POV-Ray : Newsgroups : povray.unofficial.patches : UberPOV is back : Re: UberPOV is back Server Time
21 Apr 2024 08:33:00 EDT (-0400)
  Re: UberPOV is back  
From: clipka
Date: 1 Sep 2016 07:00:58
Message: <57c80a6a$1@news.povray.org>
Am 01.09.2016 um 10:54 schrieb Mr:

> Though I already spent quite some time to emulate Blender Shaders in POV as they
> get exported, I would like to try to link the exporter to these new shading
> models to find out if it gives closer results.
> 
> Oren Nayar seems straightforward because Blender has its implementation of that
> very model, but what would you say is the closest to Lommel-Seeliger in Blender
> shading models?

Since I'm no Blender user, I need to do a bit of guesswork here...

> -Fresnel?

Nope. Fresnel isn't actually a shading model, but rather a mathematical
term that appears in various shading models. From what I gather on the
internet, it seems that in Blender you can slap it onto every proper
shading model as an option (whether it makes sense or not; I suspect
some models may be sophisticated enough to already account for the
fresnel effect by themselves), and you can do the same in POV-Ray 3.7.1
by using the `fresnel` keyword directly in the `finish` block.

To the contrary: As you view the surface at a more shallow angle,
applying a Fresnel term will make the diffuse component fall off; using
the Lommel-Seeliger model, the diffuse brightness instead intensifies.

> -Lambert with a Normal input ramp?

Dunno. I wasn't able to find any reference to "Normal input ramp" on the
Internet, so someone would have to explain to me what a "Normal input
ramp" is, and how it would be plugged into the Lambert model.

> -Minnaert?

Nope. Again, to the contrary: Minnaert also makes the diffuse component
fall off at shallower angles.


I suspect Lommel-Seeliger is no suitable substitute for any of the
Blender shaders. And if you were to try to go the reverse route and try
to simulate Lommel-Seeliger in Blender, a mix of Lambert and Oren-Nayar
with a particular set of parameter settings (possibly 100% roughness,
but that's just a quick guess) might actually be the best approximation.


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.