Advanced search

On other websites (10423)

• FFmpeg extracts black image from H264 stream

  8 June 2022, by massivemoisture

  I have a C# application that receives H264 stream through a socket. I want to continuously get the latest image from that stream.

  


  Here's what I did with FFmpeg 5.0.1, just a rough sample to get ONE latest image, how I start FFmpeg:

  


  var ffmpegInfo = new ProcessStartInfo(FFMPEG_PATH);
ffmpegInfo.RedirectStandardInput = true;
ffmpegInfo.RedirectStandardOutput = true;
ffmpegInfo.RedirectStandardError = true;
ffmpegInfo.UseShellExecute = false;

ffmpegInfo.Arguments = "-i pipe: -f h264 -pix_fmt bgr24 -an -sn pipe:";

ffmpegInfo.UseShellExecute = false;
ffmpegInfo.CreateNoWindow = true;

Process myFFmpeg = new Process();
myFFmpeg.StartInfo = ffmpegInfo;
myFFmpeg.EnableRaisingEvents = true;
myFFmpeg.Start();

var inStream = myFFmpeg.StandardInput.BaseStream;
FileStream baseStream = myFFmpeg.StandardOutput.BaseStream as FileStream;
myFFmpeg.BeginErrorReadLine();


  


  Then I start a new thread to receive the stream through socket:

  


  // inStream is "myFFmpeg.StandardInput.BaseStream" from the code block above
var t = Task.Run(() => ReceiveStream(inStream));


  


  Next I read the output from FFmpeg:

  


  byte[] decoded = new byte[Width * Height * 3];
int numBytesToRead = Width * Height * 3;
int numBytesRead = 0;

while (numBytesToRead > 0)
{
   int n = baseStream.Read(decoded, 0, decoded.Length);
   Console.WriteLine($"Read {n} bytes");
   if (n == 0)
   {
      break;
   }              
      numBytesRead += n;
      numBytesToRead -= n;
}


  


  Lastly, I use ImageSharp library to save decoded byte array as a jpeg file.

  


  image.Save("test.jpeg", encoder);


  


  However, test.jpeg always comes out as a black image. What did I do wrong?
Here's the stderr log that I got from ffmpeg:

  


  
  Duration: N/A, bitrate: N/A
  Stream #0:0: Video: h264 (High), yuv420p(tv, smpte170m/bt470bg/smpte170m, progressive), 1080x2256, 25 fps, 25 tbr, 1200k tbn
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264))
Incompatible pixel format 'bgr24' for codec 'libx264', auto-selecting format 'yuv444p'
[libx264 @ 0x11b9068f0] using cpu capabilities: ARMv8 NEON
[libx264 @ 0x11b9068f0] profile High 4:4:4 Predictive, level 5.0, 4:4:4, 8-bit
Output #0, h264, to 'pipe:':
  Metadata:
    encoder         : Lavf59.16.100
  Stream #0:0: Video: h264, yuv444p(tv, smpte170m/bt470bg/smpte170m, progressive), 1080x2256, q=2-31, 25 fps, 25 tbn
    Metadata:
      encoder         : Lavc59.18.100 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/0 buffer size: 0 vbv_delay: N/A
frame=    1 fps=0.0 q=0.0 size=       0kB time=00:00:00.00 bitrate=N/A speed=   0x    
frame=   56 fps=0.0 q=0.0 size=       0kB time=00:00:00.00 bitrate=N/A speed=   0x
frame=   87 fps= 83 q=28.0 size=     370kB time=00:00:01.16 bitrate=2610.6kbits/s speed=1.11x
frame=  118 fps= 75 q=28.0 size=     698kB time=00:00:02.40 bitrate=2381.4kbits/s speed=1.54x
frame=  154 fps= 75 q=28.0 size=    1083kB time=00:00:03.84 bitrate=2311.1kbits/s speed=1.86x
...


  


  Thank you!

