FFmpeg开源项目编译说明_ffmpeg6.0 编译 dxva2

来源：https://wenku.baidu.com/view/349caa7a770bf78a6429543f.html

FFmpeg开源项目编译说明

----搭建编译环境、用VS2013编译和调试


一、准备工作
1、本机环境：win7 sp1，74位，vs2013
2、ffmpeg官网上有源代码和现成的静态库和动态库可以调用。如果需要定制ffmpeg的模块，或者跟踪调试和分析研究，则需要进行编译和调试。这里采用的Ffmpeg版本为当前最新版本：2.6.2。
3、ffmpeg本身是linux下的开源项目。它在linux、windows系统中都可以编译。在windows系统，尽量选择VS2013编译工具，是因为VS2013支持大部分C99的特性，基本不需要改动代码，也不需要使用C99转C89的工具。
4、ffmpeg是一个开源的多媒体库，使用非常广泛。在linux下编译ffmpeg非常简单，而在windows下编译就不是那么容易了。一般在windows下使用MinGW的gcc toolchain进行编译，这样的话，因为ffmpeg的导出函数均是C风格，因而gcc编译的lib可以被vs链接。但是gcc的debug符号与vs的debug符号（*.pdb）是无法兼容的，因此我们用vs来开发基于ffmpeg的程序时无法深入ffmpeg内部进行debug


二、搭建编译环境
MinGW和yasm是绕不过去的，因为需要使用MinGW来生成config.h，而ffmpeg的汇编语法和VS的不一样，因此需要yasm。具体步骤如下：
1、 下载FFMPEG源码，下载地址：http://ffmpeg.zeranoe.com/builds/；（这里也可以下载：静态库static、动态库shared、开发库dev）
2、 下载MinGW安装器，下载地址：http://www.mingw.org/；下载完成后安装，安装完成后点运行，标记上以下几项：
 
然后在Installation菜单下点击Apply Changes（mingw32-gcc-g++也可不选择，因为我们用vs2013编译！）；
注意：运行下载的MinGW安装管理器，安装好MinGW，里面已经包含msys。假设安装好后MinGW路径为：C:/MinGW ，则msys路径应为：C:/MinGW/msys 。要将C:\MinGW\bin和C:\MinGW\msys\1.0\bin加到系统path环境变量中（注意：win7的path里面的目录分割符是采用反斜杠的“\”）。
3、 下载yasm.exe，下载地址： http://yasm.tortall.net/ ；下载后的文件如下图所示：（根据计算机的32位和64位选择，本机为用win64位的yasm-1.2.0-win64.exe）
 
下载后改名为yasm.exe，再复制到C:/MinGW/msys/1.0/bin目录下；
4、 复制C:/MinGW/msys/1.0/msys.bat 到同目录下，改名叫做msys_vs2013.bat（这样做是为了保留原来的文件！）。用编辑器打开C:/MinGW/msys/1.0/msys_vs2013.bat，在此文件的最前面(@echo off之后)添加一行如下内容：
call "C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin\vcvars32.bat" (要与vs2013的实际安装路径一致)；
5、 重命名 C:/MinGW/msys/1.0/bin/link.exe 为link_renamed.exe (依实际安装选择路径)，这一步是防止这个link.exe与vc的link.exe发生冲突，编译完成后可修改回来。
三、编译ffmpeg
1、双击C:/MinGW/msys/1.0/msys_vs2013.bat，运行shell（linux的shell）：转到FFMPEG源代码根目录下，可运行 ./configure --help查看编译配置选项（用./configure –help>>lisq.txt，在lisq.txt中查看方便一些），可以使用的命令及选项为：
静态库：./configure  --enable-static   --prefix=./vs2013_build  --enable-debug  --toolchain=msvc 
动态库：./configure  --enable-shared  --prefix=./vs2013_build  --enable-debug  --toolchain=msvc
等待配置完成返回（大约两分钟）；
其他常用选项：--enable-avresample   
2、 输入 make编译；(提示：make clean是清理项目文件) （大约七八分钟）
3、 输入make install安装。（大约一分钟）
如果静态库的configure配置，完成后，生成编译好的ffmpe库，有4个目录，bin、include、lib、share 。生成的头文件（*.h）及库（*.a，这是带有调试信息的静态库，可以改为*.lib）已经在ffmpeg源代码下的vs2013_build目录下，使用这个库，即可在VS下编译，且可以调试单步进入FFMPEG函数的内部，跟踪代码的执行情况。
如果是动态库配置，再编译和安装，则bin目录里面有*.lib和*.dll，在lib目录里面有*.def导出文件。开发阶段只需要工程中包含include和lib，运行阶段需要对应dll（注意两部分版本要一致）.
四、我的实例：验证VS2013编译和调试ffmpeg内部库函数
在 vs2013_build/share/ffmpeg/examples 目录下有若干个示例。
把 avio_reading.c 添加到我们的test_ffmpeg工程（vs2013），配置include与lib路径、依赖库。在链接库中加入文件（静态库调用ffmpeg）： 
ws2_32.lib;libavcodec.a;libavdevice.a;libavfilter.a;libavformat.a;libavutil.a;libswresample.a;libswscale.a
编译，链接，运行，OK！---》》可以加断点和跟踪调试，进入ffmpeg函数内部，实际是跟踪到了ffmpeg源代码目录下面对应的那些*.c文件了！！
 


