It appears that when there is a scene change, Compressor (using VBR which does scene change detection on first pass) puts I frames *ahead* of the scene change instead of the first frame after the scene change.
This makes it impossible to add new chapter marks in DVDSP on the first frame after a scene change, unless the video was marked up before encoding.
Of course, I could go back and open the video in FCP and insert missing chapter marks and re-encode the video, if there is time to do this. But, I'd rather have more options for adding new chapters within DVDSP when I am authoring.
A solution might lie with a program that automatically detects scene changes on DV, as an alternative to manually marking up a video with compression markers. FCP's DV stop/start detection isn't much of a help here because it reads meta data captured from tape that indicates camera on/off, not scene change. Maybe there is a FCP plugin somewhere for scene change detection, but it seems to me a better solution would be an encoder/transcoder that put the I-frames where they could be better used later as chapter marks.
If Compressor/Quicktime 's scene change detection won't do this, might another software encodier/transcoder have control over where I frames are inserted so that the video would be marked for I on the first frame *after* the scene change during the transcoding process?
Has anyone found a solution or, is there some setting in Compressor that I haven't found?
Thanks,
Cris