On 11/7/05, Cory Papenfuss wrote: > I've got another stream that I suspect is somewhat hosed. Having > not really played with the previous one much, I'm not quite sure that this > is correct. > Opening misc_family_and_airmotor.nuv > #0 PTS:36037 Delta: 0.000000ms queue: 64 > #1 PTS:34381 Delta: 18.400000ms queue: 1 > ptsIncrement: 3003 Frame #: 2 > PTS discrepency: 20878 != 20859 on B-Type (3) > PTS discrepency: 32884 != 32871 on B-Type (7) > PTS discrepency: 44889 != 44883 on P-Type (11) > PTS discrepency: 53904 != 53892 on P-Type (14) > PTS discrepency: 62917 != 62901 on I-Type (2) > PTS discrepency: 65942 != 65904 on B-Type (3) > PTS discrepency: 80923 != 80919 on P-Type (8) > PTS discrepency: 86937 != 86925 on B-Type (10) > PTS discrepency: 92938 != 92931 on B-Type (12) > PTS discrepency: 95946 != 95934 on B-Type (13) These aren't really a problem, though I'd certainly use -f on the stream at a minimim (or you could use -t, doesn't really make a difference). All the above says is that the expected PTS doesn't match what was actualy found (actual on the left, expected on the right). A new frame is inserted whenever actual-expected > 0.8*ptsIncrement (this is regardless of what options are used)
If you specify neither -f nor -t, then nothing will be done about the above (the PTS will be exactly the same as the original (minus the fixed offset)). Most standalone players will ignore the discrepency, and will play at a fixed 29.997 fps anyway. If you specify either -f or -t, the pts will be set to the expected value. The only thing -t (--no3to2) brings to the table is to remove any telecine flags, and remove all half-frames (this is just a flag, no data is removed), which will result in inserting an additional frame every ~6 frames during telecine sections). The interesting message would be 'Buiding P-frame' (which is actually an I-frame, but whatever) telling you a new frame was being inserted. In your stream, the delta is varying, so this is just a not very clean stream. but the largest delta I see is ~17. this is 17/90000=0.18ms, you aren't going to notice any lipsync issues with that small of an offset. In the next version, I'll push that message up to a higher debug level. It isn't very useful except when something goes wrong. .Geoff _______________________________________________ mythtv-dev mailing list mythtv-dev@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
