[Mihai]

There are ofcourse still many improvements that can be done to the material system, there's no discussion about that. This is what software development is. I like to keep things in perspective, I mean with XSI's system until version 3.5 I think, people had trouble even applying a bumpmap properly, finding the correct nodes, and then connecting them in the right way. But once you get past the initial "shock" of doing materials with nodes, you realize what a great system it is, how flexible it is. Maxwell's system is much easier than that. Beta is still being mentioned but what if you wanted to make a glass with golden decorations for example? You couldn't. What about a pretty rough plastic which also has a glossy finish? You couldn't. Multi colored glass? You couldn't. Painted porcelain? You couldn't. How would you be able to make such materials with beta, without having to blend between two material types? So you would eventually end up with the ability to have several material groups, and blending between them via a weightmap. This is what we have now.

How would you make these kinds of materials for other renderers? You have no problem understanding the complexities there, which are often more difficult to grasp, yet you refuse to understand Maxwell's system, simply because it's not like beta.....it can't be like beta when we also need more flexibility. But the material wizards which will be improved is a way for you to keep working like in beta, if this is all you need. All I'm trying to do is help understand the real system because sooner or later you will need it's flexibility.

But what about the materials. If I wanted to create for example a white paint finish with 30% gloss and a total reflectance of say 70%, it would be trial and error until I get the correct look. Where's the physical correctness in this? How are you supposed to define a physically correct material if such a basic property as total material reflectance is pure guesswork.

But how do you decide which part of light is glossy, and which is simply diffusely reflected? When does one part become the other? Is that how they measure materials in real life? Do you have a certain steel being officially defined as, 30% glossy and 70% reflective? I don't think so. I guess the total material reflectance is the reflection color chip. With the roughness setting you decide in what way that light is reflected. But seeing as how many materials which we think are homogenous are really a mix of different materials, with different reflective qualities, you have to use a mix of different bsdf's. You won't get a totally accurate material, but at least now you have a chance to get much closer than with the beta system. If we really wanted total accuracy we would need to get a bsdf measurement of the surface with a gonioreflectometer, then use that measurement to drive a bsdf material in Maxwell....not many people have access to such a device, and Maxwell can't currently use custom bsdf measurements, but who knows in the future that might be yet another posibility.

[giacob]

as to me i dont intend to say that this material system is bad but that this bsdf layer thingy dont seems much related to real phisics behaviour of materials ...
sure that this system give more freedoom that the old one...though this freedom is limited to possibility of making one material composed of two material blended together trough a bitmap mask... just what in 3d max u can call a blended material... good --but not that much freedom in my opinio...
much better would be if instead we would have some procedurals maps that could be set without resorting to ps or whatever.... besides there should be the possility to mask two or blended material ... the u would have a real freedoom

[Mihai]

Sure, all great ideas. It would be really cool to have procedurals in Maxwell, I think there's no doubt it's in the developers plans. Better blending possibilities , basic image editing options, like contrast, brightness, ability to turn off/on layers....many things could be improved and they will be. Maxwell 1 has been out 3-4 weeks though and people are already saying it's long overdue for an update....I don't think it's realistic to expect huge improvements every 2 weeks which such a software, it's not a word processor.

[Frances]

well Frances i must say i think you are not right.

Um. You know I was joking, right?

[tom]

Thank you for providing this helpful information Mihai. I see some friends are discussing the approach and I suggest them reading pages 79-82 of the manual prior to making false conclusions. Also I agree the word "layer" is confusing, however you should think it more like components of the same substance rather than a classical stacked layer system. Using multiple BSDF layers doesn't mean you're topcoating other substances over one substance. It rather means simulating different characteristics of one single substance with blending them.

[Frances]

Using multiple BSDF layers doesn't mean you're topcoating other substances over one substance. It rather means simulating different characteristics of one single substance with blending them.

Thanks for the explanation. It helps.

[GM5]

I haven't had much time to use the new material editor but I noticed that some of the new a-team materials (like jdhill's AE_Butterscotch) don't follow the RC-5 "don't exceed 100% when using weightmaps" rule. Is there a post or something in the manual that I missed about this?

-Greg

[lllab]

ups, sorry frances:-)

cheers
stefan

[acquiesse]

Hi Greg,

I think you'll find that if the number exceeds 100, Maxwell uses the proportions, ie. 60 & 240 is treated the same as 20 & 80. I'm sure someone will correct me if not :)

[KRZ]

its like that if blendmode is "normal". if its "adaptive" you get the unrealistic values of simple addition of the values

[Thomas An.]

Hi Greg,

I think you'll find that if the number exceeds 100, Maxwell uses the proportions, ie. 60 & 240 is treated the same as 20 & 80. I'm sure someone will correct me if not

Yes, this is what it does (it normalizes automaticaly if you do not do it manually). So if both of your layers are 100 then each contributes 100/200 (or 50%)

[GM5]

Thanks for the quick answers!

-Greg

[Rochr]

Thanks for the effort Mihai, although i´m just as lost as i was from the beginning.
Guess it´s like Zbrush, either you can work with it or you can´t. Think i can safely say that i fit in the latter cathegory.

[wagurto]

Thanks God Maxwell is easy as taking a picture!! I need a Phd to get this thing right! Sorry guys maybe I am not ready for "phisically correct" yet!

[aitraaz]

I mean, the new material system for me is fine, quite interesting really...Plus, as pointed out, it always takes a bit of time to master new tools.

As to how the new material system is a system which can be correlated to the 'physically correct,' that's a whole other argument (probably an interesting one after about a half a bottle of scotch whiskey), but it risks running into something rather philosophical.

For all the signifigant innovations in maxwell, our great mother nature is a bit, err, more complicated and mysterious so any system that NL decided on was bound to run into some criticism, justifiably so, thats how thing improve, through criticism and feedback. The base looks solid.

*but,*

the golden question - is this daunting material system a bit more than the core can handle at this point? eg., is it causing rendertimes and noise to increase due to the much more complex calculations it necessarily requires? Could it be?