Page 3 of 4
Re: issue with glass
Posted: Fri Oct 29, 2021 7:21 am
by daros
Hi Dade, please help me with this. In production this is a huge problem because most shop displays and counters have this kind of design, a glass cage of which we want to appreciate the glass thickness and consistency, and the products behind. At the moment our user cannot deal with this kind of design because behind the glass everything get's dark and makes the glass look like a smoked glass. Thanks for your patience.
Re: issue with glass
Posted: Fri Oct 29, 2021 11:08 am
by Dade
The last scene you posted shows nothing (it is just black). Try to post a scene that show this problem:
daros wrote: ↑Tue Oct 12, 2021 1:07 pm
The problem shows up as soon as the light is bounced by a slightly glossy material.
View 02 600x600.png
The only difference between this scene and the previous ones is that the wall material, which is reverberating the light to our cage, is slightly reflective/glossy.
The luxcore scene of the last picture is exactly the one i sent you today. If you replace the material assigned to the "space" layer with a white matte you get again the correct result.
You have never posted this scene.
Re: issue with glass
Posted: Fri Oct 29, 2021 12:52 pm
by daros
Hi dade, i hope this time we see the scene in the same way.
The only difference between this two renderings is the fact that in the dark one the wall material is a glossy plastic.
Strange is that the glossy plastic version looses a lot of light compared to the matte white version.
The other problem is that in the glossy plastic version the light do not travels through the glass despite the shadow color of the glass is white.
Re: issue with glass
Posted: Sun Oct 31, 2021 10:29 am
by Dade
daros wrote: ↑Fri Oct 29, 2021 12:52 pm
The only difference between this two renderings is the fact that in the dark one the wall material is a glossy plastic.
Emm, no, the walls are "glossytranslucent", it is a transparent material:
You need to switch to a solid material or the scene doesn't make any sense (especially with the outside sky).
I reworked the test scene to make sense to me and I fixed a problem with some type of shadow ray going trough shadow transparent objects in BiDir:
It may, or may be not, related to your problem.
Re: issue with glass
Posted: Fri Nov 05, 2021 10:16 am
by daros
Hi dade, thank you very very much for this fix!
Big difference!
Before:
Now:
Re: issue with glass
Posted: Fri Nov 12, 2021 2:08 pm
by Sharlybg
Look like I have the issue back on last build : obviouslly the shadow under the glass shit is to dark.
Light tracing
ON
Light tracing
OFF
File:
Re: issue with glass
Posted: Sat Nov 13, 2021 3:28 pm
by Dade
Sharlybg wrote: ↑Fri Nov 12, 2021 2:08 pm
Look like I have the issue back on last build : obviouslly the shadow under the glass shit is to dark.
It is the very same definition of SDS paths: you need to enable caustic cache to see the light, trough glass (S), on the floor (D).
Re: issue with glass
Posted: Sat Nov 13, 2021 4:11 pm
by Sharlybg
It is the very same definition of SDS paths: you need to enable caustic cache to see the light, trough glass (S), on the floor (D).
This is even what look strange to me. In fact in cycles you obtain a very different result
And we all know how """good""" this engine play with caustics. And there is nothing special:
simple glass shader. The luxcore shadow look obviously too dark, like light rays aren't able
to pass through the opened side arround the glass.
cycles
updated with cycles shader:
Re: issue with glass
Posted: Sat Nov 13, 2021 6:03 pm
by epilectrolytics
Sharlybg wrote: ↑Fri Nov 12, 2021 2:08 pm
Looks like I have the issue back on last build : obviously the shadow under the glass sheet is too dark.
It does
not work with PathGPU + indirect cache alone.
It
does work with PathGPU + indirect + caustic caches
or
with PathGPU +indirect cache + bright shadow color (fastest way).
It does also work with BiDir Metro or PathCPU Metro without caches (slow).
Curiously it does not work with BiDir Sobol or PathGPU Sobol without caches
Re: issue with glass
Posted: Sat Nov 13, 2021 10:45 pm
by Dade
Sharlybg wrote: ↑Sat Nov 13, 2021 4:11 pm
It is the very same definition of SDS paths: you need to enable caustic cache to see the light, trough glass (S), on the floor (D).
This is even what look strange to me. In fact in cycles you obtain a very different result
It is a fake result you can obtain in LuxCore using shadow transparency (this is what all previous posts in this threads are about).
glass_bug.zip