  


  Edit: as suggested by @kesh, I have changed h264 to rawvideo, the arguments now are: -i pipe: -f rawvideo -pix_fmt bgr24 -an -sn pipe:

  


  Here's the output of ffmpeg:

  


  Input #0, h264, from 'pipe:':
  Duration: N/A, bitrate: N/A
  Stream #0:0: Video: h264 (High), yuv420p(tv, smpte170m/bt470bg/smpte170m, progressive), 1080x2256, 25 fps, 25 tbr, 1200k tbn
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> rawvideo (native))
// About 9 of these "No accelerated colorspace..." message
[swscaler @ 0x128690000] [swscaler @ 0x1286a0000] No accelerated colorspace conversion found from yuv420p to bgr24.
Output #0, rawvideo, to 'pipe:':
  Metadata:
    encoder         : Lavf59.16.100
  Stream #0:0: Video: rawvideo (BGR[24] / 0x18524742), bgr24(pc, gbr/bt470bg/smpte170m, progressive), 1080x2256, q=2-31, 1461888 kb/s, 25 fps, 25 tbn
    Metadata:
      encoder         : Lavc59.18.100 rawvideo
frame=    1 fps=0.0 q=0.0 size=       0kB time=00:00:00.00 bitrate=N/A speed=   0x
// FFmpeg outputs no more log after this


  


• Output black when I decode h264 720p with ffmpeg

  6 December 2017, by José Marqueses Saxo

  First, sorry for my english. When I decode h264 720p in ardrone2.0 my output is black and I cant see anything.

  I have try to change the value of pCodecCtx->pix_fmt = AV_PIX_FMT_BGR24; to pCodecCtx->pix_fmt = AV_PIX_FMT_YUV420P; and the value of pCodecCtxH264->pix_fmt = AV_PIX_FMT_BGR24; to pCodecCtxH264->pix_fmt = AV_PIX_FMT_YUV420P; but my program crash. What am I doing wrong?. Thank you, see part of my code:

  av_register_all();
  
  avcodec_register_all();
  
  avformat_network_init();
  
  
  
  // 1.2. Open video file
  
  if(avformat_open_input(&pFormatCtx, drone_addr, NULL, NULL) != 0) {
  
    mexPrintf("No conecct with Drone");
  
    EndVideo();
  
    return;
  
  }
  
  
  
  pCodec    = avcodec_find_decoder(AV_CODEC_ID_H264);
  
  
  
  pCodecCtx = avcodec_alloc_context3(pCodec);
  
  pCodecCtx->pix_fmt = AV_PIX_FMT_BGR24;
  
  pCodecCtx->skip_frame = AVDISCARD_DEFAULT;
  
  pCodecCtx->error_concealment = FF_EC_GUESS_MVS | FF_EC_DEBLOCK;
  
  pCodecCtx->err_recognition = AV_EF_CAREFUL;
  
  pCodecCtx->skip_loop_filter = AVDISCARD_DEFAULT;
  
  pCodecCtx->workaround_bugs = FF_BUG_AUTODETECT;
  
  pCodecCtx->codec_type = AVMEDIA_TYPE_VIDEO;
  
  pCodecCtx->codec_id = AV_CODEC_ID_H264;
  
  pCodecCtx->skip_idct = AVDISCARD_DEFAULT;
  
  pCodecCtx->width = 1280;
  
  pCodecCtx->height = 720;
  
  
  
  pCodecH264 = avcodec_find_decoder(AV_CODEC_ID_H264);
  
  pCodecCtxH264 = avcodec_alloc_context3(pCodecH264);
  
  
  
  
  
  pCodecCtxH264->pix_fmt = AV_PIX_FMT_BGR24;
  
  pCodecCtxH264->skip_frame = AVDISCARD_DEFAULT;
  
  pCodecCtxH264->error_concealment = FF_EC_GUESS_MVS | FF_EC_DEBLOCK;
  
  pCodecCtxH264->err_recognition = AV_EF_CAREFUL;
  
  pCodecCtxH264->skip_loop_filter = AVDISCARD_DEFAULT;
  
  pCodecCtxH264->workaround_bugs = FF_BUG_AUTODETECT;
  
  pCodecCtxH264->codec_type = AVMEDIA_TYPE_VIDEO;
  
  pCodecCtxH264->codec_id = AV_CODEC_ID_H264;
  
  pCodecCtxH264->skip_idct = AVDISCARD_DEFAULT;
  
  
  
  if(avcodec_open2(pCodecCtxH264, pCodecH264, &optionsDict) < 0)
  
  {
  
     mexPrintf("Error opening H264 codec");
  
     return ;
  
  }
  
  
  
  pFrame_BGR24 = av_frame_alloc();
  
  
  
  
  
  if(pFrame_BGR24 == NULL) {
  
     mexPrintf("Could not allocate pFrame_BGR24\n");
  
     return ;
  
  }
  
  
  