Ffmpeg各模块的功能


FFmpeg是一套可以用来记录、转换数字音频、视频，并能将其转化为流的开源计算机程序。它包括了领先的音/视频编码库libavcodec等。
libavformat：用于各种音视频封装格式的生成和解析，包括获取解码所需信息以生成解码上下文结构
和读取音视频帧等功能；
libavcodec：用于各种类型声音/图像编解码；
libavutil：包含一些公共的工具函数；
libswscale：用于视频场景比例缩放、色彩映射转换；
libpostproc：用于后期效果处理；
ffmpeg：该项目提供的一个工具，可用于格式转换、解码或电视卡即时编码等；
ffsever：一个 HTTP 多媒体即时广播串流服务器；
ffplay：是一个简单的播放器，使用ffmpeg 库解析和解码，通过SDL显示；




ffmpeg编译选项汇总
===========================================================


在控制台输入如下命令可看到 ffmpeg 详细的编译选项。
[root@localhost ffmpeg]# ./configure --help
Usage: configure [options]
Options: [defaults in brackets after descriptions]


Help options:
  --help                   print this message
  --list-decoders          show all available decoders
  --list-encoders          show all available encoders
  --list-hwaccels          show all available hardware accelerators
  --list-demuxers          show all available demuxers
  --list-muxers            show all available muxers
  --list-parsers           show all available parsers
  --list-protocols         show all available protocols
  --list-bsfs              show all available bitstream filters
  --list-indevs            show all available input devices
  --list-outdevs           show all available output devices
  --list-filters           show all available filters


Standard options:
  --logfile=FILE           log tests and output to FILE [config.log]
  --disable-logging        do not log configure debug information
  --fatal-warnings         fail if any configure warning is generated
  --prefix=PREFIX          install in PREFIX []
  --bindir=DIR             install binaries in DIR [PREFIX/bin]
  --datadir=DIR            install data files in DIR [PREFIX/share/ffmpeg]
  --docdir=DIR             install documentation in DIR [PREFIX/share/doc/ffmpeg]
  --libdir=DIR             install libs in DIR [PREFIX/lib]
  --shlibdir=DIR           install shared libs in DIR [LIBDIR]
  --incdir=DIR             install includes in DIR [PREFIX/include]
  --mandir=DIR             install man page in DIR [PREFIX/share/man]
  --enable-rpath           use rpath to allow installing libraries in paths
                           not part of the dynamic linker search path
                           use rpath when linking programs [USE WITH CARE]


Licensing options:
  --enable-gpl             allow use of GPL code, the resulting libs
                           and binaries will be under GPL [no]
  --enable-version3        upgrade (L)GPL to version 3 [no]
  --enable-nonfree         allow use of nonfree code, the resulting libs
                           and binaries will be unredistributable [no]


