On Thu, 2001-12-13 at 20:14, Ben Bridgwater wrote: > > > also, maybe becuase it was > > recorded by a unsteady hand using a one of those vhs cam > > recorders(?)...as such, instead of people just moving, you get the whole > > image canvas also moving becuase of the 'unsteady hand' when recording. > > Interesting point! I guess if the picture is moving around more than > desired then maybe you're killing CPU cycles by having to use a wider > motion search in the MPEG compression to achieve the same result. It'd > be interesting to take a "shaky" video, software image stabilize it, and > do an MPEG compression time/size comparison. another crazy programmer trying to take over the world! lol i guess this would be like trying to stop the earth from turning on it's axis. one would need to find one pixel which remains consistant, or map part of the video that doesn't change (outline a chair, corners of walls, etc) and then would need to submit that to an equation that equalizes and stills all motion. ...mm...thinking more on this makes it seem more easily do-able. ..interesting indeed. > > Ben >