s += 4; - } - if (dst_width & 1) { - dst[0] = (s[0] + s[1] + 1) >> 1; - } -} - -void ScaleRowDown2Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - const uint8* s = src_ptr; - const uint8* t = src_ptr + src_stride; - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = (s[0] + s[1] + t[0] + t[1] + 2) >> 2; - dst[1] = (s[2] + s[3] + t[2] + t[3] + 2) >> 2; - dst += 2; - s += 4; - t += 4; - } - if (dst_width & 1) { - dst[0] = (s[0] + s[1] + t[0] + t[1] + 2) >> 2; - } -} - -void ScaleRowDown2Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - const uint16* s = src_ptr; - const uint16* t = src_ptr + src_stride; - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = (s[0] + s[1] + t[0] + t[1] + 2) >> 2; - dst[1] = (s[2] + s[3] + t[2] + t[3] + 2) >> 2; - dst += 2; - s += 4; - t += 4; - } - if (dst_width & 1) { - dst[0] = (s[0] + s[1] + t[0] + t[1] + 2) >> 2; - } -} - -void ScaleRowDown4_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = src_ptr[2]; - dst[1] = src_ptr[6]; - dst += 2; - src_ptr += 8; - } - if (dst_width & 1) { - dst[0] = src_ptr[2]; - } -} - -void ScaleRowDown4_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = src_ptr[2]; - dst[1] = src_ptr[6]; - dst += 2; - src_ptr += 8; - } - if (dst_width & 1) { - dst[0] = src_ptr[2]; - } -} - -void ScaleRowDown4Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - intptr_t stride = src_stride; - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + src_ptr[3] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2] + src_ptr[stride + 3] + - src_ptr[stride * 2 + 0] + src_ptr[stride * 2 + 1] + - src_ptr[stride * 2 + 2] + src_ptr[stride * 2 + 3] + - src_ptr[stride * 3 + 0] + src_ptr[stride * 3 + 1] + - src_ptr[stride * 3 + 2] + src_ptr[stride * 3 + 3] + - 8) >> 4; - dst[1] = (src_ptr[4] + src_ptr[5] + src_ptr[6] + src_ptr[7] + - src_ptr[stride + 4] + src_ptr[stride + 5] + - src_ptr[stride + 6] + src_ptr[stride + 7] + - src_ptr[stride * 2 + 4] + src_ptr[stride * 2 + 5] + - src_ptr[stride * 2 + 6] + src_ptr[stride * 2 + 7] + - src_ptr[stride * 3 + 4] + src_ptr[stride * 3 + 5] + - src_ptr[stride * 3 + 6] + src_ptr[stride * 3 + 7] + - 8) >> 4; - dst += 2; - src_ptr += 8; - } - if (dst_width & 1) { - dst[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + src_ptr[3] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2] + src_ptr[stride + 3] + - src_ptr[stride * 2 + 0] + src_ptr[stride * 2 + 1] + - src_ptr[stride * 2 + 2] + src_ptr[stride * 2 + 3] + - src_ptr[stride * 3 + 0] + src_ptr[stride * 3 + 1] + - src_ptr[stride * 3 + 2] + src_ptr[stride * 3 + 3] + - 8) >> 4; - } -} - -void ScaleRowDown4Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - intptr_t stride = src_stride; - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + src_ptr[3] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2] + src_ptr[stride + 3] + - src_ptr[stride * 2 + 0] + src_ptr[stride * 2 + 1] + - src_ptr[stride * 2 + 2] + src_ptr[stride * 2 + 3] + - src_ptr[stride * 3 + 0] + src_ptr[stride * 3 + 1] + - src_ptr[stride * 3 + 2] + src_ptr[stride * 3 + 3] + - 8) >> 4; - dst[1] = (src_ptr[4] + src_ptr[5] + src_ptr[6] + src_ptr[7] + - src_ptr[stride + 4] + src_ptr[stride + 5] + - src_ptr[stride + 6] + src_ptr[stride + 7] + - src_ptr[stride * 2 + 4] + src_ptr[stride * 2 + 5] + - src_ptr[stride * 2 + 6] + src_ptr[stride * 2 + 7] + - src_ptr[stride * 3 + 4] + src_ptr[stride * 3 + 5] + - src_ptr[stride * 3 + 6] + src_ptr[stride * 3 + 7] + - 8) >> 4; - dst += 2; - src_ptr += 8; - } - if (dst_width & 1) { - dst[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + src_ptr[3] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2] + src_ptr[stride + 3] + - src_ptr[stride * 2 + 0] + src_ptr[stride * 2 + 1] + - src_ptr[stride * 2 + 2] + src_ptr[stride * 2 + 3] + - src_ptr[stride * 3 + 0] + src_ptr[stride * 3 + 1] + - src_ptr[stride * 3 + 2] + src_ptr[stride * 3 + 3] + - 8) >> 4; - } -} - -void ScaleRowDown34_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - int x; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (x = 0; x < dst_width; x += 3) { - dst[0] = src_ptr[0]; - dst[1] = src_ptr[1]; - dst[2] = src_ptr[3]; - dst += 3; - src_ptr += 4; - } -} - -void ScaleRowDown34_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - int x; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (x = 0; x < dst_width; x += 3) { - dst[0] = src_ptr[0]; - dst[1] = src_ptr[1]; - dst[2] = src_ptr[3]; - dst += 3; - src_ptr += 4; - } -} - -// Filter rows 0 and 1 together, 3 : 1 -void ScaleRowDown34_0_Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width) { - const uint8* s = src_ptr; - const uint8* t = src_ptr + src_stride; - int x; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (x = 0; x < dst_width; x += 3) { - uint8 a0 = (s[0] * 3 + s[1] * 1 + 2) >> 2; - uint8 a1 = (s[1] * 1 + s[2] * 1 + 1) >> 1; - uint8 a2 = (s[2] * 1 + s[3] * 3 + 2) >> 2; - uint8 b0 = (t[0] * 3 + t[1] * 1 + 2) >> 2; - uint8 b1 = (t[1] * 1 + t[2] * 1 + 1) >> 1; - uint8 b2 = (t[2] * 1 + t[3] * 3 + 2) >> 2; - d[0] = (a0 * 3 + b0 + 2) >> 2; - d[1] = (a1 * 3 + b1 + 2) >> 2; - d[2] = (a2 * 3 + b2 + 2) >> 2; - d += 3; - s += 4; - t += 4; - } -} - -void ScaleRowDown34_0_Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* d, int dst_width) { - const uint16* s = src_ptr; - const uint16* t = src_ptr + src_stride; - int x; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (x = 0; x < dst_width; x += 3) { - uint16 a0 = (s[0] * 3 + s[1] * 1 + 2) >> 2; - uint16 a1 = (s[1] * 1 + s[2] * 1 + 1) >> 1; - uint16 a2 = (s[2] * 1 + s[3] * 3 + 2) >> 2; - uint16 b0 = (t[0] * 3 + t[1] * 1 + 2) >> 2; - uint16 b1 = (t[1] * 1 + t[2] * 1 + 1) >> 1; - uint16 b2 = (t[2] * 1 + t[3] * 3 + 2) >> 2; - d[0] = (a0 * 3 + b0 + 2) >> 2; - d[1] = (a1 * 3 + b1 + 2) >> 2; - d[2] = (a2 * 3 + b2 + 2) >> 2; - d += 3; - s += 4; - t += 4; - } -} - -// Filter rows 1 and 2 together, 1 : 1 -void ScaleRowDown34_1_Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width) { - const uint8* s = src_ptr; - const uint8* t = src_ptr + src_stride; - int x; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (x = 0; x < dst_width; x += 3) { - uint8 a0 = (s[0] * 3 + s[1] * 1 + 2) >> 2; - uint8 a1 = (s[1] * 1 + s[2] * 1 + 1) >> 1; - uint8 a2 = (s[2] * 1 + s[3] * 3 + 2) >> 2; - uint8 b0 = (t[0] * 3 + t[1] * 1 + 2) >> 2; - uint8 b1 = (t[1] * 1 + t[2] * 1 + 1) >> 1; - uint8 b2 = (t[2] * 1 + t[3] * 3 + 2) >> 2; - d[0] = (a0 + b0 + 1) >> 1; - d[1] = (a1 + b1 + 1) >> 1; - d[2] = (a2 + b2 + 1) >> 1; - d += 3; - s += 4; - t += 4; - } -} - -void ScaleRowDown34_1_Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* d, int dst_width) { - const uint16* s = src_ptr; - const uint16* t = src_ptr + src_stride; - int x; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (x = 0; x < dst_width; x += 3) { - uint16 a0 = (s[0] * 3 + s[1] * 1 + 2) >> 2; - uint16 a1 = (s[1] * 1 + s[2] * 1 + 1) >> 1; - uint16 a2 = (s[2] * 1 + s[3] * 3 + 2) >> 2; - uint16 b0 = (t[0] * 3 + t[1] * 1 + 2) >> 2; - uint16 b1 = (t[1] * 1 + t[2] * 1 + 1) >> 1; - uint16 b2 = (t[2] * 1 + t[3] * 3 + 2) >> 2; - d[0] = (a0 + b0 + 1) >> 1; - d[1] = (a1 + b1 + 1) >> 1; - d[2] = (a2 + b2 + 1) >> 1; - d += 3; - s += 4; - t += 4; - } -} - -// Scales a single row of pixels using point sampling. -void ScaleCols_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst_ptr[0] = src_ptr[x >> 16]; - x += dx; - dst_ptr[1] = src_ptr[x >> 16]; - x += dx; - dst_ptr += 2; - } - if (dst_width & 1) { - dst_ptr[0] = src_ptr[x >> 16]; - } -} - -void ScaleCols_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx) { - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst_ptr[0] = src_ptr[x >> 16]; - x += dx; - dst_ptr[1] = src_ptr[x >> 16]; - x += dx; - dst_ptr += 2; - } - if (dst_width & 1) { - dst_ptr[0] = src_ptr[x >> 16]; - } -} - -// Scales a single row of pixels up by 2x using point sampling. -void ScaleColsUp2_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst_ptr[1] = dst_ptr[0] = src_ptr[0]; - src_ptr += 1; - dst_ptr += 2; - } - if (dst_width & 1) { - dst_ptr[0] = src_ptr[0]; - } -} - -void ScaleColsUp2_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx) { - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst_ptr[1] = dst_ptr[0] = src_ptr[0]; - src_ptr += 1; - dst_ptr += 2; - } - if (dst_width & 1) { - dst_ptr[0] = src_ptr[0]; - } -} - -// (1-f)a + fb can be replaced with a + f(b-a) -#define BLENDER(a, b, f) (uint8)((int)(a) + \ - ((int)(f) * ((int)(b) - (int)(a)) >> 16)) - -void ScaleFilterCols_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - int j; - for (j = 0; j < dst_width - 1; j += 2) { - int xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - x += dx; - xi = x >> 16; - a = src_ptr[xi]; - b = src_ptr[xi + 1]; - dst_ptr[1] = BLENDER(a, b, x & 0xffff); - x += dx; - dst_ptr += 2; - } - if (dst_width & 1) { - int xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - } -} - -void ScaleFilterCols64_C(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x32, int dx) { - int64 x = (int64)(x32); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - int64 xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - x += dx; - xi = x >> 16; - a = src_ptr[xi]; - b = src_ptr[xi + 1]; - dst_ptr[1] = BLENDER(a, b, x & 0xffff); - x += dx; - dst_ptr += 2; - } - if (dst_width & 1) { - int64 xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - } -} -#undef BLENDER - -#define BLENDER(a, b, f) (uint16)((int)(a) + \ - ((int)(f) * ((int)(b) - (int)(a)) >> 16)) - -void ScaleFilterCols_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x, int dx) { - int j; - for (j = 0; j < dst_width - 1; j += 2) { - int xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - x += dx; - xi = x >> 16; - a = src_ptr[xi]; - b = src_ptr[xi + 1]; - dst_ptr[1] = BLENDER(a, b, x & 0xffff); - x += dx; - dst_ptr += 2; - } - if (dst_width & 1) { - int xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - } -} - -void ScaleFilterCols64_16_C(uint16* dst_ptr, const uint16* src_ptr, - int dst_width, int x32, int dx) { - int64 x = (int64)(x32); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - int64 xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - x += dx; - xi = x >> 16; - a = src_ptr[xi]; - b = src_ptr[xi + 1]; - dst_ptr[1] = BLENDER(a, b, x & 0xffff); - x += dx; - dst_ptr += 2; - } - if (dst_width & 1) { - int64 xi = x >> 16; - int a = src_ptr[xi]; - int b = src_ptr[xi + 1]; - dst_ptr[0] = BLENDER(a, b, x & 0xffff); - } -} -#undef BLENDER - -void ScaleRowDown38_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - int x; - assert(dst_width % 3 == 0); - for (x = 0; x < dst_width; x += 3) { - dst[0] = src_ptr[0]; - dst[1] = src_ptr[3]; - dst[2] = src_ptr[6]; - dst += 3; - src_ptr += 8; - } -} - -void ScaleRowDown38_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst, int dst_width) { - int x; - assert(dst_width % 3 == 0); - for (x = 0; x < dst_width; x += 3) { - dst[0] = src_ptr[0]; - dst[1] = src_ptr[3]; - dst[2] = src_ptr[6]; - dst += 3; - src_ptr += 8; - } -} - -// 8x3 -> 3x1 -void ScaleRowDown38_3_Box_C(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - intptr_t stride = src_stride; - int i; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (i = 0; i < dst_width; i += 3) { - dst_ptr[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2] + src_ptr[stride * 2 + 0] + - src_ptr[stride * 2 + 1] + src_ptr[stride * 2 + 2]) * - (65536 / 9) >> 16; - dst_ptr[1] = (src_ptr[3] + src_ptr[4] + src_ptr[5] + - src_ptr[stride + 3] + src_ptr[stride + 4] + - src_ptr[stride + 5] + src_ptr[stride * 2 + 3] + - src_ptr[stride * 2 + 4] + src_ptr[stride * 2 + 5]) * - (65536 / 9) >> 16; - dst_ptr[2] = (src_ptr[6] + src_ptr[7] + - src_ptr[stride + 6] + src_ptr[stride + 7] + - src_ptr[stride * 2 + 6] + src_ptr[stride * 2 + 7]) * - (65536 / 6) >> 16; - src_ptr += 8; - dst_ptr += 3; - } -} - -void ScaleRowDown38_3_Box_16_C(const uint16* src_ptr, - ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width) { - intptr_t stride = src_stride; - int i; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (i = 0; i < dst_width; i += 3) { - dst_ptr[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2] + src_ptr[stride * 2 + 0] + - src_ptr[stride * 2 + 1] + src_ptr[stride * 2 + 2]) * - (65536 / 9) >> 16; - dst_ptr[1] = (src_ptr[3] + src_ptr[4] + src_ptr[5] + - src_ptr[stride + 3] + src_ptr[stride + 4] + - src_ptr[stride + 5] + src_ptr[stride * 2 + 3] + - src_ptr[stride * 2 + 4] + src_ptr[stride * 2 + 5]) * - (65536 / 9) >> 16; - dst_ptr[2] = (src_ptr[6] + src_ptr[7] + - src_ptr[stride + 6] + src_ptr[stride + 7] + - src_ptr[stride * 2 + 6] + src_ptr[stride * 2 + 7]) * - (65536 / 6) >> 16; - src_ptr += 8; - dst_ptr += 3; - } -} - -// 8x2 -> 3x1 -void ScaleRowDown38_2_Box_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - intptr_t stride = src_stride; - int i; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (i = 0; i < dst_width; i += 3) { - dst_ptr[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2]) * (65536 / 6) >> 16; - dst_ptr[1] = (src_ptr[3] + src_ptr[4] + src_ptr[5] + - src_ptr[stride + 3] + src_ptr[stride + 4] + - src_ptr[stride + 5]) * (65536 / 6) >> 16; - dst_ptr[2] = (src_ptr[6] + src_ptr[7] + - src_ptr[stride + 6] + src_ptr[stride + 7]) * - (65536 / 4) >> 16; - src_ptr += 8; - dst_ptr += 3; - } -} - -void ScaleRowDown38_2_Box_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int dst_width) { - intptr_t stride = src_stride; - int i; - assert((dst_width % 3 == 0) && (dst_width > 0)); - for (i = 0; i < dst_width; i += 3) { - dst_ptr[0] = (src_ptr[0] + src_ptr[1] + src_ptr[2] + - src_ptr[stride + 0] + src_ptr[stride + 1] + - src_ptr[stride + 2]) * (65536 / 6) >> 16; - dst_ptr[1] = (src_ptr[3] + src_ptr[4] + src_ptr[5] + - src_ptr[stride + 3] + src_ptr[stride + 4] + - src_ptr[stride + 5]) * (65536 / 6) >> 16; - dst_ptr[2] = (src_ptr[6] + src_ptr[7] + - src_ptr[stride + 6] + src_ptr[stride + 7]) * - (65536 / 4) >> 16; - src_ptr += 8; - dst_ptr += 3; - } -} - -void ScaleAddRows_C(const uint8* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int src_width, int src_height) { - int x; - assert(src_width > 0); - assert(src_height > 0); - for (x = 0; x < src_width; ++x) { - const uint8* s = src_ptr + x; - unsigned int sum = 0u; - int y; - for (y = 0; y < src_height; ++y) { - sum += s[0]; - s += src_stride; - } - // TODO(fbarchard): Consider limitting height to 256 to avoid overflow. - dst_ptr[x] = sum < 65535u ? sum : 65535u; - } -} - -void ScaleAddRows_16_C(const uint16* src_ptr, ptrdiff_t src_stride, - uint32* dst_ptr, int src_width, int src_height) { - int x; - assert(src_width > 0); - assert(src_height > 0); - for (x = 0; x < src_width; ++x) { - const uint16* s = src_ptr + x; - unsigned int sum = 0u; - int y; - for (y = 0; y < src_height; ++y) { - sum += s[0]; - s += src_stride; - } - // No risk of overflow here now - dst_ptr[x] = sum; - } -} - -void ScaleARGBRowDown2_C(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = src[1]; - dst[1] = src[3]; - src += 4; - dst += 2; - } - if (dst_width & 1) { - dst[0] = src[1]; - } -} - -void ScaleARGBRowDown2Linear_C(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - int x; - for (x = 0; x < dst_width; ++x) { - dst_argb[0] = (src_argb[0] + src_argb[4] + 1) >> 1; - dst_argb[1] = (src_argb[1] + src_argb[5] + 1) >> 1; - dst_argb[2] = (src_argb[2] + src_argb[6] + 1) >> 1; - dst_argb[3] = (src_argb[3] + src_argb[7] + 1) >> 1; - src_argb += 8; - dst_argb += 4; - } -} - -void ScaleARGBRowDown2Box_C(const uint8* src_argb, ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - int x; - for (x = 0; x < dst_width; ++x) { - dst_argb[0] = (src_argb[0] + src_argb[4] + - src_argb[src_stride] + src_argb[src_stride + 4] + 2) >> 2; - dst_argb[1] = (src_argb[1] + src_argb[5] + - src_argb[src_stride + 1] + src_argb[src_stride + 5] + 2) >> 2; - dst_argb[2] = (src_argb[2] + src_argb[6] + - src_argb[src_stride + 2] + src_argb[src_stride + 6] + 2) >> 2; - dst_argb[3] = (src_argb[3] + src_argb[7] + - src_argb[src_stride + 3] + src_argb[src_stride + 7] + 2) >> 2; - src_argb += 8; - dst_argb += 4; - } -} - -void ScaleARGBRowDownEven_C(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - - int x; - for (x = 0; x < dst_width - 1; x += 2) { - dst[0] = src[0]; - dst[1] = src[src_stepx]; - src += src_stepx * 2; - dst += 2; - } - if (dst_width & 1) { - dst[0] = src[0]; - } -} - -void ScaleARGBRowDownEvenBox_C(const uint8* src_argb, - ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - int x; - for (x = 0; x < dst_width; ++x) { - dst_argb[0] = (src_argb[0] + src_argb[4] + - src_argb[src_stride] + src_argb[src_stride + 4] + 2) >> 2; - dst_argb[1] = (src_argb[1] + src_argb[5] + - src_argb[src_stride + 1] + src_argb[src_stride + 5] + 2) >> 2; - dst_argb[2] = (src_argb[2] + src_argb[6] + - src_argb[src_stride + 2] + src_argb[src_stride + 6] + 2) >> 2; - dst_argb[3] = (src_argb[3] + src_argb[7] + - src_argb[src_stride + 3] + src_argb[src_stride + 7] + 2) >> 2; - src_argb += src_stepx * 4; - dst_argb += 4; - } -} - -// Scales a single row of pixels using point sampling. -void ScaleARGBCols_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst[0] = src[x >> 16]; - x += dx; - dst[1] = src[x >> 16]; - x += dx; - dst += 2; - } - if (dst_width & 1) { - dst[0] = src[x >> 16]; - } -} - -void ScaleARGBCols64_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x32, int dx) { - int64 x = (int64)(x32); - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst[0] = src[x >> 16]; - x += dx; - dst[1] = src[x >> 16]; - x += dx; - dst += 2; - } - if (dst_width & 1) { - dst[0] = src[x >> 16]; - } -} - -// Scales a single row of pixels up by 2x using point sampling. -void ScaleARGBColsUp2_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - dst[1] = dst[0] = src[0]; - src += 1; - dst += 2; - } - if (dst_width & 1) { - dst[0] = src[0]; - } -} - -// Mimics SSSE3 blender -#define BLENDER1(a, b, f) ((a) * (0x7f ^ f) + (b) * f) >> 7 -#define BLENDERC(a, b, f, s) (uint32)( \ - BLENDER1(((a) >> s) & 255, ((b) >> s) & 255, f) << s) -#define BLENDER(a, b, f) \ - BLENDERC(a, b, f, 24) | BLENDERC(a, b, f, 16) | \ - BLENDERC(a, b, f, 8) | BLENDERC(a, b, f, 0) - -void ScaleARGBFilterCols_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - int xi = x >> 16; - int xf = (x >> 9) & 0x7f; - uint32 a = src[xi]; - uint32 b = src[xi + 1]; - dst[0] = BLENDER(a, b, xf); - x += dx; - xi = x >> 16; - xf = (x >> 9) & 0x7f; - a = src[xi]; - b = src[xi + 1]; - dst[1] = BLENDER(a, b, xf); - x += dx; - dst += 2; - } - if (dst_width & 1) { - int xi = x >> 16; - int xf = (x >> 9) & 0x7f; - uint32 a = src[xi]; - uint32 b = src[xi + 1]; - dst[0] = BLENDER(a, b, xf); - } -} - -void ScaleARGBFilterCols64_C(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x32, int dx) { - int64 x = (int64)(x32); - const uint32* src = (const uint32*)(src_argb); - uint32* dst = (uint32*)(dst_argb); - int j; - for (j = 0; j < dst_width - 1; j += 2) { - int64 xi = x >> 16; - int xf = (x >> 9) & 0x7f; - uint32 a = src[xi]; - uint32 b = src[xi + 1]; - dst[0] = BLENDER(a, b, xf); - x += dx; - xi = x >> 16; - xf = (x >> 9) & 0x7f; - a = src[xi]; - b = src[xi + 1]; - dst[1] = BLENDER(a, b, xf); - x += dx; - dst += 2; - } - if (dst_width & 1) { - int64 xi = x >> 16; - int xf = (x >> 9) & 0x7f; - uint32 a = src[xi]; - uint32 b = src[xi + 1]; - dst[0] = BLENDER(a, b, xf); - } -} -#undef BLENDER1 -#undef BLENDERC -#undef BLENDER - -// Scale plane vertically with bilinear interpolation. -void ScalePlaneVertical(int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint8* src_argb, uint8* dst_argb, - int x, int y, int dy, - int bpp, enum FilterMode filtering) { - // TODO(fbarchard): Allow higher bpp. - int dst_width_bytes = dst_width * bpp; - void (*InterpolateRow)(uint8* dst_argb, const uint8* src_argb, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_C; - const int max_y = (src_height > 1) ? ((src_height - 1) << 16) - 1 : 0; - int j; - assert(bpp >= 1 && bpp <= 4); - assert(src_height != 0); - assert(dst_width > 0); - assert(dst_height > 0); - src_argb += (x >> 16) * bpp; -#if defined(HAS_INTERPOLATEROW_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && dst_width_bytes >= 16) { - InterpolateRow = InterpolateRow_Any_SSE2; - if (IS_ALIGNED(dst_width_bytes, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSE2; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && dst_width_bytes >= 16) { - InterpolateRow = InterpolateRow_Any_SSSE3; - if (IS_ALIGNED(dst_width_bytes, 16)) { - InterpolateRow = InterpolateRow_Unaligned_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && dst_width_bytes >= 32) { - InterpolateRow = InterpolateRow_Any_AVX2; - if (IS_ALIGNED(dst_width_bytes, 32)) { - InterpolateRow = InterpolateRow_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_NEON) - if (TestCpuFlag(kCpuHasNEON) && dst_width_bytes >= 16) { - InterpolateRow = InterpolateRow_Any_NEON; - if (IS_ALIGNED(dst_width_bytes, 16)) { - InterpolateRow = InterpolateRow_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROWS_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && dst_width_bytes >= 4 && - IS_ALIGNED(src_argb, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride, 4)) { - InterpolateRow = InterpolateRow_Any_MIPS_DSPR2; - if (IS_ALIGNED(dst_width_bytes, 4)) { - InterpolateRow = InterpolateRow_MIPS_DSPR2; - } - } -#endif - for (j = 0; j < dst_height; ++j) { - int yi; - int yf; - if (y > max_y) { - y = max_y; - } - yi = y >> 16; - yf = filtering ? ((y >> 8) & 255) : 0; - InterpolateRow(dst_argb, src_argb + yi * src_stride, - src_stride, dst_width_bytes, yf); - dst_argb += dst_stride; - y += dy; - } -} -void ScalePlaneVertical_16(int src_height, - int dst_width, int dst_height, - int src_stride, int dst_stride, - const uint16* src_argb, uint16* dst_argb, - int x, int y, int dy, - int wpp, enum FilterMode filtering) { - // TODO(fbarchard): Allow higher wpp. - int dst_width_words = dst_width * wpp; - void (*InterpolateRow)(uint16* dst_argb, const