Recherche avancée

Sur d’autres sites (2991)

• Hareware accelerated HEVC encode/decode in python on Intel x64 with Nvidia GPU [closed]

  24 mars 2021, par Jason M

  I am working on a python project of decoding a 4k@25fps hevc stream from a IP camera, processing it, croping some region-of-interest and encoding each into a new stream. The program runs on a Intel i7-10700 PC with NVidia Gtx1080. I see that opencv offers Intel MediaSDK support and cuda codec, while Nvidia offers PyNvCodec.

  


  What may be my best combination of hardware-accelerated encoding/decoding options ? A demo project would be nicer.

  


• How can I seek to frame No. X with ffmpeg ?

  27 août 2018, par Giumo

  I’m writing a video editor, and I need to seek to exact frame, knowing the frame number.

  Other posts on stackoverflow told me that ffmpeg may give me a few broken frames after seeking, which is not a problem for playback but a big problem for video editors.

  And I need to seek by frame number, not by time, which will become inaccurate when converted to frame number.

  I’ve read dranger’s tuts (which is outdated now), and end up with :

  av_seek_frame(fmt_ctx, video_stream_id, frame, AVSEEK_FLAG_ANY);

  It always seek to frame No. 0, and always return 0 which means success.
  Then I tried to read Blender’s source code and found it really complex(maybe I should implement an image buffer ?).

  So, is there any simple way to seek to a frame with just a simple call like seek(context, frame_number)(while getting a full frame, not a broken one) ? Or, is there any lightweight library that simplifies this ?

  EDIT :
  Thanks to praks411，I found the solution :

  void AV_seek(AV * av, size_t frame)
  
  {
  
      int frame_delta = frame - av->frame_id;
  
      if (frame_delta < 0 || frame_delta > 5)
  
          av_seek_frame(av->fmt_ctx, av->video_stream_id,
  
                  frame, AVSEEK_FLAG_BACKWARD);
  
      while (av->frame_id != frame)
  
          AV_read_frame(av);
  
  }
  
  
  
  void AV_read_frame(AV * av)
  
  {
  
      AVPacket packet;
  
      int frame_done;
  
  
  
      while (av_read_frame(av->fmt_ctx, &packet) >= 0) {
  
          if (packet.stream_index == av->video_stream_id) {
  
              avcodec_decode_video2(av->codec_ctx, av->frame, &frame_done, &packet);
  
              if (frame_done) {
  
                  ...
  
                  av->frame_id = packet.dts;
  
                  av_free_packet(&packet);
  
                  return;
  
              }
  
          }
  
          av_free_packet(&packet);
  
      }
  
  }

  EDIT2 :
  Turns out there is a library for this : FFMS2.
  It is "an FFmpeg based source library [...] for easy frame accurate access", and is portable (at least across Windows and Linux).

• Recording ip camera stream writing video file [closed]

  10 décembre 2020, par Marc

  We are implementing a camera surveillance system project using python with openCV/ffmpeg and Cuda based on client-server architecture, this system should hold up to 40-60 IP cameras for live, record and playback features.

  


  The problem is that their is delay in live view between our program and the web page of the camera and problem in writing frames into video files(recording).

  


  Does the problem due to the used language or a hardware issue, and does python represents the efficient language for this type of projects ?