Recherche avancée

Sur d’autres sites (7133)

• ffmpeg how to calculate complete frame ['I' frame ] from [ 'P' frame ]. It is conceptually correct ?

  15 février 2014, par Whoami

  I am trying to get the knowledge of ffmpeg streaming video handling.

  what I understood :

  I get from the IPed Camaera, the frames like 'IPPPPPPPPPPPPPPPPIPPPPPPP'..

  Frame 'I' is a complete frame, where as frame 'P' depends on the previous either 'P' or 'I' frame which ever is the closes.

  I get the frame by using avcodec_decode_video2

  while (av_read_frame (context, &packet) >=0)
  
  {   
  
          //LOGD (" Received PACKET...DTS and PTS  %ld and %ld ", packet.pts, packet.dts);
  
  
  
          if(packet.stream_index == videoStreamIndex ) { 
  
                  avcodec_decode_video2 (pCodecCtx, pFrame, &finished, &packet);
  
                  if ( finished) {
  
                  // Here is my frame, getting the type by av_get_picture_type_char(pFrame->pict_type).
  
            }
  
   }

  Now, When i display just the frames that i have received, looks like whenever 'I' frame received, it displays properly, when received 'P' frames, the image goes for a toss.

  1) We need to manually do any calculation to convert 'P' Frame to 'I' Frame so that it can be rendered ?

  2) If not (1), what do i have to take care ?..Does DTS/PTS calcuation do the magic here ?

• Creating .ts chunks from .mp4 file

  11 juin 2015, par kopalvich

  I need to create a .ts chunks from a .mp4 file without using ffmpeg API. I have already implemented this task using ffmpeg API, and it works, but my team lead wants me to get rid of it, and it is impossible to convince him against that. The code is supposed to run under load.

  I have already found out how to read mp4 format, where to find frames, their offsets and sizes in a file, their pts’s, for audio and video. A huge achievement for me as I haven’t been working in this area for long.

  All I managed is to create playlist where each chunk starts with I-frame. I feed the start and stop pts’s of playlist entry to the old code(slightly modified) using ffmpeg API, and it creates proper ts chunks.

  But but to create ts chunks without ffmpeg API I still cannot.

  At first I tried to read the code of nginx rtmp module, the hls part of it, but couldn’t understand anything there. It’s too complex and I lack specific knowledge.
  Now I’d like to read something on ts format. Can anyone advise where to look ?

  Thanks.

• How can i remove portions of audio instead of concatenate the portions

  16 octobre 2022, par Plainsage

  If i've got this command that concatenates audio between 115-368 seconds and 605 to the end and the output only has this audio.

  


  ffmpeg -i "abc.mp3" -filter_complex "[0:a]atrim=start=115:end=368,asetpts=PTS-STARTPTS[ba];[0:a]atrim=start=605,asetpts=PTS-STARTPTS[da]" -b:a 320k "def.mp3"


  


  How can i make this same command so that instead of the output being concatenation of these 2, it removes these two portions and the output has the remaining audio in it

  


  I know there's a way where i can just instead get the start and end of the audio i want and concatenate those, but i'd like to know for my knowledge if there is a way to remove the audio, rather than concatenate into the output.