[Maxer]

Mihai thanks so much for the explanation you gave it was something that was desperately needed for those of us who haven't extensively studied materials and the complex manufacturing processes that make up the objects in our world.

IMHO we've gotten way to technical with this new layer system and NL has taken this physically accurate theory way to far to the point that it's actually making their already crippled render engine even more difficult to use. I understand the need to be able to create realistic textures, but no one cares if it's totally realistic from every angle. I also agree that the material editor and sample materials seem to have been to heavily influenced by A-team members that work in the product design field. I can see why you would need all these different settings; however the Architectural field of illustration really has no need for such precise descriptive materials which is why most of the people who liked the beta are working in that area.

I'm also confused as to how these materials are "more physically accurate" than the beta, you've definitely complicated the material creation process but over complication doesn’t necessarily equal more realistic. For example to create a plastic material two BSDF layers are needed and each one requires that you adjust reflectance, transmittance, attenuation distance, ND, roughness, and the weighs of each. All of these parameters have their own values, but how does one determine what the roughness value should be, or the attenuation distance? The answer is you guess, and you use your eyes to tell you that yes this material looks right or no it doesn’t. At this point you have completely lost all physical accuracy and you going on what looks good to you. The material that you have just created has no relation to a real material other than it looks similar, and if that is the only parameter that matters I have to ask why this system is better than the beta? I agree you can make more materials with this system, even some that don't exist, but at what cost, and does this actually put us past where we were with the beta or farther behind?

[Mihai]

Well there isn't just one type of "yellow plastic". You can have a plastic with different surface right? So if you change the roughness it doesn't necessarily mean it's not "accurate".

You don't need to change all those parameters you mentioned just to make a plastic, and besides you still have the material wizard, so is it really that complex?

The material editor simply translates what actually makes a material look the way it does in real life as closely as possible. That's why the parameter that changes how shiny a material is perceived, is called roughness. Doesn't this seem more intuitive to you than a combination of something called Phong shader with a setting called diffuse and another called specular?

It's just seems more logical to me.....if I want a material that looks less shiny, what do I change, what really makes a material look less shiny in real life? And also in other renderers if you raise the specular, you have to take care to also lower the diffuse color because that's what would happen in real life. How is that any easier or more intuitive, or more precise?

[Maxer]

Well there isn't just one type of "yellow plastic". You can have a plastic with different surface right? So if you change the roughness it doesn't necessarily mean it's not "accurate".

True, but these roughness settings aren’t based in any scientific understanding of the material, there just best guesses. It seems to me that NL is trying to create this perfect physically accurate render engine without considering that the people using it aren’t perfect and don’t have intimate understanding of what constitutes a material and therefore are incapable of creating a physically accurate material. I think this is why we don’t see any mind blowing images coming out of V1, the understanding of materials and how to create them simply isn’t there for 99% of the people. This is why the beta method was so good; you didn’t have to know that material X has 7 different BSDF layers and what each group of settings had to be in order to accurately represent it. You didn’t have to re-educate the entire CG community on how to create a material because you were using industry standard terminology, no one cared if it was only 75% accurate, it looked correct and therefore it was correct.

You don't need to change all those parameters you mentioned just to make a plastic, and besides you still have the material wizard, so is it really that complex?

No, that’s about the only thing that isn’t complex.

The material editor simply translates what actually makes a material look the way it does in real life as closely as possible. That's why the parameter that changes how shiny a material is perceived, is called roughness. Doesn't this seem more intuitive to you than a combination of something called Phong shader with a setting called diffuse and another called specular?

I don’t think the entire problem can be summed up in a couple of settings; you have to look at the whole method as a whole. The process to create materials is to complex and technical, and then you have to deal with the speed issue while trying to get test renderings of your materials., then there are all the limitations of the plugins to deal with.

It's just seems more logical to me.....if I want a material that looks less shiny, what do I change, what really makes a material look less shiny in real life? And also in other renderers if you raise the specular, you have to take care to also lower the diffuse color because that's what would happen in real life. How is that any easier or more intuitive, or more precise?

I agree roughness isn’t that hard of a concept to understand, but other settings like Attenuation Distance, Transmittance & ND value are settings that make creating simple glass transparency a much more complicated task than is necessary.

[Rochr]

I´m not going to get into your argument cause i don´t understand half of what you´re talking about regarding Maxwells material editor, but i must say that it´s one of the least intuitive editors i´ve seen in materials creation. If it would be intuitive, people wouldn´t have a problem working with it.
How the hell do you even know what or where to press to make things happen? I can´t even guess what i´m supposed to do.

Btw, if the editor is that great, how come it doesn´t support procedural maps? I´ld love to see anyone create a nice rough concrete material without the use of image based textures.

[Mihai]

Rochr, if you took the time to read the manual it would be clearer....what do you expect exactly? This is knowledge offered to you, take it or leave it, but maybe you should consider part of the problem is you refuse to accept any kind of new knowledge.

True, but these roughness settings aren’t based in any scientific understanding of the material, there just best guesses.

I'm not sure what you mean there, how come you understood and accepted what IOR (or nd) does in beta, but now suddenly in 1.0, it's all a mystery? Same with roughness? If it's one or two parameters you have problems with, work with that, don't say everything is all too complex.

