Very cool! Just wanna point out that Mirror + Rotate is really just 3 different mirrors. Of course it may be more interesting to try to characterize the visual domains in terms of those 3 mirrors rather than trying to do so obfuscated between mirror and rotate.

Equivalent to an iterative displacement map? (Or is the displacement map dynamic as well?)

Very cool regardless.

Reminds me too of playing around with the feedback from pointing a video camera at the monitor displaying the camera's output. (Probably most famously demonstrated in the opening credits to a particular period of the TV series Doctor Who.)

Of course by both rotating the vide camera, zooming in on the display, you have two of those three transforms.

I’m a sucker for math visualizations on a website. They’re always sick.