The NVIDIA ffmpeg package supports hardware-accelerated decode on Jetson platforms


No, it is not supported on r32.4.4.
Since the implementation is based on jetson_multimedia_api, you may consider to use it and refer to 01_video_encode sample.

pi@pi-desktop:~/nvffmpeg/ffmpeg$ ffmpeg -decoders | grep h264
ffmpeg version N-103235-g1c5610824a Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 7 (Ubuntu/Linaro 7.5.0-3ubuntu1~18.04)
  configuration: --enable-nvv4l2dec --enable-libv4l2 --enable-shared --extra-libs='-L/usr/lib/aarch64-linux-gnu/tegra -lnvbuf_utils' --extra-cflags='-I /usr/src/jetson_multimedia_api/include/'
  libavutil      57.  4.100 / 57.  4.100
  libavcodec     59.  4.101 / 59.  4.101
  libavformat    59.  4.101 / 59.  4.101
  libavdevice    59.  0.100 / 59.  0.100
  libavfilter     8.  1.103 /  8.  1.103
  libswscale      6.  0.100 /  6.  0.100
  libswresample   4.  0.100 /  4.  0.100
 VFS..D h264                 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10
 V..... h264_v4l2m2m         V4L2 mem2mem H.264 decoder wrapper (codec h264)

As you can see h264_nvmpi does not exist.
You need the source code that can use h264_nvmpi .

1 Like

@DaneLLL why do you call this a supported version of ffmpeg??

This doesn’t support CUDA or hardware acceleration - or any of the important decoders… can you be clear about when NVIDIA will release a fully supported version of ffmpeg with hardware decoding?

I’ve tried everything to compile from source or use any of the NVIDIA governed repo’s - none of it today appears to support hwaccel cuda on NVIDIA Jetson Nano boards… it quite frankly has been a disappointing journey for me so far with these boards. I’ve read others have returned their boards under RMA because of the lacking hw support to ffmpeg. Even the official guide here: Using FFmpeg with NVIDIA GPU Hardware Acceleration :: NVIDIA Video Codec SDK Documentation doesn’t work!!! Why can this not be standard to your Jetson linux image distribution?! Sorry for the frustration - but I’ve been at this for a week, and also feel close to returning my board due to the lack of support to ffmpeg. Most of us just want hardware decoding or encoding using the board’s dedicated GPU so that its not hogging CPU resources… now I might as well use RPI…

ffmpge user

Disappointed with “NVIDIA Jetson nano”.
Do you have a return process?

Note that the version installed by nvidia-ffmpeg package is installed in /usr/local, so be sure to have this bin in your path before /usr/bin where standard ffmpeg package is installed. You may also run :

/usr/local/bin/ffmpeg -decoders | grep h264

In Jetson platforms, there is individual hardware decoder and it is named *_nvv4l2dec like h264_nvv4l2dec.
Please follow the steps and check if you can see the decoders:
Hardware accelerated video playback with L4T ffmpeg - #7 by DaneLLL