following some advice from the doom9 fori, i now do the audio demux with megui's hd streams extractor, as that fills corrupted audio frames with silence, thereby keeping sync. that also means that i have to do the trimming with avs cutter (also in megui), so i delegated these two tasks there instead of dginv, and now i get almost perfect sync. i still do the indexing and avs creation (and if needed, image cropping and deinterlacing) in dginv -- it's still the best for that task, not mentioning a huge performance boost
, so i just needed to adjust my workflow.and in case a video frame is dropped by dginv (i guess if dginv encounters a corrupted video frame header, you skip it just like a corrupted audio frame header -- right? but if in-frame data is corrupted (by a signal glitch), as you don't check that (as we've established before) the frame is still kept, if my understanding is correct), it is always much easier to drop a few audio frames using avs cutter and a re-encode of the track, than reencoding the entire video track.
anyways, thanks for the clarifications, it seems now my workflow is set up right at last.
just one more question: do you log dropped audio/video frames (like position, no. of dropped frames) anywhere?
