> In some commercial renderer I'm using, tweaking the BSP size and depth for every
> scene is key to getting reasonable render times. In Pov it seems to take care of
> itself, from reading what I could find about it. That would be great, but does
> it? Or would we benefit from even more acceleration if we tweak every time? if
> so, with which parameters and with what tools or statistics to set them? (In
> the renderer I talked about, I use the diagnostic modes showing some blue to
> red false color dynamic for BSP "size" or "depth", one at a time).
> 
> 

If you mean "binary space partitioning", POV-Ray 3.6.2 don't use it. It 
use a simple bounding boxes testing protocol.

The 3.7 betas do, optionaly, use an octree structure invoked by the 
"+bm2" switch. You can use it from the command line or from an .ini 
file. It can accelerate the tracing. The depth is dynamicaly set at 
parse time.


Alain

Post a reply to this message