Strange Refraction behaviour

[putyi]

During the visualization of an office space, we found that the refraction acts somewhat wired in certain situations. This image shows our color and and space setup proposal we gave to our client. You can see that despite of the low render settings there's nothing abnormal on this image.



This is the color setup requested by our client. They also wanted a sandblasted text on the glass wall, so we mapped there one. Besides that imo these colors look quite ugly, there's something strange going on with the glass wall refraction. It's almost like the glass was a single polygon surface, and so the entire office space behind the glass now rests "inside" the glass. However, the glass wall still has volume, normals are OK (only changed the shader) and also note that the amount of fog seems to be OK.



The shader is a VRayBlend shader with a sandblasted, and clear glass in two slots, and a grayscale text map controlling the blend.

Also, we found that sometimes when using refractive shaders inside a VRayblend shader the fog suddenly disappears.

regards

[vlado]

Can you show more exactly what does the Blend material look like? As for the fog, only the fog from the Base material will be considered. This is because "coat" materials are assumed to be really just coatings without volume.

You may also want to do the sand-blasting by mapping the different properties of a single VRayMtl, rather than blending two materials.

Best regards,
Vlado

[putyi]

Sure, here is how the blend shader looks like:


The two green glass shaders are almost identical except for the blasted having a refraction glossy of 0.8. The glossy on that shader is interpolated.
Both of them having the same IOR, fog, etc. values. I've decided to go with the blend shader because it's more flexible than mapping a single one, and recoloring the map when our not-so-confident client figures that he wants something more or less glossy.

So considering the fog, the base shader's fog settings will affect the entire shader? I suppose this goes for blended SSS shaders too. Now, this explains a lot, because a few days ago, I've tried to blend a glass like shader with some sss to give the material a more volumetric look.

Also I've found that there are some shaders now impossible to reproduce with the current VRay shader accurately. For example, a sub surface scattering shader should turn transparent when thin enough. This phenomenon can be widely seen in macro photographs of plants, where part of the plant is almost glass like and becomes more opaque when the stem gets thicker. And this goes for almost every material, for example, a suitably thin film of metal will also be transparent.

Anyway, the reason why I'm boring you with my shader model unification concerns, is that I've always looked at the VrayBlend shader as "The" solution to many of my problems, but I think all this should go into the wishlist. Revised blending of refractive shaders, 3D mappable fog color... unified refraction-glossy refraction-sss handling...

regards,

[vlado]

Also I've found that there are some shaders now impossible to reproduce with the current VRay shader accurately. For example, a sub surface scattering shader should turn transparent when thin enough. This phenomenon can be widely seen in macro photographs of plants, where part of the plant is almost glass like and becomes more opaque when the stem gets thicker. And this goes for almost every material, for example, a suitably thin film of metal will also be transparent.

You'd need to use the "water (soft)" sss model to get this, or the new "hybrid" one in RC4. The default "hard" model is specifically for thick non-transparent materials.

Best regards,
Vlado

[putyi]

I thought that the 'water" model is supposed to this, but I was unable to get the desired effects, maybe I did something wrong.

So there will be another model in the upcoming RC4?
Is there any chance for me to get a copy? I've written you several emails already, but for because of diplacement fixes (crashing...) in rc4. Better, more realistic sss behaviour is one thing, something I wuld really like to play with, but a crash free displacement is a must have concerning our current projects.

Regards,

[putyi]

I did a little test in an isolated scene. Here you can see the results:



The shader is the same setup: One blend wrapping two refractive shaders. Could it be, that the blended IOR's get multiplied?

Thanks for your help, Vlado!

Regards,