Post
by badmartialarts » Sun Feb 27, 2005 10:45 pm
After Effects and Premiere prefer every frame to be a keyframe in a video. XviD and DivX under standard settings do not do this. Think of it like this. What Premiere wants is a video that each frame is the whole thing, like this:
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
But in order to compress better, XviD/DivX use some fancy techniques to only keep track of the differences between frames rather than the full frame itself, like this:
|------|--|---|--------|----|------------|--|
So, if in Premiere you wanted to cut the first 3 frames off the clip, it would no longer have that first frame to get the rest of the picture from and you end up with that green mess that you see in a lot of (hate to stereotype) LBZ/Sum Naruto videos. Plus, if it's encoded weird, sometimes it'll just crash Premiere (as it's doing to you) whereas a video player will just skip the weird frame and move on.
You're best bet would be to open up the file in VirtualDub and resave with a HuffYUV codec. The file will be huge but they probably won't crash stuff anymore.