LuxCoreRender Forums

- Upgrade the Luxcore fbm texture node to match the current cycles noise node with three color channels, lacunarity, distortion etc.

For me the main obstacle has always been the slow viewport when pre-rendering. The inability to move objects when pre-rendering is in progress makes it inconvenient to customize the scene.

I also really miss Bevel-shader and Cryptomate.

Dez! wrote: Sat Apr 05, 2025 7:01 am The inability to move objects when pre-rendering is in progress makes it inconvenient to customize the scene.

I want to ask for clarification, what exactly do you mean by "pre-rendering"? The phase when it says "starting viewport render"? Maybe also the CUDA/OCL kernel compilation step?

I also suspect this is somehow related to large scenes? At least for setups like the Canelle et Fromage or Hall_Bench test scenes, I don't have issues to move objects once the render started.
(And I personally don't work so much with complex scenes.)

Replace Embree3 by Embree4

Perhaps this would also give motivation to look into deformation motionblur.

But if we really are thinking into long term development of LuxCore, I think one fundamental thing need attention.
Supported geometry formats. In reality I think we should look into native USD support.
MaterialX is also popular, but honestly I dont think that is such important one. But proper goemetry support is crucial, much more so than rendering speed.
Ply just is not cut out for modern pipeline, skeletons, instances, particles, time samples, custom attributes.
Blender is also working on USD support, not just import export, but actual USD support. This will be really important and could save a lot of conversion required.

Perhaps some smart people can share their knowledge, but I feel like this could be pretty important looking forward.

Dez! wrote: Sat Apr 05, 2025 7:01 am I also really miss Bevel-shader

Yes, seconded.
There is already a GitHub branch and also some discussion.
Another feature that was included (maybe not finished) but never exposed.

Hello,

I wonder whether it might be possible to turn luxcore into something that can be used for engineering purposes. I mean usage of lighting units, spectral rendering and materials with more physics based definitions (more complex scattering, absorption, etc. on surfaces and volumes)

Target could be to become an alternative to ANSYS Speos, that is a quite an industry standard for technical, lit up visualisations.

What do you think?

Regards
Martin

lighting_freak wrote: Fri Apr 18, 2025 1:00 pm I wonder whether it might be possible to turn luxcore into something that can be used for engineering purposes.

Generally speaking, I also see engineering applications, which I am planning to work towards, however I would not use the phrase "turn luxcore into" because I don't see the need to sacrifice any general artistic applications for it.

lighting_freak wrote: Fri Apr 18, 2025 1:00 pm I mean usage of lighting units, spectral rendering and materials with more physics based definitions (more complex scattering, absorption, etc. on surfaces and volumes)

Note that LuxCore already offers an irradiance AOV. Might be useful for illumination work already.

Spectral rendering is already on this list.

About materials with physics based definitions, this would require to explicitly name algorithms, techniques etc. for implementation.
I do have tabulated reflectance data, or likewise AR coatings for glass in mind already.

Hi,

I totally agree with you. Maybe the "turn into" has to be rephrased to "add to"...
I'm already aware of the irradiance AOV.

I'm also thinking of some tabulated data files to define the individual properties.
They should work in variable depth of detail.
Maybe we could talk in detail about the solutions that are currently available for lighting simulation.

I think with some work I could generate some initial library of those "optical" properties and support the project.

BR
Martin

- Expose russian roulette parameter (minimum light bounce) like in Cycles and LuxCoreUI.

(previously posted those on Discord, adding them here as well)

- adaptive subdivision like for Cycles in Blender (pixel-based micro-polygons). See the subdivision modifier in Blender.

- Elliptical filtering to solve high-frequency texture (bitmaps and procedural) banding and moire artifacts. This would be very useful since we can't use pixel filtering with a denoiser.

Thanks for all the great work!
mgfx