Put maps in Reflection, or in Glossiness slot?

[Torsten]

Hey Guys,

Maybe a very simple question. But i am struggling to comprehend when to put put a map in the reflection slot or when to put it in the Glossiness slot. For instance a wood or concrete floor. I want the floor to have different refelctions when light falls on it. You can either put a b/w map in the reflection or in the glossiness slot. Both generate a result mimicking the different gloss of the material. You can also use different maps in both slots.

But what is physicallly accurate? When do i use the reflection slot, and when the reflection glossiness? And what are the results using the first or the latter?

Cheers guys!

[Juraj]

For non-metallic materials, it's glossiness slot.

If Corona's material behaved correctly, that would be enough for "physically-based" mat. But you can with artistic licence, always boost the effect by placing the same map in reflection slot as well. In Corona currently, it's sometimes almost necessary because this will additionally clamp the specular which is welcomed.

Reflection slot should be used as mask for materials which differ in reflectivity.

But as you said, the effect may look similar, which is why the confusion in workflow arisen in first place. I can't describe how mad I am when I download a model from 3dSky or DesignConnected, and they are using 100perc. reflection slot to make glossiness, no map in glossines but only single number for flat effect. It's absolutely wrong, it looks completely wrong and it takes me too much time to fix.

Read up great Dubcat's guides on this forum. He describes the topic very comprehensively and provides handful of tools and advice how to fix some issues if you want to get as close to 100perc. physical based approach as possible.

[FrostKiwi]

The very first time I switched to pbr rendering with vray, this confused me aswell.
In reflection slot you control which parts of the surface reflect light in a specular fashion, as opposed to the standard diffuse model. In physics terms it only maps what parts bounce lightrays with directionality and which bounce light in complete random direction.
The glossiness map determines how directional the light gets bounced. When the map is 255 white (1.0) glosiness, then light rays bounce of the surface with the same angle as  the hit the surface. With lower glossiness, this angle gets mor random and starts to diffuse the reflective rays, hence the name
In physics 0 glossiness value = diffuse surface the same as 0 reflectance, but in rendering we approximate 0 glosiness with a fast mathematical method and due to Corona's ggx model glossiness curve, 0 glosiness doesn't actually reach that low, but this is irrelevant for you.

TL;DR example
Reflection slot = which parts of the model show metal parts and should reflect
Glosiness = how much each metal part is corroded and diffuses light in different amounts

PS: you can get confused with the levels in the materials, the reflectance value is a multiplier. If you have a map with a  max of 180 reflectancd,  then that part will never be Mord reflective than 180 ~ 70% and can only get lowered by the multiplier.

God dmamit, not first D:

[Torsten]

Thanks Juraj and SairesArt for you input on this jubject,

It helped me to choose where to put which map. What Juraj says, about not getting the desired effect only using glossiness maps, i also experience in some material setup. That's why i sometimes use a altered glossiness map into the reflection slot. But tuning the materials reflection than becomes a lot harder. So to choose which one to use was not that clear to me.

[Juraj]

Thanks Juraj and SairesArt for you input on this jubject,

It helped me to choose where to put which map. What Juraj says, about not getting the desired effect only using glossiness maps, i also experience in some material setup. That's why i sometimes use a altered glossiness map into the reflection slot. But tuning the materials reflection than becomes a lot harder. So to choose which one to use was not that clear to me.

Definitely read Dubcat's explanation to why this is. The specular reflection doesn't get dimmed enough by lower glossiness, and glossiness range of 0-1 in Corona isn't the same as 0-1 industry standard. If the result looks wrong, start lowering specular reflection (simply clamp by multiplier, or use map in reflection slot), and IOR (from 1.52 to 1.4, 1.33, 1.27..) until it looks good again.

Ondra promised solution after 1.4 is stable out. They already have it ready just need testing. After that, we can safely use glossiness exclusively :- )

[gabyanz]

Juraj,

where we can find Dubcat's explanations?

I also struggle to find the right way to use the maps in gloss/reflection slots.

thanks

[Rhodesy]

This is a really interesting topic and something I'm scratching my head about too. I'm sure I read some of Dubcats stuff a while back but I think I need to revisit so will try and search it out.

Is it OK to have 255 value in the reflectance slot? I I know we have to be well under that in diffuse for realistic results and performance, but is it a different rule for reflectance values? Does it make any difference to performance? I usually just knock a bit off but I'm unsure if I need to or not. By the sounds of it its probably helpful for more dull materials anyway which I decrease a lot anyway to get the right look, so it would appear I'm encountering the same issues as being raised in this thread.

Also to use B/W or coloured maps in the reflectance slot if used there? Im not quite sure what is correct as it does make a noticable difference to the visual output depending on the saturation.

Looking forward to these material changes filtering through to the C4D version, hopefully this sort of thing can be introduced shortly after the Max version with a few internal tweaks.

Thanks.

[maru]

Is it OK to have 255 value in the reflectance slot?

It is usually OK. Fresnel does his job and the 255 value is only in the extreme grazing angles.
I am not 100% sure about mirrory materials with Fresnel IOR like 32 or more, but I would say that:
-in terms of realism there should be little difference between 255 and making it slightly darker
-in terms of speed there should be no difference - MSI (and/or vfb clamping in 1.4) will get rid of caustics anyway

Also to use B/W or coloured maps in the reflectance slot if used there? Im not quite sure what is correct as it does make a noticable difference to the visual output depending on the saturation.

