Arnold 5 missing features

[Jason_Higgins]

Hello,

after a longer absence i tested version 5 over the last few days and was mildly shocked. The denoiser and adaptive sampling are nice, but so much has been removed without substitution.

I have shaders that make use of Ward, Ashikhmin-Shirley and Phong which are all gone! The standard material doesn't even have the GGX/Beckmann selection anymore and the controls for direct diffuse and specular are also gone! The constant decay option of the lights is gone as well which is particularly bad.

And as if that wasn't enough i had to find out, that the AL shader is not available for 5. The official reason is that it's not needed anymore because of what 5 has now - what a joke. The AL shader offered so much more compared to the old and new aiSurface. The extra samples for example were exceptionally useful.

Also the AOVs. It was so easy to create AOVs, custom groups and assigning channels to them. Now it's all LPE and nothing is handy anymore.

What were you thinking?

[lee_griggs]

Arnold 5 FAQ

[Jason_Higgins]

What is that supposed to do?

I'm well aware of the changes. They are the reason for my post.

[aaronfross]

They decided that you don't need those features anymore.

[Jason_Higgins]

The way it looks to me now, that is the sad truth.

[Christoph_Schaedl]

same with the latest renderman or vray version.
they delete features to make the renderer faster and less artist driven.

[Slava_91]

For the alShader.
1.This shaders without energy conversion.
However alSurface have a bug in all version with composite two specular layers.

2. ID AOVs in alSurface. Same feature you have in aiSurfaceStandart in 5.1 core.

3. Light group and shadow group in alSaders create only direct and indirect beauty passes, LPE can create light group per component aov.

4.diffusion and direction sss in alSurface. Randomwalk in Standart shader works with better quallity and lower noise and render speed.
- 3 sss layers in alSurface an 1 in aiStandardSurface.
alSurface never used 3 layers of SSS for the texture, and it was completely not correct to do this, all 3 layers multiplied by the diffusion layer,
take two textures and connect them through a aiMultiply node, what result will you get? and what you get if you multiply all 4 textures?

They used only to correct the color of the depth sss,
the same thing in the new shader does sss radius color per component (r,g,b).

5.Ward, Ashikhmin-Shirley, Phong, Beckmann. This is a bsdf of the times of dinosaurs, they are not used anywhere else, they never simulate the real reflections and glare, and create a fake thinking, for simulation real reflection need blend two specular layer.

6.roughness clumping parameter in alSurface.I've never have such problems with a new shader, to say that it is necessary someone in aiStandartSurface.

7.absortion in new shader(transnitance color + depth), scatter (transnitance scatter +depth).

8.Debug mode in alSurface and alLayer. Do you want say this more easy than press "isolate selection" botton in ARV on concrete node, or ARV debug mode?)

for the shaders and nodes
alLayer-> aiMix, but arnold have more greater node aiBlendShader like in vray BlendMaterials
alLayerColor->aiLayerColor
alLayerFloat->aiLayerFloat
alRemapColor or Float->aiColorCorrect and aiRange
read documentation for more information about nodes...

And you always have chance to convert you old shaders to new with Converting script in arnold panel tools.

Maybe this helps you understand why its happened

Pixar’s Foundation for Materials

Disney Extending the Disney BRDF to a BSDF with Integrated Subsurface Scattering

Revisiting Physically Based Shading at Imageworks

Sincerely,
Slava S.

[Jason_Higgins]

Thank you for the detailed explanation, but nothing of that has caused me any trouble in the past. I am an artist, i have no clue about physics and coding looks all greek to me. I also don't care about the age of bsdfs but as an artist, to me it matters that they look different so i can be more creative. The same applies to the light decay i mentioned and very useful settings of the AL shader like the extra samples, the GGX/Beckmann selection and others.

None of that i have in Arnold 5 and there is no replacement either. Arnold 5 is obviously not for me. I guess i have to look for something else or just keep using Arnold 4. Thanks for the replies!

[aaronfross · Accepted solution]

Solid Angle seems to have a religious devotion to physical and mathematical accuracy. Do they care about, and/or can they comprehend, the needs of artists? A totally fake lighting solution can be more desirable than their elegantly-constructed simulation. I am suitably impressed that they've reduced the universe to an equation, but that doesn't help me as much as I think they believe. As a user, I feel abandoned in favor of a deluded, Quixotic quest for the Platonic ideals of physical light propagation. We are in the business of creating beautiful illusions (it's *Maya*, goddess damn it), not simulation of reality down to the quantum level. Removing Constant light decay was a bad decision. But at least we can still cheat reality in a few ways in Arnold 5.

[Slava_91]

Everything moves to get user work as possible simple, this happens not only in arnold, but in all known renderers, the old algorithms no longer in trend and actually not used anywhere.
So why waste time developing and supporting for old tools and algorithms?
As for light decay you can use LightDecay Filter and spreed parameters in AreaLight for more artistic controls or Direction light for constant decay.

I think they delete the decay type for good reason, or not...)

[Slava_91]

I would suggest you adapt, 4th core will sooner or later force you to switch to newer versions, and old bsdf in another renderers will be removed with time ...