Compiling FFmpeg on the Raspberry Pi?

The Raspberry Pi is a popular single-board computer used by hobbyists and professionals alike for a variety of projects. One common use case is to set up the Pi as a low-power media server or embedded device that can encode video files. To enable advanced video encoding capabilities, the FFmpeg multimedia framework can be compiled to run efficiently on the Raspberry Pi.

Compiling FFmpeg on the Raspberry Pi?

Gather Required Software and Dependencies

FFmpeg relies on other open-source libraries like x264 and fdk-aac to enable support for encoding formats like H.264 video and AAC audio. We will need to install development packages for these libraries before compiling FFmpeg:

  • sudo apt update
  • sudo apt install autoconf automake build-essential libass-dev libfreetype6-dev libvorbis-dev libx264-dev libtheora-dev libvpx-dev libfdk-aac-dev libmp3lame-dev

This will ensure your Pi has all the necessary dependencies to build FFmpeg.

Downloading and Unpacking the FFmpeg Source Code

Now we can download and unpack the latest FFmpeg source code:


tar xjvf ffmpeg-snapshot.tar.bz2

This will extract the files into a directory like ffmpeg. We will run the compilation process inside this extracted directory.

Configuring the Compile Options

Before compiling, we need to configure FFmpeg’s build options. This allows selecting which codecs, formats, and filters should be included.

Run the configuration script with the following options:

cd ffmpeg

./configure \

  –prefix=”$HOME/ffmpeg_build” \ 

  –extra-cflags=”-I$HOME/ffmpeg_build/include” \

  –extra-ldflags=”-L$HOME/ffmpeg_build/lib” \

  –bindir=”$HOME/bin” \

  –enable-gpl \  

  –enable-libass \

  –enable-libfdk-aac \

  –enable-libfreetype \

  –enable-libmp3lame \  

  –enable-libtheora \

  –enable-libvorbis \

  –enable-libvpx \

  –enable-libx264 \


This configures FFmpeg with the dependencies we installed earlier and enables support for proprietary codecs for maximum encoding capabilities.

The –prefix option sets the install location to $HOME/ffmpeg_build where FFmpeg’s headers and libraries will reside after compiling. The –bindir option makes the FFmpeg executable available systemwide at $HOME/bin.

Running the Compile Process

Now we start the actual compilation process:

make -j4

make install

hash -r

The -j flag compiles using 4 concurrent processes to speed up building on the quad core Pi.

make install copies the finished FFmpeg executable along with headers, libs, and tools to the ffmpeg_build folder.

hash -r refreshes bash so the new FFmpeg command is detected at $HOME/bin.

Testing the Compiled FFmpeg

We can verify FFmpeg finished compiling properly and has the expected codec capabilities:

ffmpeg -encoders

This will list all the enabled encoders like libx264 for H.264 video.

Finally, let’s compress a video file as a test:

ffmpeg -i input_video.mkv -codec:v libx264 -crf 28 output_video.mp4

The Pi CPU will be pushed to 100% for a few minutes, but it should successfully encode the video in H.264 format! FFmpeg is now ready for your video processing projects on the Pi.

Key Takeaways

  • Installing development packages allows FFmpeg to benefit from libraries like x264 for maximum codec support
  • Carefully configuring FFmpeg’s compile options enables features needed for your use case
  • Using parallel compilation speeds up the build process significantly
  • Testing encoding capabilities confirms FFmpeg was compiled properly

FFmpeg compilation opens up many possibilities like compressing videos for a Pi-based media center. Understanding the compile process helps debug issues and optimize performance.


Compiling complex software like FFmpeg can seem daunting at first on Linux platforms like the Raspberry Pi. By systematically setting up dependencies, configuring options, running the compile process, and testing, we can build FFmpeg to take full advantage of the Pi’s quad-core CPU. FFmpeg enables projects like automating video encoding jobs or setting up a home media server with broad format support. Learning best practices for compilation sets a foundation for customizing other Linux software to perfectly fit your needs.


