Real time H.264 Encoding and Decoding
Using FFmpeg and x264
Introduction
Thisdocumentation serves as a quick overview on the first development phaseof our wireless video compression and transmission testbed, i.e. H.264realtime encoding, transmission, and decoding using two open sourceprojects FFmpeg and x264.
FFmpeg is a comprehensive multimediaencoding and decoding library that consists of numerous audio, video,and container formats. For H.264 video encoding, FFmpeg uses externalopen source library x264. In our testbed, FFmpeg works as an encodingand decoding front-end, while x264 is the encoding engine. At theencoder side, frames are captured by camera and then encoded into H.264bitstream in real time. Encoded packets are transmitted over UDPpackets to the receiver. At the receiver side, packets are received,decoded, and displayed in real time too. We have also pruned FFmpegsuch that source codes of unrelated codecs and formats have beenremoved from the package. It will make our future development easier.
Infollowing sections, we will give brief introduction on FFmpeg and x264,as well as necessary external libraries in Section 2. We will alsointroduce how to use our software to perform real time H.264 encoding,transmission, and decoding in Section 3. This documentation is notintended to be comprehensive. Instead, hopefully it will provide enoughinformation to interested readers to quickly set up and try out thesoftware, or even to do their own development.
Source Packages and Libraries
x264
x264 is a free library for encoding H264/AVC video streams. It is used by many other projects. Below is its official website.
Website: http://www.videolan.org/developers/x264.html
Download: git clone git://git.videolan.org/x264.git
Followingcommands are used to build debug and release version of x264. Aftersuccessful configurations, use “make” and “make install” to compile andinstall x264.
debug version: ./configure --disable-asm --enable-debug --enable-pic
--enable-shared --disable-mp4-output
release version: ./configure --enable-pic --enable-shared
--disable-mp4-output
FFmpeg
FFmpegis a complete open source solution to record, convert and stream audioand video. It includes libavcodec, the leading audio/video codeclibrary used by several other projects. It also includes libavformat,an audio/video container mux and demux library. FFmpeg is developedunder Linux, but it can compiled under most operating systems,including Windows. For detailed
information and downloading, please visit following website.
Website: http://ffmpeg.mplayerhq.hu/
Download: git clone git://git.mplayerhq.hu/ffmpeg/
cd ffmpeg
git clone git://git.mplayerhq.hu/libswscale/
Following commands are used to build debug and release version of FFmpeg. Note that we must
configureFFmpeg to use x264 library. After successful configurations, use “make”and “make install” to compile and install FFmpeg.
debug version: ./configure --enable-gpl --enable-libx264 --enable-swscale
--enable-debug --disable-optimizations --disable-stripping
release version: ./configure --enable-gpl --enable-libx264 --enable-swscale
Git
Git is an open source version control system designed to handle very large projects with speed
and efficiency, but just as well suited for small personal repositories. We need Git to download
FFmpeg and x264 source codes.
Website: http://git.or.cz/
Download: http://www.kernel.org/pub/software/scm/git/
Subversion (SVN)
Subversion is an open source version control system. It is used to maintain current and historical
versions of files such as source code, web pages, and documentation. Subversion is well-known in
the open source community and is used on many open source projects and we need Subversion to down Yasm and SDL.
Website: http://subversion.tigris.org/
Download: http://subversion.tigris.org/getting.html#source-release
Yasm
Yasmis a complete rewrite of the NASM assembler under the “new” BSD License(some portions are under other licenses, see COPYING for details). BothFFmpeg and x264 uses Yasm.
Website: http://www.tortall.net/projects/yasm/
Download: svn co http://www.tortall.net/svn/yasm/trunk/yasm yasm
Simple DirectMedia Layer (SDL)
SimpleDirectMedia Layer is a cross-platform multimedia library designed toprovide low level access to audio, keyboard, mouse, joystick, 3Dhardware via OpenGL, and 2D video framebuffer. It is used by MPEGplayback software, emulators, and many popular games and FFmpeg usesSDL for displaying decoded video contents.
Website: http://www.libsdl.org/
Download: svn checkout http://svn.libsdl.org/branches/SDL-1.2
To Use FFmpeg and FFplay
Encoding
Encoding is done by binary ffmpeg in our trimmed FFmpeg package.It calls x264 library to encode frames captured by camera into H.264bitstreams. We have modified ffmpeg such that encoded bitstreams aresent to receiver over UDP packets, or saved locally, or both. A commoncommand line input is shown below. In this example, the input of theencoder is camera captured frames with CIF resolution. Frame rate is10fps. The output is H.264 bitstream with bitrate at 50kbps. Intraframe interval is set to 100. The encodeded bitstream is both sent tolocalhost with port number 12349, and saved as a local file out.264 in/bs directory.
./ffmpeg -g 100 -f video4linux -b 50k -s cif -r 10 -i /dev/video0
-vcodec libx264 -y -dest_ip 127.0.0.1 -dest_port 12349 -f h264../bs/out.264
The command below only sends encoded bitstream over UDP packets:
./ffmpeg -g 100 -f video4linux -b 50k -s cif -r 10 -i /dev/video0
-vcodec libx264 -y -dest_ip 127.0.0.1 -dest_port 12349 -f h264
The command below only saves encoded bitstream to a local file:
./ffmpeg -g 100 -f video4linux -b 50k -s cif -r 10 -i /dev/video0
-vcodec libx264 -y -f h264 ../bs/out.264
In Table 1 we list some frequently used options of ffmpeg.
Decoding
Decoding is done by binary ffplay in our trimmed FFmpeg package.It decodes H.264 bitstream and displays in real time. We have modifiedffplay such that it can both decodes local H.264 file and UDP packetscontaining H.264 bitstream, but not simultaneously. The command belowdecodes H.264 bitstream in UDP packets assuming port number 12349 isused. Combined with the first ffmpeg command line example, real timeH.264 encoding and decoding will be performed on the same PC.
./ffplay -dest_port 12349 -f h264
The command below decodes local H.264 bitstream assuming the out.264 file is located in /bs directory.
./ffplay ../bs/out.264
Below we list some frequently used options and hot keys of ffplay.
Options:
- -h: show help
- -f: force format
- -s: set frame size (WxH or abbreviation)
- -dest_port: specify destination port number
- -fs: force full screen
- -x: force displayed width
- -y: force displayed height
Hot keys:
- q/ESC: quit
- f: toggle full screen
- p/SPC: pause