Recherche avancée

Sur d’autres sites (12369)

• avformat/movenc : allow writing out channel count in MP4 and 3GP

  7 février 2023, par Jan Ekström

  avformat/movenc : allow writing out channel count in MP4 and 3GP
  ISOBMFF (14496-12) made this field ('channelcount') in the
  
  AudioSampleEntry structure non-template¹ somewhere before the
  
  release of the 2022 edition. As for ETSI TS 126 244 AKA 3GPP
  
  file format (V16.1.0, 2020-10), it does not seem contain any
  
  references limiting the channelcount entry in AudioSampleEntry
  
  or in its own definition of EVSSampleEntry.
  fate-mov-mp4-chapters test had to be adjusted as it output a
  
  mono vorbis stream, which would now be properly marked as such
  
  in the container.
  1 : As per 14496-12 :
  
     Fields shown as “template” in the box descriptions are fields
  
     which are coded with a default value unless a derived
  
     specification defines their use and permits writers to use
  
     other values than the default.
  

  • [DH] libavformat/movenc.c
  • [DH] tests/ref/fate/mov-mp4-chapters

• Picking the right gpu for ffmpeg decoding/encoding with nvidia/cuda

  14 avril 2020, par Reza Mir

  I am writing code using ffmpeg/nvidia/cuda library to decode and encode some video. I am trying to pick one specific GPU for decoding/encoding among some existing GPUs installed on the system. Here is the list of existing GPUs :

  



   nvidia-smi 
Tue Apr 14 00:14:59 2020       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.64.00    Driver Version: 440.64.00    CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  On   | 00000000:1A:00.0 Off |                  N/A |
| 41%   33C    P8    20W / 260W |      0MiB / 11019MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX 108...  On   | 00000000:1B:00.0 Off |                  N/A |
| 41%   65C    P2   182W / 250W |   7167MiB / 11178MiB |     53%      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX 108...  On   | 00000000:1E:00.0 Off |                  N/A |
| 14%   55C    P2    61W / 250W |   1943MiB / 11178MiB |     26%      Default |
+-------------------------------+----------------------+----------------------+
|   3  Quadro P5000        On   | 00000000:3D:00.0 Off |                  Off |
| 26%   33C    P8     6W / 180W |      1MiB / 16278MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   4  TITAN RTX           On   | 00000000:3E:00.0 Off |                  N/A |
| 41%   38C    P0    72W / 280W |      0MiB / 24220MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   5  TITAN RTX           On   | 00000000:41:00.0 Off |                  N/A |
| 40%   32C    P8    14W / 280W |      0MiB / 24220MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+



  



  I would like to pick GPU #3 in my code and use it for decoding. I am using h264_nvenc and h264_cuvid as encoder and decoder in my program and setting the options of decoder/encoder using the following code :

  



  av_opt_set(decoder_context->priv_data, "gpu", "Quadro P5000", 0);


  



  But this code doesn't pick Quadro GPU for decoding/encoding. Instead, it always goes to GPU #4 in my test :

  



  nvidia-smi pmon | grep tx
    4      20096     C    22     3    44    16   tx            
    4      20096     C    26     4    49    17   tx            
    4      20096     C    25     4    49    17   tx     


  


• no difference between h264 and h264_cuvid

  10 juillet 2020, par Ruslan Ablyazov

  I use libav to decode an RTSP stream. I tried two code options : with auto codec selection and with hardcoded codec.

  


  pCodec = avcodec_find_decoder (pFormatCtx->streams[p_InputVideoStreamIndex]->codecpar->codec_id);&#xA;//pCodec = avcodec_find_decoder_by_name ("h264_cuvid");&#xA;&#xA;LOG (info) &lt;&lt; "Video codec name" &lt;&lt; pCodec->name &lt;&lt; ". Full name:" &lt;&lt; (pCodec->long_name? PCodec->long_name: "<null>");&#xA;</null>

  


  The avcodec_find_decoder() function returns h264 codec. The problem is that the speed and load on the CPU when using h264 is no different from h264_cuvid. What could be the reason ? When I use h264_cuvid, the video card is loaded by 2-3% more.

  


  Video stream : 1920x1080 25 FPS (main and high profile).
Video card : NVIDIA Corporation GP107 [GeForce GTX 1050 Ti] (rev a1).
Processor : Intel (R) Core (TM) i5-8400 CPU @ 2.80GHz.

  


  Compilation config ffmpeg :

  


  ffmpeg version 4.3 Copyright (c) 2000-2020 the FFmpeg developers
  built with gcc 8 (Debian 8.3.0-6)
  configuration: --enable-gpl --enable-ladspa --enable-libpulse --enable-libsoxr --enable-libspeex --enable-avfilter --enable-avresample --enable-postproc --enable-pthreads --enable -libass --enable-libfreetype --enable-libmp3lame --enable-nonfree --disable-ffplay --enable-libxvid --enable-cuda --enable-cuda-nvcc --enable-cuda-sdk --enable- cuvid --enable-nvenc --enable-nonfree --enable-libnpp --enable-shared --extra-cflags = -I/usr/local/cuda/include --extra-ldflags = -L/usr/local/cuda/lib64