migoller/ffmpeg repository overview

⁠FFmpeg Docker image

Minimalistic Docker image with FFmpeg⁠ and support for hardware acceleration⁠ based on archlinux⁠. You can install the latest build of this image by running docker pull migoller/ffmpeg.

This image can be used as a base for an encoding farm or for Shinobi CCTV⁠ hardware accelerated Docker images. The image is scheduled for weekly builds on every Sunday.

Please use GitLab issues⁠ to report any bug or missing feature.

Feel free to contribute on my GitLab FFmpeg Docker image repository⁠.

⁠Examples

⁠Encode rtsp stream into a MP4

Let's encode a rtsp stream into an MP4 out_nohwaccel.mp4 is located in the current directory.

$ docker run -it -rm -v $PWD:/tmp \
        migoller/ffmpeg \
        -loglevel info -stats \
        -i rtsp://<rtsp url> \
        -vf 'scale=640:480' \
        -preset ultrafast \
        -c:v libx264 \
        -t 20 -r 8 -f mp4 /tmp/out_nohwaccel.mp4

⁠Encode rtsp stream into a MP4 with hardware acceleration

Now let's encode a rtsp stream into an MP4 out_hwaccel.mp4 is located in the current directory with VA-API hardware acceleration of an Intel i965 compatible CPU with Intel HD GPU.

$ docker run -it -rm --device /dev/dri:/dev/dri -v $PWD:/tmp -e LIBVA_DRIVER_NAME="i965" \
        migoller/ffmpeg \
        -loglevel info -stats \
        -i rtsp://<rtsp url> \
        -vaapi_device /dev/dri/renderD128 \
        -vf 'hwupload,scale_vaapi=w=640:h=480:format=nv12' \
        -preset ultrafast \
        -c:v h264_vaapi \
        -t 20 -r 8 -f mp4 /tmp/out_hwaccel.mp4

⁠Hardware video acceleration

Hardware video acceleration makes it possible for the video card to decode/encode video, thus offloading the CPU and saving power:

• Video Acceleration API (VA-API) provides both hardware accelerated video encoding and decoding, designed for Intel CPU / GPU.
• Video Decode and Presentation API for Unix (VDPAU) allows to offload portions of the video decoding process and video post-processing to the GPU video-hardware, developed by NVIDIA. VDPAU has not been updated since September 2015.
• NVDECODE/NVENCODE - hardware video acceleration used by NVIDIA Fermi, Kepler, Maxwell and Pascal generation GPUs.

Please have a look at archlinux Hardware video acceleration⁠ for further details.

⁠Intel

⁠Intel graphics open-source drivers support VA-API:

• HD Graphics series starting from CannonLake (or optionally from Broadwell) and newer are supported by intel-media-driver.
• GMA 4500 series and newer GPUs up to Coffee Lake are supported by libva-intel-driver.

⁠NVIDIA

⁠Nouveau open-source driver supports both VA-API and VDPAU:

• GeForce 8 series and newer GPUs up until GeForce GTX 750 are supported by libva-mesa-driver and mesa-vdpau.

⁠NVIDIA proprietary driver supports via nvidia-utils:

• VDPAU on GeForce 8 series and newer GPUs
• NVDECODE on Fermi and newer GPUs
• NVENCODE on Kepler and newer GPUs

⁠ATI/AMD

⁠ATI and AMDGPU open-source drivers support both VA-API and VDPAU:

• VDPAU on Radeon R300 and newer GPUs is supported by mesa-vdpau.
• VA-API on Radeon HD 2000 and newer GPUs is supported by libva-mesa-driver.

⁠Translation layers

⁠To get VA-API support when device driver provides none:

• libva-vdpau-driver – VDPAU backend for VA-API.

⁠To get VDPAU support when device driver provides none:

• libvdpau-va-gl – VA-API backend for VDPAU, only H.264 support.

⁠Configuration

To use the hardware acceleration features you have to configure your Docker host and the container to match the required drivers.

⁠Docker host configuration

Have your GPUs´ drivers up and running on your host.

Ensure that your graphics card has been recognized correctly running vainfo, vdpauinfo or nvidia-utils.

⁠Verifying VA-API drivers

Run vainfo to check that your GPU is recognized correctly.

