BSSRDF render technique

[playmesumch00ns]

AFAIK it's not been implemented yet, it's still relatively new, tho no doubt it'llbe in the next version of Brazil or FR, given the rate the R&D guys work there.

it's a pain in the arse, but possible to implement in BMRT or prman 11, or mentalray, if you can get your head round those equations:)

[qu2k]

what is it?

[Pyro2301]

http://graphics.ucsd.edu/~henrik/papers/bssrdf/


Henrik is just friggen awesome. :D

-Victor

[erikals]

Yep,
That's the .pdf I got the pictures from. The render technique looks really awesome, can't wait to use it. The way I see it it will revolutionize the way computer generated humans are made in CG pictures/movies. Awesome stuff. I hope to see more of it soon. Maybe it's not too far away? I noticed that the document was dated back to Siggraph 2001.

playmesumch00ns:
Brazil is cool. I don't think I have heard of FR though, will have to check it out. Pixar Render Man is way too advanced for most users. But what about Lightwave? I guess maybe it will take a while before they will make it?.. if they ever will...

I wonder if it would be possible to combine HDRI rendering with BSSRDF, that would be cool...

(I don't know too much about this, so I might be saying things that doesn't make sense, like combining HDRI with BSSDRF)

Thanks for replies everyone, :)
Erik

[playmesumch00ns]

BRDF = Bi-directional Reflectance Diffusion Function = Phong, Lambert, Blinn etc etc

It's not really that difficult an extension to current practice, which is why I suggested one could code it without too much trouble The clever part comes in optimizing the thing to render quickly. The times they quote in the paper are, I feel, slightly less than accurate. I think they're quoting times for actual pixel generation and not including any pre-processing time.

Anyway won't be too long before this is in most renderers as standard. The trouble comes in tweaking the settings!

And yes it's perfectly possible to combine HDRI-based lighting and the BSSRDF technique presented in this paper: HDRI-lighting just tells the renderer how the light gets to the surface, the BSSRDF tells it what the light does when it hits the surface. Although I think it would be very, very slow.

[andrewbutts]

Hiya,

I would not be suprised if Worley Labs' (http://www.worley.com) "G2" plugin for Lightwave contained an implementation of the "fast" BSSRDF rendering method described in the paper given the past collaboration between Worley and Steve Marschner (and who knows who else at the Stanford graphics lab?). I have the impression G2 has two (more?) modes, nice mode and fast mode. I'd expect fast mode to use this method.

Keep in mind that the technique described in the paper produces only "faked" subsurface scattering.

Another Lightwave plugin, OGO_HIKARI, does actual subsurface scattering simulation for sure, but at the cost of speed.

Andrew

[erikals]

Hmm,
looks like maybe you are right, on this page
http://www.worley.com/G2/g2_scatter.html#startit
there is a (geeky details) link that tells more about it.

I just think it's strange that none of the renders quite live up to the images shown in the .pdf document Pyro2301 posted a link to earlier, not for G2 neither for Ogo Hikari's plug-in, though I think maybe Ogo's look a little better than G2's, hard to tell.

Thank you very much for the info, I didn't know G2 / OH's plug-in had such features. This will help me very much trying to find the best way to render human skin.

I guess maybe you saw this post from The Ripper, using G2. From what I understood though, he didn't use G2's SSS as he didn't find it very good neither efficient. Maybe he should check out Ogo Hikari's plug-in?
http://www.cgtalk.com/showthread.php?s=&postid=441169#post441169
(Edit: he also used HDRI)

Thanks everyone,
Erik

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.