The problem with the way I shot the blobs at the surface is that they 
are just randomly shot at the surface. Photon-mapping has a more 
efficient way of doing this. The photons are shot at the target object 
in a spiral pattern, and therefore cover the surface more completely 
with fewer samples.

Chris Huff's blob pattern would be a great benefit. Is that part of 
MegaPovPlus? If so, I could get started on something right away :).

Christoph Hormann wrote:

> 
> This 'scattering photon mapping' is definitely an interesting idea.  I 
> don't know it it will work efficiently in practice but that's difficult 
> to tell.  Note that it is a completely different approach towards 
> scattering than usual.
> 
> But your blob based technique could be much more efficient with Chris 
> Huff's blob pattern.  This way you would avoid the need for the multiple 
> object patterns and don't need a slow proximity pattern either.
> 
> Christoph
> 


-- 
Samuel Benge

stb### [at] hotmailcom
See my website@: http://www.goldrush.com/~abenge/Top/index.html

Post a reply to this message