uint16* src_argb, - ptrdiff_t src_stride, int dst_width, int source_y_fraction) = - InterpolateRow_16_C; - const int max_y = (src_height > 1) ? ((src_height - 1) << 16) - 1 : 0; - int j; - assert(wpp >= 1 && wpp <= 2); - assert(src_height != 0); - assert(dst_width > 0); - assert(dst_height > 0); - src_argb += (x >> 16) * wpp; -#if defined(HAS_INTERPOLATEROW_16_SSE2) - if (TestCpuFlag(kCpuHasSSE2) && dst_width_bytes >= 16) { - InterpolateRow = InterpolateRow_Any_16_SSE2; - if (IS_ALIGNED(dst_width_bytes, 16)) { - InterpolateRow = InterpolateRow_Unaligned_16_SSE2; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_16_SSE2; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_SSSE3) - if (TestCpuFlag(kCpuHasSSSE3) && dst_width_bytes >= 16) { - InterpolateRow = InterpolateRow_Any_16_SSSE3; - if (IS_ALIGNED(dst_width_bytes, 16)) { - InterpolateRow = InterpolateRow_Unaligned_16_SSSE3; - if (IS_ALIGNED(src_argb, 16) && IS_ALIGNED(src_stride, 16) && - IS_ALIGNED(dst_argb, 16) && IS_ALIGNED(dst_stride, 16)) { - InterpolateRow = InterpolateRow_16_SSSE3; - } - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_AVX2) - if (TestCpuFlag(kCpuHasAVX2) && dst_width_bytes >= 32) { - InterpolateRow = InterpolateRow_Any_16_AVX2; - if (IS_ALIGNED(dst_width_bytes, 32)) { - InterpolateRow = InterpolateRow_16_AVX2; - } - } -#endif -#if defined(HAS_INTERPOLATEROW_16_NEON) - if (TestCpuFlag(kCpuHasNEON) && dst_width_bytes >= 16) { - InterpolateRow = InterpolateRow_Any_16_NEON; - if (IS_ALIGNED(dst_width_bytes, 16)) { - InterpolateRow = InterpolateRow_16_NEON; - } - } -#endif -#if defined(HAS_INTERPOLATEROWS_16_MIPS_DSPR2) - if (TestCpuFlag(kCpuHasMIPS_DSPR2) && dst_width_bytes >= 4 && - IS_ALIGNED(src_argb, 4) && IS_ALIGNED(src_stride, 4) && - IS_ALIGNED(dst_argb, 4) && IS_ALIGNED(dst_stride, 4)) { - InterpolateRow = InterpolateRow_Any_16_MIPS_DSPR2; - if (IS_ALIGNED(dst_width_bytes, 4)) { - InterpolateRow = InterpolateRow_16_MIPS_DSPR2; - } - } -#endif - for (j = 0; j < dst_height; ++j) { - int yi; - int yf; - if (y > max_y) { - y = max_y; - } - yi = y >> 16; - yf = filtering ? ((y >> 8) & 255) : 0; - InterpolateRow(dst_argb, src_argb + yi * src_stride, - src_stride, dst_width_words, yf); - dst_argb += dst_stride; - y += dy; - } -} - -// Simplify the filtering based on scale factors. -enum FilterMode ScaleFilterReduce(int src_width, int src_height, - int dst_width, int dst_height, - enum FilterMode filtering) { - if (src_width < 0) { - src_width = -src_width; - } - if (src_height < 0) { - src_height = -src_height; - } - if (filtering == kFilterBox) { - // If scaling both axis to 0.5 or larger, switch from Box to Bilinear. - if (dst_width * 2 >= src_width && dst_height * 2 >= src_height) { - filtering = kFilterBilinear; - } - // If scaling to larger, switch from Box to Bilinear. - if (dst_width >= src_width || dst_height >= src_height) { - filtering = kFilterBilinear; - } - } - if (filtering == kFilterBilinear) { - if (src_height == 1) { - filtering = kFilterLinear; - } - // TODO(fbarchard): Detect any odd scale factor and reduce to Linear. - if (dst_height == src_height || dst_height * 3 == src_height) { - filtering = kFilterLinear; - } - // TODO(fbarchard): Remove 1 pixel wide filter restriction, which is to - // avoid reading 2 pixels horizontally that causes memory exception. - if (src_width == 1) { - filtering = kFilterNone; - } - } - if (filtering == kFilterLinear) { - if (src_width == 1) { - filtering = kFilterNone; - } - // TODO(fbarchard): Detect any odd scale factor and reduce to None. - if (dst_width == src_width || dst_width * 3 == src_width) { - filtering = kFilterNone; - } - } - return filtering; -} - -// Divide num by div and return as 16.16 fixed point result. -int FixedDiv_C(int num, int div) { - return (int)(((int64)(num) << 16) / div); -} - -// Divide num by div and return as 16.16 fixed point result. -int FixedDiv1_C(int num, int div) { - return (int)((((int64)(num) << 16) - 0x00010001) / - (div - 1)); -} - -#define CENTERSTART(dx, s) (dx < 0) ? -((-dx >> 1) + s) : ((dx >> 1) + s) - -// Compute slope values for stepping. -void ScaleSlope(int src_width, int src_height, - int dst_width, int dst_height, - enum FilterMode filtering, - int* x, int* y, int* dx, int* dy) { - assert(x != NULL); - assert(y != NULL); - assert(dx != NULL); - assert(dy != NULL); - assert(src_width != 0); - assert(src_height != 0); - assert(dst_width > 0); - assert(dst_height > 0); - // Check for 1 pixel and avoid FixedDiv overflow. - if (dst_width == 1 && src_width >= 32768) { - dst_width = src_width; - } - if (dst_height == 1 && src_height >= 32768) { - dst_height = src_height; - } - if (filtering == kFilterBox) { - // Scale step for point sampling duplicates all pixels equally. - *dx = FixedDiv(Abs(src_width), dst_width); - *dy = FixedDiv(src_height, dst_height); - *x = 0; - *y = 0; - } else if (filtering == kFilterBilinear) { - // Scale step for bilinear sampling renders last pixel once for upsample. - if (dst_width <= Abs(src_width)) { - *dx = FixedDiv(Abs(src_width), dst_width); - *x = CENTERSTART(*dx, -32768); // Subtract 0.5 (32768) to center filter. - } else if (dst_width > 1) { - *dx = FixedDiv1(Abs(src_width), dst_width); - *x = 0; - } - if (dst_height <= src_height) { - *dy = FixedDiv(src_height, dst_height); - *y = CENTERSTART(*dy, -32768); // Subtract 0.5 (32768) to center filter. - } else if (dst_height > 1) { - *dy = FixedDiv1(src_height, dst_height); - *y = 0; - } - } else if (filtering == kFilterLinear) { - // Scale step for bilinear sampling renders last pixel once for upsample. - if (dst_width <= Abs(src_width)) { - *dx = FixedDiv(Abs(src_width), dst_width); - *x = CENTERSTART(*dx, -32768); // Subtract 0.5 (32768) to center filter. - } else if (dst_width > 1) { - *dx = FixedDiv1(Abs(src_width), dst_width); - *x = 0; - } - *dy = FixedDiv(src_height, dst_height); - *y = *dy >> 1; - } else { - // Scale step for point sampling duplicates all pixels equally. - *dx = FixedDiv(Abs(src_width), dst_width); - *dy = FixedDiv(src_height, dst_height); - *x = CENTERSTART(*dx, 0); - *y = CENTERSTART(*dy, 0); - } - // Negative src_width means horizontally mirror. - if (src_width < 0) { - *x += (dst_width - 1) * *dx; - *dx = -*dx; - // src_width = -src_width; // Caller must do this. - } -} -#undef CENTERSTART - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/scale_mips.cc b/src/main/jni/libyuv/source/scale_mips.cc deleted file mode 100644 index 3eb4f27c4..000000000 --- a/src/main/jni/libyuv/source/scale_mips.cc +++ /dev/null @@ -1,654 +0,0 @@ -/* - * Copyright 2012 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/basic_types.h" -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC MIPS DSPR2 -#if !defined(LIBYUV_DISABLE_MIPS) && \ - defined(__mips_dsp) && (__mips_dsp_rev >= 2) && \ - (_MIPS_SIM == _MIPS_SIM_ABI32) - -void ScaleRowDown2_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - __asm__ __volatile__( - ".set push \n" - ".set noreorder \n" - - "srl $t9, %[dst_width], 4 \n" // iterations -> by 16 - "beqz $t9, 2f \n" - " nop \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t1, 4(%[src_ptr]) \n" // |7|6|5|4| - "lw $t2, 8(%[src_ptr]) \n" // |11|10|9|8| - "lw $t3, 12(%[src_ptr]) \n" // |15|14|13|12| - "lw $t4, 16(%[src_ptr]) \n" // |19|18|17|16| - "lw $t5, 20(%[src_ptr]) \n" // |23|22|21|20| - "lw $t6, 24(%[src_ptr]) \n" // |27|26|25|24| - "lw $t7, 28(%[src_ptr]) \n" // |31|30|29|28| - // TODO(fbarchard): Use odd pixels instead of even. - "precr.qb.ph $t8, $t1, $t0 \n" // |6|4|2|0| - "precr.qb.ph $t0, $t3, $t2 \n" // |14|12|10|8| - "precr.qb.ph $t1, $t5, $t4 \n" // |22|20|18|16| - "precr.qb.ph $t2, $t7, $t6 \n" // |30|28|26|24| - "addiu %[src_ptr], %[src_ptr], 32 \n" - "addiu $t9, $t9, -1 \n" - "sw $t8, 0(%[dst]) \n" - "sw $t0, 4(%[dst]) \n" - "sw $t1, 8(%[dst]) \n" - "sw $t2, 12(%[dst]) \n" - "bgtz $t9, 1b \n" - " addiu %[dst], %[dst], 16 \n" - - "2: \n" - "andi $t9, %[dst_width], 0xf \n" // residue - "beqz $t9, 3f \n" - " nop \n" - - "21: \n" - "lbu $t0, 0(%[src_ptr]) \n" - "addiu %[src_ptr], %[src_ptr], 2 \n" - "addiu $t9, $t9, -1 \n" - "sb $t0, 0(%[dst]) \n" - "bgtz $t9, 21b \n" - " addiu %[dst], %[dst], 1 \n" - - "3: \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [dst] "+r" (dst) - : [dst_width] "r" (dst_width) - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9" - ); -} - -void ScaleRowDown2Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - const uint8* t = src_ptr + src_stride; - - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - "srl $t9, %[dst_width], 3 \n" // iterations -> step 8 - "bltz $t9, 2f \n" - " nop \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t1, 4(%[src_ptr]) \n" // |7|6|5|4| - "lw $t2, 8(%[src_ptr]) \n" // |11|10|9|8| - "lw $t3, 12(%[src_ptr]) \n" // |15|14|13|12| - "lw $t4, 0(%[t]) \n" // |19|18|17|16| - "lw $t5, 4(%[t]) \n" // |23|22|21|20| - "lw $t6, 8(%[t]) \n" // |27|26|25|24| - "lw $t7, 12(%[t]) \n" // |31|30|29|28| - "addiu $t9, $t9, -1 \n" - "srl $t8, $t0, 16 \n" // |X|X|3|2| - "ins $t0, $t4, 16, 16 \n" // |17|16|1|0| - "ins $t4, $t8, 0, 16 \n" // |19|18|3|2| - "raddu.w.qb $t0, $t0 \n" // |17+16+1+0| - "raddu.w.qb $t4, $t4 \n" // |19+18+3+2| - "shra_r.w $t0, $t0, 2 \n" // |t0+2|>>2 - "shra_r.w $t4, $t4, 2 \n" // |t4+2|>>2 - "srl $t8, $t1, 16 \n" // |X|X|7|6| - "ins $t1, $t5, 16, 16 \n" // |21|20|5|4| - "ins $t5, $t8, 0, 16 \n" // |22|23|7|6| - "raddu.w.qb $t1, $t1 \n" // |21+20+5+4| - "raddu.w.qb $t5, $t5 \n" // |23+22+7+6| - "shra_r.w $t1, $t1, 2 \n" // |t1+2|>>2 - "shra_r.w $t5, $t5, 2 \n" // |t5+2|>>2 - "srl $t8, $t2, 16 \n" // |X|X|11|10| - "ins $t2, $t6, 16, 16 \n" // |25|24|9|8| - "ins $t6, $t8, 0, 16 \n" // |27|26|11|10| - "raddu.w.qb $t2, $t2 \n" // |25+24+9+8| - "raddu.w.qb $t6, $t6 \n" // |27+26+11+10| - "shra_r.w $t2, $t2, 2 \n" // |t2+2|>>2 - "shra_r.w $t6, $t6, 2 \n" // |t5+2|>>2 - "srl $t8, $t3, 16 \n" // |X|X|15|14| - "ins $t3, $t7, 16, 16 \n" // |29|28|13|12| - "ins $t7, $t8, 0, 16 \n" // |31|30|15|14| - "raddu.w.qb $t3, $t3 \n" // |29+28+13+12| - "raddu.w.qb $t7, $t7 \n" // |31+30+15+14| - "shra_r.w $t3, $t3, 2 \n" // |t3+2|>>2 - "shra_r.w $t7, $t7, 2 \n" // |t7+2|>>2 - "addiu %[src_ptr], %[src_ptr], 16 \n" - "addiu %[t], %[t], 16 \n" - "sb $t0, 0(%[dst]) \n" - "sb $t4, 1(%[dst]) \n" - "sb $t1, 2(%[dst]) \n" - "sb $t5, 3(%[dst]) \n" - "sb $t2, 4(%[dst]) \n" - "sb $t6, 5(%[dst]) \n" - "sb $t3, 6(%[dst]) \n" - "sb $t7, 7(%[dst]) \n" - "bgtz $t9, 1b \n" - " addiu %[dst], %[dst], 8 \n" - - "2: \n" - "andi $t9, %[dst_width], 0x7 \n" // x = residue - "beqz $t9, 3f \n" - " nop \n" - - "21: \n" - "lwr $t1, 0(%[src_ptr]) \n" - "lwl $t1, 3(%[src_ptr]) \n" - "lwr $t2, 0(%[t]) \n" - "lwl $t2, 3(%[t]) \n" - "srl $t8, $t1, 16 \n" - "ins $t1, $t2, 16, 16 \n" - "ins $t2, $t8, 0, 16 \n" - "raddu.w.qb $t1, $t1 \n" - "raddu.w.qb $t2, $t2 \n" - "shra_r.w $t1, $t1, 2 \n" - "shra_r.w $t2, $t2, 2 \n" - "sb $t1, 0(%[dst]) \n" - "sb $t2, 1(%[dst]) \n" - "addiu %[src_ptr], %[src_ptr], 4 \n" - "addiu $t9, $t9, -2 \n" - "addiu %[t], %[t], 4 \n" - "bgtz $t9, 21b \n" - " addiu %[dst], %[dst], 2 \n" - - "3: \n" - ".set pop \n" - - : [src_ptr] "+r" (src_ptr), - [dst] "+r" (dst), [t] "+r" (t) - : [dst_width] "r" (dst_width) - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9" - ); -} - -void ScaleRowDown4_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - "srl $t9, %[dst_width], 3 \n" - "beqz $t9, 2f \n" - " nop \n" - - ".p2align 2 \n" - "1: \n" - "lw $t1, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t2, 4(%[src_ptr]) \n" // |7|6|5|4| - "lw $t3, 8(%[src_ptr]) \n" // |11|10|9|8| - "lw $t4, 12(%[src_ptr]) \n" // |15|14|13|12| - "lw $t5, 16(%[src_ptr]) \n" // |19|18|17|16| - "lw $t6, 20(%[src_ptr]) \n" // |23|22|21|20| - "lw $t7, 24(%[src_ptr]) \n" // |27|26|25|24| - "lw $t8, 28(%[src_ptr]) \n" // |31|30|29|28| - "precr.qb.ph $t1, $t2, $t1 \n" // |6|4|2|0| - "precr.qb.ph $t2, $t4, $t3 \n" // |14|12|10|8| - "precr.qb.ph $t5, $t6, $t5 \n" // |22|20|18|16| - "precr.qb.ph $t6, $t8, $t7 \n" // |30|28|26|24| - "precr.qb.ph $t1, $t2, $t1 \n" // |12|8|4|0| - "precr.qb.ph $t5, $t6, $t5 \n" // |28|24|20|16| - "addiu %[src_ptr], %[src_ptr], 32 \n" - "addiu $t9, $t9, -1 \n" - "sw $t1, 0(%[dst]) \n" - "sw $t5, 4(%[dst]) \n" - "bgtz $t9, 1b \n" - " addiu %[dst], %[dst], 8 \n" - - "2: \n" - "andi $t9, %[dst_width], 7 \n" // residue - "beqz $t9, 3f \n" - " nop \n" - - "21: \n" - "lbu $t1, 0(%[src_ptr]) \n" - "addiu %[src_ptr], %[src_ptr], 4 \n" - "addiu $t9, $t9, -1 \n" - "sb $t1, 0(%[dst]) \n" - "bgtz $t9, 21b \n" - " addiu %[dst], %[dst], 1 \n" - - "3: \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [dst] "+r" (dst) - : [dst_width] "r" (dst_width) - : "t1", "t2", "t3", "t4", "t5", - "t6", "t7", "t8", "t9" - ); -} - -void ScaleRowDown4Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - intptr_t stride = src_stride; - const uint8* s1 = src_ptr + stride; - const uint8* s2 = s1 + stride; - const uint8* s3 = s2 + stride; - - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - "srl $t9, %[dst_width], 1 \n" - "andi $t8, %[dst_width], 1 \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t1, 0(%[s1]) \n" // |7|6|5|4| - "lw $t2, 0(%[s2]) \n" // |11|10|9|8| - "lw $t3, 0(%[s3]) \n" // |15|14|13|12| - "lw $t4, 4(%[src_ptr]) \n" // |19|18|17|16| - "lw $t5, 4(%[s1]) \n" // |23|22|21|20| - "lw $t6, 4(%[s2]) \n" // |27|26|25|24| - "lw $t7, 4(%[s3]) \n" // |31|30|29|28| - "raddu.w.qb $t0, $t0 \n" // |3 + 2 + 1 + 0| - "raddu.w.qb $t1, $t1 \n" // |7 + 6 + 5 + 4| - "raddu.w.qb $t2, $t2 \n" // |11 + 10 + 9 + 8| - "raddu.w.qb $t3, $t3 \n" // |15 + 14 + 13 + 12| - "raddu.w.qb $t4, $t4 \n" // |19 + 18 + 17 + 16| - "raddu.w.qb $t5, $t5 \n" // |23 + 22 + 21 + 20| - "raddu.w.qb $t6, $t6 \n" // |27 + 26 + 25 + 24| - "raddu.w.qb $t7, $t7 \n" // |31 + 30 + 29 + 28| - "add $t0, $t0, $t1 \n" - "add $t1, $t2, $t3 \n" - "add $t0, $t0, $t1 \n" - "add $t4, $t4, $t5 \n" - "add $t6, $t6, $t7 \n" - "add $t4, $t4, $t6 \n" - "shra_r.w $t0, $t0, 4 \n" - "shra_r.w $t4, $t4, 4 \n" - "sb $t0, 0(%[dst]) \n" - "sb $t4, 1(%[dst]) \n" - "addiu %[src_ptr], %[src_ptr], 8 \n" - "addiu %[s1], %[s1], 8 \n" - "addiu %[s2], %[s2], 8 \n" - "addiu %[s3], %[s3], 8 \n" - "addiu $t9, $t9, -1 \n" - "bgtz $t9, 1b \n" - " addiu %[dst], %[dst], 2 \n" - "beqz $t8, 2f \n" - " nop \n" - - "lw $t0, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t1, 0(%[s1]) \n" // |7|6|5|4| - "lw $t2, 0(%[s2]) \n" // |11|10|9|8| - "lw $t3, 0(%[s3]) \n" // |15|14|13|12| - "raddu.w.qb $t0, $t0 \n" // |3 + 2 + 1 + 0| - "raddu.w.qb $t1, $t1 \n" // |7 + 6 + 5 + 4| - "raddu.w.qb $t2, $t2 \n" // |11 + 10 + 9 + 8| - "raddu.w.qb $t3, $t3 \n" // |15 + 14 + 13 + 12| - "add $t0, $t0, $t1 \n" - "add $t1, $t2, $t3 \n" - "add $t0, $t0, $t1 \n" - "shra_r.w $t0, $t0, 4 \n" - "sb $t0, 0(%[dst]) \n" - - "2: \n" - ".set pop \n" - - : [src_ptr] "+r" (src_ptr), - [dst] "+r" (dst), - [s1] "+r" (s1), - [s2] "+r" (s2), - [s3] "+r" (s3) - : [dst_width] "r" (dst_width) - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6","t7", "t8", "t9" - ); -} - -void ScaleRowDown34_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - ".p2align 2 \n" - "1: \n" - "lw $t1, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t2, 4(%[src_ptr]) \n" // |7|6|5|4| - "lw $t3, 8(%[src_ptr]) \n" // |11|10|9|8| - "lw $t4, 12(%[src_ptr]) \n" // |15|14|13|12| - "lw $t5, 16(%[src_ptr]) \n" // |19|18|17|16| - "lw $t6, 20(%[src_ptr]) \n" // |23|22|21|20| - "lw $t7, 24(%[src_ptr]) \n" // |27|26|25|24| - "lw $t8, 28(%[src_ptr]) \n" // |31|30|29|28| - "precrq.qb.ph $t0, $t2, $t4 \n" // |7|5|15|13| - "precrq.qb.ph $t9, $t6, $t8 \n" // |23|21|31|30| - "addiu %[dst_width], %[dst_width], -24 \n" - "ins $t1, $t1, 8, 16 \n" // |3|1|0|X| - "ins $t4, $t0, 8, 16 \n" // |X|15|13|12| - "ins $t5, $t5, 8, 16 \n" // |19|17|16|X| - "ins $t8, $t9, 8, 16 \n" // |X|31|29|28| - "addiu %[src_ptr], %[src_ptr], 32 \n" - "packrl.ph $t0, $t3, $t0 \n" // |9|8|7|5| - "packrl.ph $t9, $t7, $t9 \n" // |25|24|23|21| - "prepend $t1, $t2, 8 \n" // |4|3|1|0| - "prepend $t3, $t4, 24 \n" // |15|13|12|11| - "prepend $t5, $t6, 8 \n" // |20|19|17|16| - "prepend $t7, $t8, 24 \n" // |31|29|28|27| - "sw $t1, 0(%[dst]) \n" - "sw $t0, 4(%[dst]) \n" - "sw $t3, 8(%[dst]) \n" - "sw $t5, 12(%[dst]) \n" - "sw $t9, 16(%[dst]) \n" - "sw $t7, 20(%[dst]) \n" - "bnez %[dst_width], 1b \n" - " addiu %[dst], %[dst], 24 \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [dst] "+r" (dst), - [dst_width] "+r" (dst_width) - : - : "t0", "t1", "t2", "t3", "t4", "t5", - "t6","t7", "t8", "t9" - ); -} - -void ScaleRowDown34_0_Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "repl.ph $t3, 3 \n" // 0x00030003 - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |S3|S2|S1|S0| - "lwx $t1, %[src_stride](%[src_ptr]) \n" // |T3|T2|T1|T0| - "rotr $t2, $t0, 8 \n" // |S0|S3|S2|S1| - "rotr $t6, $t1, 8 \n" // |T0|T3|T2|T1| - "muleu_s.ph.qbl $t4, $t2, $t3 \n" // |S0*3|S3*3| - "muleu_s.ph.qbl $t5, $t6, $t3 \n" // |T0*3|T3*3| - "andi $t0, $t2, 0xFFFF \n" // |0|0|S2|S1| - "andi $t1, $t6, 0xFFFF \n" // |0|0|T2|T1| - "raddu.w.qb $t0, $t0 \n" - "raddu.w.qb $t1, $t1 \n" - "shra_r.w $t0, $t0, 1 \n" - "shra_r.w $t1, $t1, 1 \n" - "preceu.ph.qbr $t2, $t2 \n" // |0|S2|0|S1| - "preceu.ph.qbr $t6, $t6 \n" // |0|T2|0|T1| - "rotr $t2, $t2, 16 \n" // |0|S1|0|S2| - "rotr $t6, $t6, 16 \n" // |0|T1|0|T2| - "addu.ph $t2, $t2, $t4 \n" - "addu.ph $t6, $t6, $t5 \n" - "sll $t5, $t0, 1 \n" - "add $t0, $t5, $t0 \n" - "shra_r.ph $t2, $t2, 2 \n" - "shra_r.ph $t6, $t6, 2 \n" - "shll.ph $t4, $t2, 1 \n" - "addq.ph $t4, $t4, $t2 \n" - "addu $t0, $t0, $t1 \n" - "addiu %[src_ptr], %[src_ptr], 4 \n" - "shra_r.w $t0, $t0, 2 \n" - "addu.ph $t6, $t6, $t4 \n" - "shra_r.ph $t6, $t6, 2 \n" - "srl $t1, $t6, 16 \n" - "addiu %[dst_width], %[dst_width], -3 \n" - "sb $t1, 0(%[d]) \n" - "sb $t0, 1(%[d]) \n" - "sb $t6, 2(%[d]) \n" - "bgtz %[dst_width], 1b \n" - " addiu %[d], %[d], 3 \n" - "3: \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [src_stride] "+r" (src_stride), - [d] "+r" (d), - [dst_width] "+r" (dst_width) - : - : "t0", "t1", "t2", "t3", - "t4", "t5", "t6" - ); -} - -void ScaleRowDown34_1_Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* d, int dst_width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - "repl.ph $t2, 3 \n" // 0x00030003 - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |S3|S2|S1|S0| - "lwx $t1, %[src_stride](%[src_ptr]) \n" // |T3|T2|T1|T0| - "rotr $t4, $t0, 8 \n" // |S0|S3|S2|S1| - "rotr $t6, $t1, 8 \n" // |T0|T3|T2|T1| - "muleu_s.ph.qbl $t3, $t4, $t2 \n" // |S0*3|S3*3| - "muleu_s.ph.qbl $t5, $t6, $t2 \n" // |T0*3|T3*3| - "andi $t0, $t4, 0xFFFF \n" // |0|0|S2|S1| - "andi $t1, $t6, 0xFFFF \n" // |0|0|T2|T1| - "raddu.w.qb $t0, $t0 \n" - "raddu.w.qb $t1, $t1 \n" - "shra_r.w $t0, $t0, 1 \n" - "shra_r.w $t1, $t1, 1 \n" - "preceu.ph.qbr $t4, $t4 \n" // |0|S2|0|S1| - "preceu.ph.qbr $t6, $t6 \n" // |0|T2|0|T1| - "rotr $t4, $t4, 16 \n" // |0|S1|0|S2| - "rotr $t6, $t6, 16 \n" // |0|T1|0|T2| - "addu.ph $t4, $t4, $t3 \n" - "addu.ph $t6, $t6, $t5 \n" - "shra_r.ph $t6, $t6, 2 \n" - "shra_r.ph $t4, $t4, 2 \n" - "addu.ph $t6, $t6, $t4 \n" - "addiu %[src_ptr], %[src_ptr], 4 \n" - "shra_r.ph $t6, $t6, 1 \n" - "addu $t0, $t0, $t1 \n" - "addiu %[dst_width], %[dst_width], -3 \n" - "shra_r.w $t0, $t0, 1 \n" - "srl $t1, $t6, 16 \n" - "sb $t1, 0(%[d]) \n" - "sb $t0, 1(%[d]) \n" - "sb $t6, 2(%[d]) \n" - "bgtz %[dst_width], 1b \n" - " addiu %[d], %[d], 3 \n" - "3: \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [src_stride] "+r" (src_stride), - [d] "+r" (d), - [dst_width] "+r" (dst_width) - : - : "t0", "t1", "t2", "t3", - "t4", "t5", "t6" - ); -} - -void ScaleRowDown38_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |3|2|1|0| - "lw $t1, 4(%[src_ptr]) \n" // |7|6|5|4| - "lw $t2, 8(%[src_ptr]) \n" // |11|10|9|8| - "lw $t3, 12(%[src_ptr]) \n" // |15|14|13|12| - "lw $t4, 16(%[src_ptr]) \n" // |19|18|17|16| - "lw $t5, 20(%[src_ptr]) \n" // |23|22|21|20| - "lw $t6, 24(%[src_ptr]) \n" // |27|26|25|24| - "lw $t7, 28(%[src_ptr]) \n" // |31|30|29|28| - "wsbh $t0, $t0 \n" // |2|3|0|1| - "wsbh $t6, $t6 \n" // |26|27|24|25| - "srl $t0, $t0, 8 \n" // |X|2|3|0| - "srl $t3, $t3, 16 \n" // |X|X|15|14| - "srl $t5, $t5, 16 \n" // |X|X|23|22| - "srl $t7, $t7, 16 \n" // |X|X|31|30| - "ins $t1, $t2, 24, 8 \n" // |8|6|5|4| - "ins $t6, $t5, 0, 8 \n" // |26|27|24|22| - "ins $t1, $t0, 0, 16 \n" // |8|6|3|0| - "ins $t6, $t7, 24, 8 \n" // |30|27|24|22| - "prepend $t2, $t3, 24 \n" // |X|15|14|11| - "ins $t4, $t4, 16, 8 \n" // |19|16|17|X| - "ins $t4, $t2, 0, 16 \n" // |19|16|14|11| - "addiu %[src_ptr], %[src_ptr], 32 \n" - "addiu %[dst_width], %[dst_width], -12 \n" - "addiu $t8,%[dst_width], -12 \n" - "sw $t1, 0(%[dst]) \n" - "sw $t4, 4(%[dst]) \n" - "sw $t6, 8(%[dst]) \n" - "bgez $t8, 1b \n" - " addiu %[dst], %[dst], 12 \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [dst] "+r" (dst), - [dst_width] "+r" (dst_width) - : - : "t0", "t1", "t2", "t3", "t4", - "t5", "t6", "t7", "t8" - ); -} - -void ScaleRowDown38_2_Box_MIPS_DSPR2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - intptr_t stride = src_stride; - const uint8* t = src_ptr + stride; - const int c = 0x2AAA; - - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |S3|S2|S1|S0| - "lw $t1, 4(%[src_ptr]) \n" // |S7|S6|S5|S4| - "lw $t2, 0(%[t]) \n" // |T3|T2|T1|T0| - "lw $t3, 4(%[t]) \n" // |T7|T6|T5|T4| - "rotr $t1, $t1, 16 \n" // |S5|S4|S7|S6| - "packrl.ph $t4, $t1, $t3 \n" // |S7|S6|T7|T6| - "packrl.ph $t5, $t3, $t1 \n" // |T5|T4|S5|S4| - "raddu.w.qb $t4, $t4 \n" // S7+S6+T7+T6 - "raddu.w.qb $t5, $t5 \n" // T5+T4+S5+S4 - "precrq.qb.ph $t6, $t0, $t2 \n" // |S3|S1|T3|T1| - "precrq.qb.ph $t6, $t6, $t6 \n" // |S3|T3|S3|T3| - "srl $t4, $t4, 2 \n" // t4 / 4 - "srl $t6, $t6, 16 \n" // |0|0|S3|T3| - "raddu.w.qb $t6, $t6 \n" // 0+0+S3+T3 - "addu $t6, $t5, $t6 \n" - "mul $t6, $t6, %[c] \n" // t6 * 0x2AAA - "sll $t0, $t0, 8 \n" // |S2|S1|S0|0| - "sll $t2, $t2, 8 \n" // |T2|T1|T0|0| - "raddu.w.qb $t0, $t0 \n" // S2+S1+S0+0 - "raddu.w.qb $t2, $t2 \n" // T2+T1+T0+0 - "addu $t0, $t0, $t2 \n" - "mul $t0, $t0, %[c] \n" // t0 * 0x2AAA - "addiu %[src_ptr], %[src_ptr], 8 \n" - "addiu %[t], %[t], 8 \n" - "addiu %[dst_width], %[dst_width], -3 \n" - "addiu %[dst_ptr], %[dst_ptr], 3 \n" - "srl $t6, $t6, 16 \n" - "srl $t0, $t0, 16 \n" - "sb $t4, -1(%[dst_ptr]) \n" - "sb $t6, -2(%[dst_ptr]) \n" - "bgtz %[dst_width], 1b \n" - " sb $t0, -3(%[dst_ptr]) \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [dst_ptr] "+r" (dst_ptr), - [t] "+r" (t), - [dst_width] "+r" (dst_width) - : [c] "r" (c) - : "t0", "t1", "t2", "t3", "t4", "t5", "t6" - ); -} - -void ScaleRowDown38_3_Box_MIPS_DSPR2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - intptr_t stride = src_stride; - const uint8* s1 = src_ptr + stride; - stride += stride; - const uint8* s2 = src_ptr + stride; - const int c1 = 0x1C71; - const int c2 = 0x2AAA; - - __asm__ __volatile__ ( - ".set push \n" - ".set noreorder \n" - - ".p2align 2 \n" - "1: \n" - "lw $t0, 0(%[src_ptr]) \n" // |S3|S2|S1|S0| - "lw $t1, 4(%[src_ptr]) \n" // |S7|S6|S5|S4| - "lw $t2, 0(%[s1]) \n" // |T3|T2|T1|T0| - "lw $t3, 4(%[s1]) \n" // |T7|T6|T5|T4| - "lw $t4, 0(%[s2]) \n" // |R3|R2|R1|R0| - "lw $t5, 4(%[s2]) \n" // |R7|R6|R5|R4| - "rotr $t1, $t1, 16 \n" // |S5|S4|S7|S6| - "packrl.ph $t6, $t1, $t3 \n" // |S7|S6|T7|T6| - "raddu.w.qb $t6, $t6 \n" // S7+S6+T7+T6 - "packrl.ph $t7, $t3, $t1 \n" // |T5|T4|S5|S4| - "raddu.w.qb $t7, $t7 \n" // T5+T4+S5+S4 - "sll $t8, $t5, 16 \n" // |R5|R4|0|0| - "raddu.w.qb $t8, $t8 \n" // R5+R4 - "addu $t7, $t7, $t8 \n" - "srl $t8, $t5, 16 \n" // |0|0|R7|R6| - "raddu.w.qb $t8, $t8 \n" // R7 + R6 - "addu $t6, $t6, $t8 \n" - "mul $t6, $t6, %[c2] \n" // t6 * 0x2AAA - "precrq.qb.ph $t8, $t0, $t2 \n" // |S3|S1|T3|T1| - "precrq.qb.ph $t8, $t8, $t4 \n" // |S3|T3|R3|R1| - "srl $t8, $t8, 8 \n" // |0|S3|T3|R3| - "raddu.w.qb $t8, $t8 \n" // S3 + T3 + R3 - "addu $t7, $t7, $t8 \n" - "mul $t7, $t7, %[c1] \n" // t7 * 0x1C71 - "sll $t0, $t0, 8 \n" // |S2|S1|S0|0| - "sll $t2, $t2, 8 \n" // |T2|T1|T0|0| - "sll $t4, $t4, 8 \n" // |R2|R1|R0|0| - "raddu.w.qb $t0, $t0 \n" - "raddu.w.qb $t2, $t2 \n" - "raddu.w.qb $t4, $t4 \n" - "addu $t0, $t0, $t2 \n" - "addu $t0, $t0, $t4 \n" - "mul $t0, $t0, %[c1] \n" // t0 * 0x1C71 - "addiu %[src_ptr], %[src_ptr], 8 \n" - "addiu %[s1], %[s1], 8 \n" - "addiu %[s2], %[s2], 8 \n" - "addiu %[dst_width], %[dst_width], -3 \n" - "addiu %[dst_ptr], %[dst_ptr], 3 \n" - "srl $t6, $t6, 16 \n" - "srl $t7, $t7, 16 \n" - "srl $t0, $t0, 16 \n" - "sb $t6, -1(%[dst_ptr]) \n" - "sb $t7, -2(%[dst_ptr]) \n" - "bgtz %[dst_width], 1b \n" - " sb $t0, -3(%[dst_ptr]) \n" - ".set pop \n" - : [src_ptr] "+r" (src_ptr), - [dst_ptr] "+r" (dst_ptr), - [s1] "+r" (s1), - [s2] "+r" (s2), - [dst_width] "+r" (dst_width) - : [c1] "r" (c1), [c2] "r" (c2) - : "t0", "t1", "t2", "t3", "t4", - "t5", "t6", "t7", "t8" - ); -} - -#endif // defined(__mips_dsp) && (__mips_dsp_rev >= 2) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - diff --git a/src/main/jni/libyuv/source/scale_neon.cc b/src/main/jni/libyuv/source/scale_neon.cc deleted file mode 100644 index 1b8a5ba58..000000000 --- a/src/main/jni/libyuv/source/scale_neon.cc +++ /dev/null @@ -1,764 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC Neon. -#if !defined(LIBYUV_DISABLE_NEON) && defined(__ARM_NEON__) - -// NEON downscalers with interpolation. -// Provided by Fritz Koenig - -// Read 32x1 throw away even pixels, and write 16x1. -void ScaleRowDown2_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - // load even pixels into q0, odd into q1 - MEMACCESS(0) - "vld2.8 {q0, q1}, [%0]! \n" - "subs %2, %2, #16 \n" // 16 processed per loop - MEMACCESS(1) - "vst1.8 {q1}, [%1]! \n" // store odd pixels - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst), // %1 - "+r"(dst_width) // %2 - : - : "q0", "q1" // Clobber List - ); -} - -// Read 32x2 average down and write 16x1. -void ScaleRowDown2Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - // change the stride to row 2 pointer - "add %1, %0 \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0, q1}, [%0]! \n" // load row 1 and post inc - MEMACCESS(1) - "vld1.8 {q2, q3}, [%1]! \n" // load row 2 and post inc - "subs %3, %3, #16 \n" // 16 processed per loop - "vpaddl.u8 q0, q0 \n" // row 1 add adjacent - "vpaddl.u8 q1, q1 \n" - "vpadal.u8 q0, q2 \n" // row 2 add adjacent + row1 - "vpadal.u8 q1, q3 \n" - "vrshrn.u16 d0, q0, #2 \n" // downshift, round and pack - "vrshrn.u16 d1, q1, #2 \n" - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(src_stride), // %1 - "+r"(dst), // %2 - "+r"(dst_width) // %3 - : - : "q0", "q1", "q2", "q3" // Clobber List - ); -} - -void ScaleRowDown4_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // src line 0 - "subs %2, %2, #8 \n" // 8 processed per loop - MEMACCESS(1) - "vst1.8 {d2}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "q0", "q1", "memory", "cc" - ); -} - -void ScaleRowDown4Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - const uint8* src_ptr1 = src_ptr + src_stride; - const uint8* src_ptr2 = src_ptr + src_stride * 2; - const uint8* src_ptr3 = src_ptr + src_stride * 3; -asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {q0}, [%0]! \n" // load up 16x4 - MEMACCESS(3) - "vld1.8 {q1}, [%3]! \n" - MEMACCESS(4) - "vld1.8 {q2}, [%4]! \n" - MEMACCESS(5) - "vld1.8 {q3}, [%5]! \n" - "subs %2, %2, #4 \n" - "vpaddl.u8 q0, q0 \n" - "vpadal.u8 q0, q1 \n" - "vpadal.u8 q0, q2 \n" - "vpadal.u8 q0, q3 \n" - "vpaddl.u16 q0, q0 \n" - "vrshrn.u32 d0, q0, #4 \n" // divide by 16 w/rounding - "vmovn.u16 d0, q0 \n" - MEMACCESS(1) - "vst1.32 {d0[0]}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_ptr1), // %3 - "+r"(src_ptr2), // %4 - "+r"(src_ptr3) // %5 - : - : "q0", "q1", "q2", "q3", "memory", "cc" - ); -} - -// Down scale from 4 to 3 pixels. Use the neon multilane read/write -// to load up the every 4th pixel into a 4 different registers. -// Point samples 32 pixels to 24 pixels. -void ScaleRowDown34_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // src line 0 - "subs %2, %2, #24 \n" - "vmov d2, d3 \n" // order d0, d1, d2 - MEMACCESS(1) - "vst3.8 {d0, d1, d2}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "d0", "d1", "d2", "d3", "memory", "cc" - ); -} - -void ScaleRowDown34_0_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "vmov.u8 d24, #3 \n" - "add %3, %0 \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // src line 0 - MEMACCESS(3) - "vld4.8 {d4, d5, d6, d7}, [%3]! \n" // src line 1 - "subs %2, %2, #24 \n" - - // filter src line 0 with src line 1 - // expand chars to shorts to allow for room - // when adding lines together - "vmovl.u8 q8, d4 \n" - "vmovl.u8 q9, d5 \n" - "vmovl.u8 q10, d6 \n" - "vmovl.u8 q11, d7 \n" - - // 3 * line_0 + line_1 - "vmlal.u8 q8, d0, d24 \n" - "vmlal.u8 q9, d1, d24 \n" - "vmlal.u8 q10, d2, d24 \n" - "vmlal.u8 q11, d3, d24 \n" - - // (3 * line_0 + line_1) >> 2 - "vqrshrn.u16 d0, q8, #2 \n" - "vqrshrn.u16 d1, q9, #2 \n" - "vqrshrn.u16 d2, q10, #2 \n" - "vqrshrn.u16 d3, q11, #2 \n" - - // a0 = (src[0] * 3 + s[1] * 1) >> 2 - "vmovl.u8 q8, d1 \n" - "vmlal.u8 q8, d0, d24 \n" - "vqrshrn.u16 d0, q8, #2 \n" - - // a1 = (src[1] * 1 + s[2] * 1) >> 1 - "vrhadd.u8 d1, d1, d2 \n" - - // a2 = (src[2] * 1 + s[3] * 3) >> 2 - "vmovl.u8 q8, d2 \n" - "vmlal.u8 q8, d3, d24 \n" - "vqrshrn.u16 d2, q8, #2 \n" - - MEMACCESS(1) - "vst3.8 {d0, d1, d2}, [%1]! \n" - - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride) // %3 - : - : "q0", "q1", "q2", "q3", "q8", "q9", "q10", "q11", "d24", "memory", "cc" - ); -} - -void ScaleRowDown34_1_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "vmov.u8 d24, #3 \n" - "add %3, %0 \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" // src line 0 - MEMACCESS(3) - "vld4.8 {d4, d5, d6, d7}, [%3]! \n" // src line 1 - "subs %2, %2, #24 \n" - // average src line 0 with src line 1 - "vrhadd.u8 q0, q0, q2 \n" - "vrhadd.u8 q1, q1, q3 \n" - - // a0 = (src[0] * 3 + s[1] * 1) >> 2 - "vmovl.u8 q3, d1 \n" - "vmlal.u8 q3, d0, d24 \n" - "vqrshrn.u16 d0, q3, #2 \n" - - // a1 = (src[1] * 1 + s[2] * 1) >> 1 - "vrhadd.u8 d1, d1, d2 \n" - - // a2 = (src[2] * 1 + s[3] * 3) >> 2 - "vmovl.u8 q3, d2 \n" - "vmlal.u8 q3, d3, d24 \n" - "vqrshrn.u16 d2, q3, #2 \n" - - MEMACCESS(1) - "vst3.8 {d0, d1, d2}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride) // %3 - : - : "r4", "q0", "q1", "q2", "q3", "d24", "memory", "cc" - ); -} - -#define HAS_SCALEROWDOWN38_NEON -static uvec8 kShuf38 = - { 0, 3, 6, 8, 11, 14, 16, 19, 22, 24, 27, 30, 0, 0, 0, 0 }; -static uvec8 kShuf38_2 = - { 0, 8, 16, 2, 10, 17, 4, 12, 18, 6, 14, 19, 0, 0, 0, 0 }; -static vec16 kMult38_Div6 = - { 65536 / 12, 65536 / 12, 65536 / 12, 65536 / 12, - 65536 / 12, 65536 / 12, 65536 / 12, 65536 / 12 }; -static vec16 kMult38_Div9 = - { 65536 / 18, 65536 / 18, 65536 / 18, 65536 / 18, - 65536 / 18, 65536 / 18, 65536 / 18, 65536 / 18 }; - -// 32 -> 12 -void ScaleRowDown38_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - MEMACCESS(3) - "vld1.8 {q3}, [%3] \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d0, d1, d2, d3}, [%0]! \n" - "subs %2, %2, #12 \n" - "vtbl.u8 d4, {d0, d1, d2, d3}, d6 \n" - "vtbl.u8 d5, {d0, d1, d2, d3}, d7 \n" - MEMACCESS(1) - "vst1.8 {d4}, [%1]! \n" - MEMACCESS(1) - "vst1.32 {d5[0]}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"(&kShuf38) // %3 - : "d0", "d1", "d2", "d3", "d4", "d5", "memory", "cc" - ); -} - -// 32x3 -> 12x1 -void OMITFP ScaleRowDown38_3_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - const uint8* src_ptr1 = src_ptr + src_stride * 2; - - asm volatile ( - MEMACCESS(5) - "vld1.16 {q13}, [%5] \n" - MEMACCESS(6) - "vld1.8 {q14}, [%6] \n" - MEMACCESS(7) - "vld1.8 {q15}, [%7] \n" - "add %3, %0 \n" - ".p2align 2 \n" - "1: \n" - - // d0 = 00 40 01 41 02 42 03 43 - // d1 = 10 50 11 51 12 52 13 53 - // d2 = 20 60 21 61 22 62 23 63 - // d3 = 30 70 31 71 32 72 33 73 - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" - MEMACCESS(3) - "vld4.8 {d4, d5, d6, d7}, [%3]! \n" - MEMACCESS(4) - "vld4.8 {d16, d17, d18, d19}, [%4]! \n" - "subs %2, %2, #12 \n" - - // Shuffle the input data around to get align the data - // so adjacent data can be added. 0,1 - 2,3 - 4,5 - 6,7 - // d0 = 00 10 01 11 02 12 03 13 - // d1 = 40 50 41 51 42 52 43 53 - "vtrn.u8 d0, d1 \n" - "vtrn.u8 d4, d5 \n" - "vtrn.u8 d16, d17 \n" - - // d2 = 20 30 21 31 22 32 23 33 - // d3 = 60 70 61 71 62 72 63 73 - "vtrn.u8 d2, d3 \n" - "vtrn.u8 d6, d7 \n" - "vtrn.u8 d18, d19 \n" - - // d0 = 00+10 01+11 02+12 03+13 - // d2 = 40+50 41+51 42+52 43+53 - "vpaddl.u8 q0, q0 \n" - "vpaddl.u8 q2, q2 \n" - "vpaddl.u8 q8, q8 \n" - - // d3 = 60+70 61+71 62+72 63+73 - "vpaddl.u8 d3, d3 \n" - "vpaddl.u8 d7, d7 \n" - "vpaddl.u8 d19, d19 \n" - - // combine source lines - "vadd.u16 q0, q2 \n" - "vadd.u16 q0, q8 \n" - "vadd.u16 d4, d3, d7 \n" - "vadd.u16 d4, d19 \n" - - // dst_ptr[3] = (s[6 + st * 0] + s[7 + st * 0] - // + s[6 + st * 1] + s[7 + st * 1] - // + s[6 + st * 2] + s[7 + st * 2]) / 6 - "vqrdmulh.s16 q2, q2, q13 \n" - "vmovn.u16 d4, q2 \n" - - // Shuffle 2,3 reg around so that 2 can be added to the - // 0,1 reg and 3 can be added to the 4,5 reg. This - // requires expanding from u8 to u16 as the 0,1 and 4,5 - // registers are already expanded. Then do transposes - // to get aligned. - // q2 = xx 20 xx 30 xx 21 xx 31 xx 22 xx 32 xx 23 xx 33 - "vmovl.u8 q1, d2 \n" - "vmovl.u8 q3, d6 \n" - "vmovl.u8 q9, d18 \n" - - // combine source lines - "vadd.u16 q1, q3 \n" - "vadd.u16 q1, q9 \n" - - // d4 = xx 20 xx 30 xx 22 xx 32 - // d5 = xx 21 xx 31 xx 23 xx 33 - "vtrn.u32 d2, d3 \n" - - // d4 = xx 20 xx 21 xx 22 xx 23 - // d5 = xx 30 xx 31 xx 32 xx 33 - "vtrn.u16 d2, d3 \n" - - // 0+1+2, 3+4+5 - "vadd.u16 q0, q1 \n" - - // Need to divide, but can't downshift as the the value - // isn't a power of 2. So multiply by 65536 / n - // and take the upper 16 bits. - "vqrdmulh.s16 q0, q0, q15 \n" - - // Align for table lookup, vtbl requires registers to - // be adjacent - "vmov.u8 d2, d4 \n" - - "vtbl.u8 d3, {d0, d1, d2}, d28 \n" - "vtbl.u8 d4, {d0, d1, d2}, d29 \n" - - MEMACCESS(1) - "vst1.8 {d3}, [%1]! \n" - MEMACCESS(1) - "vst1.32 {d4[0]}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride), // %3 - "+r"(src_ptr1) // %4 - : "r"(&kMult38_Div6), // %5 - "r"(&kShuf38_2), // %6 - "r"(&kMult38_Div9) // %7 - : "q0", "q1", "q2", "q3", "q8", "q9", "q13", "q14", "q15", "memory", "cc" - ); -} - -// 32x2 -> 12x1 -void ScaleRowDown38_2_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - MEMACCESS(4) - "vld1.16 {q13}, [%4] \n" - MEMACCESS(5) - "vld1.8 {q14}, [%5] \n" - "add %3, %0 \n" - ".p2align 2 \n" - "1: \n" - - // d0 = 00 40 01 41 02 42 03 43 - // d1 = 10 50 11 51 12 52 13 53 - // d2 = 20 60 21 61 22 62 23 63 - // d3 = 30 70 31 71 32 72 33 73 - MEMACCESS(0) - "vld4.8 {d0, d1, d2, d3}, [%0]! \n" - MEMACCESS(3) - "vld4.8 {d4, d5, d6, d7}, [%3]! \n" - "subs %2, %2, #12 \n" - - // Shuffle the input data around to get align the data - // so adjacent data can be added. 0,1 - 2,3 - 4,5 - 6,7 - // d0 = 00 10 01 11 02 12 03 13 - // d1 = 40 50 41 51 42 52 43 53 - "vtrn.u8 d0, d1 \n" - "vtrn.u8 d4, d5 \n" - - // d2 = 20 30 21 31 22 32 23 33 - // d3 = 60 70 61 71 62 72 63 73 - "vtrn.u8 d2, d3 \n" - "vtrn.u8 d6, d7 \n" - - // d0 = 00+10 01+11 02+12 03+13 - // d2 = 40+50 41+51 42+52 43+53 - "vpaddl.u8 q0, q0 \n" - "vpaddl.u8 q2, q2 \n" - - // d3 = 60+70 61+71 62+72 63+73 - "vpaddl.u8 d3, d3 \n" - "vpaddl.u8 d7, d7 \n" - - // combine source lines - "vadd.u16 q0, q2 \n" - "vadd.u16 d4, d3, d7 \n" - - // dst_ptr[3] = (s[6] + s[7] + s[6+st] + s[7+st]) / 4 - "vqrshrn.u16 d4, q2, #2 \n" - - // Shuffle 2,3 reg around so that 2 can be added to the - // 0,1 reg and 3 can be added to the 4,5 reg. This - // requires expanding from u8 to u16 as the 0,1 and 4,5 - // registers are already expanded. Then do transposes - // to get aligned. - // q2 = xx 20 xx 30 xx 21 xx 31 xx 22 xx 32 xx 23 xx 33 - "vmovl.u8 q1, d2 \n" - "vmovl.u8 q3, d6 \n" - - // combine source lines - "vadd.u16 q1, q3 \n" - - // d4 = xx 20 xx 30 xx 22 xx 32 - // d5 = xx 21 xx 31 xx 23 xx 33 - "vtrn.u32 d2, d3 \n" - - // d4 = xx 20 xx 21 xx 22 xx 23 - // d5 = xx 30 xx 31 xx 32 xx 33 - "vtrn.u16 d2, d3 \n" - - // 0+1+2, 3+4+5 - "vadd.u16 q0, q1 \n" - - // Need to divide, but can't downshift as the the value - // isn't a power of 2. So multiply by 65536 / n - // and take the upper 16 bits. - "vqrdmulh.s16 q0, q0, q13 \n" - - // Align for table lookup, vtbl requires registers to - // be adjacent - "vmov.u8 d2, d4 \n" - - "vtbl.u8 d3, {d0, d1, d2}, d28 \n" - "vtbl.u8 d4, {d0, d1, d2}, d29 \n" - - MEMACCESS(1) - "vst1.8 {d3}, [%1]! \n" - MEMACCESS(1) - "vst1.32 {d4[0]}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride) // %3 - : "r"(&kMult38_Div6), // %4 - "r"(&kShuf38_2) // %5 - : "q0", "q1", "q2", "q3", "q13", "q14", "memory", "cc" - ); -} - -// 16x2 -> 16x1 -void ScaleFilterRows_NEON(uint8* dst_ptr, - const uint8* src_ptr, ptrdiff_t src_stride, - int dst_width, int source_y_fraction) { - asm volatile ( - "cmp %4, #0 \n" - "beq 100f \n" - "add %2, %1 \n" - "cmp %4, #64 \n" - "beq 75f \n" - "cmp %4, #128 \n" - "beq 50f \n" - "cmp %4, #192 \n" - "beq 25f \n" - - "vdup.8 d5, %4 \n" - "rsb %4, #256 \n" - "vdup.8 d4, %4 \n" - // General purpose row blend. - "1: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q1}, [%2]! \n" - "subs %3, %3, #16 \n" - "vmull.u8 q13, d0, d4 \n" - "vmull.u8 q14, d1, d4 \n" - "vmlal.u8 q13, d2, d5 \n" - "vmlal.u8 q14, d3, d5 \n" - "vrshrn.u16 d0, q13, #8 \n" - "vrshrn.u16 d1, q14, #8 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 1b \n" - "b 99f \n" - - // Blend 25 / 75. - "25: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q1}, [%2]! \n" - "subs %3, %3, #16 \n" - "vrhadd.u8 q0, q1 \n" - "vrhadd.u8 q0, q1 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 25b \n" - "b 99f \n" - - // Blend 50 / 50. - "50: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q1}, [%2]! \n" - "subs %3, %3, #16 \n" - "vrhadd.u8 q0, q1 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 50b \n" - "b 99f \n" - - // Blend 75 / 25. - "75: \n" - MEMACCESS(1) - "vld1.8 {q1}, [%1]! \n" - MEMACCESS(2) - "vld1.8 {q0}, [%2]! \n" - "subs %3, %3, #16 \n" - "vrhadd.u8 q0, q1 \n" - "vrhadd.u8 q0, q1 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 75b \n" - "b 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - "100: \n" - MEMACCESS(1) - "vld1.8 {q0}, [%1]! \n" - "subs %3, %3, #16 \n" - MEMACCESS(0) - "vst1.8 {q0}, [%0]! \n" - "bgt 100b \n" - - "99: \n" - MEMACCESS(0) - "vst1.8 {d1[7]}, [%0] \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(src_stride), // %2 - "+r"(dst_width), // %3 - "+r"(source_y_fraction) // %4 - : - : "q0", "q1", "d4", "d5", "q13", "q14", "memory", "cc" - ); -} - -void ScaleARGBRowDown2_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - ".p2align 2 \n" - "1: \n" - // load even pixels into q0, odd into q1 - MEMACCESS(0) - "vld2.32 {q0, q1}, [%0]! \n" - MEMACCESS(0) - "vld2.32 {q2, q3}, [%0]! \n" - "subs %2, %2, #8 \n" // 8 processed per loop - MEMACCESS(1) - "vst1.8 {q1}, [%1]! \n" // store odd pixels - MEMACCESS(1) - "vst1.8 {q3}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc", "q0", "q1", "q2", "q3" // Clobber List - ); -} - -void ScaleARGBRowDown2Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - // change the stride to row 2 pointer - "add %1, %1, %0 \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld4.8 {d0, d2, d4, d6}, [%0]! \n" // load 8 ARGB pixels. - MEMACCESS(0) - "vld4.8 {d1, d3, d5, d7}, [%0]! \n" // load next 8 ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "vpaddl.u8 q0, q0 \n" // B 16 bytes -> 8 shorts. - "vpaddl.u8 q1, q1 \n" // G 16 bytes -> 8 shorts. - "vpaddl.u8 q2, q2 \n" // R 16 bytes -> 8 shorts. - "vpaddl.u8 q3, q3 \n" // A 16 bytes -> 8 shorts. - MEMACCESS(1) - "vld4.8 {d16, d18, d20, d22}, [%1]! \n" // load 8 more ARGB pixels. - MEMACCESS(1) - "vld4.8 {d17, d19, d21, d23}, [%1]! \n" // load last 8 ARGB pixels. - "vpadal.u8 q0, q8 \n" // B 16 bytes -> 8 shorts. - "vpadal.u8 q1, q9 \n" // G 16 bytes -> 8 shorts. - "vpadal.u8 q2, q10 \n" // R 16 bytes -> 8 shorts. - "vpadal.u8 q3, q11 \n" // A 16 bytes -> 8 shorts. - "vrshrn.u16 d0, q0, #2 \n" // downshift, round and pack - "vrshrn.u16 d1, q1, #2 \n" - "vrshrn.u16 d2, q2, #2 \n" - "vrshrn.u16 d3, q3, #2 \n" - MEMACCESS(2) - "vst4.8 {d0, d1, d2, d3}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(src_stride), // %1 - "+r"(dst), // %2 - "+r"(dst_width) // %3 - : - : "memory", "cc", "q0", "q1", "q2", "q3", "q8", "q9", "q10", "q11" - ); -} - -// Reads 4 pixels at a time. -// Alignment requirement: src_argb 4 byte aligned. -void ScaleARGBRowDownEven_NEON(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, uint8* dst_argb, int dst_width) { - asm volatile ( - "mov r12, %3, lsl #2 \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.32 {d0[0]}, [%0], r12 \n" - MEMACCESS(0) - "vld1.32 {d0[1]}, [%0], r12 \n" - MEMACCESS(0) - "vld1.32 {d1[0]}, [%0], r12 \n" - MEMACCESS(0) - "vld1.32 {d1[1]}, [%0], r12 \n" - "subs %2, %2, #4 \n" // 4 pixels per loop. - MEMACCESS(1) - "vst1.8 {q0}, [%1]! \n" - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(dst_width) // %2 - : "r"(src_stepx) // %3 - : "memory", "cc", "r12", "q0" - ); -} - -// Reads 4 pixels at a time. -// Alignment requirement: src_argb 4 byte aligned. -void ScaleARGBRowDownEvenBox_NEON(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - asm volatile ( - "mov r12, %4, lsl #2 \n" - "add %1, %1, %0 \n" - ".p2align 2 \n" - "1: \n" - MEMACCESS(0) - "vld1.8 {d0}, [%0], r12 \n" // Read 4 2x2 blocks -> 2x1 - MEMACCESS(1) - "vld1.8 {d1}, [%1], r12 \n" - MEMACCESS(0) - "vld1.8 {d2}, [%0], r12 \n" - MEMACCESS(1) - "vld1.8 {d3}, [%1], r12 \n" - MEMACCESS(0) - "vld1.8 {d4}, [%0], r12 \n" - MEMACCESS(1) - "vld1.8 {d5}, [%1], r12 \n" - MEMACCESS(0) - "vld1.8 {d6}, [%0], r12 \n" - MEMACCESS(1) - "vld1.8 {d7}, [%1], r12 \n" - "vaddl.u8 q0, d0, d1 \n" - "vaddl.u8 q1, d2, d3 \n" - "vaddl.u8 q2, d4, d5 \n" - "vaddl.u8 q3, d6, d7 \n" - "vswp.8 d1, d2 \n" // ab_cd -> ac_bd - "vswp.8 d5, d6 \n" // ef_gh -> eg_fh - "vadd.u16 q0, q0, q1 \n" // (a+b)_(c+d) - "vadd.u16 q2, q2, q3 \n" // (e+f)_(g+h) - "vrshrn.u16 d0, q0, #2 \n" // first 2 pixels. - "vrshrn.u16 d1, q2, #2 \n" // next 2 pixels. - "subs %3, %3, #4 \n" // 4 pixels per loop. - MEMACCESS(2) - "vst1.8 {q0}, [%2]! \n" - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stride), // %1 - "+r"(dst_argb), // %2 - "+r"(dst_width) // %3 - : "r"(src_stepx) // %4 - : "memory", "cc", "r12", "q0", "q1", "q2", "q3" - ); -} - -#endif // __ARM_NEON__ - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/scale_neon64.cc b/src/main/jni/libyuv/source/scale_neon64.cc deleted file mode 100644 index 44df55c6c..000000000 --- a/src/main/jni/libyuv/source/scale_neon64.cc +++ /dev/null @@ -1,789 +0,0 @@ -/* - * Copyright 2014 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC Neon. -#if !defined(LIBYUV_DISABLE_NEON) && defined(__aarch64__) -#ifdef HAS_SCALEROWDOWN2_NEON -// Read 32x1 throw away even pixels, and write 16x1. -void ScaleRowDown2_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - "1: \n" - // load even pixels into v0, odd into v1 - MEMACCESS(0) - "ld2 {v0.16b, v1.16b}, [%0], #32 \n" - "subs %2, %2, #16 \n" // 16 processed per loop - MEMACCESS(1) - "st1 {v1.16b}, [%1], #16 \n" // store odd pixels - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst), // %1 - "+r"(dst_width) // %2 - : - : "v0", "v1" // Clobber List - ); -} -#endif //HAS_SCALEROWDOWN2_NEON - -#ifdef HAS_SCALEROWDOWN2_NEON -// Read 32x2 average down and write 16x1. -void ScaleRowDown2Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - // change the stride to row 2 pointer - "add %1, %1, %0 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b, v1.16b}, [%0], #32 \n" // load row 1 and post inc - MEMACCESS(1) - "ld1 {v2.16b, v3.16b}, [%1], #32 \n" // load