If you zoom in on the trunk to the point where you can see the blocks of pixels, VC-1 is a lot cleaner. Strange how the MPEG image seems sharper with all the garbage it is carrying. It would be interesting to see how they compare in a scene with closely spaced lines. When you look at the key slot, MPEG has it made up of several different square objects.
BTW, how do we know any difference in this scene is not simply the compressionists' choices? MPEG may have maxed their peak BW, VC-1 let it be "typical"