I always wondered about this. I have the same TV series as local MP4 files: one with English audio, and another with Greek. I thought I could just extract the audio tracks and use them to build an MKV file with multiple audio, but it always ended up with an audio sync error. One track would always be in sync at the beginning, but 20min in could be out of sync by as much as 5seconds.
How do people build multi-audio files if the audio tracks aren’t part of the original source?