New LayeredBSDF and glossycoating material

[Lord Crc]

edit: I split this into it's own thread, the old discussion is here: viewtopic.php?f=8&t=6961

Ok I'm gonna push what I have, since I'm not introducing any new files it's easy for us to shuffle things around as we see fit.

As mentioned the reason to accept a BRDF for the coating was so I wouldn't reimplement most of the SchlickBRDF in bxdf.cpp, and also hopefully make it general enough to make it accept some other BRDF's for the coating.

If we keep SchlickGlossyBRDF I was thinking we could refactor the three Schlick classes to reduce the amount of copypasta.

I didn't feel very confident implementing the LayeredBSDF, so it's not unlikely it's littered with mistakes. Especially handling all the flags and such made my brain hurt, as you can see from the various FIXME's.

However it does seem to reproduce the standard glossy material, and seems to work with other base materials, so something is at least not entirely wrong...


As mentioned earlier, I'm using a rather cheap heuristic for the sampling weights: I just use the specular reflection of wo as wi (ie H = n) when computing the coating BRDF weight. I then adjust the weight such that the coating is sampled between 50% and 100% of the time. This ensures the coating is not undersampled in case of a rough coating over a smooth base, and seems to work well in the tests I've used.

Anyway... how to use it. Consider the following plain glossy material:

Code: Select all

MakeNamedMaterial "luxball"
   "string type" ["glossy"]
   "color Kd" [0.750 0.050 0.005]
   "color Ks" [0.1 0.1 0.1]
   "float uroughness" [0.015] "float vroughness" [0.015]


then this can be recreated using the new "glossycoating" material, which uses the LayeredBSDF, as follows:

Code: Select all

MakeNamedMaterial "diffbase"
   "string type" ["matte"]
   "color Kd" [0.750 0.050 0.005]

MakeNamedMaterial "luxball"
   "string type" ["glossycoating"]
   "string basematerial" ["diffbase"]
   "color Ks" [0.1 0.1 0.1]
   "float uroughness" [0.015] "float vroughness" [0.015]


I intentionally left some commented code in the GlossyCoating material so you can switch between different BRDF's to experiment.

Comment or cry

[Lord Crc]

I'm not sure how to properly handle specular base materials, for example mirror material.

I guess we could always sample the base, and if the sampled type was specular, use that direction. However I'm not sure how to take this into account in the Pdf function.

Alternatively I guess we could simple check NumComponents of the base, and always sample the base if there's a specular component there.

Thoughts? Perhaps tackle this later?

[Lord Crc]

Just a bit more complex setup, to illustrate possibilities as well as fun

edit: oh and thanks JtheNinja for the luxblend25 support.

Code: Select all

Texture "metalior" "fresnel" "preset"
   "string name" ["amorphous carbon"]

MakeNamedMaterial "metalbase"
   "string type" ["metal2"]
   "float uroughness" [0.012] "float vroughness" [0.012]
   "texture fresnel" ["metalior"]

MakeNamedMaterial "diffbase"
   "string type" ["matte"]
   "color Kd" [0.75000000 0.05000000 0.00500000]

TransformBegin
Scale 200 200 200
Texture "mixamount" "float" "checkerboard"
   "integer dimension" [3]   
TransformEnd

MakeNamedMaterial "mixbase"
   "string type" ["mix"]
   "texture amount" ["mixamount"]
   "string namedmaterial1" ["metalbase"]
   "string namedmaterial2" ["diffbase"]

MakeNamedMaterial "luxball"
   "string type" ["glossycoating"]
   "string basematerial" ["mixbase"]
   "color Ks" [0.1 0.1 0.1]
   "float uroughness" [0.015] "float vroughness" [0.015]


[B.Y.O.B.]

Wow, that offers many new possibilities! Seems that this will be an alternative to the layered material, just for the cases when you only need a glossy coating (which are most of the cases I think, e.g. ceramics).

[Lord Crc]

Wow, that offers many new possibilities! Seems that this will be an alternative to the layered material, just for the cases when you only need a glossy coating (which are most of the cases I think, e.g. ceramics).

Yes, indeed it's a much simplified version of that. However I think it will be useful for a lot of cases, especially since it's quite a lot faster. Here's another quick render (10 mins), this time two coating layers with bump map on the middle one only. The base layer is a fairly dark blue mattetranslucent. The material definition:

Code: Select all

MakeNamedMaterial "transbase"
   "color Kr" [0.0005000000 0.01500000 0.0500000]
   "color Kt" [0.005000000 0.01000000 0.02500000]
   "bool energyconserving" ["true"]
   "string type" ["mattetranslucent"]


TransformBegin
   Scale 100.0 100.0 100.0
   Texture "bumps.src" "float" "wrinkled"
TransformEnd
Texture "bumps" "float" "scale"
   "texture tex1" ["bumps.src"]
   "float tex2" [0.001]

MakeNamedMaterial "lowerlayer"
   "texture bumpmap" ["bumps"]
   "string basematerial" ["transbase"]
   "color Ks" [0.01 0.15 0.05]
   "float uroughness" [0.15] "float vroughness" [0.15]
   "string type" ["glossycoating"]

MakeNamedMaterial "luxball"
   "string basematerial" ["lowerlayer"]
   "color Ks" [0.02 0.02 0.02]
   "float uroughness" [0.015] "float vroughness" [0.015]
   "string type" ["glossycoating"]


[moure]

That looks great LordCrc

[SATtva]

Here's another quick render (10 mins), this time two coating layers with bump map on the middle one only.

Awesome!

[jeanphi]

Hi,

I can only repeat SATtva comment: awesome! Thanks a lot.
Let the fun begin

I think we should eventually ditch the old glossy materials and just translate them to the new infrastructure for compatibility.

I'll try to have a look at it next week.

Jeanphi

[Lord Crc]

Thanks guys

I'll try to have a look at it next week.

Great

I almost expect to have screwed up something related to the "reverse" and BxDF type flags. And then there's the case of specular base materials...

[LadeHeria]

Sorry for this question, but i don't understand, is this feature now in Luxrender ?


LadeHeria