[Radiance-general] Occluder cache application?

[Gregory J. Ward]

Hi Lars,

I'm not quite sure how your scene is set up, but I'll try to respond...

> From: "Lars O. Grobe" <grobe at gmx.net>
> Date: February 13, 2007 1:56:24 AM PST
>
> Hi,
>
> as I am going to add some dedicated occluder objects to a rather  
> large scene, I would like to understand some more details of how  
> the caching works to concentrate my work on the right parts of the  
> scene.

Just generally, there's not much point in adding dedicated occluders,  
unless your scene contains a lot of octree or mesh instances, which  
aren't included in the standard occluder cache.

> I have a scene with these light sources:
>
> - daylight by gensky, but hidden behind mkillum sources
> - groups of small light sources, again hidden in bounding mkillums
>
> Now, I guess that for the direct calculation, I will get some speed  
> when the occluders block the small light sources. But what about  
> the light sources generated by mkillum?

Any sources of type "glow" are already blocked by intervening "illum"  
surfaces.  Therefore, additional occluders shouldn't make any  
difference.

> Than, my scene has one more strange feature. More or less all the  
> outer walls are instances. So putting a big occluder around the  
> closed parts of the building would speed-up because I do not have  
> to test for light sources behind the wall. But the only outer light  
> source there is the sky distribution + sun. The sky distribution is  
> an indirect source in Radiance afaik, so will I win anything?

Very little -- just the sun left as a light source, and testing  
against a single source is generally faster than any shenanigans.

-Greg

> Thanks, hope someone can bring some light into my occluded  
> understanding of the cache,
>
> Lars.