Marton Balint: > > > On Mon, 27 Apr 2020, Marton Balint wrote: > >> >> On Sun, 26 Apr 2020, Andreas Rheinhardt wrote: >> > > [...] > >>> >>> And this filter does more than just repacketizing the samples: It also >>> discards the timing of its input and makes up completely new timestamps >>> and durations. This needs to be documented. >> >> Ok, will do. > > I have given this some additional thought. Maybe it is better to keep > the original timestamps instead of creating new. Keeping the start > timestamp should definitely be supported by default and to make this > filter analogous to af_asetnsamples, I am inclined to keep other > timestamps as well. The code does not even look ugly, and > retimeinterleave does not care about the timestamps anyway, just the > durations. So I think I will resubmit a new version which will make an > effort to keep the original timestamps. We can make it switchable if > later it turns out that regenerating would be better. > Sounds very good. In fact, the only reason I was ok with that is that I thought this bsf to be a helper for retimeinterleave and I had the impression that retimeinterleave depends on the timestamps being set in the way the earlier version set them.
- Andreas _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".