row 2 and post inc - "subs %3, %3, #16 \n" // 16 processed per loop - "uaddlp v0.8h, v0.16b \n" // row 1 add adjacent - "uaddlp v1.8h, v1.16b \n" - "uadalp v0.8h, v2.16b \n" // row 2 add adjacent + row1 - "uadalp v1.8h, v3.16b \n" - "rshrn v0.8b, v0.8h, #2 \n" // downshift, round and pack - "rshrn2 v0.16b, v1.8h, #2 \n" - MEMACCESS(2) - "st1 {v0.16b}, [%2], #16 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(src_stride), // %1 - "+r"(dst), // %2 - "+r"(dst_width) // %3 - : - : "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif //HAS_SCALEROWDOWN2_NEON - -#ifdef HAS_SCALEROWDOWN4_NEON -void ScaleRowDown4_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-3.8b}, [%0], #32 \n" // src line 0 - "subs %2, %2, #8 \n" // 8 processed per loop - MEMACCESS(1) - "st1 {v2.8b}, [%1], #8 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "v0", "v1", "v2", "v3", "memory", "cc" - ); -} -#endif //HAS_SCALEROWDOWN4_NEON - -#ifdef HAS_SCALEROWDOWN4_NEON -void ScaleRowDown4Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - const uint8* src_ptr1 = src_ptr + src_stride; - const uint8* src_ptr2 = src_ptr + src_stride * 2; - const uint8* src_ptr3 = src_ptr + src_stride * 3; -asm volatile ( - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b}, [%0], #16 \n" // load up 16x4 - MEMACCESS(3) - "ld1 {v1.16b}, [%3], #16 \n" - MEMACCESS(4) - "ld1 {v2.16b}, [%4], #16 \n" - MEMACCESS(5) - "ld1 {v3.16b}, [%5], #16 \n" - "subs %2, %2, #4 \n" - "uaddlp v0.8h, v0.16b \n" - "uadalp v0.8h, v1.16b \n" - "uadalp v0.8h, v2.16b \n" - "uadalp v0.8h, v3.16b \n" - "addp v0.8h, v0.8h, v0.8h \n" - "rshrn v0.8b, v0.8h, #4 \n" // divide by 16 w/rounding - MEMACCESS(1) - "st1 {v0.s}[0], [%1], #4 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_ptr1), // %3 - "+r"(src_ptr2), // %4 - "+r"(src_ptr3) // %5 - : - : "v0", "v1", "v2", "v3", "memory", "cc" - ); -} -#endif //HAS_SCALEROWDOWN4_NEON - -#ifdef HAS_SCALEROWDOWN34_NEON -// Down scale from 4 to 3 pixels. Use the neon multilane read/write -// to load up the every 4th pixel into a 4 different registers. -// Point samples 32 pixels to 24 pixels. -void ScaleRowDown34_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // src line 0 - "subs %2, %2, #24 \n" - "mov v2.8b, v3.8b \n" // order v0, v1, v2 - MEMACCESS(1) - "st3 {v0.8b-v2.8b}, [%1], #24 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "v0", "v1", "v2", "v3", "memory", "cc" - ); -} -#endif //HAS_SCALEROWDOWN34_NEON - -#ifdef HAS_SCALEROWDOWN34_NEON -void ScaleRowDown34_0_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movi v20.8b, #3 \n" - "add %3, %3, %0 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // src line 0 - MEMACCESS(3) - "ld4 {v4.8b-v7.8b}, [%3], #32 \n" // src line 1 - "subs %2, %2, #24 \n" - - // filter src line 0 with src line 1 - // expand chars to shorts to allow for room - // when adding lines together - "ushll v16.8h, v4.8b, #0 \n" - "ushll v17.8h, v5.8b, #0 \n" - "ushll v18.8h, v6.8b, #0 \n" - "ushll v19.8h, v7.8b, #0 \n" - - // 3 * line_0 + line_1 - "umlal v16.8h, v0.8b, v20.8b \n" - "umlal v17.8h, v1.8b, v20.8b \n" - "umlal v18.8h, v2.8b, v20.8b \n" - "umlal v19.8h, v3.8b, v20.8b \n" - - // (3 * line_0 + line_1) >> 2 - "uqrshrn v0.8b, v16.8h, #2 \n" - "uqrshrn v1.8b, v17.8h, #2 \n" - "uqrshrn v2.8b, v18.8h, #2 \n" - "uqrshrn v3.8b, v19.8h, #2 \n" - - // a0 = (src[0] * 3 + s[1] * 1) >> 2 - "ushll v16.8h, v1.8b, #0 \n" - "umlal v16.8h, v0.8b, v20.8b \n" - "uqrshrn v0.8b, v16.8h, #2 \n" - - // a1 = (src[1] * 1 + s[2] * 1) >> 1 - "urhadd v1.8b, v1.8b, v2.8b \n" - - // a2 = (src[2] * 1 + s[3] * 3) >> 2 - "ushll v16.8h, v2.8b, #0 \n" - "umlal v16.8h, v3.8b, v20.8b \n" - "uqrshrn v2.8b, v16.8h, #2 \n" - - MEMACCESS(1) - "st3 {v0.8b-v2.8b}, [%1], #24 \n" - - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride) // %3 - : - : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16", "v17", "v18", "v19", - "v20", "memory", "cc" - ); -} -#endif //ScaleRowDown34_0_Box_NEON - -#ifdef HAS_SCALEROWDOWN34_NEON -void ScaleRowDown34_1_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movi v20.8b, #3 \n" - "add %3, %3, %0 \n" - "1: \n" - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" // src line 0 - MEMACCESS(3) - "ld4 {v4.8b-v7.8b}, [%3], #32 \n" // src line 1 - "subs %2, %2, #24 \n" - // average src line 0 with src line 1 - "urhadd v0.8b, v0.8b, v4.8b \n" - "urhadd v1.8b, v1.8b, v5.8b \n" - "urhadd v2.8b, v2.8b, v6.8b \n" - "urhadd v3.8b, v3.8b, v7.8b \n" - - // a0 = (src[0] * 3 + s[1] * 1) >> 2 - "ushll v4.8h, v1.8b, #0 \n" - "umlal v4.8h, v0.8b, v20.8b \n" - "uqrshrn v0.8b, v4.8h, #2 \n" - - // a1 = (src[1] * 1 + s[2] * 1) >> 1 - "urhadd v1.8b, v1.8b, v2.8b \n" - - // a2 = (src[2] * 1 + s[3] * 3) >> 2 - "ushll v4.8h, v2.8b, #0 \n" - "umlal v4.8h, v3.8b, v20.8b \n" - "uqrshrn v2.8b, v4.8h, #2 \n" - - MEMACCESS(1) - "st3 {v0.8b-v2.8b}, [%1], #24 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride) // %3 - : - : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v20", "memory", "cc" - ); -} -#endif //HAS_SCALEROWDOWN34_NEON - -#ifdef HAS_SCALEROWDOWN38_NEON -static uvec8 kShuf38 = - { 0, 3, 6, 8, 11, 14, 16, 19, 22, 24, 27, 30, 0, 0, 0, 0 }; -static uvec8 kShuf38_2 = - { 0, 16, 32, 2, 18, 33, 4, 20, 34, 6, 22, 35, 0, 0, 0, 0 }; -static vec16 kMult38_Div6 = - { 65536 / 12, 65536 / 12, 65536 / 12, 65536 / 12, - 65536 / 12, 65536 / 12, 65536 / 12, 65536 / 12 }; -static vec16 kMult38_Div9 = - { 65536 / 18, 65536 / 18, 65536 / 18, 65536 / 18, - 65536 / 18, 65536 / 18, 65536 / 18, 65536 / 18 }; - -// 32 -> 12 -void ScaleRowDown38_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - MEMACCESS(3) - "ld1 {v3.16b}, [%3] \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.16b, v1.16b}, [%0], #32 \n" - "subs %2, %2, #12 \n" - "tbl v2.16b, {v0.16b, v1.16b}, v3.16b \n" - MEMACCESS(1) - "st1 {v2.8b}, [%1], #8 \n" - MEMACCESS(1) - "st1 {v2.s}[2], [%1], #4 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"(&kShuf38) // %3 - : "v0", "v1", "v2", "v3", "memory", "cc" - ); -} - -#endif //HAS_SCALEROWDOWN38_NEON - -#ifdef HAS_SCALEROWDOWN38_NEON -// 32x3 -> 12x1 -void OMITFP ScaleRowDown38_3_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - const uint8* src_ptr1 = src_ptr + src_stride * 2; - - asm volatile ( - MEMACCESS(5) - "ld1 {v29.8h}, [%5] \n" - MEMACCESS(6) - "ld1 {v30.16b}, [%6] \n" - MEMACCESS(7) - "ld1 {v31.8h}, [%7] \n" - "add %3, %3, %0 \n" - "1: \n" - - // 00 40 01 41 02 42 03 43 - // 10 50 11 51 12 52 13 53 - // 20 60 21 61 22 62 23 63 - // 30 70 31 71 32 72 33 73 - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" - MEMACCESS(3) - "ld4 {v4.8b-v7.8b}, [%3], #32 \n" - MEMACCESS(4) - "ld4 {v16.8b-v19.8b}, [%4], #32 \n" - "subs %2, %2, #12 \n" - - // Shuffle the input data around to get align the data - // so adjacent data can be added. 0,1 - 2,3 - 4,5 - 6,7 - // 00 10 01 11 02 12 03 13 - // 40 50 41 51 42 52 43 53 - "trn1 v20.8b, v0.8b, v1.8b \n" - "trn2 v21.8b, v0.8b, v1.8b \n" - "trn1 v22.8b, v4.8b, v5.8b \n" - "trn2 v23.8b, v4.8b, v5.8b \n" - "trn1 v24.8b, v16.8b, v17.8b \n" - "trn2 v25.8b, v16.8b, v17.8b \n" - - // 20 30 21 31 22 32 23 33 - // 60 70 61 71 62 72 63 73 - "trn1 v0.8b, v2.8b, v3.8b \n" - "trn2 v1.8b, v2.8b, v3.8b \n" - "trn1 v4.8b, v6.8b, v7.8b \n" - "trn2 v5.8b, v6.8b, v7.8b \n" - "trn1 v16.8b, v18.8b, v19.8b \n" - "trn2 v17.8b, v18.8b, v19.8b \n" - - // 00+10 01+11 02+12 03+13 - // 40+50 41+51 42+52 43+53 - "uaddlp v20.4h, v20.8b \n" - "uaddlp v21.4h, v21.8b \n" - "uaddlp v22.4h, v22.8b \n" - "uaddlp v23.4h, v23.8b \n" - "uaddlp v24.4h, v24.8b \n" - "uaddlp v25.4h, v25.8b \n" - - // 60+70 61+71 62+72 63+73 - "uaddlp v1.4h, v1.8b \n" - "uaddlp v5.4h, v5.8b \n" - "uaddlp v17.4h, v17.8b \n" - - // combine source lines - "add v20.4h, v20.4h, v22.4h \n" - "add v21.4h, v21.4h, v23.4h \n" - "add v20.4h, v20.4h, v24.4h \n" - "add v21.4h, v21.4h, v25.4h \n" - "add v2.4h, v1.4h, v5.4h \n" - "add v2.4h, v2.4h, v17.4h \n" - - // dst_ptr[3] = (s[6 + st * 0] + s[7 + st * 0] - // + s[6 + st * 1] + s[7 + st * 1] - // + s[6 + st * 2] + s[7 + st * 2]) / 6 - "sqrdmulh v2.8h, v2.8h, v29.8h \n" - "xtn v2.8b, v2.8h \n" - - // Shuffle 2,3 reg around so that 2 can be added to the - // 0,1 reg and 3 can be added to the 4,5 reg. This - // requires expanding from u8 to u16 as the 0,1 and 4,5 - // registers are already expanded. Then do transposes - // to get aligned. - // xx 20 xx 30 xx 21 xx 31 xx 22 xx 32 xx 23 xx 33 - "ushll v16.8h, v16.8b, #0 \n" - "uaddl v0.8h, v0.8b, v4.8b \n" - - // combine source lines - "add v0.8h, v0.8h, v16.8h \n" - - // xx 20 xx 21 xx 22 xx 23 - // xx 30 xx 31 xx 32 xx 33 - "trn1 v1.8h, v0.8h, v0.8h \n" - "trn2 v4.8h, v0.8h, v0.8h \n" - "xtn v0.4h, v1.4s \n" - "xtn v4.4h, v4.4s \n" - - // 0+1+2, 3+4+5 - "add v20.8h, v20.8h, v0.8h \n" - "add v21.8h, v21.8h, v4.8h \n" - - // Need to divide, but can't downshift as the the value - // isn't a power of 2. So multiply by 65536 / n - // and take the upper 16 bits. - "sqrdmulh v0.8h, v20.8h, v31.8h \n" - "sqrdmulh v1.8h, v21.8h, v31.8h \n" - - // Align for table lookup, vtbl requires registers to - // be adjacent - "tbl v3.16b, {v0.16b, v1.16b, v2.16b}, v30.16b \n" - - MEMACCESS(1) - "st1 {v3.8b}, [%1], #8 \n" - MEMACCESS(1) - "st1 {v3.s}[2], [%1], #4 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride), // %3 - "+r"(src_ptr1) // %4 - : "r"(&kMult38_Div6), // %5 - "r"(&kShuf38_2), // %6 - "r"(&kMult38_Div9) // %7 - : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16", "v17", - "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25", "v29", - "v30", "v31", "memory", "cc" - ); -} -#endif //HAS_SCALEROWDOWN38_NEON - -#ifdef HAS_SCALEROWDOWN38_NEON -// 32x2 -> 12x1 -void ScaleRowDown38_2_Box_NEON(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - MEMACCESS(4) - "ld1 {v30.8h}, [%4] \n" - MEMACCESS(5) - "ld1 {v31.16b}, [%5] \n" - "add %3, %3, %0 \n" - "1: \n" - - // 00 40 01 41 02 42 03 43 - // 10 50 11 51 12 52 13 53 - // 20 60 21 61 22 62 23 63 - // 30 70 31 71 32 72 33 73 - MEMACCESS(0) - "ld4 {v0.8b-v3.8b}, [%0], #32 \n" - MEMACCESS(3) - "ld4 {v4.8b-v7.8b}, [%3], #32 \n" - "subs %2, %2, #12 \n" - - // Shuffle the input data around to get align the data - // so adjacent data can be added. 0,1 - 2,3 - 4,5 - 6,7 - // 00 10 01 11 02 12 03 13 - // 40 50 41 51 42 52 43 53 - "trn1 v16.8b, v0.8b, v1.8b \n" - "trn2 v17.8b, v0.8b, v1.8b \n" - "trn1 v18.8b, v4.8b, v5.8b \n" - "trn2 v19.8b, v4.8b, v5.8b \n" - - // 20 30 21 31 22 32 23 33 - // 60 70 61 71 62 72 63 73 - "trn1 v0.8b, v2.8b, v3.8b \n" - "trn2 v1.8b, v2.8b, v3.8b \n" - "trn1 v4.8b, v6.8b, v7.8b \n" - "trn2 v5.8b, v6.8b, v7.8b \n" - - // 00+10 01+11 02+12 03+13 - // 40+50 41+51 42+52 43+53 - "uaddlp v16.4h, v16.8b \n" - "uaddlp v17.4h, v17.8b \n" - "uaddlp v18.4h, v18.8b \n" - "uaddlp v19.4h, v19.8b \n" - - // 60+70 61+71 62+72 63+73 - "uaddlp v1.4h, v1.8b \n" - "uaddlp v5.4h, v5.8b \n" - - // combine source lines - "add v16.4h, v16.4h, v18.4h \n" - "add v17.4h, v17.4h, v19.4h \n" - "add v2.4h, v1.4h, v5.4h \n" - - // dst_ptr[3] = (s[6] + s[7] + s[6+st] + s[7+st]) / 4 - "uqrshrn v2.8b, v2.8h, #2 \n" - - // Shuffle 2,3 reg around so that 2 can be added to the - // 0,1 reg and 3 can be added to the 4,5 reg. This - // requires expanding from u8 to u16 as the 0,1 and 4,5 - // registers are already expanded. Then do transposes - // to get aligned. - // xx 20 xx 30 xx 21 xx 31 xx 22 xx 32 xx 23 xx 33 - - // combine source lines - "uaddl v0.8h, v0.8b, v4.8b \n" - - // xx 20 xx 21 xx 22 xx 23 - // xx 30 xx 31 xx 32 xx 33 - "trn1 v1.8h, v0.8h, v0.8h \n" - "trn2 v4.8h, v0.8h, v0.8h \n" - "xtn v0.4h, v1.4s \n" - "xtn v4.4h, v4.4s \n" - - // 0+1+2, 3+4+5 - "add v16.8h, v16.8h, v0.8h \n" - "add v17.8h, v17.8h, v4.8h \n" - - // Need to divide, but can't downshift as the the value - // isn't a power of 2. So multiply by 65536 / n - // and take the upper 16 bits. - "sqrdmulh v0.8h, v16.8h, v30.8h \n" - "sqrdmulh v1.8h, v17.8h, v30.8h \n" - - // Align for table lookup, vtbl requires registers to - // be adjacent - - "tbl v3.16b, {v0.16b, v1.16b, v2.16b}, v31.16b \n" - - MEMACCESS(1) - "st1 {v3.8b}, [%1], #8 \n" - MEMACCESS(1) - "st1 {v3.s}[2], [%1], #4 \n" - "bgt 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(src_stride) // %3 - : "r"(&kMult38_Div6), // %4 - "r"(&kShuf38_2) // %5 - : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16", "v17", - "v18", "v19", "v30", "v31", "memory", "cc" - ); -} -#endif //HAS_SCALEROWDOWN38_NEON - -// 16x2 -> 16x1 -void ScaleFilterRows_NEON(uint8* dst_ptr, - const uint8* src_ptr, ptrdiff_t src_stride, - int dst_width, int source_y_fraction) { - int y_fraction = 256 - source_y_fraction; - asm volatile ( - "cmp %4, #0 \n" - "beq 100f \n" - "add %2, %2, %1 \n" - "cmp %4, #64 \n" - "beq 75f \n" - "cmp %4, #128 \n" - "beq 50f \n" - "cmp %4, #192 \n" - "beq 25f \n" - - "dup v5.8b, %w4 \n" - "dup v4.8b, %w5 \n" - // General purpose row blend. - "1: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v1.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "umull v6.8h, v0.8b, v4.8b \n" - "umull2 v7.8h, v0.16b, v4.16b \n" - "umlal v6.8h, v1.8b, v5.8b \n" - "umlal2 v7.8h, v1.16b, v5.16b \n" - "rshrn v0.8b, v6.8h, #8 \n" - "rshrn2 v0.16b, v7.8h, #8 \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 1b \n" - "b 99f \n" - - // Blend 25 / 75. - "25: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v1.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 25b \n" - "b 99f \n" - - // Blend 50 / 50. - "50: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v1.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 50b \n" - "b 99f \n" - - // Blend 75 / 25. - "75: \n" - MEMACCESS(1) - "ld1 {v1.16b}, [%1], #16 \n" - MEMACCESS(2) - "ld1 {v0.16b}, [%2], #16 \n" - "subs %3, %3, #16 \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - "urhadd v0.16b, v0.16b, v1.16b \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 75b \n" - "b 99f \n" - - // Blend 100 / 0 - Copy row unchanged. - "100: \n" - MEMACCESS(1) - "ld1 {v0.16b}, [%1], #16 \n" - "subs %3, %3, #16 \n" - MEMACCESS(0) - "st1 {v0.16b}, [%0], #16 \n" - "bgt 100b \n" - - "99: \n" - MEMACCESS(0) - "st1 {v0.b}[15], [%0] \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(src_stride), // %2 - "+r"(dst_width), // %3 - "+r"(source_y_fraction),// %4 - "+r"(y_fraction) // %5 - : - : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "memory", "cc" - ); -} - -#ifdef HAS_SCALEARGBROWDOWN2_NEON -void ScaleARGBRowDown2_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - "1: \n" - // load even pixels into q0, odd into q1 - MEMACCESS (0) - "ld2 {v0.4s, v1.4s}, [%0], #32 \n" - MEMACCESS (0) - "ld2 {v2.4s, v3.4s}, [%0], #32 \n" - "subs %2, %2, #8 \n" // 8 processed per loop - MEMACCESS (1) - "st1 {v1.16b}, [%1], #16 \n" // store odd pixels - MEMACCESS (1) - "st1 {v3.16b}, [%1], #16 \n" - "bgt 1b \n" - : "+r" (src_ptr), // %0 - "+r" (dst), // %1 - "+r" (dst_width) // %2 - : - : "memory", "cc", "v0", "v1", "v2", "v3" // Clobber List - ); -} -#endif //HAS_SCALEARGBROWDOWN2_NEON - -#ifdef HAS_SCALEARGBROWDOWN2_NEON -void ScaleARGBRowDown2Box_NEON(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst, int dst_width) { - asm volatile ( - // change the stride to row 2 pointer - "add %1, %1, %0 \n" - "1: \n" - MEMACCESS (0) - "ld4 {v0.16b - v3.16b}, [%0], #64 \n" // load 8 ARGB pixels. - "subs %3, %3, #8 \n" // 8 processed per loop. - "uaddlp v0.8h, v0.16b \n" // B 16 bytes -> 8 shorts. - "uaddlp v1.8h, v1.16b \n" // G 16 bytes -> 8 shorts. - "uaddlp v2.8h, v2.16b \n" // R 16 bytes -> 8 shorts. - "uaddlp v3.8h, v3.16b \n" // A 16 bytes -> 8 shorts. - MEMACCESS (1) - "ld4 {v16.16b - v19.16b}, [%1], #64 \n" // load 8 more ARGB pixels. - "uadalp v0.8h, v16.16b \n" // B 16 bytes -> 8 shorts. - "uadalp v1.8h, v17.16b \n" // G 16 bytes -> 8 shorts. - "uadalp v2.8h, v18.16b \n" // R 16 bytes -> 8 shorts. - "uadalp v3.8h, v19.16b \n" // A 16 bytes -> 8 shorts. - "rshrn v0.8b, v0.8h, #2 \n" // downshift, round and pack - "rshrn v1.8b, v1.8h, #2 \n" - "rshrn v2.8b, v2.8h, #2 \n" - "rshrn v3.8b, v3.8h, #2 \n" - MEMACCESS (2) - "st4 {v0.8b - v3.8b}, [%2], #32 \n" - "bgt 1b \n" - : "+r" (src_ptr), // %0 - "+r" (src_stride), // %1 - "+r" (dst), // %2 - "+r" (dst_width) // %3 - : - : "memory", "cc", "v0", "v1", "v2", "v3", "v16", "v17", "v18", "v19" - ); -} -#endif //HAS_SCALEARGBROWDOWN2_NEON - -#ifdef HAS_SCALEARGBROWDOWNEVEN_NEON -// Reads 4 pixels at a time. -// Alignment requirement: src_argb 4 byte aligned. -void ScaleARGBRowDownEven_NEON(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, uint8* dst_argb, int dst_width) { - asm volatile ( - "1: \n" - MEMACCESS(0) - "ld1 {v0.s}[0], [%0], %3 \n" - MEMACCESS(0) - "ld1 {v0.s}[1], [%0], %3 \n" - MEMACCESS(0) - "ld1 {v0.s}[2], [%0], %3 \n" - MEMACCESS(0) - "ld1 {v0.s}[3], [%0], %3 \n" - "subs %2, %2, #4 \n" // 4 pixels per loop. - MEMACCESS(1) - "st1 {v0.16b}, [%1], #16 \n" - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(dst_width) // %2 - : "r"(src_stepx * 4) // %3 - : "memory", "cc", "v0" - ); -} -#endif //HAS_SCALEARGBROWDOWNEVEN_NEON - -#ifdef HAS_SCALEARGBROWDOWNEVEN_NEON -// Reads 4 pixels at a time. -// Alignment requirement: src_argb 4 byte aligned. -// TODO, might be worth another optimization pass in future. -// It could be upgraded to 8 pixels at a time to start with. -void ScaleARGBRowDownEvenBox_NEON(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - asm volatile ( - "add %1, %1, %0 \n" - "1: \n" - MEMACCESS(0) - "ld1 {v0.8b}, [%0], %4 \n" // Read 4 2x2 blocks -> 2x1 - MEMACCESS(1) - "ld1 {v1.8b}, [%1], %4 \n" - MEMACCESS(0) - "ld1 {v2.8b}, [%0], %4 \n" - MEMACCESS(1) - "ld1 {v3.8b}, [%1], %4 \n" - MEMACCESS(0) - "ld1 {v4.8b}, [%0], %4 \n" - MEMACCESS(1) - "ld1 {v5.8b}, [%1], %4 \n" - MEMACCESS(0) - "ld1 {v6.8b}, [%0], %4 \n" - MEMACCESS(1) - "ld1 {v7.8b}, [%1], %4 \n" - "uaddl v0.8h, v0.8b, v1.8b \n" - "uaddl v2.8h, v2.8b, v3.8b \n" - "uaddl v4.8h, v4.8b, v5.8b \n" - "uaddl v6.8h, v6.8b, v7.8b \n" - "mov v16.d[1], v0.d[1] \n" // ab_cd -> ac_bd - "mov v0.d[1], v2.d[0] \n" - "mov v2.d[0], v16.d[1] \n" - "mov v16.d[1], v4.d[1] \n" // ef_gh -> eg_fh - "mov v4.d[1], v6.d[0] \n" - "mov v6.d[0], v16.d[1] \n" - "add v0.8h, v0.8h, v2.8h \n" // (a+b)_(c+d) - "add v4.8h, v4.8h, v6.8h \n" // (e+f)_(g+h) - "rshrn v0.8b, v0.8h, #2 \n" // first 2 pixels. - "rshrn2 v0.16b, v4.8h, #2 \n" // next 2 pixels. - "subs %3, %3, #4 \n" // 4 pixels per loop. - MEMACCESS(2) - "st1 {v0.16b}, [%2], #16 \n" - "bgt 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stride), // %1 - "+r"(dst_argb), // %2 - "+r"(dst_width) // %3 - : "r"(src_stepx * 4) // %4 - : "memory", "cc", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v16" - ); -} -#endif // HAS_SCALEARGBROWDOWNEVEN_NEON -#endif // __aarch64__ - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/scale_posix.cc b/src/main/jni/libyuv/source/scale_posix.cc deleted file mode 100644 index 352e66782..000000000 --- a/src/main/jni/libyuv/source/scale_posix.cc +++ /dev/null @@ -1,1315 +0,0 @@ -/* - * Copyright 2013 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for GCC x86 and x64. -#if !defined(LIBYUV_DISABLE_X86) && (defined(__x86_64__) || defined(__i386__)) - -// Offsets for source bytes 0 to 9 -static uvec8 kShuf0 = - { 0, 1, 3, 4, 5, 7, 8, 9, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Offsets for source bytes 11 to 20 with 8 subtracted = 3 to 12. -static uvec8 kShuf1 = - { 3, 4, 5, 7, 8, 9, 11, 12, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Offsets for source bytes 21 to 31 with 16 subtracted = 5 to 31. -static uvec8 kShuf2 = - { 5, 7, 8, 9, 11, 12, 13, 15, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Offsets for source bytes 0 to 10 -static uvec8 kShuf01 = - { 0, 1, 1, 2, 2, 3, 4, 5, 5, 6, 6, 7, 8, 9, 9, 10 }; - -// Offsets for source bytes 10 to 21 with 8 subtracted = 3 to 13. -static uvec8 kShuf11 = - { 2, 3, 4, 5, 5, 6, 6, 7, 8, 9, 9, 10, 10, 11, 12, 13 }; - -// Offsets for source bytes 21 to 31 with 16 subtracted = 5 to 31. -static uvec8 kShuf21 = - { 5, 6, 6, 7, 8, 9, 9, 10, 10, 11, 12, 13, 13, 14, 14, 15 }; - -// Coefficients for source bytes 0 to 10 -static uvec8 kMadd01 = - { 3, 1, 2, 2, 1, 3, 3, 1, 2, 2, 1, 3, 3, 1, 2, 2 }; - -// Coefficients for source bytes 10 to 21 -static uvec8 kMadd11 = - { 1, 3, 3, 1, 2, 2, 1, 3, 3, 1, 2, 2, 1, 3, 3, 1 }; - -// Coefficients for source bytes 21 to 31 -static uvec8 kMadd21 = - { 2, 2, 1, 3, 3, 1, 2, 2, 1, 3, 3, 1, 2, 2, 1, 3 }; - -// Coefficients for source bytes 21 to 31 -static vec16 kRound34 = - { 2, 2, 2, 2, 2, 2, 2, 2 }; - -static uvec8 kShuf38a = - { 0, 3, 6, 8, 11, 14, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }; - -static uvec8 kShuf38b = - { 128, 128, 128, 128, 128, 128, 0, 3, 6, 8, 11, 14, 128, 128, 128, 128 }; - -// Arrange words 0,3,6 into 0,1,2 -static uvec8 kShufAc = - { 0, 1, 6, 7, 12, 13, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Arrange words 0,3,6 into 3,4,5 -static uvec8 kShufAc3 = - { 128, 128, 128, 128, 128, 128, 0, 1, 6, 7, 12, 13, 128, 128, 128, 128 }; - -// Scaling values for boxes of 3x3 and 2x3 -static uvec16 kScaleAc33 = - { 65536 / 9, 65536 / 9, 65536 / 6, 65536 / 9, 65536 / 9, 65536 / 6, 0, 0 }; - -// Arrange first value for pixels 0,1,2,3,4,5 -static uvec8 kShufAb0 = - { 0, 128, 3, 128, 6, 128, 8, 128, 11, 128, 14, 128, 128, 128, 128, 128 }; - -// Arrange second value for pixels 0,1,2,3,4,5 -static uvec8 kShufAb1 = - { 1, 128, 4, 128, 7, 128, 9, 128, 12, 128, 15, 128, 128, 128, 128, 128 }; - -// Arrange third value for pixels 0,1,2,3,4,5 -static uvec8 kShufAb2 = - { 2, 128, 5, 128, 128, 128, 10, 128, 13, 128, 128, 128, 128, 128, 128, 128 }; - -// Scaling values for boxes of 3x2 and 2x2 -static uvec16 kScaleAb2 = - { 65536 / 3, 65536 / 3, 65536 / 2, 65536 / 3, 65536 / 3, 65536 / 2, 0, 0 }; - -// GCC versions of row functions are verbatim conversions from Visual C. -// Generated using gcc disassembly on Visual C object file: -// objdump -D yuvscaler.obj >yuvscaler.txt - -void ScaleRowDown2_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void ScaleRowDown2Linear_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10, 0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "movdqa %%xmm1,%%xmm3 \n" - "psrlw $0x8,%%xmm1 \n" - "pand %%xmm5,%%xmm2 \n" - "pand %%xmm5,%%xmm3 \n" - "pavgw %%xmm2,%%xmm0 \n" - "pavgw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void ScaleRowDown2Box_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - MEMOPREG(movdqa,0x00,0,3,1,xmm2) // movdqa (%0,%3,1),%%xmm2 - BUNDLEALIGN - MEMOPREG(movdqa,0x10,0,3,1,xmm3) // movdqa 0x10(%0,%3,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "movdqa %%xmm1,%%xmm3 \n" - "psrlw $0x8,%%xmm1 \n" - "pand %%xmm5,%%xmm2 \n" - "pand %%xmm5,%%xmm3 \n" - "pavgw %%xmm2,%%xmm0 \n" - "pavgw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)) // %3 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void ScaleRowDown2_Unaligned_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "psrlw $0x8,%%xmm0 \n" - "psrlw $0x8,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void ScaleRowDown2Linear_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "movdqa %%xmm1,%%xmm3 \n" - "psrlw $0x8,%%xmm1 \n" - "pand %%xmm5,%%xmm2 \n" - "pand %%xmm5,%%xmm3 \n" - "pavgw %%xmm2,%%xmm0 \n" - "pavgw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void ScaleRowDown2Box_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrlw $0x8,%%xmm5 \n" - - LABELALIGN - "1: \n" - "movdqu " MEMACCESS(0) ",%%xmm0 \n" - "movdqu " MEMACCESS2(0x10,0) ",%%xmm1 \n" - MEMOPREG(movdqu,0x00,0,3,1,xmm2) // movdqu (%0,%3,1),%%xmm2 - BUNDLEALIGN - MEMOPREG(movdqu,0x10,0,3,1,xmm3) // movdqu 0x10(%0,%3,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "movdqa %%xmm1,%%xmm3 \n" - "psrlw $0x8,%%xmm1 \n" - "pand %%xmm5,%%xmm2 \n" - "pand %%xmm5,%%xmm3 \n" - "pavgw %%xmm2,%%xmm0 \n" - "pavgw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqu %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "sub $0x10,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)) // %3 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm5" -#endif - ); -} - -void ScaleRowDown4_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "pcmpeqb %%xmm5,%%xmm5 \n" - "psrld $0x18,%%xmm5 \n" - "pslld $0x10,%%xmm5 \n" - - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pand %%xmm5,%%xmm0 \n" - "pand %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "psrlw $0x8,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm5" -#endif - ); -} - -void ScaleRowDown4Box_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - intptr_t stridex3 = 0; - asm volatile ( - "pcmpeqb %%xmm7,%%xmm7 \n" - "psrlw $0x8,%%xmm7 \n" - "lea " MEMLEA4(0x00,4,4,2) ",%3 \n" - - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - MEMOPREG(movdqa,0x00,0,4,1,xmm2) // movdqa (%0,%4,1),%%xmm2 - BUNDLEALIGN - MEMOPREG(movdqa,0x10,0,4,1,xmm3) // movdqa 0x10(%0,%4,1),%%xmm3 - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - MEMOPREG(movdqa,0x00,0,4,2,xmm2) // movdqa (%0,%4,2),%%xmm2 - BUNDLEALIGN - MEMOPREG(movdqa,0x10,0,4,2,xmm3) // movdqa 0x10(%0,%4,2),%%xmm3 - MEMOPREG(movdqa,0x00,0,3,1,xmm4) // movdqa (%0,%3,1),%%xmm4 - MEMOPREG(movdqa,0x10,0,3,1,xmm5) // movdqa 0x10(%0,%3,1),%%xmm5 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm4,%%xmm2 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm5,%%xmm3 \n" - "pavgb %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "movdqa %%xmm1,%%xmm3 \n" - "psrlw $0x8,%%xmm1 \n" - "pand %%xmm7,%%xmm2 \n" - "pand %%xmm7,%%xmm3 \n" - "pavgw %%xmm2,%%xmm0 \n" - "pavgw %%xmm3,%%xmm1 \n" - "packuswb %%xmm1,%%xmm0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "psrlw $0x8,%%xmm0 \n" - "pand %%xmm7,%%xmm2 \n" - "pavgw %%xmm2,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x8,1) ",%1 \n" - "sub $0x8,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width), // %2 - "+r"(stridex3) // %3 - : "r"((intptr_t)(src_stride)) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm7" -#endif - ); -} - -void ScaleRowDown34_SSSE3(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movdqa %0,%%xmm3 \n" - "movdqa %1,%%xmm4 \n" - "movdqa %2,%%xmm5 \n" - : - : "m"(kShuf0), // %0 - "m"(kShuf1), // %1 - "m"(kShuf2) // %2 - ); - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm2 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm2,%%xmm1 \n" - "palignr $0x8,%%xmm0,%%xmm1 \n" - "pshufb %%xmm3,%%xmm0 \n" - "pshufb %%xmm4,%%xmm1 \n" - "pshufb %%xmm5,%%xmm2 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "movq %%xmm1," MEMACCESS2(0x8,1) " \n" - "movq %%xmm2," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x18,1) ",%1 \n" - "sub $0x18,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5" -#endif - ); -} - -void ScaleRowDown34_1_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movdqa %0,%%xmm2 \n" // kShuf01 - "movdqa %1,%%xmm3 \n" // kShuf11 - "movdqa %2,%%xmm4 \n" // kShuf21 - : - : "m"(kShuf01), // %0 - "m"(kShuf11), // %1 - "m"(kShuf21) // %2 - ); - asm volatile ( - "movdqa %0,%%xmm5 \n" // kMadd01 - "movdqa %1,%%xmm0 \n" // kMadd11 - "movdqa %2,%%xmm1 \n" // kRound34 - : - : "m"(kMadd01), // %0 - "m"(kMadd11), // %1 - "m"(kRound34) // %2 - ); - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm6 \n" - MEMOPREG(movdqa,0x00,0,3,1,xmm7) // movdqa (%0,%3),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "pshufb %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm5,%%xmm6 \n" - "paddsw %%xmm1,%%xmm6 \n" - "psrlw $0x2,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "movq %%xmm6," MEMACCESS(1) " \n" - "movdqu " MEMACCESS2(0x8,0) ",%%xmm6 \n" - MEMOPREG(movdqu,0x8,0,3,1,xmm7) // movdqu 0x8(%0,%3),%%xmm7 - "pavgb %%xmm7,%%xmm6 \n" - "pshufb %%xmm3,%%xmm6 \n" - "pmaddubsw %%xmm0,%%xmm6 \n" - "paddsw %%xmm1,%%xmm6 \n" - "psrlw $0x2,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "movq %%xmm6," MEMACCESS2(0x8,1) " \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm6 \n" - BUNDLEALIGN - MEMOPREG(movdqa,0x10,0,3,1,xmm7) // movdqa 0x10(%0,%3),%%xmm7 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm7,%%xmm6 \n" - "pshufb %%xmm4,%%xmm6 \n" - "pmaddubsw %4,%%xmm6 \n" - "paddsw %%xmm1,%%xmm6 \n" - "psrlw $0x2,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "movq %%xmm6," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x18,1) ",%1 \n" - "sub $0x18,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)), // %3 - "m"(kMadd21) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} - -void ScaleRowDown34_0_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movdqa %0,%%xmm2 \n" // kShuf01 - "movdqa %1,%%xmm3 \n" // kShuf11 - "movdqa %2,%%xmm4 \n" // kShuf21 - : - : "m"(kShuf01), // %0 - "m"(kShuf11), // %1 - "m"(kShuf21) // %2 - ); - asm volatile ( - "movdqa %0,%%xmm5 \n" // kMadd01 - "movdqa %1,%%xmm0 \n" // kMadd11 - "movdqa %2,%%xmm1 \n" // kRound34 - : - : "m"(kMadd01), // %0 - "m"(kMadd11), // %1 - "m"(kRound34) // %2 - ); - - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm6 \n" - MEMOPREG(movdqa,0x00,0,3,1,xmm7) // movdqa (%0,%3,1),%%xmm7 - "pavgb %%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm6 \n" - "pshufb %%xmm2,%%xmm6 \n" - "pmaddubsw %%xmm5,%%xmm6 \n" - "paddsw %%xmm1,%%xmm6 \n" - "psrlw $0x2,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "movq %%xmm6," MEMACCESS(1) " \n" - "movdqu " MEMACCESS2(0x8,0) ",%%xmm6 \n" - MEMOPREG(movdqu,0x8,0,3,1,xmm7) // movdqu 0x8(%0,%3,1),%%xmm7 - "pavgb %%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm6 \n" - "pshufb %%xmm3,%%xmm6 \n" - "pmaddubsw %%xmm0,%%xmm6 \n" - "paddsw %%xmm1,%%xmm6 \n" - "psrlw $0x2,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "movq %%xmm6," MEMACCESS2(0x8,1) " \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm6 \n" - MEMOPREG(movdqa,0x10,0,3,1,xmm7) // movdqa 0x10(%0,%3,1),%%xmm7 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm6,%%xmm7 \n" - "pavgb %%xmm7,%%xmm6 \n" - "pshufb %%xmm4,%%xmm6 \n" - "pmaddubsw %4,%%xmm6 \n" - "paddsw %%xmm1,%%xmm6 \n" - "psrlw $0x2,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "movq %%xmm6," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x18,1) ",%1 \n" - "sub $0x18,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)), // %3 - "m"(kMadd21) // %4 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} - -void ScaleRowDown38_SSSE3(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movdqa %3,%%xmm4 \n" - "movdqa %4,%%xmm5 \n" - - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "pshufb %%xmm4,%%xmm0 \n" - "pshufb %%xmm5,%%xmm1 \n" - "paddusb %%xmm1,%%xmm0 \n" - "movq %%xmm0," MEMACCESS(1) " \n" - "movhlps %%xmm0,%%xmm1 \n" - "movd %%xmm1," MEMACCESS2(0x8,1) " \n" - "lea " MEMLEA(0xc,1) ",%1 \n" - "sub $0xc,%2 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "m"(kShuf38a), // %3 - "m"(kShuf38b) // %4 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm4", "xmm5" -#endif - ); -} - -void ScaleRowDown38_2_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movdqa %0,%%xmm2 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm4 \n" - "movdqa %3,%%xmm5 \n" - : - : "m"(kShufAb0), // %0 - "m"(kShufAb1), // %1 - "m"(kShufAb2), // %2 - "m"(kScaleAb2) // %3 - ); - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(pavgb,0x00,0,3,1,xmm0) // pavgb (%0,%3,1),%%xmm0 - "lea " MEMLEA(0x10,0) ",%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "pshufb %%xmm2,%%xmm1 \n" - "movdqa %%xmm0,%%xmm6 \n" - "pshufb %%xmm3,%%xmm6 \n" - "paddusw %%xmm6,%%xmm1 \n" - "pshufb %%xmm4,%%xmm0 \n" - "paddusw %%xmm0,%%xmm1 \n" - "pmulhuw %%xmm5,%%xmm1 \n" - "packuswb %%xmm1,%%xmm1 \n" - "sub $0x6,%2 \n" - "movd %%xmm1," MEMACCESS(1) " \n" - "psrlq $0x10,%%xmm1 \n" - "movd %%xmm1," MEMACCESS2(0x2,1) " \n" - "lea " MEMLEA(0x6,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)) // %3 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -void ScaleRowDown38_3_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - asm volatile ( - "movdqa %0,%%xmm2 \n" - "movdqa %1,%%xmm3 \n" - "movdqa %2,%%xmm4 \n" - "pxor %%xmm5,%%xmm5 \n" - : - : "m"(kShufAc), // %0 - "m"(kShufAc3), // %1 - "m"(kScaleAc33) // %2 - ); - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movdqa,0x00,0,3,1,xmm6) // movdqa (%0,%3,1),%%xmm6 - "movhlps %%xmm0,%%xmm1 \n" - "movhlps %%xmm6,%%xmm7 \n" - "punpcklbw %%xmm5,%%xmm0 \n" - "punpcklbw %%xmm5,%%xmm1 \n" - "punpcklbw %%xmm5,%%xmm6 \n" - "punpcklbw %%xmm5,%%xmm7 \n" - "paddusw %%xmm6,%%xmm0 \n" - "paddusw %%xmm7,%%xmm1 \n" - MEMOPREG(movdqa,0x00,0,3,2,xmm6) // movdqa (%0,%3,2),%%xmm6 - "lea " MEMLEA(0x10,0) ",%0 \n" - "movhlps %%xmm6,%%xmm7 \n" - "punpcklbw %%xmm5,%%xmm6 \n" - "punpcklbw %%xmm5,%%xmm7 \n" - "paddusw %%xmm6,%%xmm0 \n" - "paddusw %%xmm7,%%xmm1 \n" - "movdqa %%xmm0,%%xmm6 \n" - "psrldq $0x2,%%xmm0 \n" - "paddusw %%xmm0,%%xmm6 \n" - "psrldq $0x2,%%xmm0 \n" - "paddusw %%xmm0,%%xmm6 \n" - "pshufb %%xmm2,%%xmm6 \n" - "movdqa %%xmm1,%%xmm7 \n" - "psrldq $0x2,%%xmm1 \n" - "paddusw %%xmm1,%%xmm7 \n" - "psrldq $0x2,%%xmm1 \n" - "paddusw %%xmm1,%%xmm7 \n" - "pshufb %%xmm3,%%xmm7 \n" - "paddusw %%xmm7,%%xmm6 \n" - "pmulhuw %%xmm4,%%xmm6 \n" - "packuswb %%xmm6,%%xmm6 \n" - "sub $0x6,%2 \n" - "movd %%xmm6," MEMACCESS(1) " \n" - "psrlq $0x10,%%xmm6 \n" - "movd %%xmm6," MEMACCESS2(0x2,1) " \n" - "lea " MEMLEA(0x6,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)) // %3 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7" -#endif - ); -} - -void ScaleAddRows_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int src_width, int src_height) { - int tmp_height = 0; - intptr_t tmp_src = 0; - asm volatile ( - "pxor %%xmm4,%%xmm4 \n" - "sub $0x1,%5 \n" - - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "mov %0,%3 \n" - "add %6,%0 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm4,%%xmm0 \n" - "punpckhbw %%xmm4,%%xmm1 \n" - "mov %5,%2 \n" - "test %2,%2 \n" - "je 3f \n" - - LABELALIGN - "2: \n" - "movdqa " MEMACCESS(0) ",%%xmm2 \n" - "add %6,%0 \n" - "movdqa %%xmm2,%%xmm3 \n" - "punpcklbw %%xmm4,%%xmm2 \n" - "punpckhbw %%xmm4,%%xmm3 \n" - "paddusw %%xmm2,%%xmm0 \n" - "paddusw %%xmm3,%%xmm1 \n" - "sub $0x1,%2 \n" - "jg 2b \n" - - LABELALIGN - "3: \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,1) " \n" - "lea " MEMLEA(0x10,3) ",%0 \n" - "lea " MEMLEA(0x20,1) ",%1 \n" - "sub $0x10,%4 \n" - "jg 1b \n" - : "+r"(src_ptr), // %0 - "+r"(dst_ptr), // %1 - "+r"(tmp_height), // %2 - "+r"(tmp_src), // %3 - "+r"(src_width), // %4 - "+rm"(src_height) // %5 - : "rm"((intptr_t)(src_stride)) // %6 - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4" -#endif - ); -} - -// Bilinear column filtering. SSSE3 version. -void ScaleFilterCols_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - intptr_t x0 = 0, x1 = 0, temp_pixel = 0; - asm volatile ( - "movd %6,%%xmm2 \n" - "movd %7,%%xmm3 \n" - "movl $0x04040000,%k2 \n" - "movd %k2,%%xmm5 \n" - "pcmpeqb %%xmm6,%%xmm6 \n" - "psrlw $0x9,%%xmm6 \n" - "pextrw $0x1,%%xmm2,%k3 \n" - "subl $0x2,%5 \n" - "jl 29f \n" - "movdqa %%xmm2,%%xmm0 \n" - "paddd %%xmm3,%%xmm0 \n" - "punpckldq %%xmm0,%%xmm2 \n" - "punpckldq %%xmm3,%%xmm3 \n" - "paddd %%xmm3,%%xmm3 \n" - "pextrw $0x3,%%xmm2,%k4 \n" - - LABELALIGN - "2: \n" - "movdqa %%xmm2,%%xmm1 \n" - "paddd %%xmm3,%%xmm2 \n" - MEMOPARG(movzwl,0x00,1,3,1,k2) // movzwl (%1,%3,1),%k2 - "movd %k2,%%xmm0 \n" - "psrlw $0x9,%%xmm1 \n" - BUNDLEALIGN - MEMOPARG(movzwl,0x00,1,4,1,k2) // movzwl (%1,%4,1),%k2 - "movd %k2,%%xmm4 \n" - "pshufb %%xmm5,%%xmm1 \n" - "punpcklwd %%xmm4,%%xmm0 \n" - "pxor %%xmm6,%%xmm1 \n" - "pmaddubsw %%xmm1,%%xmm0 \n" - "pextrw $0x1,%%xmm2,%k3 \n" - "pextrw $0x3,%%xmm2,%k4 \n" - "psrlw $0x7,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movd %%xmm0,%k2 \n" - "mov %w2," MEMACCESS(0) " \n" - "lea " MEMLEA(0x2,0) ",%0 \n" - "sub $0x2,%5 \n" - "jge 2b \n" - - LABELALIGN - "29: \n" - "addl $0x1,%5 \n" - "jl 99f \n" - MEMOPARG(movzwl,0x00,1,3,1,k2) // movzwl (%1,%3,1),%k2 - "movd %k2,%%xmm0 \n" - "psrlw $0x9,%%xmm2 \n" - "pshufb %%xmm5,%%xmm2 \n" - "pxor %%xmm6,%%xmm2 \n" - "pmaddubsw %%xmm2,%%xmm0 \n" - "psrlw $0x7,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movd %%xmm0,%k2 \n" - "mov %b2," MEMACCESS(0) " \n" - "99: \n" - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+a"(temp_pixel), // %2 - "+r"(x0), // %3 - "+r"(x1), // %4 - "+rm"(dst_width) // %5 - : "rm"(x), // %6 - "rm"(dx) // %7 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -// Reads 4 pixels, duplicates them and writes 8 pixels. -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -void ScaleColsUp2_SSE2(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpcklbw %%xmm0,%%xmm0 \n" - "punpckhbw %%xmm1,%%xmm1 \n" - "sub $0x20,%2 \n" - "movdqa %%xmm0," MEMACCESS(0) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,0) " \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "jg 1b \n" - - : "+r"(dst_ptr), // %0 - "+r"(src_ptr), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void ScaleARGBRowDown2_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "shufps $0xdd,%%xmm1,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void ScaleARGBRowDown2Linear_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "movdqa %%xmm0,%%xmm2 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm2 \n" - "pavgb %%xmm2,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -void ScaleARGBRowDown2Box_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(0) ",%%xmm0 \n" - "movdqa " MEMACCESS2(0x10,0) ",%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movdqa,0x00,0,3,1,xmm2) // movdqa (%0,%3,1),%%xmm2 - MEMOPREG(movdqa,0x10,0,3,1,xmm3) // movdqa 0x10(%0,%3,1),%%xmm3 - "lea " MEMLEA(0x20,0) ",%0 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm2 \n" - "pavgb %%xmm2,%%xmm0 \n" - "sub $0x4,%2 \n" - "movdqa %%xmm0," MEMACCESS(1) " \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(dst_argb), // %1 - "+r"(dst_width) // %2 - : "r"((intptr_t)(src_stride)) // %3 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3" -#endif - ); -} - -// Reads 4 pixels at a time. -// Alignment requirement: dst_argb 16 byte aligned. -void ScaleARGBRowDownEven_SSE2(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - intptr_t src_stepx_x4 = (intptr_t)(src_stepx); - intptr_t src_stepx_x12 = 0; - asm volatile ( - "lea " MEMLEA3(0x00,1,4) ",%1 \n" - "lea " MEMLEA4(0x00,1,1,2) ",%4 \n" - LABELALIGN - "1: \n" - "movd " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movd,0x00,0,1,1,xmm1) // movd (%0,%1,1),%%xmm1 - "punpckldq %%xmm1,%%xmm0 \n" - BUNDLEALIGN - MEMOPREG(movd,0x00,0,1,2,xmm2) // movd (%0,%1,2),%%xmm2 - MEMOPREG(movd,0x00,0,4,1,xmm3) // movd (%0,%4,1),%%xmm3 - "lea " MEMLEA4(0x00,0,1,4) ",%0 \n" - "punpckldq %%xmm3,%%xmm2 \n" - "punpcklqdq %%xmm2,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stepx_x4), // %1 - "+r"(dst_argb), // %2 - "+r"(dst_width), // %3 - "+r"(src_stepx_x12) // %4 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3" -#endif - ); -} - -// Blends four 2x2 to 4x1. -// Alignment requirement: dst_argb 16 byte aligned. -void ScaleARGBRowDownEvenBox_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, int src_stepx, - uint8* dst_argb, int dst_width) { - intptr_t src_stepx_x4 = (intptr_t)(src_stepx); - intptr_t src_stepx_x12 = 0; - intptr_t row1 = (intptr_t)(src_stride); - asm volatile ( - "lea " MEMLEA3(0x00,1,4) ",%1 \n" - "lea " MEMLEA4(0x00,1,1,2) ",%4 \n" - "lea " MEMLEA4(0x00,0,5,1) ",%5 \n" - - LABELALIGN - "1: \n" - "movq " MEMACCESS(0) ",%%xmm0 \n" - MEMOPREG(movhps,0x00,0,1,1,xmm0) // movhps (%0,%1,1),%%xmm0 - MEMOPREG(movq,0x00,0,1,2,xmm1) // movq (%0,%1,2),%%xmm1 - BUNDLEALIGN - MEMOPREG(movhps,0x00,0,4,1,xmm1) // movhps (%0,%4,1),%%xmm1 - "lea " MEMLEA4(0x00,0,1,4) ",%0 \n" - "movq " MEMACCESS(5) ",%%xmm2 \n" - BUNDLEALIGN - MEMOPREG(movhps,0x00,5,1,1,xmm2) // movhps (%5,%1,1),%%xmm2 - MEMOPREG(movq,0x00,5,1,2,xmm3) // movq (%5,%1,2),%%xmm3 - MEMOPREG(movhps,0x00,5,4,1,xmm3) // movhps (%5,%4,1),%%xmm3 - "lea " MEMLEA4(0x00,5,1,4) ",%5 \n" - "pavgb %%xmm2,%%xmm0 \n" - "pavgb %%xmm3,%%xmm1 \n" - "movdqa %%xmm0,%%xmm2 \n" - "shufps $0x88,%%xmm1,%%xmm0 \n" - "shufps $0xdd,%%xmm1,%%xmm2 \n" - "pavgb %%xmm2,%%xmm0 \n" - "sub $0x4,%3 \n" - "movdqa %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jg 1b \n" - : "+r"(src_argb), // %0 - "+r"(src_stepx_x4), // %1 - "+r"(dst_argb), // %2 - "+rm"(dst_width), // %3 - "+r"(src_stepx_x12), // %4 - "+r"(row1) // %5 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3" -#endif - ); -} - -void ScaleARGBCols_SSE2(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - intptr_t x0 = 0, x1 = 0; - asm volatile ( - "movd %5,%%xmm2 \n" - "movd %6,%%xmm3 \n" - "pshufd $0x0,%%xmm2,%%xmm2 \n" - "pshufd $0x11,%%xmm3,%%xmm0 \n" - "paddd %%xmm0,%%xmm2 \n" - "paddd %%xmm3,%%xmm3 \n" - "pshufd $0x5,%%xmm3,%%xmm0 \n" - "paddd %%xmm0,%%xmm2 \n" - "paddd %%xmm3,%%xmm3 \n" - "pshufd $0x0,%%xmm3,%%xmm3 \n" - "pextrw $0x1,%%xmm2,%k0 \n" - "pextrw $0x3,%%xmm2,%k1 \n" - "cmp $0x0,%4 \n" - "jl 99f \n" - "sub $0x4,%4 \n" - "jl 49f \n" - - LABELALIGN - "40: \n" - MEMOPREG(movd,0x00,3,0,4,xmm0) // movd (%3,%0,4),%%xmm0 - MEMOPREG(movd,0x00,3,1,4,xmm1) // movd (%3,%1,4),%%xmm1 - "pextrw $0x5,%%xmm2,%k0 \n" - "pextrw $0x7,%%xmm2,%k1 \n" - "paddd %%xmm3,%%xmm2 \n" - "punpckldq %%xmm1,%%xmm0 \n" - MEMOPREG(movd,0x00,3,0,4,xmm1) // movd (%3,%0,4),%%xmm1 - MEMOPREG(movd,0x00,3,1,4,xmm4) // movd (%3,%1,4),%%xmm4 - "pextrw $0x1,%%xmm2,%k0 \n" - "pextrw $0x3,%%xmm2,%k1 \n" - "punpckldq %%xmm4,%%xmm1 \n" - "punpcklqdq %%xmm1,%%xmm0 \n" - "sub $0x4,%4 \n" - "movdqu %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x10,2) ",%2 \n" - "jge 40b \n" - - "49: \n" - "test $0x2,%4 \n" - "je 29f \n" - BUNDLEALIGN - MEMOPREG(movd,0x00,3,0,4,xmm0) // movd (%3,%0,4),%%xmm0 - MEMOPREG(movd,0x00,3,1,4,xmm1) // movd (%3,%1,4),%%xmm1 - "pextrw $0x5,%%xmm2,%k0 \n" - "punpckldq %%xmm1,%%xmm0 \n" - "movq %%xmm0," MEMACCESS(2) " \n" - "lea " MEMLEA(0x8,2) ",%2 \n" - "29: \n" - "test $0x1,%4 \n" - "je 99f \n" - MEMOPREG(movd,0x00,3,0,4,xmm0) // movd (%3,%0,4),%%xmm0 - "movd %%xmm0," MEMACCESS(2) " \n" - "99: \n" - : "+a"(x0), // %0 - "+d"(x1), // %1 - "+r"(dst_argb), // %2 - "+r"(src_argb), // %3 - "+r"(dst_width) // %4 - : "rm"(x), // %5 - "rm"(dx) // %6 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4" -#endif - ); -} - -// Reads 4 pixels, duplicates them and writes 8 pixels. -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -void ScaleARGBColsUp2_SSE2(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - asm volatile ( - LABELALIGN - "1: \n" - "movdqa " MEMACCESS(1) ",%%xmm0 \n" - "lea " MEMLEA(0x10,1) ",%1 \n" - "movdqa %%xmm0,%%xmm1 \n" - "punpckldq %%xmm0,%%xmm0 \n" - "punpckhdq %%xmm1,%%xmm1 \n" - "sub $0x8,%2 \n" - "movdqa %%xmm0," MEMACCESS(0) " \n" - "movdqa %%xmm1," MEMACCESS2(0x10,0) " \n" - "lea " MEMLEA(0x20,0) ",%0 \n" - "jg 1b \n" - - : "+r"(dst_argb), // %0 - "+r"(src_argb), // %1 - "+r"(dst_width) // %2 - : - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1" -#endif - ); -} - -// Shuffle table for arranging 2 pixels into pairs for pmaddubsw -static uvec8 kShuffleColARGB = { - 0u, 4u, 1u, 5u, 2u, 6u, 3u, 7u, // bbggrraa 1st pixel - 8u, 12u, 9u, 13u, 10u, 14u, 11u, 15u // bbggrraa 2nd pixel -}; - -// Shuffle table for duplicating 2 fractions into 8 bytes each -static uvec8 kShuffleFractions = { - 0u, 0u, 0u, 0u, 0u, 0u, 0u, 0u, 4u, 4u, 4u, 4u, 4u, 4u, 4u, 4u, -}; - -// Bilinear row filtering combines 4x2 -> 4x1. SSSE3 version -void ScaleARGBFilterCols_SSSE3(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - intptr_t x0 = 0, x1 = 0; - asm volatile ( - "movdqa %0,%%xmm4 \n" - "movdqa %1,%%xmm5 \n" - : - : "m"(kShuffleColARGB), // %0 - "m"(kShuffleFractions) // %1 - ); - - asm volatile ( - "movd %5,%%xmm2 \n" - "movd %6,%%xmm3 \n" - "pcmpeqb %%xmm6,%%xmm6 \n" - "psrlw $0x9,%%xmm6 \n" - "pextrw $0x1,%%xmm2,%k3 \n" - "sub $0x2,%2 \n" - "jl 29f \n" - "movdqa %%xmm2,%%xmm0 \n" - "paddd %%xmm3,%%xmm0 \n" - "punpckldq %%xmm0,%%xmm2 \n" - "punpckldq %%xmm3,%%xmm3 \n" - "paddd %%xmm3,%%xmm3 \n" - "pextrw $0x3,%%xmm2,%k4 \n" - - LABELALIGN - "2: \n" - "movdqa %%xmm2,%%xmm1 \n" - "paddd %%xmm3,%%xmm2 \n" - MEMOPREG(movq,0x00,1,3,4,xmm0) // movq (%1,%3,4),%%xmm0 - "psrlw $0x9,%%xmm1 \n" - BUNDLEALIGN - MEMOPREG(movhps,0x00,1,4,4,xmm0) // movhps (%1,%4,4),%%xmm0 - "pshufb %%xmm5,%%xmm1 \n" - "pshufb %%xmm4,%%xmm0 \n" - "pxor %%xmm6,%%xmm1 \n" - "pmaddubsw %%xmm1,%%xmm0 \n" - "psrlw $0x7,%%xmm0 \n" - "pextrw $0x1,%%xmm2,%k3 \n" - "pextrw $0x3,%%xmm2,%k4 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movq %%xmm0," MEMACCESS(0) " \n" - "lea " MEMLEA(0x8,0) ",%0 \n" - "sub $0x2,%2 \n" - "jge 2b \n" - - LABELALIGN - "29: \n" - "add $0x1,%2 \n" - "jl 99f \n" - "psrlw $0x9,%%xmm2 \n" - BUNDLEALIGN - MEMOPREG(movq,0x00,1,3,4,xmm0) // movq (%1,%3,4),%%xmm0 - "pshufb %%xmm5,%%xmm2 \n" - "pshufb %%xmm4,%%xmm0 \n" - "pxor %%xmm6,%%xmm2 \n" - "pmaddubsw %%xmm2,%%xmm0 \n" - "psrlw $0x7,%%xmm0 \n" - "packuswb %%xmm0,%%xmm0 \n" - "movd %%xmm0," MEMACCESS(0) " \n" - - LABELALIGN - "99: \n" - : "+r"(dst_argb), // %0 - "+r"(src_argb), // %1 - "+rm"(dst_width), // %2 - "+r"(x0), // %3 - "+r"(x1) // %4 - : "rm"(x), // %5 - "rm"(dx) // %6 - : "memory", "cc" -#if defined(__native_client__) && defined(__x86_64__) - , "r14" -#endif -#if defined(__SSE2__) - , "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6" -#endif - ); -} - -// Divide num by div and return as 16.16 fixed point result. -int FixedDiv_X86(int num, int div) { - asm volatile ( - "cdq \n" - "shld $0x10,%%eax,%%edx \n" - "shl $0x10,%%eax \n" - "idiv %1 \n" - "mov %0, %%eax \n" - : "+a"(num) // %0 - : "c"(div) // %1 - : "memory", "cc", "edx" - ); - return num; -} - -// Divide num - 1 by div - 1 and return as 16.16 fixed point result. -int FixedDiv1_X86(int num, int div) { - asm volatile ( - "cdq \n" - "shld $0x10,%%eax,%%edx \n" - "shl $0x10,%%eax \n" - "sub $0x10001,%%eax \n" - "sbb $0x0,%%edx \n" - "sub $0x1,%1 \n" - "idiv %1 \n" - "mov %0, %%eax \n" - : "+a"(num) // %0 - : "c"(div) // %1 - : "memory", "cc", "edx" - ); - return num; -} - -#endif // defined(__x86_64__) || defined(__i386__) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/scale_win.cc b/src/main/jni/libyuv/source/scale_win.cc deleted file mode 100644 index 840b9738d..000000000 --- a/src/main/jni/libyuv/source/scale_win.cc +++ /dev/null @@ -1,1320 +0,0 @@ -/* - * Copyright 2013 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - -#include "libyuv/row.