> > What would be really neat is to have a bicubic patch as an area light
>
> Yeah - U would have to sample many points on the patch to determine the amount of
> light hitting the surface - bit like focal blur or use point lights distributed
> all over the patch.

You can kind of already do this with radiosity, but it takes too long to get a good
enough distribution of samples to make clear shadows visible.  What I'm talking about
is an area light just like the rectangular implimentation except with shadow ray
tested for a patch instead of a planar rectangle.

> > since it's got u and v built in
>
> What does that have to do with it - unless using point lights all over an object
> - nothing much

In block_area_light, replace the uv of the standard area light with the uv of a
bicubic patch.  You need the parameters of the shape to do more than 4 samples.

Anyway, just an idea but I think it would be neat.

-Mike

Post a reply to this message