I would compare it to using BW and colored maps in diffuse. So it depends on the case. Sometimes you just need colorful reflections (e.g. "thin film").

[Juraj]

This is a really interesting topic and something I'm scratching my head about too. I'm sure I read some of Dubcats stuff a while back but I think I need to revisit so will try and search it out.

Is it OK to have 255 value in the reflectance slot? I I know we have to be well under that in diffuse for realistic results and performance, but is it a different rule for reflectance values? Does it make any difference to performance? I usually just knock a bit off but I'm unsure if I need to or not. By the sounds of it its probably helpful for more dull materials anyway which I decrease a lot anyway to get the right look, so it would appear I'm encountering the same issues as being raised in this thread.

Also to use B/W or coloured maps in the reflectance slot if used there? Im not quite sure what is correct as it does make a noticable difference to the visual output depending on the saturation.

Looking forward to these material changes filtering through to the C4D version, hopefully this sort of thing can be introduced shortly after the Max version with a few internal tweaks.

Thanks.

255 in any slot except diffuse is completely OK, because it's representing value. 255 in RGB in reflectance is simply 100perc total reflectance in grazing angle. In rendering, value and tone are often interchangeable, sometimes representing the actual result (diffuse/albedo) ,sometimes just driving parameter.

Coloured maps in reflectance should be almost exclusively reserved for metallic materials, which have full black diffuse. If non-metallic material has colored specular (like black leather), it's almost always because of thin film effect as Maru stated. That is best used as overlay coat material.

For non-metallic material, grazing angle specular is always 100perc, but Corona doesn't compute specular dimming correctly using IOR model. If you lower glossiness, you need to also lower specular level and IOR....this is workaround until more correct model will be implemented in 1.5+.

[Rhodesy]

Thanks for the advice guys, starting to make a bit more sense. Glad to know that 255 in channels outside of the diffuse isn't going to cause any problems and B/W spec maps are the way to go outside of metals. Looking forward to seing how the 'fix' goes so we just control everything from the glossiness, that does indeed make sense and one less slot to think about. Strange the RL didn't get this as expected from the start as they are leading the way in many areas and they have been clever to strip everything back to the essentials to create beautiful mats.

[Juraj]

Thanks for the advice guys, starting to make a bit more sense. Glad to know that 255 in channels outside of the diffuse isn't going to cause any problems and B/W spec maps are the way to go outside of metals. Looking forward to seing how the 'fix' goes so we just control everything from the glossiness, that does indeed make sense and one less slot to think about. Strange the RL didn't get this as expected from the start as they are leading the way in many areas and they have been clever to strip everything back to the essentials to create beautiful mats.

Kind of weird, yes.. I think everybody hoped this would be fixed by migrating to GGX brdf as it happened in Vray, but this was done somehow by merging with the older behaviour, so the materials do look nicer, but are still complicated to create. But Ondra says they have this working internally and that we could perhaps get our hand on it in daily's after 1.4 is out.

Because of this my library looks bit absurd.  Base material of glossiness 0.9-1.0 has 1.52 IOR and 1.0/255 reflectance and then I gradually lower both as I lower glossiness. I originally just kept lowering reflectance, which clamps intensity all across, but the IOR also dims the fresnel effect, courtesy of Dubcat's suggestion. Only after this I got rid of the "shining" effect on matte materials. I could never quite put my finger on what was wrong all these years :- D The correct behavior that guarantees all materials are real-world plausible would rid me of some mild OCD :- )

It will also tremendously increase teaching workflow across studios all around. Imagine you can teach beginners how to construct materials in minutes rather than hours, and give them concrete instructions, instead of vague "make it till it looks good".

[Rhodesy]

Yes I agree with that, getting the reflection right is really satisfying on the shader ball but sometimes it doesn't look as good in use, so simplifying it makes it easier to understand when trying to teach staff / students. It is a bit muddy at the moment. Also standard IOR seems to vary from program to program and forget looking them up online - thats even more confusing when trying to teach someone. If IOR really was to be user/learner friendly then I think it needs a rethink in general. The C4D version has a slider that goes up to 3 but you obviously need to spin it way past that for metals - not that you would know that from looking at it as a learner. But we dont want it too easy, otherwise we'd all be out of a job! I'm already a bit concerned about that with the quality and simplicity of corona out of the box ;-)

[Juraj]

Imho the 'metalness' workflow does this best, it hides the IOR from the user :- ). For example in Unreal (which uses simplified Disney convention), has specular level from 0 to 1. By default it's 0.5 which is IOR 1.52 :- ). 0 is I think 1.27-1.33 or something, for those few liquids like water and so. 1.0 is something like IOR2.2 for some very reflective plastics.
But if you select metal input, the IOR is bypassed and it moves into some 20-999 scale. Of course, that isn't as physical as using ComplexFresnel plugin but imho it works very user-friendly.

I am not concerned about too easy :- ) When some technicalities became easier, the focus will simply shift elsewhere. Already today I spent very little time on such things, most of them time we spent on styling and communication with clients, not 3D stuff.

[Rhodesy]

Sure, it's better for us too. Much better! And don't want to derail the thread any more than I have done as that is a whole can of worms that topic.

I like the idea of a two option strategy for metals and non metals. Makes sense. Its kind of like when Maxwell introduced its material wizard years ago as most of us had never heard about IOR at the time. I remember creating those first blurry metals and was blown away coming from the comparatively primitive C4D internal renderer.