[MPlayer-dev-eng] Re: [PATCH] dvr-ms fixes for pts, key frame detection and seeking

[John Donaghy]

>  if the container gives constant-distance timestamps that
> will produce the same fps. The value of sh_video->fps doesn't matter all
> that much anyway. Video timing is done based on the timestamps.


Hi

I've got a prototype build of the asf demuxer now that plays HD dvr-ms fies
ok except in the occasional situation where there is a significant jump in
the timestamps between 2 video frames. This only appears to be happening in
progressive ATSC recordings at the end of a commercial break. At this point
the AV sync goes out and doesnt recover. Below are some timestamps from a
sample I have.

Can anyone suggest what I can do in the demuxer to make this stay in sync?

AUDIO: vid_frame_num=690, pts=235.39157, diff=0.12800
VIDEO: vid_frame_num=691, pts=236.01378, diff=0.01669
VIDEO: vid_frame_num=692, pts=236.03047, diff=0.01669
VIDEO: vid_frame_num=693, pts=236.04713, diff=0.01667
VIDEO: vid_frame_num=694, pts=236.06382, diff=0.01669
VIDEO: vid_frame_num=695, pts=236.08051, diff=0.01669
VIDEO: vid_frame_num=696, pts=236.09721, diff=0.01669
VIDEO: vid_frame_num=697, pts=236.11387, diff=0.01667
VIDEO: vid_frame_num=698, pts=236.13056, diff=0.01669
VIDEO: vid_frame_num=699, pts=236.14725, diff=0.01669
VIDEO: vid_frame_num=700, pts=236.16392, diff=0.01667
AUDIO: vid_frame_num=700, pts=235.95597, diff=0.56440
VIDEO: vid_frame_num=701, pts=236.18061, diff=0.01669
VIDEO: vid_frame_num=702, pts=236.19730, diff=0.01669

VIDEO PTS JUMP HERE
VIDEO: vid_frame_num=703, pts=236.61399, diff=0.41669
VIDEO: vid_frame_num=704, pts=236.63068, diff=0.01669
VIDEO: vid_frame_num=705, pts=236.64735, diff=0.01667
VIDEO: vid_frame_num=706, pts=236.66404, diff=0.01669
AUDIO: vid_frame_num=706, pts=236.08397, diff=0.12800
VIDEO: vid_frame_num=707, pts=236.68073, diff=0.01669
VIDEO: vid_frame_num=708, pts=236.69742, diff=0.01669
VIDEO: vid_frame_num=709, pts=236.71409, diff=0.01667
VIDEO: vid_frame_num=710, pts=236.73078, diff=0.01669