LuxCoreRender Forums

Tried standalone, same result.
Also tried setting the cpu thread count to 0. Again, no change to the shadow, only the reflection by the mirror cube disappears (as I would expect)

This is very strange, not sure what is wrong on my end...

Is there anything else I should try to fix this?

jgrover110 wrote: ↑Sun Apr 12, 2020 9:28 pm Is there anything else I should try to fix this?

Does it happen if you disable cache periodic update ? And does it happen only with this particular scene ?

Dade wrote: ↑Sun Apr 12, 2020 11:40 pm

jgrover110 wrote: ↑Sun Apr 12, 2020 9:28 pm Is there anything else I should try to fix this?

Does it happen if you disable cache periodic update ? And does it happen only with this particular scene ?

Yes this still happens with periodic update disabled, and also with caustic cache disabled entirely.

Not just this particular scene, any scene with HDRi lighting.

Do you have an Intel GPU or are you using Intel OpenCL in general ?

Dade wrote: ↑Mon Apr 13, 2020 3:46 pm Do you have an Intel GPU or are you using Intel OpenCL in general ?

Nope, OpenCL on an RX 570 4gb, not overclocked

Adjusting some of the settings of the scene, I have found significant differences between CPU and OpenCL when setting the importance value of the HDRi to 0, can anyone else confirm this?

Could this be linked to my problem, as they both seem to affect the direct lighting from the HDRi?

jgrover110 wrote: ↑Mon Apr 13, 2020 7:13 pm Adjusting some of the settings of the scene, I have found significant differences between CPU and OpenCL when setting the importance value of the HDRi to 0, can anyone else confirm this?

Between CPU and OCL, when importance = 0 for both of them? No, they look the same here.

The output looks very different the regular scene, but then again it does not surprise me that a setting of 0 on the only light source of the scene results in undefined behaviour...

If we need further tests of this, I would suggest you export the scene from blender into luxcore format and provide us those files. Then we can compare with the same standalone build and at the same time rule out a problem coming from Blender and the addon. Does this sound like a good idea to you?

From my testing, I have the same issue with the standalone build as the blender addon. If you are unable to reproduce it, then this must be an issue with my system. The best I can do it describe the issue I am having and investigate what I believe to be the problem.

From what I can see, the problem seems to be the sampling of the HDRi in direct lighting, as the brightness of the shadows appears to be correct in the reflection of the metal cube. This was confirmed to me when I tried it with importance of 0.

What this did was effectively disabled importance sampling of the HDRi and was sampling it purely at random. In OpenCL the direct lighting was almost completely black apart from the fireflies where it had sampled the sun. The reflections of the floor however, were a lot brighter, meaning that the sky was being sampled properly. In CPU, the floor was the same brightness as the reflections of the floor, which means that the sky was being sampled properly even in direct light.

If my system is the only way of producing this bug, then best I can do is investigate myself and rely my findings in hope that something gets found.