//=time() ?>
This is a side-by-side comparison - the light comes from close to where the camera is, you can see where there is a sort of projection of the faces. It is subtle, and I've been thinking for a while how to fix it - as a plus, this solution should extend easily to point lights.
Fixed the aliasing in my previous solution for directional lighting. Went from a 2d solution that closely mirrored the raycasting of the display function to a 3d solution that looks at the amount of absorptive material between each voxel and the light source.
seeing a major improvement in aliasing with the new tile-based approach (there is supersampling involved)
This has actually been a pretty major fix - I've got a lot more usable range with the transparency now, things look way more dynamic
@R4_Unit Got that working, kind of interesting how its casting such defined shadows - I think it has to do with the order in which the upward rays are evaluated - I am working with rgb data instead of states, too, so I have an alpha threshold instead of checking for state=0