Shader Suggestions

Generic forum to discuss Octane Render, post ideas and suggest improvements.
Forum rules
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
User avatar
Silverwing
Licensed Customer
Posts: 287
Joined: Wed Jun 15, 2011 8:36 pm
Location: Ludwigsburg Germany
Contact:

Th there deer Octane Team,
First of all thanks for your hard work on Octane. I really like it and its quality is really outstanding!

I am a artist really focused on shading and I hereby wanted to give a couple of suggestions for shader implementation that came into my mind for eventual further releases of octane:

A fresnel Shader:
A shader that is like the Falloff Shader in behaviour only it gives out a Fresnel curve.
You should be able to tweak the Index / Nd value of course.
Reason:
Fresnel is a really important part of shading. You already have implemented it with the "Index" you can adjust to control the reflection amount in the "Glossy" material model and the refraction amount in the "Specular" material model.
If you want to do a car paint material you would use a blend material to blend materials together with 2 different reflection roughness levels.
The amount they are blended together is provided by the Fresnel normally.
Lacking a Fresnel shader you have to tweak the falloff the way it comes close to an Fresnel and use it in the mix amount slot. But this way you never can mace sure you have nailed down the Falloff to a physical correct amount. A Fresnel shader would make this much more simple by letting you just put in the right index and therefore producing a photo real result by itself.

Additional:
Integration of a Complex Fresnel (with a Extinction Coefficient)
Reason:
As you might now the Fresnel model implemented in most 3D applications is called "real part of complex index of refraction" and is driven with the Index Number / IOR / Nd derived from a materials meassured index at a light wavelength of 0.5893 μm.
This model is however only intended for non metal materials (as plastic glass etc)
To calculate the view depended reflection curve for metals you would need to add a extinction coefficient "K" (which is the imaginary part of the Fresnel formula)to the Index value. They both then dive the function and you will be able to calculate physically accurate metal reflections.
P.s. the only renderer I know of that has this implemented at the time is Maxwell.

GGX Distribution Functions
Recently Vray (in the 3.0 Built) and C4D R16 both have implemented the GGX "shader" to allow more flexibility with shading.
I´d really love to see this implemented in Octane as well! https://vimeo.com/88605345
Reason:
Different types of material have different distribution behaviours of reflections. To get even more realistic Reflection behavior this will further enhance the palette of what you can do with octane.

In my opinion this would improve Octanes feature set further and give even more flexibility in to the hand of the artists!
Thanks for reading and considering!

Cheers,
Raphael
WIN 10 PRO 64 | ASUS X99-E WS | 3 X GTX 1080Ti | i7 5960X 8X 3,33GHz | 64GB RAM
Vue2Octane
Licensed Customer
Posts: 88
Joined: Thu Jun 26, 2014 8:16 am

I am actually not convinced by the refractive index that is provided in Octane.
I mean you can set the index AND seperately transmission and reflection, which is just nonsense. Any given index of a material results in a certain reflection and transmission just by physics. To then be able to change them again....I do not know. The questions pops up what to input for transmission and reflection once the index is set to be physically accurate! Both 1, Both 0... I do not know. Also angle dependence of reflection is impossible with this setup. Also that the sum of transmission and reflection can be greater than 1 is just artificial. That would mean new energy is produced once a ray hits a surface. I am quite sceptical about all refractive index things in Octane. Especially because it is supposed to be physically accurate pathtracer. If I want to tinker with settings and shaders then I can use any biased render engine.
Phil_RA
Licensed Customer
Posts: 75
Joined: Fri Jul 25, 2014 2:30 am

I agree, I thought reflection should always be 1 and you only set the Index instead, but when I tried to make smooth black plastic it came off looking too gray, I had to lower the reflection value to get the proper look. This seems wrong. Also, the database should have a separate section of physically correct materials. There should be an "official" plastic material that was made by the devs, which people can use as a reference.
User avatar
Silverwing
Licensed Customer
Posts: 287
Joined: Wed Jun 15, 2011 8:36 pm
Location: Ludwigsburg Germany
Contact:

P.s. also I would love to see a anisotropic shader model in there!
I know you can use bump to do it (as in treal world) but sometimes it gives you moire artifacts this way.

Thanks for your consideration!

Cheers,
Raphael
WIN 10 PRO 64 | ASUS X99-E WS | 3 X GTX 1080Ti | i7 5960X 8X 3,33GHz | 64GB RAM
User avatar
glimpse
Licensed Customer
Posts: 3740
Joined: Wed Jan 26, 2011 2:17 pm
Contact:

Raphael, agree with You - these would be really nice additions! now it's like hit & miss (test, test, test..)..just trying to make as close as possible (from the looks), but it would be way easier just to drop a value & forget..
Rickky
Licensed Customer
Posts: 127
Joined: Wed Jun 02, 2010 11:09 am

You can look at thea render with its gpu render presto. The material setting is well done ;)
I saw the Grant Warwick vray shader. Interesting approach. He is as well an octane user. Maybe I can give his feedback and if he succeed to build this shader in octane
raylab
Licensed Customer
Posts: 41
Joined: Sun Mar 21, 2010 5:35 pm

Really great suggestions Raphael! I think these looks interesting also...

Shifted Gamma micro-facet distribution
http://hal.inria.fr/docs/00/70/23/04/PDF/paper.pdf

ABC Shader
http://vcl.itn.liu.se/publications/2012 ... 012/LKYU12
- http://about.me/raylab_sv - 1st Supermicro 3*Titan - Netstor Turbobox 4*Titan - 2nd 2*TitanZ - Netstor Turbobox 4*780
CurseStudio
Licensed Customer
Posts: 28
Joined: Tue May 08, 2012 5:01 pm

+1 Bump!
realise
Licensed Customer
Posts: 33
Joined: Thu Mar 26, 2015 1:24 pm

+1 (x1000)

Huge bump!
itsallgoode9
Licensed Customer
Posts: 893
Joined: Thu Apr 03, 2014 9:04 am
Location: New York City
Contact:

i'm not convinced about their model either. it actually turns out that the code for the fresnel for the specular materials was broken and it was giving full reflectivity all the time. they fixed that in the newest update though. still though, i feel like something is off and I have to reduce the reflectivity on most all of my materials.
Intel i7-3930K, 64gb RAM, Asus X79 Deluxe mobo, 2x EVGA 780 6gb (for rendering), 1x PNY quaddro k4000 (for display)
Windows 8.1 x64, Maya 2014, Octane Render v2
Post Reply

Return to “General Discussion”