cos weighted, with what appears to be more directed fades from light

Fixed hemisphere distribution to use a cocentric based approach, scene is far better lit now

Cos weighted sampling, noting the more dim reflection of the still malfunctioning emissive sphere.

sample rate for previous image, excluding GI bounces

Is light falloff 1/sqrtdist or 1/dist^2? This is 1/dist^2 which looks exceedingly dark except near the light sources.Emissive material now works, except not directly into the camera. Fixed a bias issue. Changed point lights material to apply a fade for a bloom-like effect.

Max 128 samples, took 20 minutes to render. the lights intensity was turned down by a lot.

No changes from project 10. I have a question about the need to calculate the light directly hitting surfaces as GI light, rather than as direct light from previously.