This Young Rival video is a 3D movie done as moving random-dot stereograms. It pretty much only works if you full-screen it, or get your nose right up against the glass. It's probably impossible to see on a phone. They made it with a Kinect. I'm curious about what the limitations are here. Is the chunkiness of the models because of the scanner, or because the "pixels" have to be large for the stereogram trick to work? Does it have to be black and white because color would spoil the illusion? Also, I wonder if you could encode two different videos in here at once, depending on which focus spots your eyes locked on.
This Night Bus video lets you move a slider left and right to see two different versions of the video. The cool part is that it also changes the audio mix proportionally: all the way to the left and you hear more of the girl's voice, and vice versa:
It's hard for me to watch this, though, because I always wonder what I'm missing in the other one. (I'm also the kind of person who read text-adventure books from start to finish.)
A few years ago Rihanna released a video in two versions, one all goth and one all kawaii, shot on the same set, that would have benefitted from this treatment. (Some people put together side-by-side edits of it.) There are a lot of videos out there where there are multiple versions with censorship edits, or extended versions, or alternate endings, or movie-sountrack scenes dropped in, but it's rare for the versions to be both different and similar enough for panning to be interesting.