Configuration options:
  --disable-static         do not build static libraries [no]
  --enable-shared          build shared libraries [no]
  --enable-small           optimize for size instead of speed
  --disable-runtime-cpudetect disable detecting cpu capabilities at runtime (smaller binary)
  --enable-gray            enable full grayscale support (slower color)
  --disable-swscale-alpha  disable alpha channel support in swscale
  --disable-all            disable building components, libraries and programs
  --enable-incompatible-libav-abi enable incompatible Libav fork ABI [no]
  --enable-raise-major     increase major version numbers in sonames [no]


Program options:
  --disable-programs       do not build command line programs
  --disable-ffmpeg         disable ffmpeg build
  --disable-ffplay         disable ffplay build
  --disable-ffprobe        disable ffprobe build
  --disable-ffserver       disable ffserver build


Documentation options:
  --disable-doc            do not build documentation
  --disable-htmlpages      do not build HTML documentation pages
  --disable-manpages       do not build man documentation pages
  --disable-podpages       do not build POD documentation pages
  --disable-txtpages       do not build text documentation pages


Component options:
  --disable-avdevice       disable libavdevice build
  --disable-avcodec        disable libavcodec build
  --disable-avformat       disable libavformat build
  --disable-avutil         disable libavutil build
  --disable-swresample     disable libswresample build
  --disable-swscale        disable libswscale build
  --disable-postproc       disable libpostproc build
  --disable-avfilter       disable libavfilter build
  --enable-avresample      enable libavresample build [no]
  --disable-pthreads       disable pthreads [autodetect]
  --disable-w32threads     disable Win32 threads [autodetect]
  --disable-os2threads     disable OS/2 threads [autodetect]
  --disable-network        disable network support [no]
  --disable-dct            disable DCT code
  --disable-dwt            disable DWT code
  --disable-error-resilience disable error resilience code
  --disable-lsp            disable LSP code
  --disable-lzo            disable LZO decoder code
  --disable-mdct           disable MDCT code
  --disable-rdft           disable RDFT code
  --disable-fft            disable FFT code
  --disable-faan           disable floating point AAN (I)DCT code
  --disable-pixelutils     disable pixel utils in libavutil


Hardware accelerators:
  --disable-dxva2          disable DXVA2 code [autodetect]
  --disable-vaapi          disable VAAPI code [autodetect]
  --disable-vda            disable VDA code [autodetect]
  --disable-vdpau          disable VDPAU code [autodetect]


Individual component options:
  --disable-everything     disable all components listed below
  --disable-encoder=NAME   disable encoder NAME
  --enable-encoder=NAME    enable encoder NAME
  --disable-encoders       disable all encoders
  --disable-decoder=NAME   disable decoder NAME
  --enable-decoder=NAME    enable decoder NAME
  --disable-decoders       disable all decoders
  --disable-hwaccel=NAME   disable hwaccel NAME
  --enable-hwaccel=NAME    enable hwaccel NAME
  --disable-hwaccels       disable all hwaccels
  --disable-muxer=NAME     disable muxer NAME
  --enable-muxer=NAME      enable muxer NAME
  --disable-muxers         disable all muxers
  --disable-demuxer=NAME   disable demuxer NAME
  --enable-demuxer=NAME    enable demuxer NAME
  --disable-demuxers       disable all demuxers
  --enable-parser=NAME     enable parser NAME
  --disable-parser=NAME    disable parser NAME
  --disable-parsers        disable all parsers
  --enable-bsf=NAME        enable bitstream filter NAME
  --disable-bsf=NAME       disable bitstream filter NAME
  --disable-bsfs           disable all bitstream filters
  --enable-protocol=NAME   enable protocol NAME
  --disable-protocol=NAME  disable protocol NAME
  --disable-protocols      disable all protocols
  --enable-indev=NAME      enable input device NAME
  --disable-indev=NAME     disable input device NAME
  --disable-indevs         disable input devices
  --enable-outdev=NAME     enable output device NAME
  --disable-outdev=NAME    disable output device NAME
  --disable-outdevs        disable output devices
  --disable-devices        disable all devices
  --enable-filter=NAME     enable filter NAME
  --disable-filter=NAME    disable filter NAME
  --disable-filters        disable all filters