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -// This module is for Visual C x86. -#if !defined(LIBYUV_DISABLE_X86) && defined(_M_IX86) && defined(_MSC_VER) - -// Offsets for source bytes 0 to 9 -static uvec8 kShuf0 = - { 0, 1, 3, 4, 5, 7, 8, 9, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Offsets for source bytes 11 to 20 with 8 subtracted = 3 to 12. -static uvec8 kShuf1 = - { 3, 4, 5, 7, 8, 9, 11, 12, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Offsets for source bytes 21 to 31 with 16 subtracted = 5 to 31. -static uvec8 kShuf2 = - { 5, 7, 8, 9, 11, 12, 13, 15, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Offsets for source bytes 0 to 10 -static uvec8 kShuf01 = - { 0, 1, 1, 2, 2, 3, 4, 5, 5, 6, 6, 7, 8, 9, 9, 10 }; - -// Offsets for source bytes 10 to 21 with 8 subtracted = 3 to 13. -static uvec8 kShuf11 = - { 2, 3, 4, 5, 5, 6, 6, 7, 8, 9, 9, 10, 10, 11, 12, 13 }; - -// Offsets for source bytes 21 to 31 with 16 subtracted = 5 to 31. -static uvec8 kShuf21 = - { 5, 6, 6, 7, 8, 9, 9, 10, 10, 11, 12, 13, 13, 14, 14, 15 }; - -// Coefficients for source bytes 0 to 10 -static uvec8 kMadd01 = - { 3, 1, 2, 2, 1, 3, 3, 1, 2, 2, 1, 3, 3, 1, 2, 2 }; - -// Coefficients for source bytes 10 to 21 -static uvec8 kMadd11 = - { 1, 3, 3, 1, 2, 2, 1, 3, 3, 1, 2, 2, 1, 3, 3, 1 }; - -// Coefficients for source bytes 21 to 31 -static uvec8 kMadd21 = - { 2, 2, 1, 3, 3, 1, 2, 2, 1, 3, 3, 1, 2, 2, 1, 3 }; - -// Coefficients for source bytes 21 to 31 -static vec16 kRound34 = - { 2, 2, 2, 2, 2, 2, 2, 2 }; - -static uvec8 kShuf38a = - { 0, 3, 6, 8, 11, 14, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }; - -static uvec8 kShuf38b = - { 128, 128, 128, 128, 128, 128, 0, 3, 6, 8, 11, 14, 128, 128, 128, 128 }; - -// Arrange words 0,3,6 into 0,1,2 -static uvec8 kShufAc = - { 0, 1, 6, 7, 12, 13, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 }; - -// Arrange words 0,3,6 into 3,4,5 -static uvec8 kShufAc3 = - { 128, 128, 128, 128, 128, 128, 0, 1, 6, 7, 12, 13, 128, 128, 128, 128 }; - -// Scaling values for boxes of 3x3 and 2x3 -static uvec16 kScaleAc33 = - { 65536 / 9, 65536 / 9, 65536 / 6, 65536 / 9, 65536 / 9, 65536 / 6, 0, 0 }; - -// Arrange first value for pixels 0,1,2,3,4,5 -static uvec8 kShufAb0 = - { 0, 128, 3, 128, 6, 128, 8, 128, 11, 128, 14, 128, 128, 128, 128, 128 }; - -// Arrange second value for pixels 0,1,2,3,4,5 -static uvec8 kShufAb1 = - { 1, 128, 4, 128, 7, 128, 9, 128, 12, 128, 15, 128, 128, 128, 128, 128 }; - -// Arrange third value for pixels 0,1,2,3,4,5 -static uvec8 kShufAb2 = - { 2, 128, 5, 128, 128, 128, 10, 128, 13, 128, 128, 128, 128, 128, 128, 128 }; - -// Scaling values for boxes of 3x2 and 2x2 -static uvec16 kScaleAb2 = - { 65536 / 3, 65536 / 3, 65536 / 2, 65536 / 3, 65536 / 3, 65536 / 2, 0, 0 }; - -// Reads 32 pixels, throws half away and writes 16 pixels. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown2_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride ignored - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - psrlw xmm0, 8 // isolate odd pixels. - psrlw xmm1, 8 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - ret - } -} - -// Blends 32x1 rectangle to 16x1. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown2Linear_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - - movdqa xmm2, xmm0 // average columns (32 to 16 pixels) - psrlw xmm0, 8 - movdqa xmm3, xmm1 - psrlw xmm1, 8 - pand xmm2, xmm5 - pand xmm3, xmm5 - pavgw xmm0, xmm2 - pavgw xmm1, xmm3 - packuswb xmm0, xmm1 - - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - ret - } -} - -// Blends 32x2 rectangle to 16x1. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown2Box_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_ptr - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_ptr - mov ecx, [esp + 4 + 16] // dst_width - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + esi] - movdqa xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 // average rows - pavgb xmm1, xmm3 - - movdqa xmm2, xmm0 // average columns (32 to 16 pixels) - psrlw xmm0, 8 - movdqa xmm3, xmm1 - psrlw xmm1, 8 - pand xmm2, xmm5 - pand xmm3, xmm5 - pavgw xmm0, xmm2 - pavgw xmm1, xmm3 - packuswb xmm0, xmm1 - - sub ecx, 16 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - pop esi - ret - } -} - -// Reads 32 pixels, throws half away and writes 16 pixels. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown2_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride ignored - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - - align 4 - wloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - psrlw xmm0, 8 // isolate odd pixels. - psrlw xmm1, 8 - packuswb xmm0, xmm1 - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - ret - } -} - -// Blends 32x1 rectangle to 16x1. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown2Linear_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - - align 4 - wloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - lea eax, [eax + 32] - - movdqa xmm2, xmm0 // average columns (32 to 16 pixels) - psrlw xmm0, 8 - movdqa xmm3, xmm1 - psrlw xmm1, 8 - pand xmm2, xmm5 - pand xmm3, xmm5 - pavgw xmm0, xmm2 - pavgw xmm1, xmm3 - packuswb xmm0, xmm1 - - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - ret - } -} - -// Blends 32x2 rectangle to 16x1. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown2Box_Unaligned_SSE2(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_ptr - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_ptr - mov ecx, [esp + 4 + 16] // dst_width - pcmpeqb xmm5, xmm5 // generate mask 0x00ff00ff - psrlw xmm5, 8 - - align 4 - wloop: - movdqu xmm0, [eax] - movdqu xmm1, [eax + 16] - movdqu xmm2, [eax + esi] - movdqu xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 // average rows - pavgb xmm1, xmm3 - - movdqa xmm2, xmm0 // average columns (32 to 16 pixels) - psrlw xmm0, 8 - movdqa xmm3, xmm1 - psrlw xmm1, 8 - pand xmm2, xmm5 - pand xmm3, xmm5 - pavgw xmm0, xmm2 - pavgw xmm1, xmm3 - packuswb xmm0, xmm1 - - sub ecx, 16 - movdqu [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - pop esi - ret - } -} - -// Point samples 32 pixels to 8 pixels. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 8 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown4_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride ignored - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - pcmpeqb xmm5, xmm5 // generate mask 0x00ff0000 - psrld xmm5, 24 - pslld xmm5, 16 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - pand xmm0, xmm5 - pand xmm1, xmm5 - packuswb xmm0, xmm1 - psrlw xmm0, 8 - packuswb xmm0, xmm0 - sub ecx, 8 - movq qword ptr [edx], xmm0 - lea edx, [edx + 8] - jg wloop - - ret - } -} - -// Blends 32x4 rectangle to 8x1. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 8 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown4Box_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - push edi - mov eax, [esp + 8 + 4] // src_ptr - mov esi, [esp + 8 + 8] // src_stride - mov edx, [esp + 8 + 12] // dst_ptr - mov ecx, [esp + 8 + 16] // dst_width - lea edi, [esi + esi * 2] // src_stride * 3 - pcmpeqb xmm7, xmm7 // generate mask 0x00ff00ff - psrlw xmm7, 8 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + esi] - movdqa xmm3, [eax + esi + 16] - pavgb xmm0, xmm2 // average rows - pavgb xmm1, xmm3 - movdqa xmm2, [eax + esi * 2] - movdqa xmm3, [eax + esi * 2 + 16] - movdqa xmm4, [eax + edi] - movdqa xmm5, [eax + edi + 16] - lea eax, [eax + 32] - pavgb xmm2, xmm4 - pavgb xmm3, xmm5 - pavgb xmm0, xmm2 - pavgb xmm1, xmm3 - - movdqa xmm2, xmm0 // average columns (32 to 16 pixels) - psrlw xmm0, 8 - movdqa xmm3, xmm1 - psrlw xmm1, 8 - pand xmm2, xmm7 - pand xmm3, xmm7 - pavgw xmm0, xmm2 - pavgw xmm1, xmm3 - packuswb xmm0, xmm1 - - movdqa xmm2, xmm0 // average columns (16 to 8 pixels) - psrlw xmm0, 8 - pand xmm2, xmm7 - pavgw xmm0, xmm2 - packuswb xmm0, xmm0 - - sub ecx, 8 - movq qword ptr [edx], xmm0 - lea edx, [edx + 8] - jg wloop - - pop edi - pop esi - ret - } -} - -// Point samples 32 pixels to 24 pixels. -// Produces three 8 byte values. For each 8 bytes, 16 bytes are read. -// Then shuffled to do the scaling. - -// Note that movdqa+palign may be better than movdqu. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 8 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown34_SSSE3(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride ignored - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - movdqa xmm3, kShuf0 - movdqa xmm4, kShuf1 - movdqa xmm5, kShuf2 - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - movdqa xmm2, xmm1 - palignr xmm1, xmm0, 8 - pshufb xmm0, xmm3 - pshufb xmm1, xmm4 - pshufb xmm2, xmm5 - movq qword ptr [edx], xmm0 - movq qword ptr [edx + 8], xmm1 - movq qword ptr [edx + 16], xmm2 - lea edx, [edx + 24] - sub ecx, 24 - jg wloop - - ret - } -} - -// Blends 32x2 rectangle to 24x1 -// Produces three 8 byte values. For each 8 bytes, 16 bytes are read. -// Then shuffled to do the scaling. - -// Register usage: -// xmm0 src_row 0 -// xmm1 src_row 1 -// xmm2 shuf 0 -// xmm3 shuf 1 -// xmm4 shuf 2 -// xmm5 madd 0 -// xmm6 madd 1 -// xmm7 kRound34 - -// Note that movdqa+palign may be better than movdqu. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 8 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown34_1_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_ptr - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_ptr - mov ecx, [esp + 4 + 16] // dst_width - movdqa xmm2, kShuf01 - movdqa xmm3, kShuf11 - movdqa xmm4, kShuf21 - movdqa xmm5, kMadd01 - movdqa xmm6, kMadd11 - movdqa xmm7, kRound34 - - align 4 - wloop: - movdqa xmm0, [eax] // pixels 0..7 - movdqa xmm1, [eax + esi] - pavgb xmm0, xmm1 - pshufb xmm0, xmm2 - pmaddubsw xmm0, xmm5 - paddsw xmm0, xmm7 - psrlw xmm0, 2 - packuswb xmm0, xmm0 - movq qword ptr [edx], xmm0 - movdqu xmm0, [eax + 8] // pixels 8..15 - movdqu xmm1, [eax + esi + 8] - pavgb xmm0, xmm1 - pshufb xmm0, xmm3 - pmaddubsw xmm0, xmm6 - paddsw xmm0, xmm7 - psrlw xmm0, 2 - packuswb xmm0, xmm0 - movq qword ptr [edx + 8], xmm0 - movdqa xmm0, [eax + 16] // pixels 16..23 - movdqa xmm1, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm1 - pshufb xmm0, xmm4 - movdqa xmm1, kMadd21 - pmaddubsw xmm0, xmm1 - paddsw xmm0, xmm7 - psrlw xmm0, 2 - packuswb xmm0, xmm0 - sub ecx, 24 - movq qword ptr [edx + 16], xmm0 - lea edx, [edx + 24] - jg wloop - - pop esi - ret - } -} - -// Note that movdqa+palign may be better than movdqu. -// Alignment requirement: src_ptr 16 byte aligned, dst_ptr 8 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleRowDown34_0_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_ptr - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_ptr - mov ecx, [esp + 4 + 16] // dst_width - movdqa xmm2, kShuf01 - movdqa xmm3, kShuf11 - movdqa xmm4, kShuf21 - movdqa xmm5, kMadd01 - movdqa xmm6, kMadd11 - movdqa xmm7, kRound34 - - align 4 - wloop: - movdqa xmm0, [eax] // pixels 0..7 - movdqa xmm1, [eax + esi] - pavgb xmm1, xmm0 - pavgb xmm0, xmm1 - pshufb xmm0, xmm2 - pmaddubsw xmm0, xmm5 - paddsw xmm0, xmm7 - psrlw xmm0, 2 - packuswb xmm0, xmm0 - movq qword ptr [edx], xmm0 - movdqu xmm0, [eax + 8] // pixels 8..15 - movdqu xmm1, [eax + esi + 8] - pavgb xmm1, xmm0 - pavgb xmm0, xmm1 - pshufb xmm0, xmm3 - pmaddubsw xmm0, xmm6 - paddsw xmm0, xmm7 - psrlw xmm0, 2 - packuswb xmm0, xmm0 - movq qword ptr [edx + 8], xmm0 - movdqa xmm0, [eax + 16] // pixels 16..23 - movdqa xmm1, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm1, xmm0 - pavgb xmm0, xmm1 - pshufb xmm0, xmm4 - movdqa xmm1, kMadd21 - pmaddubsw xmm0, xmm1 - paddsw xmm0, xmm7 - psrlw xmm0, 2 - packuswb xmm0, xmm0 - sub ecx, 24 - movq qword ptr [edx + 16], xmm0 - lea edx, [edx+24] - jg wloop - - pop esi - ret - } -} - -// 3/8 point sampler - -// Scale 32 pixels to 12 -__declspec(naked) __declspec(align(16)) -void ScaleRowDown38_SSSE3(const uint8* src_ptr, ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - mov eax, [esp + 4] // src_ptr - // src_stride ignored - mov edx, [esp + 12] // dst_ptr - mov ecx, [esp + 16] // dst_width - movdqa xmm4, kShuf38a - movdqa xmm5, kShuf38b - - align 4 - xloop: - movdqa xmm0, [eax] // 16 pixels -> 0,1,2,3,4,5 - movdqa xmm1, [eax + 16] // 16 pixels -> 6,7,8,9,10,11 - lea eax, [eax + 32] - pshufb xmm0, xmm4 - pshufb xmm1, xmm5 - paddusb xmm0, xmm1 - - sub ecx, 12 - movq qword ptr [edx], xmm0 // write 12 pixels - movhlps xmm1, xmm0 - movd [edx + 8], xmm1 - lea edx, [edx + 12] - jg xloop - - ret - } -} - -// Scale 16x3 pixels to 6x1 with interpolation -__declspec(naked) __declspec(align(16)) -void ScaleRowDown38_3_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_ptr - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_ptr - mov ecx, [esp + 4 + 16] // dst_width - movdqa xmm2, kShufAc - movdqa xmm3, kShufAc3 - movdqa xmm4, kScaleAc33 - pxor xmm5, xmm5 - - align 4 - xloop: - movdqa xmm0, [eax] // sum up 3 rows into xmm0/1 - movdqa xmm6, [eax + esi] - movhlps xmm1, xmm0 - movhlps xmm7, xmm6 - punpcklbw xmm0, xmm5 - punpcklbw xmm1, xmm5 - punpcklbw xmm6, xmm5 - punpcklbw xmm7, xmm5 - paddusw xmm0, xmm6 - paddusw xmm1, xmm7 - movdqa xmm6, [eax + esi * 2] - lea eax, [eax + 16] - movhlps xmm7, xmm6 - punpcklbw xmm6, xmm5 - punpcklbw xmm7, xmm5 - paddusw xmm0, xmm6 - paddusw xmm1, xmm7 - - movdqa xmm6, xmm0 // 8 pixels -> 0,1,2 of xmm6 - psrldq xmm0, 2 - paddusw xmm6, xmm0 - psrldq xmm0, 2 - paddusw xmm6, xmm0 - pshufb xmm6, xmm2 - - movdqa xmm7, xmm1 // 8 pixels -> 3,4,5 of xmm6 - psrldq xmm1, 2 - paddusw xmm7, xmm1 - psrldq xmm1, 2 - paddusw xmm7, xmm1 - pshufb xmm7, xmm3 - paddusw xmm6, xmm7 - - pmulhuw xmm6, xmm4 // divide by 9,9,6, 9,9,6 - packuswb xmm6, xmm6 - - sub ecx, 6 - movd [edx], xmm6 // write 6 pixels - psrlq xmm6, 16 - movd [edx + 2], xmm6 - lea edx, [edx + 6] - jg xloop - - pop esi - ret - } -} - -// Scale 16x2 pixels to 6x1 with interpolation -__declspec(naked) __declspec(align(16)) -void ScaleRowDown38_2_Box_SSSE3(const uint8* src_ptr, - ptrdiff_t src_stride, - uint8* dst_ptr, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_ptr - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_ptr - mov ecx, [esp + 4 + 16] // dst_width - movdqa xmm2, kShufAb0 - movdqa xmm3, kShufAb1 - movdqa xmm4, kShufAb2 - movdqa xmm5, kScaleAb2 - - align 4 - xloop: - movdqa xmm0, [eax] // average 2 rows into xmm0 - pavgb xmm0, [eax + esi] - lea eax, [eax + 16] - - movdqa xmm1, xmm0 // 16 pixels -> 0,1,2,3,4,5 of xmm1 - pshufb xmm1, xmm2 - movdqa xmm6, xmm0 - pshufb xmm6, xmm3 - paddusw xmm1, xmm6 - pshufb xmm0, xmm4 - paddusw xmm1, xmm0 - - pmulhuw xmm1, xmm5 // divide by 3,3,2, 3,3,2 - packuswb xmm1, xmm1 - - sub ecx, 6 - movd [edx], xmm1 // write 6 pixels - psrlq xmm1, 16 - movd [edx + 2], xmm1 - lea edx, [edx + 6] - jg xloop - - pop esi - ret - } -} - -// Reads 16xN bytes and produces 16 shorts at a time. -// TODO(fbarchard): Make this handle 4xN bytes for any width ARGB. -__declspec(naked) __declspec(align(16)) -void ScaleAddRows_SSE2(const uint8* src_ptr, ptrdiff_t src_stride, - uint16* dst_ptr, int src_width, - int src_height) { - __asm { - push esi - push edi - push ebx - push ebp - mov esi, [esp + 16 + 4] // src_ptr - mov edx, [esp + 16 + 8] // src_stride - mov edi, [esp + 16 + 12] // dst_ptr - mov ecx, [esp + 16 + 16] // dst_width - mov ebx, [esp + 16 + 20] // height - pxor xmm4, xmm4 - dec ebx - - align 4 - xloop: - // first row - movdqa xmm0, [esi] - lea eax, [esi + edx] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm4 - punpckhbw xmm1, xmm4 - lea esi, [esi + 16] - mov ebp, ebx - test ebp, ebp - je ydone - - // sum remaining rows - align 4 - yloop: - movdqa xmm2, [eax] // read 16 pixels - lea eax, [eax + edx] // advance to next row - movdqa xmm3, xmm2 - punpcklbw xmm2, xmm4 - punpckhbw xmm3, xmm4 - paddusw xmm0, xmm2 // sum 16 words - paddusw xmm1, xmm3 - sub ebp, 1 - jg yloop - - align 4 - ydone: - movdqa [edi], xmm0 - movdqa [edi + 16], xmm1 - lea edi, [edi + 32] - - sub ecx, 16 - jg xloop - - pop ebp - pop ebx - pop edi - pop esi - ret - } -} - -// Bilinear column filtering. SSSE3 version. -// TODO(fbarchard): Port to Neon -// TODO(fbarchard): Switch the following: -// xor ebx, ebx -// mov bx, word ptr [esi + eax] // 2 source x0 pixels -// To -// movzx ebx, word ptr [esi + eax] // 2 source x0 pixels -// when drmemory bug fixed. -// https://code.google.com/p/drmemory/issues/detail?id=1396 - -__declspec(naked) __declspec(align(16)) -void ScaleFilterCols_SSSE3(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - __asm { - push ebx - push esi - push edi - mov edi, [esp + 12 + 4] // dst_ptr - mov esi, [esp + 12 + 8] // src_ptr - mov ecx, [esp + 12 + 12] // dst_width - movd xmm2, [esp + 12 + 16] // x - movd xmm3, [esp + 12 + 20] // dx - mov eax, 0x04040000 // shuffle to line up fractions with pixel. - movd xmm5, eax - pcmpeqb xmm6, xmm6 // generate 0x007f for inverting fraction. - psrlw xmm6, 9 - pextrw eax, xmm2, 1 // get x0 integer. preroll - sub ecx, 2 - jl xloop29 - - movdqa xmm0, xmm2 // x1 = x0 + dx - paddd xmm0, xmm3 - punpckldq xmm2, xmm0 // x0 x1 - punpckldq xmm3, xmm3 // dx dx - paddd xmm3, xmm3 // dx * 2, dx * 2 - pextrw edx, xmm2, 3 // get x1 integer. preroll - - // 2 Pixel loop. - align 4 - xloop2: - movdqa xmm1, xmm2 // x0, x1 fractions. - paddd xmm2, xmm3 // x += dx - movzx ebx, word ptr [esi + eax] // 2 source x0 pixels - movd xmm0, ebx - psrlw xmm1, 9 // 7 bit fractions. - movzx ebx, word ptr [esi + edx] // 2 source x1 pixels - movd xmm4, ebx - pshufb xmm1, xmm5 // 0011 - punpcklwd xmm0, xmm4 - pxor xmm1, xmm6 // 0..7f and 7f..0 - pmaddubsw xmm0, xmm1 // 16 bit, 2 pixels. - pextrw eax, xmm2, 1 // get x0 integer. next iteration. - pextrw edx, xmm2, 3 // get x1 integer. next iteration. - psrlw xmm0, 7 // 8.7 fixed point to low 8 bits. - packuswb xmm0, xmm0 // 8 bits, 2 pixels. - movd ebx, xmm0 - mov [edi], bx - lea edi, [edi + 2] - sub ecx, 2 // 2 pixels - jge xloop2 - - align 4 - xloop29: - - add ecx, 2 - 1 - jl xloop99 - - // 1 pixel remainder - movzx ebx, word ptr [esi + eax] // 2 source x0 pixels - movd xmm0, ebx - psrlw xmm2, 9 // 7 bit fractions. - pshufb xmm2, xmm5 // 0011 - pxor xmm2, xmm6 // 0..7f and 7f..0 - pmaddubsw xmm0, xmm2 // 16 bit - psrlw xmm0, 7 // 8.7 fixed point to low 8 bits. - packuswb xmm0, xmm0 // 8 bits - movd ebx, xmm0 - mov [edi], bl - - align 4 - xloop99: - - pop edi - pop esi - pop ebx - ret - } -} - -// Reads 16 pixels, duplicates them and writes 32 pixels. -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleColsUp2_SSE2(uint8* dst_ptr, const uint8* src_ptr, - int dst_width, int x, int dx) { - __asm { - mov edx, [esp + 4] // dst_ptr - mov eax, [esp + 8] // src_ptr - mov ecx, [esp + 12] // dst_width - - align 4 - wloop: - movdqa xmm0, [eax] - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpcklbw xmm0, xmm0 - punpckhbw xmm1, xmm1 - sub ecx, 32 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - jg wloop - - ret - } -} - -// Reads 8 pixels, throws half away and writes 4 even pixels (0, 2, 4, 6) -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleARGBRowDown2_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - __asm { - mov eax, [esp + 4] // src_argb - // src_stride ignored - mov edx, [esp + 12] // dst_argb - mov ecx, [esp + 16] // dst_width - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - shufps xmm0, xmm1, 0xdd - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - ret - } -} - -// Blends 8x1 rectangle to 4x1. -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleARGBRowDown2Linear_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - __asm { - mov eax, [esp + 4] // src_argb - // src_stride ignored - mov edx, [esp + 12] // dst_argb - mov ecx, [esp + 16] // dst_width - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - lea eax, [eax + 32] - movdqa xmm2, xmm0 - shufps xmm0, xmm1, 0x88 // even pixels - shufps xmm2, xmm1, 0xdd // odd pixels - pavgb xmm0, xmm2 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - ret - } -} - -// Blends 8x2 rectangle to 4x1. -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleARGBRowDown2Box_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - uint8* dst_argb, int dst_width) { - __asm { - push esi - mov eax, [esp + 4 + 4] // src_argb - mov esi, [esp + 4 + 8] // src_stride - mov edx, [esp + 4 + 12] // dst_argb - mov ecx, [esp + 4 + 16] // dst_width - - align 4 - wloop: - movdqa xmm0, [eax] - movdqa xmm1, [eax + 16] - movdqa xmm2, [eax + esi] - movdqa xmm3, [eax + esi + 16] - lea eax, [eax + 32] - pavgb xmm0, xmm2 // average rows - pavgb xmm1, xmm3 - movdqa xmm2, xmm0 // average columns (8 to 4 pixels) - shufps xmm0, xmm1, 0x88 // even pixels - shufps xmm2, xmm1, 0xdd // odd pixels - pavgb xmm0, xmm2 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - pop esi - ret - } -} - -// Reads 4 pixels at a time. -// Alignment requirement: dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleARGBRowDownEven_SSE2(const uint8* src_argb, ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - __asm { - push ebx - push edi - mov eax, [esp + 8 + 4] // src_argb - // src_stride ignored - mov ebx, [esp + 8 + 12] // src_stepx - mov edx, [esp + 8 + 16] // dst_argb - mov ecx, [esp + 8 + 20] // dst_width - lea ebx, [ebx * 4] - lea edi, [ebx + ebx * 2] - - align 4 - wloop: - movd xmm0, [eax] - movd xmm1, [eax + ebx] - punpckldq xmm0, xmm1 - movd xmm2, [eax + ebx * 2] - movd xmm3, [eax + edi] - lea eax, [eax + ebx * 4] - punpckldq xmm2, xmm3 - punpcklqdq xmm0, xmm2 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - pop edi - pop ebx - ret - } -} - -// Blends four 2x2 to 4x1. -// Alignment requirement: dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleARGBRowDownEvenBox_SSE2(const uint8* src_argb, - ptrdiff_t src_stride, - int src_stepx, - uint8* dst_argb, int dst_width) { - __asm { - push ebx - push esi - push edi - mov eax, [esp + 12 + 4] // src_argb - mov esi, [esp + 12 + 8] // src_stride - mov ebx, [esp + 12 + 12] // src_stepx - mov edx, [esp + 12 + 16] // dst_argb - mov ecx, [esp + 12 + 20] // dst_width - lea esi, [eax + esi] // row1 pointer - lea ebx, [ebx * 4] - lea edi, [ebx + ebx * 2] - - align 4 - wloop: - movq xmm0, qword ptr [eax] // row0 4 pairs - movhps xmm0, qword ptr [eax + ebx] - movq xmm1, qword ptr [eax + ebx * 2] - movhps xmm1, qword ptr [eax + edi] - lea eax, [eax + ebx * 4] - movq xmm2, qword ptr [esi] // row1 4 pairs - movhps xmm2, qword ptr [esi + ebx] - movq xmm3, qword ptr [esi + ebx * 2] - movhps xmm3, qword ptr [esi + edi] - lea esi, [esi + ebx * 4] - pavgb xmm0, xmm2 // average rows - pavgb xmm1, xmm3 - movdqa xmm2, xmm0 // average columns (8 to 4 pixels) - shufps xmm0, xmm1, 0x88 // even pixels - shufps xmm2, xmm1, 0xdd // odd pixels - pavgb xmm0, xmm2 - sub ecx, 4 - movdqa [edx], xmm0 - lea edx, [edx + 16] - jg wloop - - pop edi - pop esi - pop ebx - ret - } -} - -// Column scaling unfiltered. SSE2 version. -__declspec(naked) __declspec(align(16)) -void ScaleARGBCols_SSE2(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - __asm { - push edi - push esi - mov edi, [esp + 8 + 4] // dst_argb - mov esi, [esp + 8 + 8] // src_argb - mov ecx, [esp + 8 + 12] // dst_width - movd xmm2, [esp + 8 + 16] // x - movd xmm3, [esp + 8 + 20] // dx - - pshufd xmm2, xmm2, 0 // x0 x0 x0 x0 - pshufd xmm0, xmm3, 0x11 // dx 0 dx 0 - paddd xmm2, xmm0 - paddd xmm3, xmm3 // 0, 0, 0, dx * 2 - pshufd xmm0, xmm3, 0x05 // dx * 2, dx * 2, 0, 0 - paddd xmm2, xmm0 // x3 x2 x1 x0 - paddd xmm3, xmm3 // 0, 0, 0, dx * 4 - pshufd xmm3, xmm3, 0 // dx * 4, dx * 4, dx * 4, dx * 4 - - pextrw eax, xmm2, 1 // get x0 integer. - pextrw edx, xmm2, 3 // get x1 integer. - - cmp ecx, 0 - jle xloop99 - sub ecx, 4 - jl xloop49 - - // 4 Pixel loop. - align 4 - xloop4: - movd xmm0, [esi + eax * 4] // 1 source x0 pixels - movd xmm1, [esi + edx * 4] // 1 source x1 pixels - pextrw eax, xmm2, 5 // get x2 integer. - pextrw edx, xmm2, 7 // get x3 integer. - paddd xmm2, xmm3 // x += dx - punpckldq xmm0, xmm1 // x0 x1 - - movd xmm1, [esi + eax * 4] // 1 source x2 pixels - movd xmm4, [esi + edx * 4] // 1 source x3 pixels - pextrw eax, xmm2, 1 // get x0 integer. next iteration. - pextrw edx, xmm2, 3 // get x1 integer. next iteration. - punpckldq xmm1, xmm4 // x2 x3 - punpcklqdq xmm0, xmm1 // x0 x1 x2 x3 - sub ecx, 4 // 4 pixels - movdqu [edi], xmm0 - lea edi, [edi + 16] - jge xloop4 - - align 4 - xloop49: - test ecx, 2 - je xloop29 - - // 2 Pixels. - movd xmm0, [esi + eax * 4] // 1 source x0 pixels - movd xmm1, [esi + edx * 4] // 1 source x1 pixels - pextrw eax, xmm2, 5 // get x2 integer. - punpckldq xmm0, xmm1 // x0 x1 - - movq qword ptr [edi], xmm0 - lea edi, [edi + 8] - - xloop29: - test ecx, 1 - je xloop99 - - // 1 Pixels. - movd xmm0, [esi + eax * 4] // 1 source x2 pixels - movd dword ptr [edi], xmm0 - align 4 - xloop99: - - pop esi - pop edi - ret - } -} - -// Bilinear row filtering combines 2x1 -> 1x1. SSSE3 version. -// TODO(fbarchard): Port to Neon - -// Shuffle table for arranging 2 pixels into pairs for pmaddubsw -static uvec8 kShuffleColARGB = { - 0u, 4u, 1u, 5u, 2u, 6u, 3u, 7u, // bbggrraa 1st pixel - 8u, 12u, 9u, 13u, 10u, 14u, 11u, 15u // bbggrraa 2nd pixel -}; - -// Shuffle table for duplicating 2 fractions into 8 bytes each -static uvec8 kShuffleFractions = { - 0u, 0u, 0u, 0u, 0u, 0u, 0u, 0u, 4u, 4u, 4u, 4u, 4u, 4u, 4u, 4u, -}; - -__declspec(naked) __declspec(align(16)) -void ScaleARGBFilterCols_SSSE3(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - __asm { - push esi - push edi - mov edi, [esp + 8 + 4] // dst_argb - mov esi, [esp + 8 + 8] // src_argb - mov ecx, [esp + 8 + 12] // dst_width - movd xmm2, [esp + 8 + 16] // x - movd xmm3, [esp + 8 + 20] // dx - movdqa xmm4, kShuffleColARGB - movdqa xmm5, kShuffleFractions - pcmpeqb xmm6, xmm6 // generate 0x007f for inverting fraction. - psrlw xmm6, 9 - pextrw eax, xmm2, 1 // get x0 integer. preroll - sub ecx, 2 - jl xloop29 - - movdqa xmm0, xmm2 // x1 = x0 + dx - paddd xmm0, xmm3 - punpckldq xmm2, xmm0 // x0 x1 - punpckldq xmm3, xmm3 // dx dx - paddd xmm3, xmm3 // dx * 2, dx * 2 - pextrw edx, xmm2, 3 // get x1 integer. preroll - - // 2 Pixel loop. - align 4 - xloop2: - movdqa xmm1, xmm2 // x0, x1 fractions. - paddd xmm2, xmm3 // x += dx - movq xmm0, qword ptr [esi + eax * 4] // 2 source x0 pixels - psrlw xmm1, 9 // 7 bit fractions. - movhps xmm0, qword ptr [esi + edx * 4] // 2 source x1 pixels - pshufb xmm1, xmm5 // 0000000011111111 - pshufb xmm0, xmm4 // arrange pixels into pairs - pxor xmm1, xmm6 // 0..7f and 7f..0 - pmaddubsw xmm0, xmm1 // argb_argb 16 bit, 2 pixels. - pextrw eax, xmm2, 1 // get x0 integer. next iteration. - pextrw edx, xmm2, 3 // get x1 integer. next iteration. - psrlw xmm0, 7 // argb 8.7 fixed point to low 8 bits. - packuswb xmm0, xmm0 // argb_argb 8 bits, 2 pixels. - movq qword ptr [edi], xmm0 - lea edi, [edi + 8] - sub ecx, 2 // 2 pixels - jge xloop2 - - align 4 - xloop29: - - add ecx, 2 - 1 - jl xloop99 - - // 1 pixel remainder - psrlw xmm2, 9 // 7 bit fractions. - movq xmm0, qword ptr [esi + eax * 4] // 2 source x0 pixels - pshufb xmm2, xmm5 // 00000000 - pshufb xmm0, xmm4 // arrange pixels into pairs - pxor xmm2, xmm6 // 0..7f and 7f..0 - pmaddubsw xmm0, xmm2 // argb 16 bit, 1 pixel. - psrlw xmm0, 7 - packuswb xmm0, xmm0 // argb 8 bits, 1 pixel. - movd [edi], xmm0 - - align 4 - xloop99: - - pop edi - pop esi - ret - } -} - -// Reads 4 pixels, duplicates them and writes 8 pixels. -// Alignment requirement: src_argb 16 byte aligned, dst_argb 16 byte aligned. -__declspec(naked) __declspec(align(16)) -void ScaleARGBColsUp2_SSE2(uint8* dst_argb, const uint8* src_argb, - int dst_width, int x, int dx) { - __asm { - mov edx, [esp + 4] // dst_argb - mov eax, [esp + 8] // src_argb - mov ecx, [esp + 12] // dst_width - - align 4 - wloop: - movdqa xmm0, [eax] - lea eax, [eax + 16] - movdqa xmm1, xmm0 - punpckldq xmm0, xmm0 - punpckhdq xmm1, xmm1 - sub ecx, 8 - movdqa [edx], xmm0 - movdqa [edx + 16], xmm1 - lea edx, [edx + 32] - jg wloop - - ret - } -} - -// Divide num by div and return as 16.16 fixed point result. -__declspec(naked) __declspec(align(16)) -int FixedDiv_X86(int num, int div) { - __asm { - mov eax, [esp + 4] // num - cdq // extend num to 64 bits - shld edx, eax, 16 // 32.16 - shl eax, 16 - idiv dword ptr [esp + 8] - ret - } -} - -// Divide num by div and return as 16.16 fixed point result. -__declspec(naked) __declspec(align(16)) -int FixedDiv1_X86(int num, int div) { - __asm { - mov eax, [esp + 4] // num - mov ecx, [esp + 8] // denom - cdq // extend num to 64 bits - shld edx, eax, 16 // 32.16 - shl eax, 16 - sub eax, 0x00010001 - sbb edx, 0 - sub ecx, 1 - idiv ecx - ret - } -} - -#endif // !defined(LIBYUV_DISABLE_X86) && defined(_M_IX86) && defined(_MSC_VER) - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif diff --git a/src/main/jni/libyuv/source/video_common.cc b/src/main/jni/libyuv/source/video_common.cc deleted file mode 100644 index efbedf46e..000000000 --- a/src/main/jni/libyuv/source/video_common.cc +++ /dev/null @@ -1,64 +0,0 @@ -/* - * Copyright 2011 The LibYuv Project Authors. All rights reserved. - * - * Use of this source code is governed by a BSD-style license - * that can be found in the LICENSE file in the root of the source - * tree. An additional intellectual property rights grant can be found - * in the file PATENTS. All contributing project authors may - * be found in the AUTHORS file in the root of the source tree. - */ - - -#include "libyuv/video_common.h" - -#ifdef __cplusplus -namespace libyuv { -extern "C" { -#endif - -#define ARRAY_SIZE(x) (int)(sizeof(x) / sizeof(x[0])) - -struct FourCCAliasEntry { - uint32 alias; - uint32 canonical; -}; - -static const struct FourCCAliasEntry kFourCCAliases[] = { - {FOURCC_IYUV, FOURCC_I420}, - {FOURCC_YU16, FOURCC_I422}, - {FOURCC_YU24, FOURCC_I444}, - {FOURCC_YUYV, FOURCC_YUY2}, - {FOURCC_YUVS, FOURCC_YUY2}, // kCMPixelFormat_422YpCbCr8_yuvs - {FOURCC_HDYC, FOURCC_UYVY}, - {FOURCC_2VUY, FOURCC_UYVY}, // kCMPixelFormat_422YpCbCr8 - {FOURCC_JPEG, FOURCC_MJPG}, // Note: JPEG has DHT while MJPG does not. - {FOURCC_DMB1, FOURCC_MJPG}, - {FOURCC_BA81, FOURCC_BGGR}, - {FOURCC_RGB3, FOURCC_RAW }, - {FOURCC_BGR3, FOURCC_24BG}, - {FOURCC_CM32, FOURCC_BGRA}, // kCMPixelFormat_32ARGB - {FOURCC_CM24, FOURCC_RAW }, // kCMPixelFormat_24RGB - {FOURCC_L555, FOURCC_RGBO}, // kCMPixelFormat_16LE555 - {FOURCC_L565, FOURCC_RGBP}, // kCMPixelFormat_16LE565 - {FOURCC_5551, FOURCC_RGBO}, // kCMPixelFormat_16LE5551 -}; -// TODO(fbarchard): Consider mapping kCMPixelFormat_32BGRA to FOURCC_ARGB. -// {FOURCC_BGRA, FOURCC_ARGB}, // kCMPixelFormat_32BGRA - -LIBYUV_API -uint32 CanonicalFourCC(uint32 fourcc) { - int i; - for (i = 0; i < ARRAY_SIZE(kFourCCAliases); ++i) { - if (kFourCCAliases[i].alias == fourcc) { - return kFourCCAliases[i].canonical; - } - } - // Not an alias, so return it as-is. - return fourcc; -} - -#ifdef __cplusplus -} // extern "C" -} // namespace libyuv -#endif - diff --git a/src/main/jni/libyuv/source/x86inc.asm b/src/main/jni/libyuv/source/x86inc.asm deleted file mode 100644 index cb5c32df3..000000000 --- a/src/main/jni/libyuv/source/x86inc.asm +++ /dev/null @@ -1,1136 +0,0 @@ -;***************************************************************************** -;* x86inc.asm: x264asm abstraction layer -;***************************************************************************** -;* Copyright (C) 2005-2012 x264 project -;* -;* Authors: Loren Merritt <lorenm@u.washington.edu> -;* Anton Mitrofanov <BugMaster@narod.ru> -;* Jason Garrett-Glaser <darkshikari@gmail.com> -;* Henrik Gramner <hengar-6@student.ltu.se> -;* -;* Permission to use, copy, modify, and/or distribute this software for any -;* purpose with or without fee is hereby granted, provided that the above -;* copyright notice and this permission notice appear in all copies. -;* -;* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES -;* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF -;* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR -;* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES -;* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN -;* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF -;* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. -;***************************************************************************** - -; This is a header file for the x264ASM assembly language, which uses -; NASM/YASM syntax combined with a large number of macros to provide easy -; abstraction between different calling conventions (x86_32, win64, linux64). -; It also has various other useful features to simplify writing the kind of -; DSP functions that are most often used in x264. - -; Unlike the rest of x264, this file is available under an ISC license, as it -; has significant usefulness outside of x264 and we want it to be available -; to the largest audience possible. Of course, if you modify it for your own -; purposes to add a new feature, we strongly encourage contributing a patch -; as this feature might be useful for others as well. Send patches or ideas -; to x264-devel@videolan.org . - -; Local changes for libyuv: -; remove %define program_name and references in labels -; rename cpus to uppercase - -%define WIN64 0 -%define UNIX64 0 -%if ARCH_X86_64 - %ifidn __OUTPUT_FORMAT__,win32 - %define WIN64 1 - %elifidn __OUTPUT_FORMAT__,win64 - %define WIN64 1 - %else - %define UNIX64 1 - %endif -%endif - -%ifdef PREFIX - %define mangle(x) _ %+ x -%else - %define mangle(x) x -%endif - -; Name of the .rodata section. -; Kludge: Something on OS X fails to align .rodata even given an align attribute, -; so use a different read-only section. -%macro SECTION_RODATA 0-1 16 - %ifidn __OUTPUT_FORMAT__,macho64 - SECTION .text align=%1 - %elifidn __OUTPUT_FORMAT__,macho - SECTION .text align=%1 - fakegot: - %elifidn __OUTPUT_FORMAT__,aout - section .text - %else - SECTION .rodata align=%1 - %endif -%endmacro - -; aout does not support align= -%macro SECTION_TEXT 0-1 16 - %ifidn __OUTPUT_FORMAT__,aout - SECTION .text - %else - SECTION .text align=%1 - %endif -%endmacro - -%if WIN64 - %define PIC -%elif ARCH_X86_64 == 0 -; x86_32 doesn't require PIC. -; Some distros prefer shared objects to be PIC, but nothing breaks if -; the code contains a few textrels, so we'll skip that complexity. - %undef PIC -%endif -%ifdef PIC - default rel -%endif - -; Always use long nops (reduces 0x90 spam in disassembly on x86_32) -CPU amdnop - -; Macros to eliminate most code duplication between x86_32 and x86_64: -; Currently this works only for leaf functions which load all their arguments -; into registers at the start, and make no other use of the stack. Luckily that -; covers most of x264's asm. - -; PROLOGUE: -; %1 = number of arguments. loads them from stack if needed. -; %2 = number of registers used. pushes callee-saved regs if needed. -; %3 = number of xmm registers used. pushes callee-saved xmm regs if needed. -; %4 = list of names to define to registers -; PROLOGUE can also be invoked by adding the same options to cglobal - -; e.g. -; cglobal foo, 2,3,0, dst, src, tmp -; declares a function (foo), taking two args (dst and src) and one local variable (tmp) - -; TODO Some functions can use some args directly from the stack. If they're the -; last args then you can just not declare them, but if they're in the middle -; we need more flexible macro. - -; RET: -; Pops anything that was pushed by PROLOGUE, and returns. - -; REP_RET: -; Same, but if it doesn't pop anything it becomes a 2-byte ret, for athlons -; which are slow when a normal ret follows a branch. - -; registers: -; rN and rNq are the native-size register holding function argument N -; rNd, rNw, rNb are dword, word, and byte size -; rNh is the high 8 bits of the word size -; rNm is the original location of arg N (a register or on the stack), dword -; rNmp is native size - -%macro DECLARE_REG 2-3 - %define r%1q %2 - %define r%1d %2d - %define r%1w %2w - %define r%1b %2b - %define r%1h %2h - %if %0 == 2 - %define r%1m %2d - %define r%1mp %2 - %elif ARCH_X86_64 ; memory - %define r%1m [rsp + stack_offset + %3] - %define r%1mp qword r %+ %1m - %else - %define r%1m [esp + stack_offset + %3] - %define r%1mp dword r %+ %1m - %endif - %define r%1 %2 -%endmacro - -%macro DECLARE_REG_SIZE 3 - %define r%1q r%1 - %define e%1q r%1 - %define r%1d e%1 - %define e%1d e%1 - %define r%1w %1 - %define e%1w %1 - %define r%1h %3 - %define e%1h %3 - %define r%1b %2 - %define e%1b %2 -%if ARCH_X86_64 == 0 - %define r%1 e%1 -%endif -%endmacro - -DECLARE_REG_SIZE ax, al, ah -DECLARE_REG_SIZE bx, bl, bh -DECLARE_REG_SIZE cx, cl, ch -DECLARE_REG_SIZE dx, dl, dh -DECLARE_REG_SIZE si, sil, null -DECLARE_REG_SIZE di, dil, null -DECLARE_REG_SIZE bp, bpl, null - -; t# defines for when per-arch register allocation is more complex than just function arguments - -%macro DECLARE_REG_TMP 1-* - %assign %%i 0 - %rep %0 - CAT_XDEFINE t, %%i, r%1 - %assign %%i %%i+1 - %rotate 1 - %endrep -%endmacro - -%macro DECLARE_REG_TMP_SIZE 0-* - %rep %0 - %define t%1q t%1 %+ q - %define t%1d t%1 %+ d - %define t%1w t%1 %+ w - %define t%1h t%1 %+ h - %define t%1b t%1 %+ b - %rotate 1 - %endrep -%endmacro - -DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14 - -%if ARCH_X86_64 - %define gprsize 8 -%else - %define gprsize 4 -%endif - -%macro PUSH 1 - push %1 - %assign stack_offset stack_offset+gprsize -%endmacro - -%macro POP 1 - pop %1 - %assign stack_offset stack_offset-gprsize -%endmacro - -%macro PUSH_IF_USED 1-* - %rep %0 - %if %1 < regs_used - PUSH r%1 - %endif - %rotate 1 - %endrep -%endmacro - -%macro POP_IF_USED 1-* - %rep %0 - %if %1 < regs_used - pop r%1 - %endif - %rotate 1 - %endrep -%endmacro - -%macro LOAD_IF_USED 1-* - %rep %0 - %if %1 < num_args - mov r%1, r %+ %1 %+ mp - %endif - %rotate 1 - %endrep -%endmacro - -%macro SUB 2 - sub %1, %2 - %ifidn %1, rsp - %assign stack_offset stack_offset+(%2) - %endif -%endmacro - -%macro ADD 2 - add %1, %2 - %ifidn %1, rsp - %assign stack_offset stack_offset-(%2) - %endif -%endmacro - -%macro movifnidn 2 - %ifnidn %1, %2 - mov %1, %2 - %endif -%endmacro - -%macro movsxdifnidn 2 - %ifnidn %1, %2 - movsxd %1, %2 - %endif -%endmacro - -%macro ASSERT 1 - %if (%1) == 0 - %error assert failed - %endif -%endmacro - -%macro DEFINE_ARGS 0-* - %ifdef n_arg_names - %assign %%i 0 - %rep n_arg_names - CAT_UNDEF arg_name %+ %%i, q - CAT_UNDEF arg_name %+ %%i, d - CAT_UNDEF arg_name %+ %%i, w - CAT_UNDEF arg_name %+ %%i, h - CAT_UNDEF arg_name %+ %%i, b - CAT_UNDEF arg_name %+ %%i, m - CAT_UNDEF arg_name %+ %%i, mp - CAT_UNDEF arg_name, %%i - %assign %%i %%i+1 - %endrep - %endif - - %xdefine %%stack_offset stack_offset - %undef stack_offset ; so that the current value of stack_offset doesn't get baked in by xdefine - %assign %%i 0 - %rep %0 - %xdefine %1q r %+ %%i %+ q - %xdefine %1d r %+ %%i %+ d - %xdefine %1w r %+ %%i %+ w - %xdefine %1h r %+ %%i %+ h - %xdefine %1b r %+ %%i %+ b - %xdefine %1m r %+ %%i %+ m - %xdefine %1mp r %+ %%i %+ mp - CAT_XDEFINE arg_name, %%i, %1 - %assign %%i %%i+1 - %rotate 1 - %endrep - %xdefine stack_offset %%stack_offset - %assign n_arg_names %0 -%endmacro - -%if WIN64 ; Windows x64 ;================================================= - -DECLARE_REG 0, rcx -DECLARE_REG 1, rdx -DECLARE_REG 2, R8 -DECLARE_REG 3, R9 -DECLARE_REG 4, R10, 40 -DECLARE_REG 5, R11, 48 -DECLARE_REG 6, rax, 56 -DECLARE_REG 7, rdi, 64 -DECLARE_REG 8, rsi, 72 -DECLARE_REG 9, rbx, 80 -DECLARE_REG 10, rbp, 88 -DECLARE_REG 11, R12, 96 -DECLARE_REG 12, R13, 104 -DECLARE_REG 13, R14, 112 -DECLARE_REG 14, R15, 120 - -%macro PROLOGUE 2-4+ 0 ; #args, #regs, #xmm_regs, arg_names... - %assign num_args %1 - %assign regs_used %2 - ASSERT regs_used >= num_args - ASSERT regs_used <= 15 - PUSH_IF_USED 7, 8, 9, 10, 11, 12, 13, 14 - %if mmsize == 8 - %assign xmm_regs_used 0 - %else - WIN64_SPILL_XMM %3 - %endif - LOAD_IF_USED 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 - DEFINE_ARGS %4 -%endmacro - -%macro WIN64_SPILL_XMM 1 - %assign xmm_regs_used %1 - ASSERT xmm_regs_used <= 16 - %if xmm_regs_used > 6 - SUB rsp, (xmm_regs_used-6)*16+16 - %assign %%i xmm_regs_used - %rep (xmm_regs_used-6) - %assign %%i %%i-1 - movdqa [rsp + (%%i-6)*16+(~stack_offset&8)], xmm %+ %%i - %endrep - %endif -%endmacro - -%macro WIN64_RESTORE_XMM_INTERNAL 1 - %if xmm_regs_used > 6 - %assign %%i xmm_regs_used - %rep (xmm_regs_used-6) - %assign %%i %%i-1 - movdqa xmm %+ %%i, [%1 + (%%i-6)*16+(~stack_offset&8)] - %endrep - add %1, (xmm_regs_used-6)*16+16 - %endif -%endmacro - -%macro WIN64_RESTORE_XMM 1 - WIN64_RESTORE_XMM_INTERNAL %1 - %assign stack_offset stack_offset-(xmm_regs_used-6)*16+16 - %assign xmm_regs_used 0 -%endmacro - -%define has_epilogue regs_used > 7 || xmm_regs_used > 6 || mmsize == 32 - -%macro RET 0 - WIN64_RESTORE_XMM_INTERNAL rsp - POP_IF_USED 14, 13, 12, 11, 10, 9, 8, 7 -%if mmsize == 32 - vzeroupper -%endif - ret -%endmacro - -%elif ARCH_X86_64 ; *nix x64 ;============================================= - -DECLARE_REG 0, rdi -DECLARE_REG 1, rsi -DECLARE_REG 2, rdx -DECLARE_REG 3, rcx -DECLARE_REG 4, R8 -DECLARE_REG 5, R9 -DECLARE_REG 6, rax, 8 -DECLARE_REG 7, R10, 16 -DECLARE_REG 8, R11, 24 -DECLARE_REG 9, rbx, 32 -DECLARE_REG 10, rbp, 40 -DECLARE_REG 11, R12, 48 -DECLARE_REG 12, R13, 56 -DECLARE_REG 13, R14, 64 -DECLARE_REG 14, R15, 72 - -%macro PROLOGUE 