My particular problem is related to editing real-time Mpeg-2 caps, have been trying various combinations to maintain synch after edit.

It occurred to me that a better understanding of how audio synch is created/maintained in the first place would be helpful.

GOP structure, Timestamps, scan offsets, I basically understand what these structures are, but am looking for methods to correct them and also to analyze or just look at them to determine differences created by various methods.

Also, if anyone knows what information is contained in a Padding Stream and how it is used, this Padding Stream seems to be a key element in my problem.