  // Determine required buffer size and allocate buffer
  
  
  
  buffer_BGR24 = 
  
  (uint8_t *)av_mallocz(av_image_get_buffer_size(AV_PIX_FMT_BGR24, 
  
  pCodecCtx->width, ((pCodecCtx->height == 720) ? 720 : pCodecCtx->height) * 
  
  sizeof(uint8_t)*3,1));
  
  
  
  // Assign buffer to image planes
  
  
  
  av_image_fill_arrays(pFrame_BGR24->data, pFrame_BGR24->linesize,
  
  buffer_BGR24,AV_PIX_FMT_BGR24, pCodecCtx->width, pCodecCtx->height,1);
  
  
  
  // format conversion context
  
  pConvertCtx_BGR24 = sws_getContext(pCodecCtx->width, pCodecCtx->height, 
  
  pCodecCtx->pix_fmt, pCodecCtx->width, pCodecCtx->height,  AV_PIX_FMT_BGR24, 
  
                                   SWS_BILINEAR | SWS_ACCURATE_RND, 0, 0, 0);
  
  
  
  // 1.6. get video frames
  
  pFrame = av_frame_alloc();
  
  
  
  av_init_packet(&packet);
  
  packet.data = NULL;
  
  packet.size = 0;
  
  }
  
  
  
  //Captura un frame
  
  void video::capture(mxArray *plhs[]) {
  
  
  
    if(av_read_frame(pFormatCtx, &packet) < 0){
  
        mexPrintf("Error al leer frame");
  
        return;
  
    }
  
     do {
  
         do {
  
            rest = avcodec_send_packet(pCodecCtxH264, &packet);
  
         } while(rest == AVERROR(EAGAIN));
  
  
  
         if(rest == AVERROR_EOF || rest == AVERROR(EINVAL)) {
  
                  printf("AVERROR(EAGAIN): %d, AVERROR_EOF: %d, 
  
                  AVERROR(EINVAL): %d\n", AVERROR(EAGAIN), AVERROR_EOF, 
  
                  AVERROR(EINVAL));
  
              printf("fe_read_frame: Frame getting error (%d)!\n", rest);
  
              return;
  
         }
  
  
  
         rest = avcodec_receive_frame(pCodecCtxH264, pFrame);
  
     } while(rest == AVERROR(EAGAIN));
  
  
  
     if(rest == AVERROR_EOF || rest == AVERROR(EINVAL)) {
  
  
  
      // An error or EOF occured,index break out and return what
  
      // we have so far.
  
        printf("AVERROR(EAGAIN): %d, AVERROR_EOF: %d, AVERROR(EINVAL): %d\n", 
  
        AVERROR(EAGAIN), AVERROR_EOF, AVERROR(EINVAL));
  
          printf("fe_read_frame: EOF or some othere decoding error (%d)!\n", 
  
          rest);
  
          return;
  
     }
  
  
  
  
  
     // 2.1.1. convert frame to GRAYSCALE [or BGR] for OpenCV
  
     sws_scale(pConvertCtx_BGR24,   (const uint8_t* const*)pFrame->data, 
  
         pFrame->linesize, 0,pCodecCtx->height,   pFrame_BGR24->data,   
  
               pFrame_BGR24->linesize);
  
  //}
  
     av_packet_unref(&packet);
  
     av_init_packet(&packet);
  
     mwSize dims[] = {(pCodecCtx->width)*((pCodecCtx->height == 720) ? 720 : 
  
     pCodecCtx->height)*sizeof(uint8_t)*3};
  
     plhs[0] = mxCreateNumericArray(1,dims,mxUINT8_CLASS, mxREAL);
  
      //plhs[0]=mxCreateDoubleMatrix(pCodecCtx->height,pCodecCtx-
  
      >width,mxREAL);
  
     point=mxGetPr(plhs[0]);
  
     memcpy(point, pFrame_BGR24->data[0],(pCodecCtx->width)*(pCodecCtx-
  
      >height)*sizeof(uint8_t)*3);
  
  }

• Anomalie #3708: .select dans menu lateral

  21 February 2016, by b b

  Le changement vient du fait qu’en 3.0 on avait un overflow: hidden; sur les .formulaire_spip li ici :

  https://core.spip.net/projects/spip/repository/entry/branches/spip-3.0/prive/themes/spip/forms.css.html#L61

  Alors qu’en 3.1, on applique cette règle uniquement sur les .editer :

  https://core.spip.net/projects/spip/repository/entry/spip/prive/themes/spip/forms.css.html#L66

  Du coup, il faut soit ajouter une classe editer sur tous les forms présents dans la colonne de gauche, ou compléter la règle.