But yeah, definitely a good idea to do the rendering step in cops. Wouldn’t have thought of that :)

Haha, I’ve been there too. I think I quickly hacked together some cops thing in a solver that would add a slice of each frame onto the previous. Took about 10min to ‘scan’ a single image that way 🫠

That’s super cool Jakob! I definitely need to have a look too at sampling the data straight into cops. Seems much more optimised than rendering each frame in karma :D