OctaneRender™ Standalone 3.08 RC 1

[smicha]

Thank you OTOY so much for you fantastic work. I'll test it on my 11x 1080 TI machine this week

[funk]

Can someone explain what's going on with sheen?

Why does it draw over reflections on the glossy material, but only draws on the diffuse part of the metallic material?

I am not exactly sure what you meant here. But basically the reflection from sheen is a cosine lobe oriented towards the direction that is 90 degrees from the viewing angle, and it peaks at this grazing angle. The roughness parameter controls how spread out the lobe is, just like your specular BRDF model, it can look really diffuse when you have high roughness, and you can have really sharp sheen like satin cloth for low roughness.

Here is a metallic material that is 100% reflective. No sheen can be seen (you explained this is intentional)

Here is a glossy material, made 100% reflective. The sheen seems to tint the reflection.

Here is the OCS

sheen_metallic_vs_glossy.ocs

(28.42 KiB) Downloaded 276 times

Which leads to another question... Why dont we have sheen on the diffuse material? This would be nice for velvet/fabrics

Is there any reason why you cannot achieve this effect with the metallic node and glossy node? You can decrease the glossiness and increase sheen.

No reason. I was just curious why it wasnt part of the diffuse material too. In fact I wish we just had a single material that combined everything.

[roeland]

I can confirm there's an issue with dispersion. This will be fixed in the next release.

In fact I wish we just had a single material that combined everything.

Yes we know — you're not the first one to make this point.

In the meantime we can try to approximate some common cases with node graphs. Does this work for a metallic + diffuse with dielectric mix?

PBR Metallic + Glossy.orbx

(72.92 KiB) Downloaded 286 times

Roeland

[funk]

I can confirm there's an issue with dispersion. This will be fixed in the next release.

In fact I wish we just had a single material that combined everything.

Yes we know — you're not the first one to make this point.

In the meantime we can try to approximate some common cases with node graphs. Does this work for a metallic + diffuse with dielectric mix?

PBR Metallic + Glossy.orbx

Roeland

I only had a quick look, but yes this seems quite close. The only thing I noticed is you have a fixed IOR (1.45) on the glossy mat.

In general I think its ok, but if people are using actual specular maps from other apps, then it wont match. Disney remap specular 0-1 to IOR 1.0 - 1.8. Specular 0.5 (the default) = IOR 1.5

EDIT: Its discussed on pg 15 of this pdf https://disney-animation.s3.amazonaws.c ... tes_v2.pdf

[Renart]

Thanks for all the fixes!
The displacement works great now and the Volume passes are very useful.

I can't wait for the next release. Any ETA?

[Master]

Hello we have a machine running with win 10 64 , C4D R19.053.
We have two titan XP and a Titan V used as display waiting for it to be supported in Octane render .
I was running octane render 3.08 test 6 .
This morning I updated the cuda to 9.1 and installed 3.08RC1.
I can't select the titan V in octane settings devices, the card is unselected when I close the settings window.
Then I updated the graphic driver to the last provided by Nvidia .
Same issue.
Another issue with the the 3.08 RC1 is that I want to preview a scene with many polygons it does not render. It takes 2 minutes 30 with 3,08 test6 and never computes with the 3,08RC1 except if you remove polygons computing.
Thanks for your help.

[andw]

There is still some weird issue (values clipping / saturation ?) with environment HDRI with high dynamic range
( https://hdrihaven.com/hdri/?h=flower_road )

Moreover, the result differs in 3.08 RC1 version

[abstrax]

Thanks for all the fixes!
The displacement works great now and the Volume passes are very useful.

I can't wait for the next release. Any ETA?

Ideally there would be another release this week with all reported issues fixed and if everything goes well, we might have a stable build by the end of next week. But it all depends on the issues that are found in the coming days.

[abstrax]

Hello we have a machine running with win 10 64 , C4D R19.053.
We have two titan XP and a Titan V used as display waiting for it to be supported in Octane render .
I was running octane render 3.08 test 6 .
This morning I updated the cuda to 9.1 and installed 3.08RC1.
I can't select the titan V in octane settings devices, the card is unselected when I close the settings window.
Then I updated the graphic driver to the last provided by Nvidia .
Same issue.
Another issue with the the 3.08 RC1 is that I want to preview a scene with many polygons it does not render. It takes 2 minutes 30 with 3,08 test6 and never computes with the 3,08RC1 except if you remove polygons computing.
Thanks for your help.

Regarding the support for Titan V: Could you send a screenshot of the Standalone and the device selection window? It sounds as if you are running the wrong version. FYI: I'm currently running Win 7 with GeForce driver 390.77 and the Titan V works as expected.

Regarding the scene that doesn't render: Could you send it to me so I can check what is going on? Thanks a lot.

[abstrax]

There is still some weird issue (values clipping / saturation ?) with environment HDRI with high dynamic range
( https://hdrihaven.com/hdri/?h=flower_road )

Compare-308.jpg

Moreover, the result differs in 3.08 RC1 version

The "halos" around the cube are an issue in the PMC kernel we are currently trying to fix.

Unfortunately, the archive didn't contain the HDR image you are using. Could you send it to me please? Thanks.