Advanced search

On other websites (5141)

• extracting HEVC bit stream data from MPEG-TS

  18 February 2015, by userDtrm

  I’m building a custom MPEG-2 TS demuxer to extract HEVC encoded bit streams. My requirement is to extract the PES data streams encapsulated within the MPEG TS packets, by stripping off the MPES TS and PES header information. I cannot use ffmpeg in this case since I need custom manipulations within the process.

  I was able to extract TS header information, but the I facing issues in extracting the data elements. Specifically, how the PES header can be removed to get the data.

  For example, when the payload unit indicator is 1 and payload available flag is 1, it doesn’t appear that any payload is available.

  Please let me know how this can be done and really appreciate if anyone can point me to some appropriate resources.

  Thanks

• How to calculate sizes for AVPicture.data[x] pointers (YUV420p) when using libffmpeg

  13 August 2013, by bradenV2

  I'm trying to get Y,U,V values separately in order to pass them to openGL and map to a texture. I know these values can be found in AVPicture.data[0] (Y) and AVPicture.data[1] (U) and AVPicture.data[2] (V)

  avcodec_decode_video2(ctx, frame, &frameFinished, packet_data->packet);
  
  AVPicture _avPicture;
  
  picSize = avpicture_get_size(ctx->pix_fmt, ctx->width, ctx->height);
  
  avpicture_alloc(&_avPicture, ctx->pix_fmt,ctx->width, ctx->height );
  
  avpicture_fill(&_avPicture, packet_data->packet, ctx->pix_fmt,ctx->width,ctx->height);

  ^^ That's working fine.
  The issue I run into is passing the Y,U,V values back to Java via JNI. I have to know the size of the data the AVPicture.data[x] pointers point to. I've tried AVPicture.linesize to no avail, as well as:

             for (y = 0; y < ctx->height; y++){
  
                  for (x = 0; x < ctx->width; x++){
  
                      yDataSize++;
  
                  }
  
              }
  
              /* Cb and Cr */
  
              for (y = 0; y < ctx->height / 2; y++) {
  
                  for (x = 0; x < ctx->width / 2; x++) {
  
                      uDataSize++;
  
                      vDataSize++;
  
                  }
  
              }

  I'm really stuck, thanks!

• How is decoded audio data stored in ffmpeg AVFrame?

  23 February 2023, by necrosato

  I'm looking for clarification on how ffmpeg stores decoded audio data in frames before I start writing code to do audio mixing. AVFrame has int format and uint8_t* data[] members. If my understanding is correct, then the bytes in data should be cast to the proper type for format before working with it. So to do a simple 2x level boost if format == AV_SAMPLE_FMT_S16, I would:

  



  int16_t* audio_samples = frame->data[0];
int num_samples = frame->nb_samples * frame->channels;
for (int i = 0; i < num_samples; ++i) {
  audio_samples[i] = audio_samples[i] * 2;
}


  



  Is this the correct way of going about things?