Author Topic: New Corona Physical Material (PBR) playground!  (Read 40000 times)

2020-12-12, 17:38:06
Reply #30

joonior1985

  • Active Users
  • **
  • Posts: 21
    • View Profile
Hey there,
is there a difference between sheen parameter and regular falloff mask? Same question could be applied to a potential difference between using new clearcoat option and regular CoronaLayeredMtl where clearcoat is a regular glossy material masked with base layer using fresnel. I'm just curious because for instance velvet material is a pain in the ass as it's way more complex than just falloff and/or "sheen" :)

2020-12-12, 18:24:56
Reply #31

Feodor

  • Active Users
  • **
  • Posts: 113
    • View Profile
As always, you are on top!

But!
When testing the new Shader, I found two problems.
First: in a refractive material, when the minimum roughness value is set, the IOR changes significantly. Shown in the picture. Especially the picture becomes different when using glass with the Shell modifier
Second: if you set the roughness value to 0 in the Sheen layer, 3ds max 2021 will hang dead. (There must be a division by zero) :))
« Last Edit: 2020-12-12, 18:29:00 by Feodor »

2020-12-12, 22:56:57
Reply #32

pokoy

  • Active Users
  • **
  • Posts: 1631
    • View Profile
As always, you are on top!

But!
When testing the new Shader, I found two problems.
First: in a refractive material, when the minimum roughness value is set, the IOR changes significantly. Shown in the picture. Especially the picture becomes different when using glass with the Shell modifier
Second: if you set the roughness value to 0 in the Sheen layer, 3ds max 2021 will hang dead. (There must be a division by zero) :))
I think the same happened with the old material, too. Could you test this to verify?

2020-12-13, 11:28:46
Reply #33

Feodor

  • Active Users
  • **
  • Posts: 113
    • View Profile
Yes, the old material has the same problem. It is worth changing the roughness value in the reflection, how much does the IOR change in refraction, but how do I understand this applies only to the background?

2020-12-13, 14:44:12
Reply #34

Juraj

  • Active Users
  • **
  • Posts: 4312
    • View Profile
    • studio website
Hey there,
is there a difference between sheen parameter and regular falloff mask? Same question could be applied to a potential difference between using new clearcoat option and regular CoronaLayeredMtl where clearcoat is a regular glossy material masked with base layer using fresnel. I'm just curious because for instance velvet material is a pain in the ass as it's way more complex than just falloff and/or "sheen" :)

From my testing, esp. Sheen doesn't seem to act much differently than Fallof, although it automatizes the setup with simpler setup ("Roughness"). I didn't test it too extensively but doesn't seem like it reacts to light or anything of that sort.

The best velvet setup is still mystery to me :- ).
talcikdemovicova.com  Website and blog
be.net/jurajtalcik   Our studio Behance portfolio
Instagram   Our studio Instagram, managed by Veronika

2020-12-13, 17:19:39
Reply #35

Feodor

  • Active Users
  • **
  • Posts: 113
    • View Profile


From my testing, esp. Sheen doesn't seem to act much differently than Fallof, although it automatizes the setup with simpler setup ("Roughness"). I didn't test it too extensively but doesn't seem like it reacts to light or anything of that sort.

The best velvet setup is still mystery to me :- ).

I don't think it's just Fallof

« Last Edit: 2020-12-13, 19:40:21 by Feodor »

2020-12-13, 18:57:05
Reply #36

joonior1985

  • Active Users
  • **
  • Posts: 21
    • View Profile

From my testing, esp. Sheen doesn't seem to act much differently than Fallof, although it automatizes the setup with simpler setup ("Roughness"). I didn't test it too extensively but doesn't seem like it reacts to light or anything of that sort.

The best velvet setup is still mystery to me :- ).

