[PCONCA]

This tip about the additive layers is true for V3 of Maxwell Render? because I've seen this in V2 but not on V3 documentation and I wonder if it does not as a warning now in V3 or simply a matter of differences in the documentation between V2 and V3...?



Thanks in advance

[eric nixon]

If your diffuse bsdf has a very light colour, the limit in maxwell3 currently is about 120%, i.e. If I want a white glossy plastic with additive reflections, I wouldn't go above 120% layer power.

If your diffuse bsdf is very dark 200% will probably be fine, it just depends whats in the bsdf's. i.e. an R2 type bsdf at 100% may only be actually contributing around 25%, it depends.

This rule applies to opaque materials only, and its possible a future version of maxwell3 will be more tolerant.

[Mihai]

See the first thing on this page:

http://www.maxwellrender.com/forum/view ... 35&t=41535

You don't have to worry any more for opaque materials how many additive layers you stack or what strength they have.

[eric nixon]

Have you tested this?

A white glossy additive above 120-125% will render, but it will render more slowly and take more sl to clear the noise, as the power increases.

[PCONCA]

My question was this one not only because the mismatch between versions documentation... It was produced too... because I see the link you said me Mihai

http://www.maxwellrender.com/forum/view ... 35&t=41535

but I see another link... doing a search on forums "additive mode" and finding this answer on the seventh post there, from dariolanza this last Fri 8 of august 2014:

"....like our suggestion of avoiding any material in Additive mode and use always Normal layers to minimize the noise..."

It turns me confusing about...

And for me is not so easy be so sure what´s happening because I can read "PRO" opinions in both senses.... and the same in the maxwell reference...

[Mihai]

Have you tested this?

A white glossy additive above 120-125% will render, but it will render more slowly and take more sl to clear the noise, as the power increases.

I'm not sure what you mean by 120-125%?

The point with the fix in V3 is that you will reach a "max" whiteness for the material now, and it won't go past that. It won't start to almost glow like it did before and cause specks that never cleared. Try stacking 10 additive layers all set to 100% in V2 vs V3.

[PCONCA]

"....like our suggestion of avoiding any material in Additive mode and use always Normal layers to minimize the noise..." but this is true at the end? is solved but gets you more noise than a Normal Layer?

[eric nixon]

I get noisy materials above a certain threshold. The noise will occur in the recesses of the geo, especially in very narrow gaps, (the recesses are glowing a little bit). This is why I have been saving mw3 mxm's in duplicate versions, my preferred version, and then a duller version, to which I append the suffix 'SAFE'.

I wasnt aware of this max value concept before now, is it in the docs?. It would be good to get a thorough explanation of how this relates to real world materials.

I havent spent much time on materials in MW3, because the UI is really unpleasant to use now, and other new features were buggy such as proc. So I thought it would be best to wait.

[photomg1]

but I see another link... doing a search on forums "additive mode" and finding this answer on the seventh post there, from dariolanza this last Fri 8 of august 2014:

"....like our suggestion of avoiding any material in Additive mode and use always Normal layers to minimize the noise..."

It turns me confusing about...

And for me is not so easy be so sure what´s happening because I can read "PRO" opinions in both senses.... and the same in the maxwell reference...[/quote]


I'd like to see an answer to that as well , I noticed it at the time so I'm glad you have asked this PCONCA !

So far I'm taking it as Normal is the optimal route , and Additive is vastly better than V2 but still ultimately inferior in terms of render time / noise than normal layer stacking.

Is that right ??

[Mihai]

I wouldn't say that's correct because many materials don't look right without using Additive. I would like to see some of these examples you're talking about though. Of course if you have a brighter material, in small recesses it can take longer to clear because of more bounces, but this would happen with a material in Normal blending mode as well. Since additive will effectively brighten the material more than it is possible with Normal blending, it can take a bit longer in some cases, but it's nowhere near as bad as it could be in V2 if you weren't careful.

So saying this loud and clear again: Don't worry anymore in V3 how many additive layers you stack. The possible additional noise is simply due to the fact that it's brighter than a normal blended one. The noise is not due to the fact that you are simply using additive. A simple test would be: create two versions of a material, one with normal blending, one with additive which results in a material with the same brightness and test it.

[dariolanza]

Hello all,

Mihai and me are saying basically the same, although I use to be more conservative.

Additive materials have been dramatically improved since v2, where excess in reflected energy generated problems like bright noise and sometimes even black bars (due numerical errors caused by that energy excess).
Now those problems are mainly solved, and we don't see that noise and black bars that were frequent until now. He is right with this.

Said that, keep in mind that with Additives you are summing the result from one layer to another, so they easily get brighter than you get with Normals. That extra brightness is the reason why people like Additives. But this extra brightness could cause that, above certain subtle undefined blending threshold (undefined because it obviously depends on the lighting of the scene, the material and even the geometry itself, like Nixon says), the result becomes too bright (brighter than a normal blended) and even appear noise in some areas (bright energy noise, not black SL noise) that is very hard to clean.

But Additives are not noisier per se, they are not working badly, but with them you can end up easier on a risky bright surface.

This is why I use to recommend using Normal materials (that don't get that brighter that easy), so you don't have to be that careful with the blend weights. There wouldn't be a problem using Additives instead, but you have to be more careful with the weights and the lighting, while Normals keeps you more on the safer area. This is why my frequent recommendation on Normals.

Let me know if this clarifies the question.

Cheers

Dario Lanza

[tom]

The additives are working perfectly correct as long as one of the components is not SSS or dielectrics and that's all. Further discussions are urban legends.

[photomg1]

Thanks for all the replies chaps !

I wouldn't say that's correct because many materials don't look right without using Additive.

would it be possible for a simple visual example of this ? I've tended to avoid additive at all costs due to past warnings but now after Mihai's comment feel I've been missing out on something all along .

[Mihai]

Plastic with shiny top coat for one very usual example....

Anything that should have a shiny top coat on it and you don't want to use Coatings, need Additive.

[photomg1]

cant find it in the docs now , but I do remember some simple examples of wood concrete plastic etc

The plastic was built just blending two bsdfs a lambert plus a shiny layer just weighted for plastic in normal mode . How is that incorrect in comparison to your additive example Mihai . I'm just looking for perfect clarity in my head on this Mihai , so sorry for my repeated question on this ( or in other words please excuse my ignorance on this )