OctaneRender™ Standalone 3.08 TEST 4

[FrankPooleFloating]

THE EAGLE HAS LANDED.

Congratulations guys, this is a huge step forward!

For a while I've been trying to make a "window shader" using emission and a texture in distribution slot + projector linked to the camera with some expressions but it was a mess. And then I found THIS, and since then I was hoping this would one day, maybe, just maybe, work in Octane.

But I did not expect that a simple copy/paste would work straight 'out of the box' in the first beta

Two polygons and one texture:


...Five minutes, 128 triangles and 5 textures later:


Here's a basic setup. Obviously, I am NOT an advanced user, so I get a crash if I try to embed a texture ( input_tex ) in the Orbx. Textures can be downloaded from the author's blog here and they should go in the "input_tex:" slot.

WindowBoxOctane.zip

Have fun!

Regards
Milan

I sat bolt upright in bed last night while thinking about this... Could OSL be used to add (perceived) thickness to 2d geo, for making fences, distant buildings, parts of spaceship hulls, etc, etc?.. Or can this tech only make the inside of something appear to have depth? I haven't had a chance to dive into the mechanics of what makes this work, so I am feeling a tad Amish with this work of the devil I see before me here...

[Jolbertoquini]

I sat bolt upright in bed last night while thinking about this... Could OSL be used to add (perceived) thickness to 2d geo, for making fences, distant buildings, parts of spaceship hulls, etc, etc?.. Or can this tech only make the inside of something appear to have depth? I haven't had a chance to dive into the mechanics of what makes this work, so I am feeling a tad Amish with this work of the devil I see before me here...

Yes, the Parallax mapping is great you can find OSL at blender but we was looking at with Pascal, here some links:

http://apoorvaj.io/exploring-bump-mappi ... .html#toc4
https://www.youtube.com/watch?v=9j9oGMDto2I

But the problem is we may need the OSL materials to be supported. for such thing or maybe the Guys could help us to create something to support parallax maps this would save some space on memory polygons and not use displacement.

the OSL version there is not really the same jiWindowBox script doesn't reacts with the light like normal and etc. cause the guys who made didn't add this options such the webgl link (parallax occlusion mapping with max steps).

Developers could someone tell us if is a possibility to make work?

Cheers,
JO

[funk]

Shouldn't a "metallic material" with n = 1.5, k = 0, look the same as a glossy material with IOR = 1.5?

The metallic material is shinier. Is this a case of using the wrong fresnel formula again? You've done this in the past and had to switch to non polarised

EDIT: Setting n = 0, and specular = 0.04 (4% = IOR 1.5), is getting much closer to the glossy material (although fresnel is a bit different due to schlick's approximation). So something is wrong with your n/k formula

[funk]

I'm also finding the Ward BRDF looks wrong. It doesnt seem to have fresnel reflections. Is this correct?

[wallace]

I'm also finding the Ward BRDF looks wrong. It doesnt seem to have fresnel reflections. Is this correct?

The implemented Ward BRDF is the original Ward model, which has a darker black rim compared to other microfacet models.

We are looking to improve the Ward BRDF to make it match closer to the other BRDF models, which will remove the effect of this black rim.

Thanks

[funk]

I'm curious why you decided to implement a single set of n/k values, rather than a set for R,G and B wavelengths? eg like the complex fresnel osl for v-ray here https://docs.chaosgroup.com/display/OSL ... nel+shader (I'm not sure how accurate this OSL shader is anyway)

[funk]

Another thing I would like to mention is roughness. I'm doing some quick comparisons to GGX in modo and v-ray, and your roughness values don't seem to match

eg. 40% roughness in modo GGX and 0.6 glossiness (the inverse) in v-ray GGX, look the same as eachother, but Octane's GGX does not match them

This means roughness textures will produce different results between render engines.

I have to wait till we get 3.08 for modo so I can do direct comparisons. In the past, the modo plugin used a formula to convert modo GGX roughness to an approximate Octane roughness eg. (ggx roughness * 2) * 2

We now need to figure out another formula for Octane GGX. It would be much better if your GGX roughness just matched the other applications.


EDIT: I seem to remember something in previous conversations, about Disney remapping GGX roughness. Found it on pg 15
https://disney-animation.s3.amazonaws.c ... tes_v2.pdf

"For roughness, we found that mapping (a = roughness^2) results in a more perceptually linear change in the roughness."

I'm not a programmer, but maybe this means something to you guys

[wallace]

Another thing I would like to mention is roughness. I'm doing some quick comparisons to GGX in modo and v-ray, and your roughness values don't seem to match

eg. 40% roughness in modo GGX and 0.6 glossiness (the inverse) in v-ray GGX, look the same as eachother, but Octane's GGX does not match them

This means roughness textures will produce different results between render engines.

I have to wait till we get 3.08 for modo so I can do direct comparisons. In the past, the modo plugin used a formula to convert modo GGX roughness to an approximate Octane roughness eg. (ggx roughness * 2) * 2

We now need to figure out another formula for Octane GGX. It would be much better if your GGX roughness just matched the other applications.


EDIT: I seem to remember something in previous conversations, about Disney remapping GGX roughness. Found it on pg 15
https://disney-animation.s3.amazonaws.c ... tes_v2.pdf

"For roughness, we found that mapping (a = roughness^2) results in a more perceptually linear change in the roughness."

I'm not a programmer, but maybe this means something to you guys

Hi,

I am aware of different roughness mappings, but as you see, Disney is just one of the many different GGX mapping’s, it’s just done to give a non linear finer control for the slider in GUI, Octane’s isn’t necessarily wrong.

I will investigate different applications to compare their mappings and adopt the most common alpha value.

About the metallic node, I did not contribute anything to it so I cannot answer your question regarding complex fresnel, my apologies.

[funk]

I am aware of different roughness mappings, but as you see, Disney is just one of the many different GGX mapping’s, it’s just done to give a non linear finer control for the slider in GUI, Octane’s isn’t necessarily wrong.

I will investigate different applications to compare their mappings and adopt the most common alpha value.

About the metallic node, I did not contribute anything to it so I cannot answer your question, my apologies.

Thanks for checking it out Wallace

Another reason to consider this is, many commercial/free "PBR" textures will end up looking very different in Octane when applying roughness textures. If we can match roughness mapping, it would make things easier

[wallace]

Absolutely, that’s a very valid reason, will get that in for next release.