I want to join two mp4 files to create a single one. The video streams are encoded in h264 and the audio in aac. I can not re-encode the videos to another format due to computational reasons. Also, I cannot use any GUI programs, all processing must be performed with Linux command line utilities. FFmpeg cannot do this for mpeg4 files so instead I used MP4Box:
MP4Box -add video1.mp4 -cat video2.mp4 newvideo.mp4
Unfortunately the audio gets all mixed up. I thought that the problem was that the audio was in aac so I transcoded it in mp3 and used again MP4Box. In this case the audio is fine for the first half of newvideo.mp4
(corresponding to video1.mp4
) but then their is no audio and I cannot navigate in the video also.
My next thought was that the audio and video streams had some small discrepancies in their lengths that I should fix. So for each input video I splitted the video and audio streams and then joined them with the -shortest option in FFmpeg.
Thus for the first video I ran:
avconv -y -i video1.mp4 -c copy -map 0:0 videostream1.mp4
avconv -y -i video1.mp4 -c copy -map 0:1 audiostream1.m4a
avconv -y -i videostream1.mp4 -i audiostream1.m4a -c copy -shortest video1_aligned.mp4
Similarly for the second video and then used MP4Box as previously. Unfortunately this didn't work either. The only success I had was when I joined the video streams separately (i.e. videostream1.mp4 and videostream2.mp4) and the audio streams (i.e. audiostream1.m4a and audiostream2.m4a) and then joined the video and audio in a final file. However, the synchronization is lost for the second half of the video. Concretely, there is a 1 sec delay of audio and video. Any suggestions are really welcome.