my ambient occlusion experiment turned out way more real-time than I thought it would lol
since the outermost faces never get occluded, there's no need to make them part of the lightmap
in the .gif above they just used whatever color happened to be at (0,0), but, that's fixed now
in the .gif above they just used whatever color happened to be at (0,0), but, that's fixed now
the sharp highlights in corners happen bc my algorithm takes shortcuts, creating some artifacts
for example, point P will be too dark because of the fact that A doesn't occlude B, and so this region C should technically not be part of the shadow casting, but it is
for example, point P will be too dark because of the fact that A doesn't occlude B, and so this region C should technically not be part of the shadow casting, but it is
solving this "occlusion of occluders" problem is, very hard, if I want it to be as fast as it is now
I'd have to resolve overlap for all shadow casters from every lightmap texel
it's a classic O(n²) problem where it goes from real-time to taking an hour or something >.>
I'd have to resolve overlap for all shadow casters from every lightmap texel
it's a classic O(n²) problem where it goes from real-time to taking an hour or something >.>
Loading suggestions...