Apparently it really does work differently. Regular Fallof doesn't take light source directional vector into account but sheen does. Check out my previews :)
P.S. I figured out (partially) a relatively good velvet shader but still it's not the general solution that works in every scenario :D

2020-12-13, 19:57:37
Reply #37

Epilog studio

  • Primary Certified Instructor
  • Active Users
  • ***
  • Posts: 47
    • View Profile
    • Epilog studio
Hello, The overall concept of new material is looking great But I am missing one big feature- ability to define BRDF more precisely. Is it planned? Tail fallof etc?

2020-12-14, 08:21:39
Reply #38

Juraj

  • Active Users
  • **
  • Posts: 4312
    • View Profile
    • studio website

From my testing, esp. Sheen doesn't seem to act much differently than Fallof, although it automatizes the setup with simpler setup ("Roughness"). I didn't test it too extensively but doesn't seem like it reacts to light or anything of that sort.

The best velvet setup is still mystery to me :- ).

Apparently it really does work differently. Regular Fallof doesn't take light source directional vector into account but sheen does. Check out my previews :)
P.S. I figured out (partially) a relatively good velvet shader but still it's not the general solution that works in every scenario :D

I downloaded this into Photoshop but I am still not sure what I am looking at :- ) The light spread the sheen "specular" lobes? The lobes react do directionality/size of light?

talcikdemovicova.com  Website and blog
be.net/jurajtalcik   Our studio Behance portfolio
Instagram   Our studio Instagram, managed by Veronika

2020-12-14, 09:35:38
Reply #39

rowmanns

  • Corona Team
  • Active Users
  • ****
  • Posts: 1346
    • View Profile
As always, you are on top!

But!
When testing the new Shader, I found two problems.
First: in a refractive material, when the minimum roughness value is set, the IOR changes significantly. Shown in the picture. Especially the picture becomes different when using glass with the Shell modifier
Second: if you set the roughness value to 0 in the Sheen layer, 3ds max 2021 will hang dead. (There must be a division by zero) :))

Hi,

Can you send over a scene for the second issue you described? I am unable to reproduce this crash/hang.

Cheers,

Rowan
Please read this before reporting bugs:
How to report issues to us!

2020-12-14, 09:44:08
Reply #40

mvshabeer

  • Active Users
  • **
  • Posts: 65
    • View Profile
    • Portfolio
This is not specifically to PBR material. Here is the comparison b/n default max physical material and Corona physical material.
It will be nice to have viewport preview (reflections) in high quality mode. Latest max  even supports bloom. For those who argue about the use case, some projects we don't even render due to time restrictions and last minute changes. We use viewport previews. So far we have been creating 2 scene files, one with max physical material setup for previews and one with corona for rendering if needed. It will be really helpful if Corona have an option of high quality preview in viewports

« Last Edit: 2020-12-14, 09:47:51 by mvshabeer »

2020-12-14, 09:47:49
Reply #41

joonior1985

  • Active Users
  • **
  • Posts: 21
    • View Profile

From my testing, esp. Sheen doesn't seem to act much differently than Fallof, although it automatizes the setup with simpler setup ("Roughness"). I didn't test it too extensively but doesn't seem like it reacts to light or anything of that sort.

The best velvet setup is still mystery to me :- ).

Apparently it really does work differently. Regular Falloff doesn't take light source directional vector into account but sheen does. Check out my previews :)
P.S. I figured out (partially) a relatively good velvet shader but still it's not the general solution that works in every scenario :D

I downloaded this into Photoshop but I am still not sure what I am looking at :- ) The light spread the sheen "specular" lobes? The lobes react do directionality/size of light?

I prepared more obvious sample that explains how it works :) The Sheen somehow reacts with count of the light sources and their direction creating this radial gradient instead of linear (when using regular falloff). I still got no clue if there is any groundbreaking advantage over well customized falloff but I'll do some more testing anyway. It may works with satin materials but velvet from my adventures with that particular material is (as said in previous post) more complex. Velvet must take into account few variables (third is most important):
1. Falloff/Sheen (done)
2. Specular with anisotropy (done)
3. Handling a 3D map/bitmap that allows to recognize what particular direction and angle the surface is to the light and camera. The real physical velvet piece (lets say a square 50x50 cm) looks different for every angle. It is caused by tiny hairs that leans always to a particular direction (in world coordinates) and once they lean towards You but if You rotate this piece of fabric they don't change their direction according to surface but they do so for the viewer/camera. Thats the basic principle that (from my knowledge) nobody handled yet :D

