LuxCoreRender Forums

Comparison of render with caustic cache (first turn) versus light tracing with SDS-cache (second turn). Caustics from cache seem too bright and there are fireflies dancing in the shadows that OIDN cannot catch, also render times are way longer. Caustics from Hybrid engine seem more natural but brig...

It is not yet implemented.

Thanks @FarbigeWelt for testing and confirming!
Good to know I'm not the only one experiencing this (and it is not an nVidia issue).
Also the caustic brightness variations are visible in your table.

But when Dade cannot reproduce this on his system he cannot fix it.

Works great with textures and bump, but I can't get normal maps to work (wrong intput warning and no visible effect).

I did various test animations last year with v2.1 and 2.2 and the combination of persistent PGI cache and OIDN (and a RTX) works just great. Problems arise with ELVC/DLSC which lack persistence implementation and caustics, because light tracing has a brightness issue and caustic cache produces firef...

Yes, great composition, simple but effective.
Mushrooms need some SSS though.

FarbigeWelt wrote: ↑

Wed Jan 01, 2020 3:15 pm

Next, a rather different View of the very Source
Cube of Dispersions 3.png

I'm unable to figure out what your set-up is but this high contrast image is some great piece of art!

2500 samples PathCPU Metro + Light tracing, 70% OIDN

Does it happen if you use a single GPU ? And CPU only ? Today I tested it on my caustic cube scene (render 50 frames of animation, Blender 2.81a + Luxcore build from around Christmas) with these results: It is a Windows PathOCL Hybrid + SDS-cache problem. - it does not happen on my Mac - it does no...

Hi and welcome to the forum,

the horizontal noise patterns are an artifact of OpenCL Sobol sampling on the GPU.
OIDN is not trained for this and then gets creative with it
To avoid that just use the random sampler (Render properties/light paths/advanced/Sampler).