2-4+ ; #args, #regs, #xmm_regs, arg_names... - %assign num_args %1 - %assign regs_used %2 - ASSERT regs_used >= num_args - ASSERT regs_used <= 15 - PUSH_IF_USED 9, 10, 11, 12, 13, 14 - LOAD_IF_USED 6, 7, 8, 9, 10, 11, 12, 13, 14 - DEFINE_ARGS %4 -%endmacro - -%define has_epilogue regs_used > 9 || mmsize == 32 - -%macro RET 0 - POP_IF_USED 14, 13, 12, 11, 10, 9 -%if mmsize == 32 - vzeroupper -%endif - ret -%endmacro - -%else ; X86_32 ;============================================================== - -DECLARE_REG 0, eax, 4 -DECLARE_REG 1, ecx, 8 -DECLARE_REG 2, edx, 12 -DECLARE_REG 3, ebx, 16 -DECLARE_REG 4, esi, 20 -DECLARE_REG 5, edi, 24 -DECLARE_REG 6, ebp, 28 -%define rsp esp - -%macro DECLARE_ARG 1-* - %rep %0 - %define r%1m [esp + stack_offset + 4*%1 + 4] - %define r%1mp dword r%1m - %rotate 1 - %endrep -%endmacro - -DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14 - -%macro PROLOGUE 2-4+ ; #args, #regs, #xmm_regs, arg_names... - %assign num_args %1 - %assign regs_used %2 - %if regs_used > 7 - %assign regs_used 7 - %endif - ASSERT regs_used >= num_args - PUSH_IF_USED 3, 4, 5, 6 - LOAD_IF_USED 0, 1, 2, 3, 4, 5, 6 - DEFINE_ARGS %4 -%endmacro - -%define has_epilogue regs_used > 3 || mmsize == 32 - -%macro RET 0 - POP_IF_USED 6, 5, 4, 3 -%if mmsize == 32 - vzeroupper -%endif - ret -%endmacro - -%endif ;====================================================================== - -%if WIN64 == 0 -%macro WIN64_SPILL_XMM 1 -%endmacro -%macro WIN64_RESTORE_XMM 1 -%endmacro -%endif - -%macro REP_RET 0 - %if has_epilogue - RET - %else - rep ret - %endif -%endmacro - -%macro TAIL_CALL 2 ; callee, is_nonadjacent - %if has_epilogue - call %1 - RET - %elif %2 - jmp %1 - %endif -%endmacro - -;============================================================================= -; arch-independent part -;============================================================================= - -%assign function_align 16 - -; Begin a function. -; Applies any symbol mangling needed for C linkage, and sets up a define such that -; subsequent uses of the function name automatically refer to the mangled version. -; Appends cpuflags to the function name if cpuflags has been specified. -%macro cglobal 1-2+ ; name, [PROLOGUE args] -%if %0 == 1 - cglobal_internal %1 %+ SUFFIX -%else - cglobal_internal %1 %+ SUFFIX, %2 -%endif -%endmacro -%macro cglobal_internal 1-2+ - %ifndef cglobaled_%1 - %xdefine %1 mangle(%1) - %xdefine %1.skip_prologue %1 %+ .skip_prologue - CAT_XDEFINE cglobaled_, %1, 1 - %endif - %xdefine current_function %1 - %ifidn __OUTPUT_FORMAT__,elf - global %1:function hidden - %else - global %1 - %endif - align function_align - %1: - RESET_MM_PERMUTATION ; not really needed, but makes disassembly somewhat nicer - %assign stack_offset 0 - %if %0 > 1 - PROLOGUE %2 - %endif -%endmacro - -%macro cextern 1 - %xdefine %1 mangle(%1) - CAT_XDEFINE cglobaled_, %1, 1 - extern %1 -%endmacro - -; like cextern, but without the prefix -%macro cextern_naked 1 - %xdefine %1 mangle(%1) - CAT_XDEFINE cglobaled_, %1, 1 - extern %1 -%endmacro - -%macro const 2+ - %xdefine %1 mangle(%1) - global %1 - %1: %2 -%endmacro - -; This is needed for ELF, otherwise the GNU linker assumes the stack is -; executable by default. -%ifidn __OUTPUT_FORMAT__,elf -SECTION .note.GNU-stack noalloc noexec nowrite progbits -%endif -%ifidn __OUTPUT_FORMAT__,elf32 -section .note.GNU-stack noalloc noexec nowrite progbits -%endif -%ifidn __OUTPUT_FORMAT__,elf64 -section .note.GNU-stack noalloc noexec nowrite progbits -%endif - -; cpuflags - -%assign cpuflags_MMX (1<<0) -%assign cpuflags_MMX2 (1<<1) | cpuflags_MMX -%assign cpuflags_3dnow (1<<2) | cpuflags_MMX -%assign cpuflags_3dnow2 (1<<3) | cpuflags_3dnow -%assign cpuflags_SSE (1<<4) | cpuflags_MMX2 -%assign cpuflags_SSE2 (1<<5) | cpuflags_SSE -%assign cpuflags_SSE2slow (1<<6) | cpuflags_SSE2 -%assign cpuflags_SSE3 (1<<7) | cpuflags_SSE2 -%assign cpuflags_SSSE3 (1<<8) | cpuflags_SSE3 -%assign cpuflags_SSE4 (1<<9) | cpuflags_SSSE3 -%assign cpuflags_SSE42 (1<<10)| cpuflags_SSE4 -%assign cpuflags_AVX (1<<11)| cpuflags_SSE42 -%assign cpuflags_xop (1<<12)| cpuflags_AVX -%assign cpuflags_fma4 (1<<13)| cpuflags_AVX -%assign cpuflags_AVX2 (1<<14)| cpuflags_AVX -%assign cpuflags_fma3 (1<<15)| cpuflags_AVX - -%assign cpuflags_cache32 (1<<16) -%assign cpuflags_cache64 (1<<17) -%assign cpuflags_slowctz (1<<18) -%assign cpuflags_lzcnt (1<<19) -%assign cpuflags_misalign (1<<20) -%assign cpuflags_aligned (1<<21) ; not a cpu feature, but a function variant -%assign cpuflags_atom (1<<22) -%assign cpuflags_bmi1 (1<<23) -%assign cpuflags_bmi2 (1<<24)|cpuflags_bmi1 -%assign cpuflags_tbm (1<<25)|cpuflags_bmi1 - -%define cpuflag(x) ((cpuflags & (cpuflags_ %+ x)) == (cpuflags_ %+ x)) -%define notcpuflag(x) ((cpuflags & (cpuflags_ %+ x)) != (cpuflags_ %+ x)) - -; Takes up to 2 cpuflags from the above list. -; All subsequent functions (up to the next INIT_CPUFLAGS) is built for the specified cpu. -; You shouldn't need to invoke this macro directly, it's a subroutine for INIT_MMX &co. -%macro INIT_CPUFLAGS 0-2 - %if %0 >= 1 - %xdefine cpuname %1 - %assign cpuflags cpuflags_%1 - %if %0 >= 2 - %xdefine cpuname %1_%2 - %assign cpuflags cpuflags | cpuflags_%2 - %endif - %xdefine SUFFIX _ %+ cpuname - %if cpuflag(AVX) - %assign AVX_enabled 1 - %endif - %if mmsize == 16 && notcpuflag(SSE2) - %define mova movaps - %define movu movups - %define movnta movntps - %endif - %if cpuflag(aligned) - %define movu mova - %elifidn %1, SSE3 - %define movu lddqu - %endif - %else - %xdefine SUFFIX - %undef cpuname - %undef cpuflags - %endif -%endmacro - -; merge MMX and SSE* - -%macro CAT_XDEFINE 3 - %xdefine %1%2 %3 -%endmacro - -%macro CAT_UNDEF 2 - %undef %1%2 -%endmacro - -%macro INIT_MMX 0-1+ - %assign AVX_enabled 0 - %define RESET_MM_PERMUTATION INIT_MMX %1 - %define mmsize 8 - %define num_mmregs 8 - %define mova movq - %define movu movq - %define movh movd - %define movnta movntq - %assign %%i 0 - %rep 8 - CAT_XDEFINE m, %%i, mm %+ %%i - CAT_XDEFINE nmm, %%i, %%i - %assign %%i %%i+1 - %endrep - %rep 8 - CAT_UNDEF m, %%i - CAT_UNDEF nmm, %%i - %assign %%i %%i+1 - %endrep - INIT_CPUFLAGS %1 -%endmacro - -%macro INIT_XMM 0-1+ - %assign AVX_enabled 0 - %define RESET_MM_PERMUTATION INIT_XMM %1 - %define mmsize 16 - %define num_mmregs 8 - %if ARCH_X86_64 - %define num_mmregs 16 - %endif - %define mova movdqa - %define movu movdqu - %define movh movq - %define movnta movntdq - %assign %%i 0 - %rep num_mmregs - CAT_XDEFINE m, %%i, xmm %+ %%i - CAT_XDEFINE nxmm, %%i, %%i - %assign %%i %%i+1 - %endrep - INIT_CPUFLAGS %1 -%endmacro - -%macro INIT_YMM 0-1+ - %assign AVX_enabled 1 - %define RESET_MM_PERMUTATION INIT_YMM %1 - %define mmsize 32 - %define num_mmregs 8 - %if ARCH_X86_64 - %define num_mmregs 16 - %endif - %define mova vmovaps - %define movu vmovups - %undef movh - %define movnta vmovntps - %assign %%i 0 - %rep num_mmregs - CAT_XDEFINE m, %%i, ymm %+ %%i - CAT_XDEFINE nymm, %%i, %%i - %assign %%i %%i+1 - %endrep - INIT_CPUFLAGS %1 -%endmacro - -INIT_XMM - -; I often want to use macros that permute their arguments. e.g. there's no -; efficient way to implement butterfly or transpose or dct without swapping some -; arguments. -; -; I would like to not have to manually keep track of the permutations: -; If I insert a permutation in the middle of a function, it should automatically -; change everything that follows. For more complex macros I may also have multiple -; implementations, e.g. the SSE2 and SSSE3 versions may have different permutations. -; -; Hence these macros. Insert a PERMUTE or some SWAPs at the end of a macro that -; permutes its arguments. It's equivalent to exchanging the contents of the -; registers, except that this way you exchange the register names instead, so it -; doesn't cost any cycles. - -%macro PERMUTE 2-* ; takes a list of pairs to swap -%rep %0/2 - %xdefine tmp%2 m%2 - %xdefine ntmp%2 nm%2 - %rotate 2 -%endrep -%rep %0/2 - %xdefine m%1 tmp%2 - %xdefine nm%1 ntmp%2 - %undef tmp%2 - %undef ntmp%2 - %rotate 2 -%endrep -%endmacro - -%macro SWAP 2-* ; swaps a single chain (sometimes more concise than pairs) -%rep %0-1 -%ifdef m%1 - %xdefine tmp m%1 - %xdefine m%1 m%2 - %xdefine m%2 tmp - CAT_XDEFINE n, m%1, %1 - CAT_XDEFINE n, m%2, %2 -%else - ; If we were called as "SWAP m0,m1" rather than "SWAP 0,1" infer the original numbers here. - ; Be careful using this mode in nested macros though, as in some cases there may be - ; other copies of m# that have already been dereferenced and don't get updated correctly. - %xdefine %%n1 n %+ %1 - %xdefine %%n2 n %+ %2 - %xdefine tmp m %+ %%n1 - CAT_XDEFINE m, %%n1, m %+ %%n2 - CAT_XDEFINE m, %%n2, tmp - CAT_XDEFINE n, m %+ %%n1, %%n1 - CAT_XDEFINE n, m %+ %%n2, %%n2 -%endif - %undef tmp - %rotate 1 -%endrep -%endmacro - -; If SAVE_MM_PERMUTATION is placed at the end of a function, then any later -; calls to that function will automatically load the permutation, so values can -; be returned in mmregs. -%macro SAVE_MM_PERMUTATION 0-1 - %if %0 - %xdefine %%f %1_m - %else - %xdefine %%f current_function %+ _m - %endif - %assign %%i 0 - %rep num_mmregs - CAT_XDEFINE %%f, %%i, m %+ %%i - %assign %%i %%i+1 - %endrep -%endmacro - -%macro LOAD_MM_PERMUTATION 1 ; name to load from - %ifdef %1_m0 - %assign %%i 0 - %rep num_mmregs - CAT_XDEFINE m, %%i, %1_m %+ %%i - CAT_XDEFINE n, m %+ %%i, %%i - %assign %%i %%i+1 - %endrep - %endif -%endmacro - -; Append cpuflags to the callee's name iff the appended name is known and the plain name isn't -%macro call 1 - call_internal %1, %1 %+ SUFFIX -%endmacro -%macro call_internal 2 - %xdefine %%i %1 - %ifndef cglobaled_%1 - %ifdef cglobaled_%2 - %xdefine %%i %2 - %endif - %endif - call %%i - LOAD_MM_PERMUTATION %%i -%endmacro - -; Substitutions that reduce instruction size but are functionally equivalent -%macro add 2 - %ifnum %2 - %if %2==128 - sub %1, -128 - %else - add %1, %2 - %endif - %else - add %1, %2 - %endif -%endmacro - -%macro sub 2 - %ifnum %2 - %if %2==128 - add %1, -128 - %else - sub %1, %2 - %endif - %else - sub %1, %2 - %endif -%endmacro - -;============================================================================= -; AVX abstraction layer -;============================================================================= - -%assign i 0 -%rep 16 - %if i < 8 - CAT_XDEFINE sizeofmm, i, 8 - %endif - CAT_XDEFINE sizeofxmm, i, 16 - CAT_XDEFINE sizeofymm, i, 32 -%assign i i+1 -%endrep -%undef i - -%macro CHECK_AVX_INSTR_EMU 3-* - %xdefine %%opcode %1 - %xdefine %%dst %2 - %rep %0-2 - %ifidn %%dst, %3 - %error non-AVX emulation of ``%%opcode'' is not supported - %endif - %rotate 1 - %endrep -%endmacro - -;%1 == instruction -;%2 == 1 if float, 0 if int -;%3 == 1 if 4-operand (xmm, xmm, xmm, imm), 0 if 2- or 3-operand (xmm, xmm, xmm) -;%4 == number of operands given -;%5+: operands -%macro RUN_AVX_INSTR 6-7+ - %ifid %6 - %define %%sizeofreg sizeof%6 - %elifid %5 - %define %%sizeofreg sizeof%5 - %else - %define %%sizeofreg mmsize - %endif - %if %%sizeofreg==32 - %if %4>=3 - v%1 %5, %6, %7 - %else - v%1 %5, %6 - %endif - %else - %if %%sizeofreg==8 - %define %%regmov movq - %elif %2 - %define %%regmov movaps - %else - %define %%regmov movdqa - %endif - - %if %4>=3+%3 - %ifnidn %5, %6 - %if AVX_enabled && %%sizeofreg==16 - v%1 %5, %6, %7 - %else - CHECK_AVX_INSTR_EMU {%1 %5, %6, %7}, %5, %7 - %%regmov %5, %6 - %1 %5, %7 - %endif - %else - %1 %5, %7 - %endif - %elif %4>=3 - %1 %5, %6, %7 - %else - %1 %5, %6 - %endif - %endif -%endmacro - -; 3arg AVX ops with a memory arg can only have it in src2, -; whereas SSE emulation of 3arg prefers to have it in src1 (i.e. the mov). -; So, if the op is symmetric and the wrong one is memory, swap them. -%macro RUN_AVX_INSTR1 8 - %assign %%swap 0 - %if AVX_enabled - %ifnid %6 - %assign %%swap 1 - %endif - %elifnidn %5, %6 - %ifnid %7 - %assign %%swap 1 - %endif - %endif - %if %%swap && %3 == 0 && %8 == 1 - RUN_AVX_INSTR %1, %2, %3, %4, %5, %7, %6 - %else - RUN_AVX_INSTR %1, %2, %3, %4, %5, %6, %7 - %endif -%endmacro - -;%1 == instruction -;%2 == 1 if float, 0 if int -;%3 == 1 if 4-operand (xmm, xmm, xmm, imm), 0 if 2- or 3-operand (xmm, xmm, xmm) -;%4 == 1 if symmetric (i.e. doesn't matter which src arg is which), 0 if not -%macro AVX_INSTR 4 - %macro %1 2-9 fnord, fnord, fnord, %1, %2, %3, %4 - %ifidn %3, fnord - RUN_AVX_INSTR %6, %7, %8, 2, %1, %2 - %elifidn %4, fnord - RUN_AVX_INSTR1 %6, %7, %8, 3, %1, %2, %3, %9 - %elifidn %5, fnord - RUN_AVX_INSTR %6, %7, %8, 4, %1, %2, %3, %4 - %else - RUN_AVX_INSTR %6, %7, %8, 5, %1, %2, %3, %4, %5 - %endif - %endmacro -%endmacro - -AVX_INSTR addpd, 1, 0, 1 -AVX_INSTR addps, 1, 0, 1 -AVX_INSTR addsd, 1, 0, 1 -AVX_INSTR addss, 1, 0, 1 -AVX_INSTR addsubpd, 1, 0, 0 -AVX_INSTR addsubps, 1, 0, 0 -AVX_INSTR andpd, 1, 0, 1 -AVX_INSTR andps, 1, 0, 1 -AVX_INSTR andnpd, 1, 0, 0 -AVX_INSTR andnps, 1, 0, 0 -AVX_INSTR blendpd, 1, 0, 0 -AVX_INSTR blendps, 1, 0, 0 -AVX_INSTR blendvpd, 1, 0, 0 -AVX_INSTR blendvps, 1, 0, 0 -AVX_INSTR cmppd, 1, 0, 0 -AVX_INSTR cmpps, 1, 0, 0 -AVX_INSTR cmpsd, 1, 0, 0 -AVX_INSTR cmpss, 1, 0, 0 -AVX_INSTR cvtdq2ps, 1, 0, 0 -AVX_INSTR cvtps2dq, 1, 0, 0 -AVX_INSTR divpd, 1, 0, 0 -AVX_INSTR divps, 1, 0, 0 -AVX_INSTR divsd, 1, 0, 0 -AVX_INSTR divss, 1, 0, 0 -AVX_INSTR dppd, 1, 1, 0 -AVX_INSTR dpps, 1, 1, 0 -AVX_INSTR haddpd, 1, 0, 0 -AVX_INSTR haddps, 1, 0, 0 -AVX_INSTR hsubpd, 1, 0, 0 -AVX_INSTR hsubps, 1, 0, 0 -AVX_INSTR maxpd, 1, 0, 1 -AVX_INSTR maxps, 1, 0, 1 -AVX_INSTR maxsd, 1, 0, 1 -AVX_INSTR maxss, 1, 0, 1 -AVX_INSTR minpd, 1, 0, 1 -AVX_INSTR minps, 1, 0, 1 -AVX_INSTR minsd, 1, 0, 1 -AVX_INSTR minss, 1, 0, 1 -AVX_INSTR movhlps, 1, 0, 0 -AVX_INSTR movlhps, 1, 0, 0 -AVX_INSTR movsd, 1, 0, 0 -AVX_INSTR movss, 1, 0, 0 -AVX_INSTR mpsadbw, 0, 1, 0 -AVX_INSTR mulpd, 1, 0, 1 -AVX_INSTR mulps, 1, 0, 1 -AVX_INSTR mulsd, 1, 0, 1 -AVX_INSTR mulss, 1, 0, 1 -AVX_INSTR orpd, 1, 0, 1 -AVX_INSTR orps, 1, 0, 1 -AVX_INSTR pabsb, 0, 0, 0 -AVX_INSTR pabsw, 0, 0, 0 -AVX_INSTR pabsd, 0, 0, 0 -AVX_INSTR packsswb, 0, 0, 0 -AVX_INSTR packssdw, 0, 0, 0 -AVX_INSTR packuswb, 0, 0, 0 -AVX_INSTR packusdw, 0, 0, 0 -AVX_INSTR paddb, 0, 0, 1 -AVX_INSTR paddw, 0, 0, 1 -AVX_INSTR paddd, 0, 0, 1 -AVX_INSTR paddq, 0, 0, 1 -AVX_INSTR paddsb, 0, 0, 1 -AVX_INSTR paddsw, 0, 0, 1 -AVX_INSTR paddusb, 0, 0, 1 -AVX_INSTR paddusw, 0, 0, 1 -AVX_INSTR palignr, 0, 1, 0 -AVX_INSTR pand, 0, 0, 1 -AVX_INSTR pandn, 0, 0, 0 -AVX_INSTR pavgb, 0, 0, 1 -AVX_INSTR pavgw, 0, 0, 1 -AVX_INSTR pblendvb, 0, 0, 0 -AVX_INSTR pblendw, 0, 1, 0 -AVX_INSTR pcmpestri, 0, 0, 0 -AVX_INSTR pcmpestrm, 0, 0, 0 -AVX_INSTR pcmpistri, 0, 0, 0 -AVX_INSTR pcmpistrm, 0, 0, 0 -AVX_INSTR pcmpeqb, 0, 0, 1 -AVX_INSTR pcmpeqw, 0, 0, 1 -AVX_INSTR pcmpeqd, 0, 0, 1 -AVX_INSTR pcmpeqq, 0, 0, 1 -AVX_INSTR pcmpgtb, 0, 0, 0 -AVX_INSTR pcmpgtw, 0, 0, 0 -AVX_INSTR pcmpgtd, 0, 0, 0 -AVX_INSTR pcmpgtq, 0, 0, 0 -AVX_INSTR phaddw, 0, 0, 0 -AVX_INSTR phaddd, 0, 0, 0 -AVX_INSTR phaddsw, 0, 0, 0 -AVX_INSTR phsubw, 0, 0, 0 -AVX_INSTR phsubd, 0, 0, 0 -AVX_INSTR phsubsw, 0, 0, 0 -AVX_INSTR pmaddwd, 0, 0, 1 -AVX_INSTR pmaddubsw, 0, 0, 0 -AVX_INSTR pmaxsb, 0, 0, 1 -AVX_INSTR pmaxsw, 0, 0, 1 -AVX_INSTR pmaxsd, 0, 0, 1 -AVX_INSTR pmaxub, 0, 0, 1 -AVX_INSTR pmaxuw, 0, 0, 1 -AVX_INSTR pmaxud, 0, 0, 1 -AVX_INSTR pminsb, 0, 0, 1 -AVX_INSTR pminsw, 0, 0, 1 -AVX_INSTR pminsd, 0, 0, 1 -AVX_INSTR pminub, 0, 0, 1 -AVX_INSTR pminuw, 0, 0, 1 -AVX_INSTR pminud, 0, 0, 1 -AVX_INSTR pmovmskb, 0, 0, 0 -AVX_INSTR pmulhuw, 0, 0, 1 -AVX_INSTR pmulhrsw, 0, 0, 1 -AVX_INSTR pmulhw, 0, 0, 1 -AVX_INSTR pmullw, 0, 0, 1 -AVX_INSTR pmulld, 0, 0, 1 -AVX_INSTR pmuludq, 0, 0, 1 -AVX_INSTR pmuldq, 0, 0, 1 -AVX_INSTR por, 0, 0, 1 -AVX_INSTR psadbw, 0, 0, 1 -AVX_INSTR pshufb, 0, 0, 0 -AVX_INSTR pshufd, 0, 1, 0 -AVX_INSTR pshufhw, 0, 1, 0 -AVX_INSTR pshuflw, 0, 1, 0 -AVX_INSTR psignb, 0, 0, 0 -AVX_INSTR psignw, 0, 0, 0 -AVX_INSTR psignd, 0, 0, 0 -AVX_INSTR psllw, 0, 0, 0 -AVX_INSTR pslld, 0, 0, 0 -AVX_INSTR psllq, 0, 0, 0 -AVX_INSTR pslldq, 0, 0, 0 -AVX_INSTR psraw, 0, 0, 0 -AVX_INSTR psrad, 0, 0, 0 -AVX_INSTR psrlw, 0, 0, 0 -AVX_INSTR psrld, 0, 0, 0 -AVX_INSTR psrlq, 0, 0, 0 -AVX_INSTR psrldq, 0, 0, 0 -AVX_INSTR psubb, 0, 0, 0 -AVX_INSTR psubw, 0, 0, 0 -AVX_INSTR psubd, 0, 0, 0 -AVX_INSTR psubq, 0, 0, 0 -AVX_INSTR psubsb, 0, 0, 0 -AVX_INSTR psubsw, 0, 0, 0 -AVX_INSTR psubusb, 0, 0, 0 -AVX_INSTR psubusw, 0, 0, 0 -AVX_INSTR ptest, 0, 0, 0 -AVX_INSTR punpckhbw, 0, 0, 0 -AVX_INSTR punpckhwd, 0, 0, 0 -AVX_INSTR punpckhdq, 0, 0, 0 -AVX_INSTR punpckhqdq, 0, 0, 0 -AVX_INSTR punpcklbw, 0, 0, 0 -AVX_INSTR punpcklwd, 0, 0, 0 -AVX_INSTR punpckldq, 0, 0, 0 -AVX_INSTR punpcklqdq, 0, 0, 0 -AVX_INSTR pxor, 0, 0, 1 -AVX_INSTR shufps, 1, 1, 0 -AVX_INSTR subpd, 1, 0, 0 -AVX_INSTR subps, 1, 0, 0 -AVX_INSTR subsd, 1, 0, 0 -AVX_INSTR subss, 1, 0, 0 -AVX_INSTR unpckhpd, 1, 0, 0 -AVX_INSTR unpckhps, 1, 0, 0 -AVX_INSTR unpcklpd, 1, 0, 0 -AVX_INSTR unpcklps, 1, 0, 0 -AVX_INSTR xorpd, 1, 0, 1 -AVX_INSTR xorps, 1, 0, 1 - -; 3DNow instructions, for sharing code between AVX, SSE and 3DN -AVX_INSTR pfadd, 1, 0, 1 -AVX_INSTR pfsub, 1, 0, 0 -AVX_INSTR pfmul, 1, 0, 1 - -; base-4 constants for shuffles -%assign i 0 -%rep 256 - %assign j ((i>>6)&3)*1000 + ((i>>4)&3)*100 + ((i>>2)&3)*10 + (i&3) - %if j < 10 - CAT_XDEFINE q000, j, i - %elif j < 100 - CAT_XDEFINE q00, j, i - %elif j < 1000 - CAT_XDEFINE q0, j, i - %else - CAT_XDEFINE q, j, i - %endif -%assign i i+1 -%endrep -%undef i -%undef j - -%macro FMA_INSTR 3 - %macro %1 4-7 %1, %2, %3 - %if cpuflag(xop) - v%5 %1, %2, %3, %4 - %else - %6 %1, %2, %3 - %7 %1, %4 - %endif - %endmacro -%endmacro - -FMA_INSTR pmacsdd, pmulld, paddd -FMA_INSTR pmacsww, pmullw, paddw -FMA_INSTR pmadcswd, pmaddwd, paddd - -; tzcnt is equivalent to "rep bsf" and is backwards-compatible with bsf. -; This lets us use tzcnt without bumping the yasm version requirement yet. -%define tzcnt rep bsf diff --git a/src/main/jni/video.c b/src/main/jni/video.c deleted file mode 100644 index e71f38bb9..000000000 --- a/src/main/jni/video.c +++ /dev/null @@ -1,109 +0,0 @@ -#include <jni.h> -#include <libyuv.h> -#include <utils.h> - -enum COLOR_FORMATTYPE { - COLOR_FormatMonochrome = 1, - COLOR_Format8bitRGB332 = 2, - COLOR_Format12bitRGB444 = 3, - COLOR_Format16bitARGB4444 = 4, - COLOR_Format16bitARGB1555 = 5, - COLOR_Format16bitRGB565 = 6, - COLOR_Format16bitBGR565 = 7, - COLOR_Format18bitRGB666 = 8, - COLOR_Format18bitARGB1665 = 9, - COLOR_Format19bitARGB1666 = 10, - COLOR_Format24bitRGB888 = 11, - COLOR_Format24bitBGR888 = 12, - COLOR_Format24bitARGB1887 = 13, - COLOR_Format25bitARGB1888 = 14, - COLOR_Format32bitBGRA8888 = 15, - COLOR_Format32bitARGB8888 = 16, - COLOR_FormatYUV411Planar = 17, - COLOR_FormatYUV411PackedPlanar = 18, - COLOR_FormatYUV420Planar = 19, - COLOR_FormatYUV420PackedPlanar = 20, - COLOR_FormatYUV420SemiPlanar = 21, - COLOR_FormatYUV422Planar = 22, - COLOR_FormatYUV422PackedPlanar = 23, - COLOR_FormatYUV422SemiPlanar = 24, - COLOR_FormatYCbYCr = 25, - COLOR_FormatYCrYCb = 26, - COLOR_FormatCbYCrY = 27, - COLOR_FormatCrYCbY = 28, - COLOR_FormatYUV444Interleaved = 29, - COLOR_FormatRawBayer8bit = 30, - COLOR_FormatRawBayer10bit = 31, - COLOR_FormatRawBayer8bitcompressed = 32, - COLOR_FormatL2 = 33, - COLOR_FormatL4 = 34, - COLOR_FormatL8 = 35, - COLOR_FormatL16 = 36, - COLOR_FormatL24 = 37, - COLOR_FormatL32 = 38, - COLOR_FormatYUV420PackedSemiPlanar = 39, - COLOR_FormatYUV422PackedSemiPlanar = 40, - COLOR_Format18BitBGR666 = 41, - COLOR_Format24BitARGB6666 = 42, - COLOR_Format24BitABGR6666 = 43, - - COLOR_TI_FormatYUV420PackedSemiPlanar = 0x7f000100, - COLOR_FormatSurface = 0x7F000789, - COLOR_QCOM_FormatYUV420SemiPlanar = 0x7fa30c00 -}; - -int isSemiPlanarYUV(int colorFormat) { - switch (colorFormat) { - case COLOR_FormatYUV420Planar: - case COLOR_FormatYUV420PackedPlanar: - return 0; - case COLOR_FormatYUV420SemiPlanar: - case COLOR_FormatYUV420PackedSemiPlanar: - case COLOR_TI_FormatYUV420PackedSemiPlanar: - return 1; - default: - return 0; - } -} - -JNIEXPORT int Java_de_pixart_messenger_utils_video_convertVideoFrame(JNIEnv *env, jclass class, jobject src, jobject dest, int destFormat, int width, int height, int padding, int swap) { - if (!src || !dest || !destFormat) { - return 0; - } - - jbyte *srcBuff = (*env)->GetDirectBufferAddress(env, src); - jbyte *destBuff = (*env)->GetDirectBufferAddress(env, dest); - - int half_width = (width + 1) / 2; - int half_height = (height + 1) / 2; - - if (!isSemiPlanarYUV(destFormat)) { - if (!swap) { - ARGBToI420(srcBuff, width * 4, - destBuff, width, - destBuff + width * height + half_width * half_height + padding * 5 / 4, half_width, - destBuff + width * height + padding, half_width, - width, height); - } else { - ARGBToI420(srcBuff, width * 4, - destBuff, width, - destBuff + width * height + padding, half_width, - destBuff + width * height + half_width * half_height + padding * 5 / 4, half_width, - width, height); - } - } else { - if (!swap) { - ARGBToNV21(srcBuff, width * 4, - destBuff, width, - destBuff + width * height + padding, half_width * 2, - width, height); - } else { - ARGBToNV12(srcBuff, width * 4, - destBuff, width, - destBuff + width * height + padding, half_width * 2, - width, height); - } - } - - return 1; -} diff --git a/src/main/jniLibs/armeabi-v7a/libtmessages.7.so b/src/main/jniLibs/armeabi-v7a/libtmessages.7.so Binary files differdeleted file mode 100644 index c086f4a98..000000000 --- a/src/main/jniLibs/armeabi-v7a/libtmessages.7.so +++ /dev/null diff --git a/src/main/jniLibs/armeabi/libtmessages.7.so b/src/main/jniLibs/armeabi/libtmessages.7.so Binary files differdeleted file mode 100644 index 42ad265ae..000000000 --- a/src/main/jniLibs/armeabi/libtmessages.7.so +++ /dev/null diff --git a/src/main/jniLibs/x86/libtmessages.7.so b/src/main/jniLibs/x86/libtmessages.7.so Binary files differdeleted file mode 100644 index aaa768eac..000000000 --- a/src/main/jniLibs/x86/libtmessages.7.so +++ /dev/null |