I'm processing a load of videos to concat together using Java on a Linux server.
First, I mix in an audio track with this process, depending on if the video already has audio or not.
ffmpeg -y -i video.mp4 -i audio.mp3
-filter_complex "[0:a]aresample=async=1,volume=1.0[a0];
[1:a]aresample=async=1,volume=1.0[a1];
[a0][a1]amix=inputs=2[a] -map 0:v -map "[a]" -c:v copy
-c:a aac -strict -2 -shortest output.mp4
Or if there is no existing audio
ffmpeg -y -i video.mp4 -i audio.mp3 -af apad
-map 0:v -map 1:a -c:v copy -shortest output.mp4
With this code below in the Java application
if(getVideoAttribute(video, "a:0", "duration") != null)
pb = new ProcessBuilder(
"ffmpeg", "-y", "-i", video.getAbsolutePath(),
"-i", audio.getAbsolutePath(),
"-filter_complex", "[0:a]aresample=async=1,volume=" + videoVolume + "[a0]; [1:a]aresample=async=1,volume=" + audioVolume + "[a1]; [a0][a1]amix=inputs=2[a]", "-map", "0:v", "-map", "[a]", "-c:v", "copy", "-c:a", "aac", "-strict", "-2", "-shortest",
outFile.getAbsolutePath()
);
else
pb = new ProcessBuilder(
"ffmpeg", "-y", "-i", video.getAbsolutePath(),
"-i", audio.getAbsolutePath(),
"-af", "apad", "-map", "0:v", "-map", "1:a", "-c:v", "copy", "-shortest",
outFile.getAbsolutePath()
);
I then add all of the video files into a textfile, and concat that text file using
ffmpeg -f concat -i list.txt -safe 0
-c:v copy -c:a aac -strict experimental output.mp4
Using this Java below
ProcessBuilder pb = new ProcessBuilder(
"ffmpeg", "-f",
"concat", "-i", Constants.TEMP_DIR_PATH + listFile.getName(),
"-safe", "0",
"-c:v","copy","-c:a","aac","-strict","experimental",
Constants.TEMP_DIR_PATH + out.getName()
);
In the final output, the audio is slightly out of sync, i.e. audio from one video runs over into the next video etc.
I checked the audio/video durations for each video and they are indeed slightly different (video duration is the first number, audio track second)
6.480000
6.528000
0e9e449b-f53e-473b-a91a-72e38ecfc789_3840.mp4
12.760000
12.736000
06bd1e70-b9aa-4c26-86d2-98fc421f6729_3840.mp4
17.520000
17.493000
f2dc7fc0-d4ac-4a82-88cb-777a16a1f818_3840.mp4
7.880000
7.850000
03681d53-4532-48e0-b3e4-cc1e7170ae65_3840.mp4
6.625000
6.592000
219d174d-9df6-41ec-b95e-2ec4137e07a4_3840.mp4
14.541992
14.485000
Which I assume is the cause of the problem.
Can anyone help me either change the way I mix them so that the video and audio is always the same length, or show me a way to use concat so it adds silence at the end of the audio so they are the same length?