2020-12-14, 11:15:40
Reply #42

Juraj

  • Active Users
  • **
  • Posts: 4312
    • View Profile
    • studio website
I wonder if what you are showing isn't just the diffuse retro-reflection very rough materials now have from Oren-Nayar shading. Is there a sheen pass to check? I would check the diffuse pass to see.

Fabric effect is extremely convincing from GeoPattern feature but sampling from normal maps not much so yet.
talcikdemovicova.com  Website and blog
be.net/jurajtalcik   Our studio Behance portfolio
Instagram   Our studio Instagram, managed by Veronika

2020-12-14, 11:58:08
Reply #43

GeorgeK

  • Corona Team
  • Active Users
  • ****
  • Posts: 566
  • George
    • View Profile
As for fabrics and sheen, if possible I would like to see an example of a photorealistic silk or velvet/crushed velvet shaders using as said detailed bump maps, for example these

Satin is a bit tough to work on bump+sheen alone, it is somewhat reproducible but the lack of detailed maps, proper UVs won't make it look nice in most angles, so you'll have to resort to using a custom fresnel.

Thus you can apply a falloff curve to sheen itself in order to emulate more satin surfaces, this works great and the effect is more subtle than diffuse+falloff and seems to react better with scene lighting overall.


For silk-velvet, sheen alone is enough, comparison: https://corona-renderer.com/comparer/RugNph
For solutions, troubleshooting, ticket requests please visit - [link]

2020-12-14, 12:58:30
Reply #44

marchik

  • Active Users
  • **
  • Posts: 111
    • View Profile

Satin is a bit tough to work on bump+sheen alone, it is somewhat reproducible but the lack of detailed maps, proper UVs won't make it look nice in most angles, so you'll have to resort to using a custom fresnel.

Thus you can apply a falloff curve to sheen itself in order to emulate more satin surfaces, this works great and the effect is more subtle than diffuse+falloff and seems to react better with scene lighting overall.


For silk-velvet, sheen alone is enough, comparison: https://corona-renderer.com/comparer/RugNph
Thx for the reply, I mainly work with fabrics and have made hundreds of attempts to recreate complex shaders such as velvet, satin, silk, wool  and so on, including growing millions of fibers in Ornatrix and combing them and then using CoronaHairMtl  to get proper anisotropy reflection for the correct velvet, and basically I can create any texture in substance designer e.g.

But at the moment, if I understand everything correctly, we have no way to get the correct direction of the anisotropic reflection along the single fabric thread using height maps or normal maps, no matter how detailed they are, we can only fake it using the anisotropy rotation map.

And we also don't have a geopattern to mimic the real fabric structure. So I'll probably have to tweak the faloffs again and wait for the next release :D

PS
please guys, pay attention to the implementation of coating, which we wrote about above. let it be possible for coating to smooth the surface, creating mirror-like glossy coat with displaying the height details of the surface lying below, as Bormax mentioned here
New material is really great! Thank you!

The Clearcoat layer is the thing I was missing for very long time, nice to get it now. Some idea came to my mind about it.
Now the Base bump affects the Clearcoat layer and Clearcoat bump gives possibility to Add bump to this layer. Would it be nice (if it's possible) to have kinda lock between Base bump and Clearcoat bump in order to imitate the bumpy surface covered by the polished transparent layer of lacquer? Something like on attached picture where I've created the geometry of the covering layer and assigned different material to the inner and covering surfaces.
According to this idea if the Lock is active Base bump affects the Coatlayer, if it's unlocked only Clearcoat layer's bump affects the Coat layer.

PPS
and maybe it is worth lowering the metalness map slot below the others, because now I habitually attach base color or diffuse to it as it is the first in the list :D