Recherche avancée

Sur d’autres sites (7091)

• ffmpeg split video H264 stutters on Windows Media Player

  9 mars 2019, par Juan Foegen

  I am converting and splitting wmv and mp4 into new mp4 files. I am using ffmpeg to do the conversion with simply :
  ffmpeg -i inputfile.wmv -vcodec libx264 -ss 00:00:10 -t 00:00:15 outputfile.mp4

  So when I create my split video it plays fine on other video players, like VLC, but stutters when initially playing on Windows Media Player and even shows a blank screen for a moment and then after about 5 seconds, the video will play normally. If I then select play again without reloading, Windows Media Player will play the video fine, with no stuttering.

  If I REMOVE the start parameter when doing the conversion, then the resulting video will play without issues on all my players, INCLUDING Windows Media Player.

  I have tried changing the frame rate, the bit rate and altered different optimization options mentioned here - https://trac.ffmpeg.org/wiki/Encode/H.264

  I need the split video to play on Windows machines on Windows Media Player and on the web streamed using the video element. This is why I chose H264 codec and MP4.

  Has anyone had issues playing ffmpeg split videos using H264 codec with Windows Media Player ? I am using the latest version of ffmpeg. Anyone have any other suggestions on any different settings I could try to adjust ?

  Thanks

• Error audio loading when runing Whisper Open AI model

  9 juin 2024, par John mick

  The problem I'm trying to solve is that I can't run Whisper model for some audio, it says something related to audio decoding.

  


  payload.wav: Invalid data found when processing input.
raise RuntimeError(f"Failed to load audio: {e.stderr.decode()}") from e


  


  I tried using the micro-machines.wav and it works fine but when i used other audio it gives me an error.

  


  import whisper

model = whisper.load_model("base")
text=model.transcribe('micro-machines.wav',fp16=False)
print(text)
text=model.transcribe('payload.wav',fp16=False)
print(text)


  


  Error I'm getting for payload :

  


  d:\...\venv\lib\site-packages\whisper\transcribe.py:79: UserWarning: FP16 is not supported on CPU; using FP32 instead&#xA;  warnings.warn("FP16 is not supported on CPU; using FP32 instead")                                                                                        &#xA;Traceback (most recent call last):&#xA;  File "d:\...\venv\lib\site-packages\whisper\audio.py", line 42, in load_audio&#xA;    ffmpeg.input(file, threads=0)                                                                                    &#xA;  File "d:\...\venv\lib\site-packages\ffmpeg\_run.py", line 325, in run        &#xA;    raise Error(&#x27;ffmpeg&#x27;, out, err)                                                                                  &#xA;ffmpeg._run.Error: ffmpeg error (see stderr output for detail)                                                       &#xA;&#xA;The above exception was the direct cause of the following exception:&#xA;&#xA;Traceback (most recent call last):&#xA;  File "C:\....\Python\Python39\lib\runpy.py", line 197, in _run_module_as_main&#xA;    return _run_code(code, main_globals, None,&#xA;  File "C:\.....\Python\Python39\lib\runpy.py", line 87, in _run_code&#xA;    exec(code, run_globals)&#xA;  File "D:\...\venv\Scripts\whisper.exe\__main__.py", line 7, in <module>&#xA;  File "d:\...\venv\lib\site-packages\whisper\transcribe.py", line 314, in cli&#xA;    result = transcribe(model, audio_path, temperature=temperature, **args)&#xA;  File "d:\...\venv\lib\site-packages\whisper\transcribe.py", line 85, in transcribe&#xA;    mel = log_mel_spectrogram(audio)&#xA;  File "d:\...\venv\lib\site-packages\whisper\audio.py", line 111, in log_mel_spectrogram&#xA;    audio = load_audio(audio)&#xA;  File "d:\...\venv\lib\site-packages\whisper\audio.py", line 47, in load_audio&#xA;    raise RuntimeError(f"Failed to load audio: {e.stderr.decode()}") from e&#xA;RuntimeError: Failed to load audio: ffmpeg version 6.0-essentials_build-www.gyan.dev Copyright (c) 2000-2023 the FFmpeg developers&#xA;  built with gcc 12.2.0 (Rev10, Built by MSYS2 project)&#xA;  configuration: --enable-gpl --enable-version3 --enable-static --disable-w32threads --disable-autodetect --enable-fontconfig --enable-iconv --enable-gnutls --enable-libxml2 --enab&#xA;le-gmp --enable-lzma --enable-zlib --enable-libsrt --enable-libssh --enable-libzmq --enable-avisynth --enable-sdl2 --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxv&#xA;id --enable-libaom --enable-libopenjpeg --enable-libvpx --enable-libass --enable-libfreetype --enable-libfribidi --enable-libvidstab --enable-libvmaf --enable-libzimg --enable-amf &#xA;--enable-cuda-llvm --enable-cuvid --enable-ffnvcodec --enable-nvdec --enable-nvenc --enable-d3d11va --enable-dxva2 --enable-libmfx --enable-libgme --enable-libopenmpt --enable-libo&#xA;pencore-amrwb --enable-libmp3lame --enable-libtheora --enable-libvo-amrwbenc --enable-libgsm --enable-libopencore-amrnb --enable-libopus --enable-libspeex --enable-libvorbis --enab&#xA;le-librubberband&#xA;  libavutil      58.  2.100 / 58.  2.100&#xA;  libavcodec     60.  3.100 / 60.  3.100&#xA;  libavformat    60.  3.100 / 60.  3.100&#xA;  libavdevice    60.  1.100 / 60.  1.100&#xA;  libavfilter     9.  3.100 /  9.  3.100&#xA;  libswscale      7.  1.100 /  7.  1.100&#xA;  libswresample   4. 10.100 /  4. 10.100&#xA;  libpostproc    57.  1.100 / 57.  1.100&#xA;payload.wav: Invalid data found when processing input&#xA;</module>

