[Mihai]

Perhaps this example will make things clearer, as to what Force fresnel does at low roughness values:



Check the refl 0/90 colors in both examples. Now it's weird that there is no difference in the material isn't it? Well it isn't because with force fresnel on, the brightness, or value if you prefer, of the color you have set has no influence at all at low roughness values. It is only the chroma that can change the look of your material, but again, that won't change how reflective the surface is. That will only change its tint, to a certain degree when roughness is low. The reflectivity is entirely controlled by the nD, while FF is on and the roughness is low. As the roughness increases, even with FF on, THEN your refl colors start to matter again.

[photomg1]

If you have a too high ND for the shiny BSDF (as I think photomg1 did in his additive test), then of course you get a much too strong reflective "coating", more noise, and practically no fresnel effect (meaning your material will be equally reflective across all viewing angles).
.

Hi Mihai,
I just used the default opaque settings on that mat ( altered roughness only). Beyond that really appreciate the shared info. The fog is starting to clear for me now on this.

best

[eric nixon]

Ok here is a 3rd version with a bit less spec (using nd3 + 70% additive layer) and below is the theoretically correct approach (nd1.4 + 100%).
Both are rendered to sl21, they do look different, but which looks more real is hard to say, because the halo effect seems a bit odd to me on the '100%' one.
The '100%' has more noise ESPECIALLY AT LOWER SL, and the '100%' still has bright noise in the preview at sl21.
Around sl16-18 which I often use for final images, the noise difference is more pronounced than in these images;




The fresnel is brighter at grazing angles with the second image, and that must add some render time. Also to tell the difference its best to download the pics and view them sequentially, because the white levels do vary across the screen on most monitors.

[Mihai]

is the theoretically correct approach (nd4 + 100%)

Why is nD 4 the theoretically correct approach? It's much too strong in my opinion. You will not find such a reflective plastic or top coating for that matter.

There is another point about lowering the additive layer weight percentage, instead of using a more normal ND and keeping the additive layer at 100%. If you just lower the additive layer weight, you are taking away from the strength of the reflection equally, even at grazing angles, which is not what you want. You want to keep that reflection strong, while having a nice fresnel falloff from 0 to 90 viewing angle. If you set such a high ND 4 and even 3, you get practically very little fresnel effect, plus also too strong reflections, thus you feel it's needed to lower the layer weight. I prefer though to just start with the "correct" nd of 1.4-1.6 (and having FF on!), and just keep the layer weight at 100%.

[eric nixon]

sry dude it was 5am, its a typo, should read ND 1.4, FF was on.

[tom]

Eric, you're not comparing equal materials. In normal blending, the reflectance of your white BSDF is compressed with its component weight. In additive blending the base white is not compressed. So, your additive version comes brighter and this is perfectly normal. The noise or glow you are complaining has nothing to do with a possible energy leak or incorrectness in additive blending. Compare the same additive setup to v2 (which had energy leak for additives) and you will see the difference. On the other hand you can never accurately match the same fresnel because in normal blending, the fresnel reflection is getting blended into diffuse while in additive blending it's intact on the surface.

[eric nixon]

Yes thats a good summary and worth mentioning, (I did know all that).

So whats going wrong here? looks quite biased to me, diffuse was set to 97.5 and 245.



Look at the sl and still noise in the preview, that is always an indication that somethings physically incorrect. Look at the halo and the glows...

The bottom line is that an experienced user, who actually takes the time to test stuff has a gut feeling somethings wrong. I think you should investigate.

[Tea_Bag]

I did notice something last week while making a few new materials. When using roughness 0 & Force Fresnel (on) with a light/bright diffuse color causes the glow/halo effect but if the roughness is set to 1 it fixes this issue - Eric on your second layer (Additive) set the roughness to 1 and re-render! I know the NL Team will have this fixed in no time! - Overall I enjoy using additive layers for almost all my material setups!

Here is a greasy red plastic using additive layers:

[eric nixon]

Awesome solution! The low sl noise is mostly gone now!

Thanks Tea-bag. I have a hunch the minimum roughness should be 1.34, will test that...

EDIT: It seems that roughness 0.01 (and up) is good. Just roughness 0 is buggy.

[tom]

Now, this was not expected for sure. Is it happening since v3.0.0.0 or newly introduced?

[eric nixon]

Hard to be sure, I had problems that were similar back then, but I was also using a coating in the material, and another time the deadline was rushed so I just avoided the additive route without testing further.

I dont always use roughness 0 for opaques, but I can see how that idea about 120-5% could have formed, if that was effectively reducing the issue on such a material.

Sorry to go off-topic but part of the reason I still use mw2 mxm's inside mw3 and dont play much with mw3 material-wise, is the annoying OK prompts in mxed.
Is there any news on restoring mxed UI to its former logicality? < really do miss the old dialogue. My XP machine seems all buggy with the new one....sometimes jerky... and with no screen refresh, so those dialogue windows pile up on top of each other and are a real headache after a while.