OctaneRender™ 2020.1 XB2 [superseded by 2020.1 XB3] Tweet

[Notiusweb]

If I try and run the Fun with Vectron scenes in this 2020 XB2, some don't work at all.
Some like Surface displacement render so ridiculously slow that they are unusable.

[Lewis]

Hi OTOY/Devs

Can you please check this scene (ORBX) and short video about Denoising issues (BUG ?) i have with 2020 XB2.

I'ts very simple scene (you already have it but for sake of conversation uploading it again).
Basically Glass behind glass confuses denoiser so it completely destroys/removes my glass railing behind my glass doors. Seems like glass behind glass is big problem for denoiser?
Can this be fixed/trained better in denoiser ?

Here is ORBX file and short video upload.

thanks

[FrankLIU]

When 2x Upsampling is used, Imager sliders' effect gets laggy/choppy, whereas when no upscaling is active, slider response is smooth and 'realtime'.

can you please share some test scene and let me have a check?

Thanks

[vijay]

Auto bump map in displacement doesn't work.

001.JPG

002.JPG

Thanks for the sample scene. This is now fixed and it will be included in the next release. Thanks!.

[Studio21]

Hi devs,

Is it possible to have a simple scene that demonstrates the advantage of using the new quad lights. ive been testing with my own scene and didnt notice and benefits in terms of resolving noise any faster then a normal emission plane. im reall confused by this

Cheers,

G

[Notiusweb]

When 2x Upsampling is used, Imager sliders' effect gets laggy/choppy, whereas when no upscaling is active, slider response is smooth and 'realtime'.

can you please share some test scene and let me have a check?

Thanks

Now I am trying on new and different scenes, and it is not happening. Then on scene where it was happening, now it is not...
Scratch that!

[mojave]

If I try and run the Fun with Vectron scenes in this 2020 XB2, some don't work at all.
Some like Surface displacement render so ridiculously slow that they are unusable.

Thank you for reporting this. We have been able to reproduce it and it will be fixed in the next build.

[Notiusweb]

If I try and run the Fun with Vectron scenes in this 2020 XB2, some don't work at all.
Some like Surface displacement render so ridiculously slow that they are unusable.

Thank you for reporting this. We have been able to reproduce it and it will be fixed in the next build.

Awesome!

In the intro post for next build can there be a sample Vectron-Volume scene as well?
It's not too intuitive using OSL and Volume SDF instances inside a Vectron.
If we had a template scene that the average Noob user, who is not very advanced (ie - Funk, Divasoft, Rohandalvi, Despot, Galleon27, Rajib, Ff7darkcloud, VVG, Frankmci, Rikk The Gaijin, Lewis, Video4), could use, it would be great!

THX!

[Notiusweb]

I found another bug for this build-

If you add Layers to a Material
ie - Material Layer>Layer 1:Diffuse Layer

And then rename them
ie - Material Layer>Layer 1:Blend Color

The name will revert back to the default "Diffuse Layer" upon saving, closing, and re-opening file.

Ideally upon shutting down scene and reopening, the Layer should still have the name "Blend Color"

[vijay]

Hi OTOY/Devs

Can you please check this scene (ORBX) and short video about Denoising issues (BUG ?) i have with 2020 XB2.

I'ts very simple scene (you already have it but for sake of conversation uploading it again).
Basically Glass behind glass confuses denoiser so it completely destroys/removes my glass railing behind my glass doors. Seems like glass behind glass is big problem for denoiser?
Can this be fixed/trained better in denoiser ?

Here is ORBX file and short video upload.

thanks

The way the denoiser currently works is that it denoises diffuse, reflection, refraction..etc individually and composited together including environment to get the denoised output. The main problem here is the affect alpha in specular materials along with keep environment and also transmission channel is not white. When you turn on the affect alpha then the pixels that get refracted are recorded as transparent. Also, we never record colour info in the alpha channel, so once you start changing transmission colour that information is lost and only environment background is blended into the denoiser final result.

Actually the denoiser result seems to be correct, when the affect alpha option is on then we shouldn't be changing transmission color. So currently there are a couple of options here, either leave as it as a limitation or [edit]change the main pass to look same as denoiser output(which is not easy and have to discuss with the team more).