I try to let ffmpeg.exe encode and mux a raw video stream (input) and an PCM audio file (input) to a MP4 file (output). The raw video stream is supplied by a named pipe as it is generated in realtime.
The command line is like: ffmpeg.exe -f rawvideo -pix_fmt rgb24 -r 30 -s 1280x720 -i \\.\pipe\pipename -i audiofile.wav -c:a ac3 -ab 128k -ar 44100 -c:v libx264 -preset slow -crf 20 -pix-fmt yuv420p output.mp4 All seems to work fine, there is no error message, and the file resulting file runs without problems in Windows Media Player and Quicktime Player. But with VLC player, the sound seems to be messed up, packets are mixed during playback. A closer look to the file resp. to the logfile of ffmpeg (options -report -debug_ts) tells me that all audio packets of the whole audio input file are processed between the first few video frames. This means that the last audio packet (timestamp 300s) is already written before video frame 30 (1s) is processed. This probably leads to an improberly muxed result file. When I simply try to write the same to an MTS oder MOV container, other problems with all players seem to occur (seeking does not work properly), so I think this is a general muxing problem. How can I get audio and video processing synchronized? Is it necessary to supply video and audio in a balanced datarate, thus control the audio stream through a pipe as well as the video stream? If so: How many audio preload would be useful/necessary/common? Thank you! _______________________________________________ ffmpeg-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-user
