[mythtv] Mythtranscode .mpg => .mpg not .mpg => .nuv

[Geoffrey Hausheer]

On 11/7/05, Cory Papenfuss wrote:
>         So, essentially this is just a message saying that the PTS doesn't
> match what's expected based on the fps interval?  In order to play it
> "correctly," the player would have to present the frame in mid-otherframe
> if I'm understanding correctly.
right.

>         This was at no debug, but I agree.  If it's mostly harmless, then
> just another case of too much info being bad... :)  FWIW, around some of
> the actual "Building" lines, I've got some much more serious jumps:
> PTS discrepency: 1077926 != 1077915 on B-Type (10)
> PTS discrepency: 1080942 != 1080918 on P-Type (11)
> PTS discrepency: 1083932 != 1083921 on B-Type (12)
> PTS discrepency: 1099504 != 1086924 on B-Type (13)
> PTS discrepency: 1102564 != 1089927 on P-Type (14)
> Building P-Frame: 15
> PTS discrepency: 1105624 != 1107945 on B-Type (0)
> PTS discrepency: 1108684 != 1110948 on B-Type (1)
> PTS discrepency: 1111746 != 1113951 on I-Type (2)
> PTS discrepency: 1114843 != 1116954 on B-Type (3)
> PTS discrepency: 1117864 != 1119957 on B-Type (4)
Now this is interesting, and shows a minor bug.

basically, between frame 10 and 14 you lost 4.2 frames, and the PTS is
being skewed to cover them.
at the 'Building P-frame: 15' line, mpeg2fix inserted 5 duplicate
frames (all copies of frame 14).  It doesn't print out how many frames
it inserted, but I can tell from the PTS jump.  The bug is that it
should have only inserted 4 frames rather than 5.  This would have
left the deltaPTS smaller than it is in the next frame (0).  But
ideally, we'd actually spread out the inserton over several frames, to
minimize the video stuttering.  Anyhow, it is doing what it is
designed to.  Be aware, as I said in a previous mail, if you actually
play back the video, you may notice some weird artifacts at the bottom
of the screen where the frames are inserted (in the above case, about
12sec from the beginning).  I found a partial work-around to this, but
it appears to be a bug in libmpeg2 when low_delay is set.  So don't
start pitching your original videos quite yet :)

.Geoff