  1. How to compile FFmpeg on Ubuntu?
    The process for compiling FFmpeg on Ubuntu is very similar as on the Raspberry Pi. Simply install the required development packages, download and unpack the FFmpeg source code, configure it with the desired options, run make, then copy the ffmpeg executable to a folder in your PATH like /usr/local/bin. Ubuntu’s beefier hardware can compile FFmpeg even faster.

  2. How to enable h264 encoding in FFmpeg?
    To enable h264 video encoding support in FFmpeg, pass the –enable-libx264 flag during the ./configure step. This enables FFmpeg to leverage hardware acceleration abilities of the x264 video coding library during h264 encoding and decoding. Make sure the libx264-dev package is installed first.

  3. How long does it take to compile FFmpeg?
    The time to compile FFmpeg can range from 10 minutes to a couple hours depending on factors like: CPU cores/speed, enable codecs/formats, optimization flags passed to the compiler, and if distributing the build job across multiple machines. On a Raspberry Pi 4, compiling FFmpeg with common encoding options takes 20-30 minutes. Building only the base ffmpeg executable is faster than enabling additional codec capabilities.

  4. Does compiling improve FFmpeg performance?
    In most cases, compiling FFmpeg specifically for your hardware and with custom options enabled does improve performance compared to using a generic build. Optimization flags tailored for your CPU architecture can be passed during compile configuration. Unnecessary bloat can be avoided by only enabling needed codecs/formats. And threading/hardware acceleration from libraries likes fdkaac, x264, or libvpx can be taken full advantage of.

  5. Can I cross-compile FFmpeg for the Raspberry Pi?
    Yes, FFmpeg can be cross-compiled instead of needing to compile directly on the Pi. This allows leveraging a faster desktop machine to build FFmpeg targeting the RPi CPU architecture. The essential steps are installing the Pi toolchain on your build system, then configuring FFmpeg using the –target-os=linux –arch=arm or similar cross-compile oriented flags.

  6. What is the difference between shared and static FFmpeg builds?
    The shared FFmpeg build links dynamically to the external libraries it relies on like fontconfig, x264, or libass. This results in a smaller ffmpeg binary size, but the required dependency libs must be present on the target system. The static build bundles the external libraries into the ffmpeg executable itself, allowing it to be portable across machines without needing to install the dependencies separately. But this increases binary size.

  7. How to enable GPU hardware acceleration with FFmpeg?
    FFmpeg can offload aspects of video encoding, decoding, and filtering to the GPU hardware acceleration abilities on modern systems – both integrated graphics on Intel chips as well as dedicated Nvidia/AMD graphics cards. Pass the –enable-cuda flag during configuration to leverage Nvidia GPUs. GPU processing depends on the underlying encoder being used. For example, h264_nvenc uses Nvidia GPUs while h264_vaapi uses Intel iGPUs.

  8. Can I use FFmpeg without installing it?
    FFmpeg provides static builds available for download that do not require installation. These contain all the internal libraries bundled with the ffmpeg program so external dependencies do not need to be present on the system. This allows using ffmpeg conveniently without compilers/dev tools needed to install from source. The tradeoff is binary size is larger.

  9. How much disk space does compiling FFmpeg use?
    Compiling FFmpeg and all its dependencies can use over 10 GB of temporary disk space during the build process. This allows simultaneously juggling extracted source code, compiling source into object files, linking object files into libraries, etc across multiple concurrent build jobs. Use df -h periodically to monitor disk space during compilation to ensure enough free space exists on the volume.

  10. Does FFplay require FFmpeg?
    Yes, FFplay is not standalone program and depends on a compiled FFmpeg for access to the protocol parsing, codecs, filters, and other pipeline software libraries that make up the full FFmpeg framework. So an FFmpeg compilation must be present in order for FFplay to function properly during video playback. FFmpeg handles all the complex media processing while FFplay simply provides the user interface.

  11. What codec does YouTube use?
    As of 2023, YouTube defaults to the AV1 video codec for newly uploaded content. AV1 can deliver better quality than AVC/H.264 at lower bitrates thanks to improved compression efficiency. However, many older videos are still encoded in AVC/H.264 or VP9. YouTube utilizes the libvpx library for encoding in VP9, and rav1e for slower encoding times with AV1. Playback uses the user’s browser and hardware capabilities to decode whichever codec the video is stored in.

