At 1:23 PM -0400 9/15/05, Manfredi, Albert E wrote: >Gary Hughes wrote: > >> I don't see the connection between better (presumably >> finer grained) motion vectors and GOP length. Could >> you expand on that? > >The B and P frames created between I frames must >accurately portray the updated images as motion occurs. >The better they do this, the greater the gop can be, >were it not for other considerations such as synch time. > Nice try Bert, but there is no connection. Improved motion prediction ONLY means that the encoded differences will be smaller. In a rather back-ass way this could allow longer GOPs, since you might have a few more bits to play with as the predictions grow less accurate because of the lack of accurate anchor frames. But this is not a feature! You need more I and P frames when there is rapid motion and new scene content is being introduced or revealed. GOPs of ANY size are influenced by these scene changes. H.264 does allow for partial updates (partial I frames) in regions where there is new content - but this has NOTHING to do with GOP size. Longer GOPs are possible when there is little change in a scene. I have also seen situations (with MPEG-2) where you can get away with longer GOPs using more P-frame, updates especially for higher frame rate source like 60 and 72p. Regards Craig ---------------------------------------------------------------------- You can UNSUBSCRIBE from the OpenDTV list in two ways: - Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word unsubscribe in the subject line.