Material library

[mgibeault]

I've yet to try V-Ray Next, I changed job and we don't use V-Ray here (yet), is the material library the same as in 3.6?
I think the library is very good for architecture and interior design but is lacking for industrial design. I wish for materials like Mold-Tech textures, more metal finishes and more plastic types.

Thanks!

[Iain]

I agree 100%. The materials library is pretty limited for industrial design. I considered switching to Keyshot at one point when I saw their Mold-Tech materials, but decided against it (too invested in V-Ray). For me at least, rendering is always done against a deadline and I don't have time to perfect materials. Being able to grab useable materials straight from the library would make an enormous difference. Even the VR Scan materials are limited, or they were when I trialled it some time ago (I didn't subscribe).

And to answer the specific question about the V-Ray Next library - it looks much the same as that from 3.6 to me. I haven't delved deeply yet, but the plastic options look very familiar. One thing that is new (new to me at least, not sure if it was in 3.6), is the Custom Library. You can now save your own materials and use them across projects. This is definitely a step in the right direction.

[mgibeault]

Yes, exactly.
At my new job I cannot push for V-Ray to replace what they've been using (Keyshot) because of that.

[Grigio18]

Is it just me who thinks that the material library is very old and made with a dated conception?
Most materials have too low resolution textures that are unlikely to hold up with 4k renderings which are now a standard.
I offer to give you a hand to make more credible and realistic materials. Let me know.

What I'd like to see in the next release is the possibility to save a material without having to save the textures, because in this way I find myself with a duplicate file that takes up uselessly space in my hard disk.
Give us the possibility to chose.

[Iain]

Grigio18, I suspect you're right about the materials in the library being a bit out of date. Or the method used to create and control them. They are complicated, and non intuitive unless you really know what you're doing (I don't).

I've just been dabbling with the new Metallic material in Next and the whole PBR thing is new to me. I've been learning what I can and it seems like an easier way to create realistic materials, I've managed to find good maps online that give me the best plastic material I've ever had (the bar was quite low). That's with very little effort, I'm basically just using them as is. Then I watched a Substance Painter tutorial - wow! I have much to learn.

The whole texturing thing is a new world. I've just used very basic texture mapping in Rhino, and never really been very satisfied.

I'd certainly be very interested to learn how you approach making more credible and realistic materials, particularly those typically used in industrial design. Thanks for your kind offer.

[Grigio18]

Grigio18, I suspect you're right about the materials in the library being a bit out of date. Or the method used to create and control them. They are complicated, and non intuitive unless you really know what you're doing (I don't).

ok, so we are on the two opposite sides, I know exactly what I'm doing: D

They are not particularly complicated, at least they aren't compared to the ones I usually do.

Grigio18,I'd certainly be very interested to learn how you approach making more credible and realistic materials, particularly those typically used in industrial design. Thanks for your kind offer.

My proposal was aimed at chosgroup, in the sense of being available to review the materials and implement them according to current standards and the guidelines that the chaosgroup itself suggests ...

However, if you want you can create a discussion in the gallery where you put your renderings and I can give you some suggestions and guidelines

[Iain]

ok, so we are on the two opposite sides, I know exactly what I'm doing: D

They are not particularly complicated, at least they aren't compared to the ones I usually do.

You’re absolutely right, materials don’t have to be complicated. I have got good results in the past keeping it simple but it often seems to be more by luck than judgement. The online help is much better now than it used to be, so the information is out there. I just need to find the time.

My proposal was aimed at chosgroup, in the sense of being available to review the materials and implement them according to current standards and the guidelines that the chaosgroup itself suggests ...

However, if you want you can create a discussion in the gallery where you put your renderings and I can give you some suggestions and guidelines

Regardless of whom the offer was aimed at, if it improves the materials I will benefit eventually so thanks again. Also for the offer of critique. Right now, I don’t have any non commercial work I can share but I’ll find something to render and post the result. I frequently need to render white plastic enclosures, and they tend to look rather grey and lifeless. Any tips would be most welcome.

[mgibeault]

Anyway, it would be good to know from ChaosGroup if there's something in the works or at least a plan for something in the future.
The first step was to get a library, the second was the possibility to add to this library and share it with colleagues. Now is time to complete it for the desired specialties.
I was about to get on this myself by using Dabarti Capture to build a real-world library but I have no spare time and I'm paid to do something else...

[konstantin_chaos]

Hi guys,

You are all right here.
The library hasn't been updated since V-Ray 3.6 in terms of the materials appearance or texture quality.
The Library management workflow, UI and functionality were our main focus for the initial Next release.

This was entirely based on feedback we got from users.
Everyone was liking the library but some workflow improvements have been requested.
Additionally all of the library shaders have been updated to use the new texture instancing functionality in Next.

