Recherche avancée

Sur d’autres sites (6046)

• Concatination of images and videos in FFMpeg [closed]

  29 octobre 2023, par Juliano David Hilario

  I've been trying to concat pieces of videos with mix of images in FFMpeg and I can not achieve desired results. I'm trying to create a mock interview project for a friend, she answers the questions from a script. The question should be flashed in the screen before the footage of her answering is shown. The videos are encoded in H.265, with frame rate of 30fps, (via mediainfo) and when I try to concat the videos and images using the concat demuxer, it shows this message in rapid succession and only output the first image in the given duration in the output file : (meaning it failed when concating the video, I assume this is due to the fact that demuxed images doesn't have audio streams.)

  


  [png @ 0x55b4a4cb3540] Invalid PNG signature 0x3DB0201D430.
Error while decoding stream #0:0: Invalid data found when processing input


  


  the command I've used :

  


  ffmpeg -hide_banner -y -f concat -safe 0 -i concat.con -r 30 -c:v libx265 -c:a aac -pix_fmt yuv420p out.mp4


  


  the slice of concat.con file :

  


  file slides/1.png
duration 6
file parts/1.mp4
file slides/2.png
duration 8
file parts/2.mp4
file slides/3.png
duration 9
file parts/3.mp4
. . .
duration 12
file parts/9.mp4
file slides/10.png
duration 9
file parts/10.mp4



  


  Since, I need to get this done if possible today, I tried this route of turning the images into videos by this command :

  


  #!/bin/sh

ffmpeg -f lavfi -i anullsrc=channel_layout=stereo:sample_rate=44100 -loop 1 -i "${1}.png" -c:v libx265 -r 30 -c:a aac -t "$2" -pix_fmt yuv420p "${1}_slide.mp4"


  


  This does create the desired result of a static video without sound but still has an audio channel (which is a pre-requisite for the concat demuxer), but when I concat it with the same command above with videos, the videos' audio desyncs and advances, now the video delays with the audio, which is not obviously feasable for the project. I suspect it has to do with frames and PTS, as MPV (a media player) logs when the duration of the video enters the video that has been concatinated that the PTS is invalid, and says :

  


  Invalid audio PTS: 6.037188 -> 6.553832

Audio/Video desynchronisation detected! Possible reasons include too slow
hardware, temporary CPU spikes, broken drivers, and broken files. Audio
position will not match to the video (see A-V status field).


  


  Tried turning the pictures into an image, and was expecting it to atleast sync with the audio

  


• video streaming using recent frames only

  30 juillet 2019, par Max Paython

  I am trying to create a live video chat application, I am dealing with latency problems but these problems are not produced by encoding or overhead, rather from the definition of streaming itself.

  For example if I start the client program 15 seconds after the server program (start listening the port 15 seconds late), the client tries the play the 15 second old stream. (and fails because of the lack of I-frames introduced by x264 zero-latency, but this is not important)

  But I am trying to make it live. Old frames should be discarded and and the most recent frame should be showing (to a 1 second buffer maybe). I failed at doing this so I wanted to wait streaming until the peer is connected, but my goal is actually the first one.

  I am using ffmpeg for streaming. Can ffmpeg wait streaming until a single client starts listening (or connects) the port.

  How does applications like Skype or Hangouts handle this ? In these applications, if a user suddenly stops listening to the port (internet error for example) for 5 seconds, the chat will continue regularly after the connections is restored, and that user will not see the 5 second old frame but the new one. However I was not able to achieve this.

• ffmpeg single output container with 4 audio channel to AWS IVS

  28 juillet 2021, par Yusufu

  Trying to send 1 video and 4 different audios to AWS IVS in single container. Which container or muxer should I use ?
FLV doesn't support multi audio, matroska doesn't support rtmp I guess link.
3GP and mp4 containers doesn't give me error but neither video showing on IVS

  


  For simple try I am using this command.

  


  ffmpeg -re -stream_loop -1 -i sample.mkv -r 30 -c:v libx264 -pix_fmt yuv420p -profile:v main -preset veryfast -x264opts "nal-hrd=cbr:no-scenecut" -minrate 3000 -maxrate 3000 -g 
60 -c:a aac -ac 2 -ar 44100 -vb 400k -maxrate 400k -minrate 400k -bufsize 800k -movflags frag_keyframe+empty_moov -f mp4 rtmps:someurls


  


  the above command doesn't include multi audio output so just trying to send except the flv format