External library support:
  --enable-avisynth        enable reading of AviSynth script files [no]
  --disable-bzlib          disable bzlib [autodetect]
  --enable-fontconfig      enable fontconfig, useful for drawtext filter [no]
  --enable-frei0r          enable frei0r video filtering [no]
  --enable-gnutls          enable gnutls, needed for https support
                           if openssl is not used [no]
  --disable-iconv          disable iconv [autodetect]
  --enable-ladspa          enable LADSPA audio filtering [no]
  --enable-libaacplus      enable AAC+ encoding via libaacplus [no]
  --enable-libass          enable libass subtitles rendering,
                           needed for subtitles and ass filter [no]
  --enable-libbluray       enable BluRay reading using libbluray [no]
  --enable-libbs2b         enable bs2b DSP library [no]
  --enable-libcaca         enable textual display using libcaca [no]
  --enable-libcelt         enable CELT decoding via libcelt [no]
  --enable-libcdio         enable audio CD grabbing with libcdio [no]
  --enable-libdc1394       enable IIDC-1394 grabbing using libdc1394
                           and libraw1394 [no]
  --enable-libfaac         enable AAC encoding via libfaac [no]
  --enable-libfdk-aac      enable AAC de/encoding via libfdk-aac [no]
  --enable-libflite        enable flite (voice synthesis) support via libflite [no]
  --enable-libfreetype     enable libfreetype, needed for drawtext filter [no]
  --enable-libfribidi      enable libfribidi, improves drawtext filter [no]
  --enable-libgme          enable Game Music Emu via libgme [no]
  --enable-libgsm          enable GSM de/encoding via libgsm [no]
  --enable-libiec61883     enable iec61883 via libiec61883 [no]
  --enable-libilbc         enable iLBC de/encoding via libilbc [no]
  --enable-libmfx          enable HW acceleration through libmfx
  --enable-libmodplug      enable ModPlug via libmodplug [no]
  --enable-libmp3lame      enable MP3 encoding via libmp3lame [no]
  --enable-libnut          enable NUT (de)muxing via libnut,
                           native (de)muxer exists [no]
  --enable-libopencore-amrnb enable AMR-NB de/encoding via libopencore-amrnb [no]
  --enable-libopencore-amrwb enable AMR-WB decoding via libopencore-amrwb [no]
  --enable-libopencv       enable video filtering via libopencv [no]
  --enable-libopenh264     enable H.264 encoding via OpenH264 [no]
  --enable-libopenjpeg     enable JPEG 2000 de/encoding via OpenJPEG [no]
  --enable-libopus         enable Opus de/encoding via libopus [no]
  --enable-libpulse        enable Pulseaudio input via libpulse [no]
  --enable-libquvi         enable quvi input via libquvi [no]
  --enable-librtmp         enable RTMP[E] support via librtmp [no]
  --enable-libschroedinger enable Dirac de/encoding via libschroedinger [no]
  --enable-libshine        enable fixed-point MP3 encoding via libshine [no]
  --enable-libsmbclient    enable Samba protocol via libsmbclient [no]
  --enable-libsoxr         enable Include libsoxr resampling [no]
  --enable-libspeex        enable Speex de/encoding via libspeex [no]
  --enable-libssh          enable SFTP protocol via libssh [no]
  --enable-libstagefright-h264  enable H.264 decoding via libstagefright [no]
  --enable-libtheora       enable Theora encoding via libtheora [no]
  --enable-libtwolame      enable MP2 encoding via libtwolame [no]
  --enable-libutvideo      enable Ut Video encoding and decoding via libutvideo [no]
  --enable-libv4l2         enable libv4l2/v4l-utils [no]
  --enable-libvidstab      enable video stabilization using vid.stab [no]
  --enable-libvo-aacenc    enable AAC encoding via libvo-aacenc [no]
  --enable-libvo-amrwbenc  enable AMR-WB encoding via libvo-amrwbenc [no]
  --enable-libvorbis       enable Vorbis en/decoding via libvorbis,
                           native implementation exists [no]
  --enable-libvpx          enable VP8 and VP9 de/encoding via libvpx [no]
  --enable-libwavpack      enable wavpack encoding via libwavpack [no]
  --enable-libwebp         enable WebP encoding via libwebp [no]
  --enable-libx264         enable H.264 encoding via x264 [no]
  --enable-libx265         enable HEVC encoding via x265 [no]
  --enable-libxavs         enable AVS encoding via xavs [no]
  --enable-libxcb          enable X11 grabbing using XCB [autodetect]
  --enable-libxcb-shm      enable X11 grabbing shm communication [autodetect]
  --enable-libxcb-xfixes   enable X11 grabbing mouse rendering [autodetect]
  --enable-libxcb-shape    enable X11 grabbing shape rendering [autodetect]
  --enable-libxvid         enable Xvid encoding via xvidcore,
                           native MPEG-4/Xvid encoder exists [no]
  --enable-libzmq          enable message passing via libzmq [no]
  --enable-libzvbi         enable teletext support via libzvbi [no]
  --disable-lzma           disable lzma [autodetect]
  --enable-decklink        enable Blackmagick DeckLink I/O support [no]
  --enable-nvenc           enable NVIDIA NVENC support [no]
  --enable-openal          enable OpenAL 1.1 capture support [no]
  --enable-opencl          enable OpenCL code
  --enable-opengl          enable OpenGL rendering [no]
  --enable-openssl         enable openssl, needed for https support
                           if gnutls is not used [no]
  --disable-sdl            disable sdl [autodetect]
  --enable-x11grab         enable X11 grabbing (legacy) [no]
  --disable-xlib           disable xlib [autodetect]
  --disable-zlib           disable zlib [autodetect]


