I have an old movie with two versions of same soundtrack, both are encoded with OGG Vorbis, encoding parameters are nearly identical, and both tracks are encoded from damaged VHS copies (on different places hopefully) of same movie. I want to cut both tracks in several pieces and re-assemble one track from them (but I don't want to re-encoding them because their quality is quite low already).
But, when I try to cut-and-merge OGG Vorbis soundtrack (by vcut from vorbis-tools 1.0.1, precisely to within sample), I got OGG file with multiple logical streams (chained streams). And then, when I try to mux it with video (into OGM - by VirtualDubMod, or into Matroska - by mkvtoolnix), I got video OK but sound ends at first merge-point.
Then, I found several methods that, at first looking, allow me to merge multiple OGG files into one logical stream. All of that method consist mainly of muxing each OGG Vorbis stream into separate file of some kind of a/v container (OGM, Matroska or mp4), merging that files together and demuxing merged OGG Vorbis stream out of merged file. But, in each attempt, resulting OGG stream becomes slightly longer - about 10-30 ms per merge-point, and contains small but audible (and quite annoying) gaps and claps near each merge-point. I suspect that merging process someway violate some trimming labels of first and last packets in each fragment, or something similar occur.
So, please help me - how can I paste Vorbis chained streams (vcut'ed precisely to within sample) into one logical stream, with same precision? Or, how can I mux .OGG soundtrack with chained streams into OGM or MKV a/v container?
Sorry for my bad English.
PS. I searched this forum, and I found few topics about similar problems (merging of chained streams) but there was said nothing about cut precision. Maybe there is some well-known technique I don't know about...