$ vainfo
error: XDG_RUNTIME_DIR not set in the environment.
error: can't connect to X server!
libva info: VA-API version 0.39.4
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_39
libva info: va_openDriver() returns 0
vainfo: VA-API version: 0.39 (libva 1.7.3)
vainfo: Driver version: Intel i965 driver for Intel(R) Skylake - 1.7.0
vainfo: Supported profile and entrypoints
      VAProfileMPEG2Simple            : VAEntrypointVLD
      VAProfileMPEG2Simple            : VAEntrypointEncSlice
      VAProfileMPEG2Main              : VAEntrypointVLD
      VAProfileMPEG2Main              : VAEntrypointEncSlice
      VAProfileH264ConstrainedBaseline: VAEntrypointVLD
      VAProfileH264ConstrainedBaseline: VAEntrypointEncSlice
      VAProfileH264Main               : VAEntrypointVLD
      VAProfileH264Main               : VAEntrypointEncSlice
      VAProfileH264High               : VAEntrypointVLD
      VAProfileH264High               : VAEntrypointEncSlice
      VAProfileH264MultiviewHigh      : VAEntrypointVLD
      VAProfileH264MultiviewHigh      : VAEntrypointEncSlice
      VAProfileH264StereoHigh         : VAEntrypointVLD
      VAProfileH264StereoHigh         : VAEntrypointEncSlice
      VAProfileVC1Simple              : VAEntrypointVLD
      VAProfileVC1Main                : VAEntrypointVLD
      VAProfileVC1Advanced            : VAEntrypointVLD
      VAProfileNone                   : VAEntrypointVideoProc
      VAProfileJPEGBaseline           : VAEntrypointVLD
      VAProfileJPEGBaseline           : VAEntrypointEncPicture
      VAProfileVP8Version0_3          : VAEntrypointVLD
      VAProfileVP8Version0_3          : VAEntrypointEncSlice
      VAProfileHEVCMain               : VAEntrypointVLD
      VAProfileHEVCMain               : VAEntrypointEncSlice

⁠Verifying VDPAU

Run vdpauinfo to check that your GPU is recognized correctly.

$ vdpauinfo
display: :0   screen: 0
API version: 1
Information string: G3DVL VDPAU Driver Shared Library version 1.0

Video surface:

name   width height types

420    16384 16384  NV12 YV12 
422    16384 16384  UYVY YUYV 
444    16384 16384  Y8U8V8A8 V8U8Y8A8 

Decoder capabilities:

name                        level macbs width height

MPEG1                          --- not supported ---
MPEG2_SIMPLE                    3  9216  2048  1152
MPEG2_MAIN                      3  9216  2048  1152
H264_BASELINE                  41  9216  2048  1152
H264_MAIN                      41  9216  2048  1152
H264_HIGH                      41  9216  2048  1152
VC1_SIMPLE                      1  9216  2048  1152
VC1_MAIN                        2  9216  2048  1152
VC1_ADVANCED                    4  9216  2048  1152

⁠FFmpeg container configuration

Run the container with the device attached /dev/dri from your Docker host into the container and set the environment accordingly to the GPU driver.

$ docker run --device /dev/dri:/dev/dri -e LIBVA_DRIVER_NAME="i965" [...] migoller/ffmpeg [...]`

Run the FFmpeg container with the correct parameters like running ffmpeg natively⁠.

⁠Configuring VA-API

You can override the driver for VA-API by using the LIBVA_DRIVER_NAME environment variable; the default value is set to i965. The correct driver name depends on your setup.

⁠Intel graphics:

• For libva-intel-driver set to i965.
• For intel-media-driver set to iHD.

⁠NVIDIA:

• For Nouveau set to nouveau.
• For NVIDIA set toe vdpau.

⁠ATI/AMD:

• For AMDGPU driver set to radeonsi.

⁠Configuring VDPAU

You need to set VDPAU_DRIVER environment variable to point to correct driver. The correct driver name depends on your setup.

⁠Intel graphics:

• For Intel Graphics or AMD Catalyst you need to set it to va_gl.

⁠NVIDIA:

• For Nouveau driver set it to nouveau.
• For NVIDIA's proprietary version set it to nvidia.

⁠ATI/AMD:

• For AMDGPU driver set it to the proper driver version depending on your GPU. Run vdpauinfo to determine the driver version of your GPU.