Which feature do you need the most?

[pegot]

The feature I need most is an updated manual

Especially for Octane Blender.

[adrianr]

Displacements are by far Octanes most obvious weakness. The implementation of them is terrible, and lag behind even ancient renderers like Modos. Like already stated above, vertex displacements just aren't an option for 99% of people interested in speed and efficiency.

I remember the most I realised how bad they were. I needed to add two displacements together to create slightly embossed text on a logo, and when using vertex, it added 3 million triangles to a scene that was only 820k to begin with.

Compare that to Modo where you just drop a displacement in the shader tree and you can stack them all day.

I actually don't understand why displacements are so bad in this renderer.

I want to +1 this more times than would be sensible. Displacement in Octane is such a miserable experience. Are there any hints on a roadmap that this is due an overhaul?

[fbf]

Something that has been #1 on my wish-list for a long time... is an alpha channel / matte for the reflection pass when using 'render layer'.

Essentially (particularly on light-colored backgrounds) setting the reflection to 'add blend mode' is not enough. Part of the compositing is obscuring other reflections from behind the object as well -- where the 'alpha channel' comes in, or even the AOV (including reflection)

Render Layer AOV Alpha Comparison.jpg

alpha reflec example.jpg

Thank you! I am hoping this shows how setting the reflection pass to add mode is not enough -- and why an alpha channel for the reflection would be INCREDIBLY helpful, or even the ability to render the reflection matte for that object (without multiple render settings/exports) Although it is currently technically possible to get the passes necessary... 'render layer' removes the reflection matte that is necessary from the AOV pass, making this a two-step render process

I appreciate all of your teams' work on improving Octane

Yes! This. Render Layer Reflection Alpha would be incredibly useful for VFX compositing. Number one feature request for me.

[p3taotoy]

PLEASE if its possible :>))
Can you have a center rotation on any mesh brought in. (Even if the mesh was off somewhere in XYZ in the modeling software)

When you bring in a mesh that was not at the origin in the original modeling software the axis of rotation in the standalone is way off!

It would be a life saver :>)))

Thanks

Oh yeah a shortcut for placement and group node :>)))

[wangdanyang1120]

Feature request: Add a node to the Output AOV Compositor that works like the 'Highlight Compression' in the OctaneCameraTag.

[e-s]

PRG Clear Sky Daylight model would be so nice. That's what Vray and Corona use, and it looks great

[e-s]

It's seemingly not much different from Nishita / Planetary. Any paper available? Documentations from these mentioned renderers are not informative.

Admittedly, I don't understand all of the technical nitty-gritty, but here is a link to the paper and presentation from SIGGRAPH:
https://cgg.mff.cuni.cz/publications/skymodel-2021/

[e-s]

I had posted this request in the C4D section but was informed that it's a limitation of OctaneCore, so I'm reposting here.

It would be great to have a streamlined workflow / UI for light linking. I think Redshift is a good reference here, where each light has a tab that can exclude/include specific objects - it's fast and straightforward to use with a couple of clicks, and easy to understand which lights are affecting each object.

[noldo]

It would be very interesting to implement a function that generates a Gaussian splat cloud from a 3D model.
The advantage would be real time navigation including reflections and transparencies, which, especially in VR environments, would be great.
I think that Gaussian splat works by adding color information to each individual point of the cloud in relation to the viewing angle, thus allowing a point to take on multiple aspects depending on how you look at it.
Obviously this information, which current software generates from photos or videos captured from different angles, could be generated by the rendering engine.
In essence, the rendering engine would have to divide the model into many small equidistant points lying on the faces and calculate for each of them the information coming from the various directions.
This would lead to a Gaussian splat environment that is enormously more precise than anything that can be generated with photos and above all observable from any point of view.

[leave2]

Displacements are by far Octanes most obvious weakness. The implementation of them is terrible, and lag behind even ancient renderers like Modos. Like already stated above, vertex displacements just aren't an option for 99% of people interested in speed and efficiency.

I remember the most I realised how bad they were. I needed to add two displacements together to create slightly embossed text on a logo, and when using vertex, it added 3 million triangles to a scene that was only 820k to begin with.

Compare that to Modo where you just drop a displacement in the shader tree and you can stack them all day.

I actually don't understand why displacements are so bad in this renderer.

I want to +1 this more times than would be sensible. Displacement in Octane is such a miserable experience. Are there any hints on a roadmap that this is due an overhaul?

And one year later, displacements are still miserable