So it turns out that for 0.19, frame accurate comercial cutting just isn't realistic. While there isn't anything wrong with the code in mythtranscode (that I'm aware of), it turns out that myth just doesn't count frames in exactly the same way. In addition, the way the index is generated in myth isn't very reliable (a cutpoint will happen at a different point depending on whether you watch the video from the beginning or whether you use fast-forward/rewind), and myth's different ways of storing them (the dvbrecorder and ivtv recorders don't use exactly the same algorithm to generate the index, and using mythcommflag to rebuild the index results in yet another answer). the above only applies to MPEG2 as far as I can tell.
I'll take a look at this after 0.19 is out the door, but it is likely that to get everything is sync may need some invasive changes, so I'm not going to start now. For the time being, the following solutions may work: 1) regenerate the key-frame index using 'mythtranscode -b' before selecting cut points 2) run mythtranscode without a cutlist before generating the cut points 3) leave at least 1 sec extra to prevent excessive cutting Either 1) or 2) should result in frame accurate cutting, though I still can't gaurantee it. 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. .Geoff _______________________________________________ mythtv-dev mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