A material system gives you certain parameters that are ment to mimic lighting effects in real life. It has become a standard from the beginning of CG to make pretty crude approximations using different shading models, because back then we just didn't have enough computing power.

The bad thing is in order to grasp how they work, you have to see it from a programmatic point of view, instead of how materials and light behaves in real life. For example to understand photon maps, or shadowmaps, the programmer has to explain to you how their program works, what it's doing in order for it to achieve a certain look, which often has little to do with real life behavior.

So you move away from a reality based explanation to a programmatic one. I've just tried to explain how it works in real life and that Maxwell tries as much as possible to mimic that. It may sound complicated at first but if you really want to do nice renders, and you don't get the look you want, you have to understand what is going on. You make much better renders if you understand these real life concepts.

Don't think the material editor will remain exactly as it is now, ofcourse it will be further developed with the rest of Maxwell.

[Maxer]

I'm not sure what you mean there, how come you understood and accepted what IOR (or nd) does in beta, but now suddenly in 1.0, it's all a mystery? Same with roughness? If it's one or two parameters you have problems with, work with that, don't say everything is all too complex.

Basically what I'm saying is were all living under the illusion that what were creating in Maxwell are physically accurate materials, when in fact their accuracy completely depends on our understanding of how that material is created and our ability to manipulate each setting. If I want to create a plastic with a transmittance value of 150,150,150 and you create one with a value of 160,160,160 for the same material, who's to say yours is more accurate than mine or vice versa. My point is that these numbers are arbitrary only now instead of 3 parameters to mess with in the beta we've now got 5 or more for every BSDF layer and the job just got more difficult.

A material system gives you certain parameters that are ment to mimic lighting effects in real life. It has become a standard from the beginning of CG to make pretty crude approximations using different shading models, because back then we just didn't have enough computing power.

We still don't

Mihai I understand you argument, heck I even agree with most of it, I know were stuck with this system so I'm doing my best to understand it. I hope in the future NL can find a way to simplify this process and bring back the beta look that were all missing.

[Maximus3D]

Sorry for cutting in here in your discussion but didn't they say there were modifying the Wizard to make it better and even easier to use for the upcoming 1.x patch. I think they did, let's see what that'll one will be maybe we could get some hints or suggestions about what's to come from the team about the updated Material Wizard.

/ Max

[Frances]

deleted

[Mihai]

If I want to create a plastic with a transmittance value of 150,150,150 and you create one with a value of 160,160,160 for the same material, who's to say yours is more accurate than mine or vice versa.

They are both equally accurate. Just different materials. Perhaps in the future Maxwell could use real life measured materials and you can just apply that, but until then don't get too stuck on what is a physically accurate material, because even with beta you still had to eyeball it to get close to a certain material. Even if you used the aluminium preset for example, how would you decide exactly what roughness setting would be "accurate" to mimic a material you saw in a picture? Btw, you can use the complex ior files for metals if you want to, no settings to take care of there except roughness and anisotropy.

Nobody will ever be satisfied if they expect to simply open up Maxwell and immediately understand what all the settings do. There's always a learning curve. With Maxwell it's not really so steep, especially considering the kind of quality you get. Besides the material settings are tied to how light really behaves, not some abstract programmers code.

Regarding the beta look, there have been examples posted already that show you any differences are due to different material settings or/and slightly different lighting. If you're willing to post a scene you did in beta, we'll make it look exactly the same. You still have to consider emitter strength, iso, gamma all have slightly changed from beta. This contributes to the differences, it doesn't necessarily mean it's less realistic.

[Frances]

deleted

[JCAddy]

Mihai I understand you argument, heck I even agree with most of it, I know were stuck with this system so I'm doing my best to understand it. I hope in the future NL can find a way to simplify this process and bring back the beta look that were all missing.

I don't think simplifying the material creation process will bring back the look we're missing. That engine is dead and that is that.

Isn't that the truth.

[Mihai]

And Devin's point about CG industry standard terminology - pay attention. We will use small words and speak slowly to you if that will help you to understand the importance of that concept.

Ok, I'm listening. Explain to me also exactly what is a photon map. What is QMC sampling? If I set samples to 5 on an area light, what does that mean exactly? These are after all important terms.......better stay with the same concepts than bring something new to the table right?

[Rochr]

Rochr, if you took the time to read the manual it would be clearer....what do you expect exactly? This is knowledge offered to you, take it or leave it, but maybe you should consider part of the problem is you refuse to accept any kind of new knowledge.

This is by no means any ranting towards you, and i really don´t mind learning new things, as long as they speed up my workflow. A materials editor that doesn´t even support procedurals, nor the ability to scale clip maps without the use of Photoshop is prolonging that workflow. So you´re right, perhaps i don´t really feel motivated to waste time on it.

But my point was merely that the materials editor is anything but intuitive, and i have no problem sticking by that statement.

[rickyinmotion]

Yes we know that the material editor is very powerful.
But i think that what people are waiting is that NL give us some
good material database (plastic, wood, cement, etc.) that we can tweak.

[Frances]

deleted