Bluedan
29th May 2002, 20:13
Some days ago I tried to edit an Xvid encode using Virtualdub 1.4.10.
When I tried to find keyframe positions to point to in the corresponding chapter list, I noticed that keyframes where situated one frame preceding the actual scene change, means: [B]not[\B] the first frame of the new scene. Only some SC showed right constellation.
At first I thought it was VD not correctly displaying info on frames beacuse it showed strange behaviour as I moved through the stream: jumping from i-frame position backwards always resulted in going frames ahead with first button hit, then backwards as intended with the next strikes.
Also the keyframe positions delivered from chapterXtractor (->synch to keyframes)
showed the same: i-frame preceding scene change.
I wonder how keyframes are detected by those programmes. There is some kind of frame header, right?
I cannot imagine a movie with that false frame sequence to playback smoothly. But it does.
Unfortunately I cannot tell exactly the version of XVid I used then, because I encoded it more than a month ago.
Also I don't know if VD needs special skills to edit XVid- video, but I think not.
Maybe this is outdated anyhow and the error corrected if it's XVid to blame, I looked at changes made in CVS core on UManics site, but understood little, unfamiliar with that invironment and I am not a programmer.
Anyone had this issue before??
When I tried to find keyframe positions to point to in the corresponding chapter list, I noticed that keyframes where situated one frame preceding the actual scene change, means: [B]not[\B] the first frame of the new scene. Only some SC showed right constellation.
At first I thought it was VD not correctly displaying info on frames beacuse it showed strange behaviour as I moved through the stream: jumping from i-frame position backwards always resulted in going frames ahead with first button hit, then backwards as intended with the next strikes.
Also the keyframe positions delivered from chapterXtractor (->synch to keyframes)
showed the same: i-frame preceding scene change.
I wonder how keyframes are detected by those programmes. There is some kind of frame header, right?
I cannot imagine a movie with that false frame sequence to playback smoothly. But it does.
Unfortunately I cannot tell exactly the version of XVid I used then, because I encoded it more than a month ago.
Also I don't know if VD needs special skills to edit XVid- video, but I think not.
Maybe this is outdated anyhow and the error corrected if it's XVid to blame, I looked at changes made in CVS core on UManics site, but understood little, unfamiliar with that invironment and I am not a programmer.
Anyone had this issue before??