On Thu, 2006-01-12 at 20:58 -0800, Geoffrey Hausheer wrote: > If you observe cases where the cuttig appears to be more than ~1sec > off, that would likely indeicate a bug, which I'd be interested in > looking at.
There is a known bug where that would be the case, see #799. DTVRecorder just scans the packets for sequence and GOP headers, which usually more or less works, but sometimes there are other sequences of 4 bytes that match in the packet so we over-count. We do the same thing in mpegrecorder, but there we never get the structured error of a frame count doubling because the encoder simply doesn't insert such data in the packets. FYI As far as I know there is no deep reason why MARK_GOP_START is still used instead of MARK_GOP_BYFRAME. The rebuilt maps should probably always use MARK_GOP_BYFRAME, since we don't know if the keyframe distance is fixed unless we first scan the entire file. And it would probably make life easier if MARK_GOP_START just went away... -- Daniel _______________________________________________ mythtv-dev mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