  


  I tried searching for solutions and I found one which says It appears that the code failed to load the audio file for some reason and even failed to display that error because e.stderr did not contain a valid UTF-8 string

  


• Detect if AVStream is broken

  9 juillet 2018, par Eugene Alexeev

  In my project I’m using an FFMpeg and I wrote a muxer which works with multi-video container (it’s an .mp4 with few video streams). And sometimes I have to deal with a situation when my code was fed up with file which has broken video stream. And the problem is - I need to find a way how I can detect broken stream and abandon it from usage.

  Here’s my code which I’m using for reading streams out of the file :

  - (BOOL)openFile:(NSString *)filePath options:(AVDictionary *)options
  
  {
  
      _pFormatCtx = NULL;
  
  
  
      if (avformat_open_input(&_pFormatCtx, [filePath UTF8String], NULL, &options) != 0) {
  
          [self raiseExceptionWithMessage:@"Couldn't open the file!"];
  
          return NO;
  
      }
  
  
  
      if (avformat_find_stream_info(_pFormatCtx, NULL) < 0) {
  
          //[self raiseExceptionWithMessage:@"Couldn't find stream info!"];
  
          return NO;
  
      }
  
  
  
      int stream_index = 0;
  
      for (int i = 0; i < _pFormatCtx->nb_streams ; i++)
  
      {
  
          AVStream *st = _pFormatCtx->streams[i];
  
          if (_pFormatCtx->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_VIDEO)
  
          {
  
              AVCodecParameters *params = st->codecpar;
  
              st->index = stream_index;
  
              _streams[@(st->index)] = [[VideoStream alloc] initWithStream:st];
  
              stream_index++;
  
          }
  
      }
  
  
  
      _firstStream = [_streams.allValues firstObject].stream;
  
  
  
      if (_streams.count == 0) {
  
          [self raiseExceptionWithMessage:@"Couldn't find videostream!"];
  
          return NO;
  
      }
  
  
  
      _pCodecParams = [self firstStream]->codecpar;
  
      _pCodec = avcodec_find_decoder(_pCodecParams->codec_id);
  
  
  
      if (!_pCodec) {
  
          [self raiseExceptionWithMessage:@"Unsupported codec!"];
  
          return NO;
  
      }
  
  
  
      _pCodecCtx = avcodec_alloc_context3(_pCodec);
  
      if (avcodec_parameters_to_context(_pCodecCtx, _pCodecParams) < 0) {
  
          [self raiseExceptionWithMessage:@"Couldn't copy params into codec"];
  
          return NO;
  
      }
  
  
  
      if (avcodec_open2(_pCodecCtx, _pCodec, 0) < 0) {
  
          [self raiseExceptionWithMessage:@"Couldn't open the Codec!"];
  
          return NO;
  
      }
  
  
  
      return YES;
  
  }

  When this method is opening video with broken video-stream, avformat_find_stream_info gives such output in logs :

  [h264 @ 0x10684e200] top block unavailable for requested intra mode -1
  
  [h264 @ 0x10684e200] error while decoding MB 45 0, bytestream 76665
  
  [h264 @ 0x10684e200] concealing 3600 DC, 3600 AC, 3600 MV errors in I frame
  
  [h264 @ 0x106854a00] top block unavailable for requested intra mode -1
  
  [h264 @ 0x106854a00] error while decoding MB 47 0, bytestream 69501
  
  [h264 @ 0x106854a00] concealing 3600 DC, 3600 AC, 3600 MV errors in I frame
  
  [h264 @ 0x10683de00] left block unavailable for requested intra4x4 mode -1
  
  [h264 @ 0x10683de00] error while decoding MB 0 1, bytestream 73290
  
  [h264 @ 0x10683de00] concealing 3600 DC, 3600 AC, 3600 MV errors in I frame
  
  [h264 @ 0x106860800] left block unavailable for requested intra4x4 mode -1
  
  [h264 @ 0x106860800] error while decoding MB 0 2, bytestream 23072
  
  [h264 @ 0x106860800] concealing 3559 DC, 3559 AC, 3559 MV errors in I frame

  Good news is avformat_find_stream_info clearly sees that something is wrong with the file. But bad news is - output of avformat_find_stream_info isn’t affected at all. After this method execution my AVFormatContext variable which is _pFormatCtx will have all streams file provides whether it’s valid or not.

  And my question - how can I differ bad stream from a good one ? What kind of data inside the stream I can rely on ? What kind of method I could use to test stream out ? Any help would be appreciated !