Ik its very simple, the technique used in the shot you showed of chewie uses a conversion process which takes information from the frames each side of the one being processed and uses it in the same maner for the whole image... This early type of conversion makes panning static objects into 3D ones because the program utilizes the object with the most motion as its point of reference.

Although this works it can be very confusing to the eye.

The trick is to take information away thus reducing processing time, simple techniques like adding a slight motion blur then using a sharpening tool to re-enhance the trailing side of an object can artificially create the illusion of an object being layerized.

The layers can be cropped to give depth of field to the whole scene

Then a subtle layer depth can be achieved using focal effects and motion sensing

Note the highly crystalized trailing edge of objects in the left hand image make the frame look like it's layers are offset...

eg:

http://i1046.photobucket.com/albums/b461/gagobaa/vlcsnap-2011-01-11-15h26m26s245.png