Recherche avancée

Sur d’autres sites (11955)

• FFmpeg overwrite an AVPacket

  26 janvier 2019, par Alejandro Ramírez

  Hello guys I am working with video. I wan to encrypt the I-frames of one video compressed in H.264. So I get the AVPacket from the video and compare if (AVPacket.flags & AV_PKT_FLAG_KEY) to know if the packet has an I frame, but when I try to print the AVPacket.data I don’t have any information to encrypt.
  Where can I get the information regarding to I-frame. ahead a put my code, thank you.

  #include <iostream>
  
  extern "C" {
  
  #include <libavcodec></libavcodec>avcodec.h>
  
  #include <libavformat></libavformat>avformat.h>
  
  #include <libswscale></libswscale>swscale.h>
  
  #include <libavutil></libavutil>avutil.h>
  
  }
  
  #define INBUF_SIZE 4096
  
  //#define AV_INPUT_BUFFER_PADDING_SIZE 32
  
  int main (int argc, char * argv[])
  
  {
  
      //av_register_all();//omit
  
      //revisar el número de argumentos en el video
  
      AVFormatContext *pFormatCtx = NULL;
  
      AVCodec *dec = NULL;
  
      AVCodecContext *pCodecCtx = NULL;
  
      AVStream *st = NULL;
  
      AVDictionary *opts = NULL;
  
      AVFrame *frame;
  
      AVPacket avpkt;
  
      uint8_t inbuf [INBUF_SIZE + AV_INPUT_BUFFER_PADDING_SIZE];
  
      FILE *f;
  
      int frame_count;
  
      int video_stream_index = -1;
  
  
  
      //av_init_packet (&amp;avpkt);
  
  
  
      memset (inbuf + INBUF_SIZE, 0, AV_INPUT_BUFFER_PADDING_SIZE);
  
      if (avformat_open_input (&amp;pFormatCtx, argv[1], NULL, NULL) != 0)
  
          return -1;
  
      if (avformat_find_stream_info (pFormatCtx, NULL) &lt; 0)
  
          return -1;
  
      //video_stream_index = av_find_best_stream (pFormatCtx,AVMEDIA_TYPE_VIDEO, -1, -1, &amp;dec,0);
  
      video_stream_index = av_find_best_stream (pFormatCtx,AVMEDIA_TYPE_VIDEO, -1, -1, &amp;dec,0);
  
      if ( video_stream_index &lt; 0)
  
          return -1;
  
      std::cout &lt;&lt; "video_stream; " &lt;&lt; video_stream_index &lt;&lt; "\n";
  
      st = pFormatCtx -> streams [video_stream_index];
  
      std::cout &lt;&lt; "number of frames " &lt;&lt; st -> nb_frames &lt;&lt; "\n";
  
      std::cout &lt;&lt; "event_flags " &lt;&lt; st -> event_flags &lt;&lt; "\n";
  
      //pCodecCtx = st -> codec;//deprecated
  
  
  
      dec = avcodec_find_decoder (st -> codecpar -> codec_id);
  
      std::cout &lt;&lt; "codec_id: " &lt;&lt; st -> codecpar -> codec_id &lt;&lt; "\n";
  
      std::cout &lt;&lt; "AV_CODEC_ID_H264: " &lt;&lt; AV_CODEC_ID_H264 &lt;&lt; "\n";
  
      if (!dec)
  
          return -1;
  
      pCodecCtx = avcodec_alloc_context3 (dec);
  
      if (!pCodecCtx)
  
          return -1;
  
      //av_dict_set (&amp;opts, "refcounted_frames", "0", 0);
  
      avcodec_parameters_to_context (pCodecCtx, st -> codecpar);
  
  
  
  
  
      std::cout &lt;&lt; "todo bien \n";    
  
      if (avcodec_open2 (pCodecCtx, dec, &amp;opts) &lt; 0)
  
          return -1;
  
      /*************hasta aqui buen codigo*********************************************/ 
  
      frame = av_frame_alloc ();
  
      if (!frame)
  
          return -1;
  
      av_init_packet (&amp;avpkt);
  
      avpkt.data = NULL;
  
      avpkt.size = 0;
  
      f = fopen (argv[1], "r");
  
      int times = 1;
  
      while (av_read_frame (pFormatCtx, &amp;avpkt) >= 0){
  
          AVPacket oripkt = avpkt;
  
          if (oripkt.stream_index == video_stream_index){
  
              if (oripkt.flags &amp; AV_PKT_FLAG_KEY){
  
                  std::cout &lt;&lt; "times: " &lt;&lt; times ++ &lt;&lt; "\n";
  
                  std::cout &lt;&lt; "avpkt.flags: " &lt;&lt; oripkt.flags &lt;&lt; "\n";
  
                  std::cout &lt;&lt; "tam avpkt.data: " &lt;&lt; sizeof(oripkt.data) &lt;&lt; "\n";
  
                  std::cout &lt;&lt; "tam avpkt.data: " &lt;&lt; oripkt.data &lt;&lt; "\n";
  
                  std::cout &lt;&lt; "oripkt.size: " &lt;&lt; oripkt.size &lt;&lt; "\n";
  
                  std::cout &lt;&lt; "oripkt.side_date_eme: " &lt;&lt; oripkt.side_data_elems &lt;&lt; "\n";
  
                  if (!oripkt.data)
  
                      std::cout &lt;&lt; "no tengo dinero \n";
  
              }
  
          }
  
      }
  
  
  
      std::cout &lt;&lt; "Fin del programa " &lt;&lt; "\n";
  
  }
  
  </iostream>

• Permanent fix for Opencv videocapture

  19 février 2018, par Ishan Sharma

  This question has been posed numerous times on many websites, but not definitive solution. I am trying to run Opencv with a video using function :

  import cv2
  
  cap = cv2.VideoCapture('video.mp4')
  
  if(cap.isOpened()==False):
  
  print "Error opening camera"

  But it fails every time. I have tried almost all steps from various sites, but couldn’t get it to work (including rebuilding ffmpeg separately).

  Any help would be much appreciated.

  My platform is Ubuntu17 and Python3.

• convert images to video with differing time ranges

  11 juin 2013, par FaultyJuggler

  I have images coming in at random times labeled with their epoch time they were taken. I want to create a video that shows their real time creation in order. FFMPEG (far as I can tell) only allows you to set the framerate.

  For now I'm looking at creating a video file per image that is as long as the gap between the current image and the next image's timestamp, then concatenating all videos together after.

  Is there a better way to do this ?