Toolchain options:
  --arch=ARCH              select architecture []
  --cpu=CPU                select the minimum required CPU (affects
                           instruction selection, may crash on older CPUs)
  --cross-prefix=PREFIX    use PREFIX for compilation tools []
  --progs-suffix=SUFFIX    program name suffix []
  --enable-cross-compile   assume a cross-compiler is used
  --sysroot=PATH           root of cross-build tree
  --sysinclude=PATH        location of cross-build system headers
  --target-os=OS           compiler targets OS []
  --target-exec=CMD        command to run executables on target
  --target-path=DIR        path to view of build directory on target
  --target-samples=DIR     path to samples directory on target
  --tempprefix=PATH        force fixed dir/prefix instead of mktemp for checks
  --toolchain=NAME         set tool defaults according to NAME
  --nm=NM                  use nm tool NM [nm -g]
  --ar=AR                  use archive tool AR [ar]
  --as=AS                  use assembler AS []
  --windres=WINDRES        use windows resource compiler WINDRES [windres]
  --yasmexe=EXE            use yasm-compatible assembler EXE [yasm]
  --cc=CC                  use C compiler CC [gcc]
  --cxx=CXX                use C compiler CXX [g++]
  --dep-cc=DEPCC           use dependency generator DEPCC [gcc]
  --ld=LD                  use linker LD []
  --pkg-config=PKGCONFIG   use pkg-config tool PKGCONFIG [pkg-config]
  --pkg-config-flags=FLAGS pass additional flags to pkgconf []
  --ranlib=RANLIB          use ranlib RANLIB [ranlib -D]
  --doxygen=DOXYGEN        use DOXYGEN to generate API doc [doxygen]
  --host-cc=HOSTCC         use host C compiler HOSTCC
  --host-cflags=HCFLAGS    use HCFLAGS when compiling for host
  --host-cppflags=HCPPFLAGS use HCPPFLAGS when compiling for host
  --host-ld=HOSTLD         use host linker HOSTLD
  --host-ldflags=HLDFLAGS  use HLDFLAGS when linking for host
  --host-libs=HLIBS        use libs HLIBS when linking for host
  --host-os=OS             compiler host OS []
  --extra-cflags=ECFLAGS   add ECFLAGS to CFLAGS []
  --extra-cxxflags=ECFLAGS add ECFLAGS to CXXFLAGS []
  --extra-ldflags=ELDFLAGS add ELDFLAGS to LDFLAGS []
  --extra-ldexeflags=ELDFLAGS add ELDFLAGS to LDEXEFLAGS []
  --extra-libs=ELIBS       add ELIBS []
  --extra-version=STRING   version string suffix []
  --optflags=OPTFLAGS      override optimization-related compiler flags
  --build-suffix=SUFFIX    library name suffix []
  --enable-pic             build position-independent code
  --enable-thumb           compile for Thumb instruction set
  --enable-lto             use link-time optimization