  12. How to enable experimental codecs?
    Enabling experimental and nascent codec implementations that are not yet mature can reveal more encoding options to test and evaluate. Pass the –enable-experimental flag during configuration before compiling FFmpeg to include cutting-edge implementations that are still in development, but can provide speed or efficiency benefits. Being experimental means quality and stability are riskier.

  13. Does multithreaded encoding speed up FFmpeg?
    Yes, running FFmpeg’s encoding pipelines across multiple threads can significantly expedite media compression and video transcoding. By default FFmpeg only uses 1 thread. Use the -threads command line flag or set the FFMPEG_THREADS environment variable to a reasonable value based on your CPU cores before launching CPU-intensive FFmpeg operations. Benchmark various thread counts to find the optimal value.

  14. How to enable hardware decoding with FFmpeg?
    FFmpeg is capable of offloading aspects of video decoding to available hardware accelerator engines, rather than only using software CPU decoding. For Intel integrated graphics add –enable-vaapi configure option. For AMD Radeon and Nvidia graphics card support –enable-vdpau and –enable-cuda respectively. Verify decoder offload engagement by passing -hwaccel cuda -hwaccel_output_format cuda as parameters during playback.

  15. How to configure FFmpeg for streaming?
    To configure FFmpeg specifically for video streaming scenarios, key configuration flags to pass are:–enable-gmp –enable-librtmp –enable-libssh –enable-libxcb –enable-libcdio –enable-gnutls –enable-libxml2 –enable-libspeex –enable-libopenjpeg –enable-libvorbis –enable-libopus –enable-libtheora –enable-frei0r –enable-libass –enable-libfreetype –disable-indev=v4l2 –enable-openssl –enable-libvpx –enable-libx264 –enable-libpulse –enable-libfontconfig –enable-libfribidi

  16. How to find hardware acceleration support on Linux?
    Use vdpauinfo, vdpaudec, or vainfo commands to probe GPU video decoding abilities. For Intel QuickSync acceleration confirmation check for supported encode/decode profiles. The FFmpeg log prints enabled hardware acceleration capabilities or can force software fallback if desired codecs are unavailable in hardware.

  17. Can I cross-compile FFmpeg for Android?
    Yes, FFmpeg can be cross-compiled for Android by using the Android NDK and stand-alone toolchain targeting ARM architectures. Key configuration flags are –target-os=linux –arch=arm –enable-cross-compile –sysroot=$ANDROID_SYSROOT –cross-prefix=$TARGET- where $TARGET points to the Android toolchain path. This allows leveraging desktop machines for quicker builds.

  18. How to enable ProRes on FFmpeg?
    To add Apple ProRes video codec support during FFmpeg compilation, the options are:–enable-libaom –enable-gnutls –enable-libx264 –enable-libx265 –enable-libvpx –enable-libtheora –enable-libvorbis –enable-libvidstab –enable-libras –enable-sdl2 –enable-libopencore-amrnb –enable-libkvazaar –enable-libopenh264 –enable-encoder=libaom_av1 –enable-librav1e –enable-version3 –enable-libsvtav1 –enable-libdav1d  –enable-librtmp –enable-openssl –enable-libfdk-aac.

  19. What hardware works best for streaming with FFmpeg?
    For software like OBS streaming using FFmpeg as a backend, an Nvidia GPU provides excellent NVENC integration for optimal quality and performance. Otherwise any modern multi-core Intel or AMD CPU with integrated graphics can efficiently handle real-time FFmpeg transcoding workloads too by enabling hardware acceleration options during compilation.

  20. Does multithreading compress faster?
    Yes, using all available CPU cores via -threads parameter leveraging hyperthreading when encoding inter-frame codecs like H.264/H.265 can expedite compression time significantly with minimal quality impact. Single-threaded encoding is preferable for intra-frame only codecs like ProRes and DNxHR. Thoroughly benchmark various thread values against target codec and hardware.

Leave a Comment