The Beam Radiance Estimate for Volumetric Photon Mapping

Oiram · Post by **Oiram** » Fri Apr 18, 2008 8:44 pm

A new method for efficiently simulating the scattering of light within participating media.

http://graphics.ucsd.edu/~wjarosz/publi ... 8beam.html

SmartDen · Post by **SmartDen** » Fri Apr 18, 2008 8:53 pm

Wow! that whould be cool to have

Post by **fused** » Fri Apr 18, 2008 10:12 pm

I think the word "estimate" in the title and description makes it a bit improbable, that we get that...

lycium · Post by **lycium** » Sat Apr 19, 2008 2:06 am

nice paper, it's cool to see how far jarosz has come since his early renderbitch days :) i'm looking forward to his siggy08 papers too.

@fused, there's nothing wrong with estimates, in an unbiased sense; all monte carlo methods are just estimators.

i really dislike photon mapping but its strength in efficient volumetric rendering is undeniable, which this paper extends...

Big Fan · Post by **Big Fan** » Sat Apr 19, 2008 8:48 am

nice speed up too

PureSpider · Post by **PureSpider** » Sat Apr 19, 2008 9:28 am

Would be great for SSS, if I'm right?

lycium · Post by **lycium** » Sat Apr 19, 2008 5:35 pm

yesindeed. however, indigo seems to be quite efficient with volumetrics (including sss) so it's probably not necessary. a nice thing about unbiased monte carlo methods is that they give the result that everything else wants to produce ;)

Azazeo · Post by **Azazeo** » Thu May 01, 2008 8:01 pm

SSS give nice result, but take a LOT of time...

2 IES Sourses, mirror, lens, prysm...
Better have some special methods to do such things (at least global dusty-air parameters

)

PS Sorry for my English

lycium · Post by **lycium** » Thu May 01, 2008 8:39 pm

fantastic render! any pc/time infos?

Azazeo · Post by **Azazeo** » Thu May 01, 2008 9:01 pm

notebook - Acer 5520G (Turion X2 TK-55 (1.8 GHz), 3GB RAM)
Time - about an 1 h (don't remember exactly)