[pluMmet]

Falloff is not a surface property and will never be. Instead, falloff is a result(behaviour) of surface's microstructure. Being more able to define the surface in small scales means the more control you have about characteristics of scattering. So it will result with different falloffs just like in velvet, polished metal etc.

So it would take a very fine displacement to acheive?

[tom]

No, it's not by displacement, it's the measured/simulated BSDF.

[Micha]

Here a new example - the light come from the side, but the center of the bright diffuse area is not at the side of the object. It follow the law of reflection and is placed at the same place like the brightest clear reflection. The edge in light dir is dark, because this area show a diffuse reflection of the dark background. So, we can not use a plastic shader.

Why do I call it "metal diffuse" and not blury reflection? The reflections on my examples are not blur. The materials are not rough. Only reflections of a rough surface I will call blur reflections. This is an other effect here.



We need no layer material, because the metal I like to get is like the plastic shader with two components - (metal) diffuse and reflection. And if Maxwell try to be physical corect, than we should get a material like this. It is not so difficult, but a fallof would be only a fake here. And don't forget, there is no dielectric layer on the metals here.

@Grim: Sorry, I don't understand. This is a lamp of a bicycle.

[Mihai]

Ok, what I have learned about metals is that since they contain so many free (or almost free) electrons, these electrons form an electromagnetic "barrier" so that photons reaching it do not penetrate it, like with plastics, but instead metals have the property that they reflect almost all the light they recieve. This and the fact that any specular highlights will be tinted in the metals color, unless the lightsource is very bright, then you will get a burn-out, or white specular.

Plastics on the other hand can absorb some of the light, light can penetrate plastic and the specular it returns is the color of the lightsource.

Now, what I'm trying to tell you is that in real life metals don't have a "metal" diffuse and then a "normal" specular reflection.

In your last example you say the reflections are not blurred. Yes they are! ALL materials have roughness. We have yet to construct a perfect mirror surface. They show a bit of blur, because of the micro grooves in that metal. What gives metal the ability to still look like it has a pretty sharp reflection is the fact that it reflects almost all of the light it recieves, there is no deep penetration and scattering like you get with plastics, which contribute to the effect that plastics will tend to look blurrier than metals even if they are very smooth.

A big part of that glow you mentioned can be attributed to the coloration of the reflected light that metals give you, except if the lightsource is strong enough like I said. You can see it in your last example as well, look at the edges of the white spot, they start to tint yellow.

The other part of that glow effect you see on worn metals is due to the deposits of very thin sheets of other substances which reflect some of the light back (in different directions, thus making a blurry reflection), but letting some light penetrate through to the metal, where it will give you that little bit sharper reflection. Or the metal suffers from oxidation, in which case I think that outer layer of oxidized metal contains less free electrons which lets light penetrate it deeper thus scattering it around before being reflected back=blurrier reflections. Polish the metal and you remove that thin sheet, you expose the clean metal=sharper reflections. But with polishing you will also increase anisotropic effect.

So there is no separate component in clean metals that give it a sharp reflection, and another one that give it blurrier reflections, or "glow". This is how we must think in Maxwell as well.

Why not take advantage of describing these properties accurately instead of creating fakes which do not exist? I think doing that we will quickly find ourselves in an impasse.

I don't see a big problem with layering a material, so you have a base, very reflective metal, and if you want on top of that to have a worn out blurrier look, add another layer.

[Micha]

Good, I could make a render test. I render two times the same metal object in a HDRI environment, first clear reflection and second blur reflection. A combination of both should look like the photo examples. Hmm, today my wife need some 3D work, but I will test it tomorrow.

[Mihai]

Good, I could make a render test. I render two times the same metal object in a HDRI environment, first clear reflection and second blur reflection. A combination of both should look like the photo examples.

If the renderer you use has a correct model for metallic surfaces

You should use something like Cook-Torrance to get correct specular tinting and reflective properties. But that model will not work well for blurrier metals I think.

I found this great paper explaining the common shading models:

http://courses.dce.harvard.edu/~cscie23 ... Models.pdf

[Hervé]

Yes, and also micro imperfectness of the surface makes it glow.

That's what I think too, and the fact that those imperfections run in random directions makes the reflection blurrier. When you polish the metal, you are "lining up" those imperfections in a dominant direction as well as diminishing the depth of those micro grooves, giving you a sharper reflection.

eeeexactly!

it all depend on how you polish... if it is like me... then no lining up... t would go all the ways...

In fact I said in an earlier post it all depend on bump... and even a super micro mini thin structure is for me a bump....



Bump bump bump bump.... beignets aux bumps....

I've made the best aluminium sink (most difficult to do) with super tiny bump maps.... unfortunatly the present beta took sooooo long to compute it... bu the result was 100 right !

here is a photo of an extremely hardcore challenge for Maxwell, because grooves and crevasses are sooooo tiny.....

[j_petrucci]

eheh but surely Tom can also do this one...

[Hervé]

mmmm, maybe.... but really hardcore.... I mean sure it can be done.... but it all depend on how far you realistcally want to go... trying to really mimic real life is "almost impossible" for some really complex stuff...

I do not say it is impossible.... but I would not even try...

[Hervé]

here are some shoys really close from that same sink...

[Olivier Cugniet]

Hervé, to be totally honest, I don't really understand your problem on metals, so I post a little test. What do you think about ?

[Hervé]

no problemos here, I was not the thread starter.....

good looking render !

what kind of technique do you use ? bump map..?

[Hervé]

http://www.maxwellrender.com/forum/view ... ht=cuisine

[Olivier Cugniet]

I used a little noise bumpmap and a specular map. I should maybe try the same scene with an other light setup, just to see...

[Hervé]

thanks ! yes maybe another light....