I don’t feel like writing a big post explaining all of it, so the summary is that I spent a few hours playing with colorspaces, Hilbert curves, and image generation, and produced some stuff that looks kind of neat. I’ve posted everything in one big wad, and you’re welcome to poke at this ugly undocumented code if you like. I also posted prior Python art scraps.
All the images in this post are shrunken down — the full size loss-less PNGs are on the github with the code. This is the kind of thing it kicks out:
Basically, I walk a 3D color space via a 3D Hilbert curve, mapping it to a 1D scalar. I then walk that 1D scalar on a 2D Hilbert curve, rendering each color in place in an image. Here’s an example of a 1D sequence of colors walked:
From there, I fiddled with the various outputs in Paint.NET to make pretty pictures, e.g.:
(Original here) I was able to render the above one at a ridiculous resolution (8192×4608), so it looks awesome spanning four monitors.
Check out the github dump for more.