After writing the spiral image normalization algorithm I promptly forgot how I did it. I then had to re-teach myself to update for double the resolution. My math totally broke down on the final iterations. I had to brute-force parts of this version but now it accepts a 33px image. If you’re wondering why it’s 33 and not 32, it’s because the center pixel is ignored. The sphere pole is composed of 16 facets which use the first spiral iteration.