Keep in mind that the idea behind the built in library was to present examples of various materials encouraging artists to modify and customize them.
As an example - the plastic materials come in only 2 colors letting users customize them according to the specific project needs.

...the material library is very old and made with a dated conception

I'm not sure what you mean here?
The texture size is intentionally limited to 1K since the library is part of the installer and it greatly affects its size.
All the textures were originally 4K in size and we could make them available at one point.

As a whole, we certainly have plans to update the library.
If it gets extended with new materials or bigger textures it would potentially require an additional download.

Any suggestion or ideas on how we can improve the library are welcome and will be considered.

Regards,
Konstantin

[Grigio18]

I'm not sure what you mean here?

Hi konstantin_chaos !

I'll give you just an example, sufficent to explain what I mean:

Let's look at metal materials from the library.
From what I learned, a physically correct metal sould not have a diffuse component, instead those in the library all have a diffuse...
Most of them use ward instead of ggx and all materials don't have active glossy fresnel...

in other words, you should provide basic materials in the library, as correct and realistic as possible, mainly:

- metals, perhaps made with the new metallic roughness workflow, since they are more realistic ...
- dieletric with a calibrated diffuse component to use as a reference, without using 1k textures, which are useless today for a commercial job.
- glass with a good enough max refl and refr depth and backside reflect active ect--

Then you can also put not physically correct materials, but be sure to give those a clear name, so that inexperienced users do not think that those materials are a quality reference but they know that those are made with shortcuts to speed up rendering times.

this I think could be a good service for your customers

[konstantin_chaos]

From what I learned, a physically correct metal sould not have a diffuse component, instead those in the library all have a diffuse...

The Diffuse portion of every material is an approximation of very blurry reflections and in that regard is not 'physical'.
Everything in the CG materials and light evaluation is some kind of approximation. The physical reality is too complex to be simulated as is.
We also have to take render speed into consideration.
I usually prefer to have some kind of a diffuse contribution for every material instead of layering extremely blurry reflections or increasing the ray depth significantly.
Aluminum is a good example of that. In my opinion using diffuse combined with reflections produces very natural results without adding too much to the render time.

Most of them use ward instead of ggx

The more blurry reflective ones use Ward once again making the material faster to render.
For some materials this is also accompanied by a decrease of the reflection depth.

all materials don't have active glossy fresnel

The option was introduced after the library was created.
Additionally I don't think that it'll change the visual appearance of the materials to the point you notice it.

in other words, you should provide basic materials in the library, as correct and realistic as possible, mainly:
- metals, perhaps made with the new metallic roughness workflow, since they are more realistic ...

The new workflow will not make materials more realistic.
The only benefit that we might gain is using more accurate IOR values for metals.
But then again, you'll hardly notice any difference.

- dieletric with a calibrated diffuse component to use as a reference

Can you give me more details about that?

without using 1k textures, which are useless today for a commercial job.

Every time a texture occupies less than 1k image space in the final render, the texture size is sufficient.
Even when it comes to 4k rendering each individual texture rarely exceeds 1000px in the image.
In that regard, the statement that all the textures that we provide are useless is incorrect.

- glass with a good enough max refl and refr depth and backside reflect

How would you define a 'good enough' depth? It's different in every situation.
The backside reflections are also a natural effect which adds to the render times and thus has been disabled for most materials.
In the case of the window materials the reflection IOR has been increased to account for the missing back-side reflections.

Then you can also put not physically correct materials, but be sure to give those a clear name

I like this suggestion.
We still have to be careful if we go there since none of the computer generated materials is 'physically correct'.

Regards,
Konstantin

[Grigio18]

I work trying to achieve the most realistic materials, that works perfectly in any light condition. render speed is not my first goal, especially now that the hardware power is not missing.

Can you give me more details about that?

as you know most of the issues with renderings are related with wrong albedo values,(especially from really bright or really dark materials) that generate issues with gi, color bleeding,fireflies long render times and more...

It would be really useful, especially for novice people to have some reference materials, really simple, to use as guide for white walls, asphalt, etc... most used and often most problematic materials... You have tons of data coming from vrscans R&d...

For example, another idea would be to give an alert message based on the name of the material.
I any parameter goes out of a range of values that are specific to that material (then one is free to ignore it)

let's say that I create a white stucco material, I set the diffuse to 255, and you find that the value is out of the reference range (180-200), you can show me an allert.

I don't know, I'm assuming, it's not easy to do, and maybe this is not the best solution and I am blathering, but I advise you to pay attention to these aspects if you want to lower the support requests and somehow train your users, they will be grateful to you

For sure this would save yourself a lot of identical questions on the forum, and you would educate users to create more realistic materials, always leaving the possibility to create even unrealistic materials, but with awareness of what they are doing.