How's the decoder going to know what to put in the middle of the picture? The use of 'motion vectors' like this are a large part of video compression. The compressed video might recognize the ball is the same and simply say 'copy this block of pixels (that contains the ball) from that X,Y coordinate in the last frame to this X',Y' coordinate in this frame.' If you crop away the edge of the frame there's no ball in the prior frame anymore. This makes the job much more difficult.įor example, say a ball was at the very edge of the picture in one frame and at the middle of the picture in the next. High compression video codecs start with similar compression techniques but on top of that frames can reference other frames. The compressed data encodes the entire picture. See Yes, that's possible because a JPG image is a single entity. It is possible for some lossy compression format to have lossless transform, for instance on JPEG it is possible to do certain crop or 90° rotate in a lossless way, viz without re-encoding.

Cut video lossless