Recherche avancée

Sur d’autres sites (10735)

• How to extract encoded transport stream frames from a video file ?

  3 mai 2022, par Ivy Growing

  Searching for a method to extract from video stream (.m2t or .ts file) a certain frames as is, encoded. OpenCV also extracts frames easily but decodes them immediately.
Given :

  


  

  • A .ts or .m2t file with H.264/MPEG-4 encoded stream.

  


  • Starting point in time for extraction like h:m:s.f (example : 0:2:1.12).

  


  • Ending point in time in the same format.

  


  


  I need to read from the file all frames in the given interval and provide to another program as buffer frame by frame as they are. The catch here is to keep the frames encoded as they are, do not decode/encode/encapsulate them.

  


  Picking a frame from the H.264 m2t to a pipe :

  


  ffmpeg -ss 0:2:1.12 -i .\my_video.ts -c:v copy -f mpegts -frames:v 1 pipe: -y -hide_banner

  


  Obviously, the time stamp is increasing for every next frame. From the pipe it is not a problem to convert it to a buffer.

  


  Questions :

  


  

  1. Is this method correct to extract a separate frame as it is without
any reference/recalculations with neighbor frames ?

  


  2. Not sure that flag -f mpegts really keeps the frame untouched. Is there better flag ? (Maybe -f null ?)

  


  3. How to know the type of extracted frame (i, P, or B) ?

  


  


  Thank you.

  


• What to pass to avcodec_decode_video2 for H.264 Transport Stream ?

  29 novembre 2016, par Danny

  I want to decode H.264 video from a collection of MPEG-2 Transport Stream packets but I am not clear what to pass to avcodec_decode_video2

  The documentation says to pass "the input AVPacket containing the input buffer."

  But what should be in the input buffer ?

  A PES packet will be spread across the payload portion of several TS packets, with NALU(s) inside the PES. So pass a TS fragment ? The entire PES ? PES payload only ?

  This Sample Code mentions :

  BUT some other codecs (msmpeg4, mpeg4) are inherently frame based, so
  you must call them with all the data for one frame exactly. You must
  also initialize ’width’ and ’height’ before initializing them.

  But I can find no info on what "all the data" means...

  Passing a fragment of a TS packet payload is not working :

  AVPacket avDecPkt;
  
  av_init_packet(&avDecPkt);
  
  avDecPkt.data = inbuf_ptr;
  
  avDecPkt.size = esBufSize;
  
  
  
  len = avcodec_decode_video2(mpDecoderContext, mpFrameDec, &got_picture, &avDecPkt);
  
  if (len < 0)
  
  {
  
      printf("  TS PKT #%.0f. Error decoding frame #%04d [rc=%d '%s']\n",
  
          tsPacket.pktNum, mDecodedFrameNum, len, av_make_error_string(errMsg, 128, len));
  
      return;
  
  }

  output

  [h264 @ 0x81cd2a0] no frame!
  
  TS PKT #2973. Error decoding frame #0001 [rc=-1094995529 'Invalid data found when processing input']

  EDIT

  Using the excellent hits from WLGfx, I made this simple program to try decoding TS packets. As input, I prepared a file containing only TS packets from the Video PID.

  It feels close but I don’t know how to set up the FormatContext. The code below segfaults at av_read_frame() (and internally at ret = s->iformat->read_packet(s, pkt)). s->iformat is zero.

  Suggestions ?

  EDIT II - Sorry, for got post source code **
  **EDIT III - Sample code updated to simulate reading TS PKT Queue

  /*
  
   * Test program for video decoder
  
   */
  
  
  
  #include 
  
  #include 
  
  #include 
  
  #include 
  
  
  
  extern "C" {
  
  
  
  #ifdef __cplusplus
  
      #define __STDC_CONSTANT_MACROS
  
      #ifdef _STDINT_H
  
          #undef _STDINT_H
  
      #endif
  
      #include 
  
  #endif
  
  }
  
  
  
  extern "C" {
  
  #include "libavcodec/avcodec.h"
  
  #include "libavformat/avformat.h"
  
  #include "libswscale/swscale.h"
  
  #include "libavutil/imgutils.h"
  
  #include "libavutil/opt.h"
  
  }
  
  
  
  
  
  class VideoDecoder
  
  {
  
  public:
  
      VideoDecoder();
  
      bool rcvTsPacket(AVPacket &inTsPacket);
  
  
  
  private:
  
      AVCodec         *mpDecoder;
  
      AVCodecContext  *mpDecoderContext;
  
      AVFrame         *mpDecodedFrame;
  
      AVFormatContext *mpFmtContext;
  
  
  
  };
  
  
  
  VideoDecoder::VideoDecoder()
  
  {
  
      av_register_all();
  
  
  
      // FORMAT CONTEXT SETUP
  
      mpFmtContext = avformat_alloc_context();
  
      mpFmtContext->flags = AVFMT_NOFILE;
  
      // ????? WHAT ELSE ???? //
  
  
  
      // DECODER SETUP
  
      mpDecoder = avcodec_find_decoder(AV_CODEC_ID_H264);
  
      if (!mpDecoder)
  
      {
  
          printf("Could not load decoder\n");
  
          exit(11);
  
      }
  
  
  
      mpDecoderContext = avcodec_alloc_context3(NULL);
  
      if (avcodec_open2(mpDecoderContext, mpDecoder, NULL) < 0)
  
      {
  
          printf("Cannot open decoder context\n");
  
          exit(1);
  
      }
  
  
  
      mpDecodedFrame = av_frame_alloc();
  
  }
  
  
  
  bool
  
  VideoDecoder::rcvTsPacket(AVPacket &inTsPkt)
  
  {
  
      bool ret = true;
  
  
  
      if ((av_read_frame(mpFmtContext, &inTsPkt)) < 0)
  
      {
  
          printf("Error in av_read_frame()\n");
  
          ret = false;
  
      }
  
      else
  
      {
  
          // success.  Decode the TS packet
  
          int got;
  
          int len = avcodec_decode_video2(mpDecoderContext, mpDecodedFrame, &got, &inTsPkt);
  
          if (len < 0)
  
              ret = false;
  
  
  
          if (got)
  
              printf("GOT A DECODED FRAME\n");
  
      }
  
  
  
      return ret;
  
  }
  
  
  
  int
  
  main(int argc, char **argv)
  
  {
  
      if (argc != 2)
  
      {
  
          printf("Usage: %s tsInFile\n", argv[0]);
  
          exit(1);
  
      }
  
  
  
      FILE *tsInFile = fopen(argv[1], "r");
  
      if (!tsInFile)
  
      {
  
          perror("Could not open TS input file");
  
          exit(2);
  
      }
  
  
  
      unsigned int tsPktNum = 0;
  
      uint8_t      tsBuffer[256];
  
      AVPacket     tsPkt;
  
      av_init_packet(&tsPkt);
  
  
  
      VideoDecoder vDecoder;
  
  
  
      while (!feof(tsInFile))
  
      {
  
          tsPktNum++;
  
  
  
          tsPkt.size = 188;
  
          tsPkt.data = tsBuffer;
  
          fread(tsPkt.data, 188, 1, tsInFile);
  
  
  
          vDecoder.rcvTsPacket(tsPkt);
  
      }
  
  }

• doc/formats : add reference to ffmpeg(1) stream specifiers as that docs is more complete

  16 février 2019, par Marton Balint

  doc/formats : add reference to ffmpeg(1) stream specifiers as that docs is more complete
  Signed-off-by : Marton Balint <cus@passwd.hu>
  

  • [DH] doc/formats.texi