There's been a pretty bad regression in compile times recently:
bounds.bounding_sphere.w == -1.0f
only means the culling is disabled (which does not always means the bounding box is invalid
At the moment disabling culling effectively means the bounding…
@Traslev Maybe? We could allow reading the render targets and AOVs from the previous frame. I can see a few use cases and it shouldn't be that hard to support.
I mean the
tilemap.viewmat
is not changed, which is theviewmat
the shadow is rendered with. So you effectively only jitter the tracing / projection direction (I think, not sure) but the…
How is this shady? This is just computing the shadow map local space. I moved it here because I found way more confusing the way it was. Also, doesn't that mess some computations in `shadow_ray_gen…
I've updated the code, but I'm not sure if it's right. The test is disabled and fails even in main if I enable it back.
Yes, I mentioned it at the last week meeting, and you seemed to agree (?). Maybe I didn't explain it properly.
I don't understand what you mean by not changing the shadow projection.
The shading…
It's better to just pass light radius
The problem is, what radius? The original or the one set by the Light module?
The compute_projection_boundaries
receives both.
I've cleaned the code a…
I can't reproduce it anymore, but I'm not sure if it's because the underlying issue has been fixed, or because the recently implemented volume bounds detection and the dither hides it. @fclem ?