Advanced options (experts only):
  --malloc-prefix=PREFIX   prefix malloc and related names with PREFIX
  --disable-symver         disable symbol versioning
  --enable-hardcoded-tables use hardcoded tables instead of runtime generation
  --disable-safe-bitstream-reader
                           disable buffer boundary checking in bitreaders
                           (faster, but may crash)
  --enable-memalign-hack   emulate memalign, interferes with memory debuggers
  --sws-max-filter-size=N  the max filter size swscale uses [256]


Optimization options (experts only):
  --disable-asm            disable all assembly optimizations
  --disable-altivec        disable AltiVec optimizations
  --disable-amd3dnow       disable 3DNow! optimizations
  --disable-amd3dnowext    disable 3DNow! extended optimizations
  --disable-mmx            disable MMX optimizations
  --disable-mmxext         disable MMXEXT optimizations
  --disable-sse            disable SSE optimizations
  --disable-sse2           disable SSE2 optimizations
  --disable-sse3           disable SSE3 optimizations
  --disable-ssse3          disable SSSE3 optimizations
  --disable-sse4           disable SSE4 optimizations
  --disable-sse42          disable SSE4.2 optimizations
  --disable-avx            disable AVX optimizations
  --disable-xop            disable XOP optimizations
  --disable-fma3           disable FMA3 optimizations
  --disable-fma4           disable FMA4 optimizations
  --disable-avx2           disable AVX2 optimizations
  --disable-armv5te        disable armv5te optimizations
  --disable-armv6          disable armv6 optimizations
  --disable-armv6t2        disable armv6t2 optimizations
  --disable-vfp            disable VFP optimizations
  --disable-neon           disable NEON optimizations
  --disable-inline-asm     disable use of inline assembly
  --disable-yasm           disable use of nasm/yasm assembly
  --disable-mipsdspr1      disable MIPS DSP ASE R1 optimizations
  --disable-mipsdspr2      disable MIPS DSP ASE R2 optimizations
  --disable-mipsfpu        disable floating point MIPS optimizations
  --disable-fast-unaligned consider unaligned accesses slow


Developer options (useful when working on FFmpeg itself):
  --disable-debug          disable debugging symbols
  --enable-debug=LEVEL     set the debug level []
  --disable-optimizations  disable compiler optimizations
  --enable-extra-warnings  enable more compiler warnings
  --disable-stripping      disable stripping of executables and shared libraries
  --assert-level=level     0(default), 1 or 2, amount of assertion testing,
                           2 causes a slowdown at runtime.
  --enable-memory-poisoning fill heap uninitialized allocated space with arbitrary data
  --valgrind=VALGRIND      run "make fate" tests through valgrind to detect memory
                           leaks and errors, using the specified valgrind binary.
                           Cannot be combined with --target-exec
  --enable-ftrapv          Trap arithmetic overflows
  --samples=PATH           location of test samples for FATE, if not set use
                           $FATE_SAMPLES at make invocation time.
  --enable-neon-clobber-test check NEON registers for clobbering (should be
                           used only for debugging purposes)
  --enable-xmm-clobber-test check XMM registers for clobbering (Win64-only;
                           should be used only for debugging purposes)
  --enable-random          randomly enable/disable components
  --disable-random
  --enable-random=LIST     randomly enable/disable specific components or
  --disable-random=LIST    component groups. LIST is a comma-separated list
                           of NAME[:PROB] entries where NAME is a component
                           (group) and PROB the probability associated with
                           NAME (default 0.5).
  --random-seed=VALUE      seed value for --enable/disable-random


NOTE: Object files are built at the place where configure is launched.
-----------------------------------------------------------------------------------