issue with glass
Posted: Fri Oct 01, 2021 9:28 pm
Hi, i have again a problem with glass.
Unfortunately i cannot provide the luxcore scene. Hopefully someone knows already this kind of problem and can give me a explaination.
i have this scene with a display which has a glass cage on top. The glass cage is modelled with thickness and with no interesections between surfaces.
I told the glass to not cast shadows. I use this workaround in order to deal with bidir's caustics behind glass problem. Normally it works so i was expeccting to see it work even in this scene. The glass is a white glass with an attenuation at two meters.
This is what i get wen i render the display inside his oroginal space. All materials are simple diffuse white (0.7) exept some light emitting surfaces and the glass itself. As you see the glass keeps being extremely dark.
here the glass is turned off
By putting the display in a simplified space the result is correct
In maxwell by the way i get the correct result even in the original scene:
So i started to turn off one by one all lights of the scene.
Te only light which makes the glass not cast shadows are the big area lights i have outside the main opening.
then i turned off everything and added just a few spots with IES data on top of the cage: still the correct result.
I don't know how to get any wiser with these tests. In Maxwell, which has the same issues as Luxcore regarding caustics behind the glass, excluding the glass from casting shadows was the solution that helped get relatively credible results because the glass really stopped casting shadows. In Luxcore it seems that under certain conditions the glass continues to cast shadows.
Is it possible that we are doing something wrong? Has anyone else experienced situations like this?
Thanks for any help.
Unfortunately i cannot provide the luxcore scene. Hopefully someone knows already this kind of problem and can give me a explaination.
i have this scene with a display which has a glass cage on top. The glass cage is modelled with thickness and with no interesections between surfaces.
I told the glass to not cast shadows. I use this workaround in order to deal with bidir's caustics behind glass problem. Normally it works so i was expeccting to see it work even in this scene. The glass is a white glass with an attenuation at two meters.
This is what i get wen i render the display inside his oroginal space. All materials are simple diffuse white (0.7) exept some light emitting surfaces and the glass itself. As you see the glass keeps being extremely dark.
here the glass is turned off
By putting the display in a simplified space the result is correct
In maxwell by the way i get the correct result even in the original scene:
So i started to turn off one by one all lights of the scene.
Te only light which makes the glass not cast shadows are the big area lights i have outside the main opening.
then i turned off everything and added just a few spots with IES data on top of the cage: still the correct result.
I don't know how to get any wiser with these tests. In Maxwell, which has the same issues as Luxcore regarding caustics behind the glass, excluding the glass from casting shadows was the solution that helped get relatively credible results because the glass really stopped casting shadows. In Luxcore it seems that under certain conditions the glass continues to cast shadows.
Is it possible that we are doing something wrong? Has anyone else experienced situations like this?
Thanks for any help.