VRay Light Disc Behind Glass Pane Renders as Black Disc

[mcquarrie1]

I haven't used VRay in a while, so I'm guessing this will be someone's easy question of the day. I have a basic bathroom scene in progress with VRayLight disc lights used as recessed can lights. Five of the lights are fine, but the sixth disc light (upper left hand corner of the shower area) renders as a black disc. I assume this is because it is behind the pane of glass. The glass pane is just a box primitive with a VRayMtl applied to it. The camera is a standard target camera, not a VRay camera. Thanks in advance for any help.

[yavor.rubenov]

I got the scene from the ticket. Here's the curve i see:



If you zoom out you can see that for values bigger than about 34 the curve goes below zero. The values of pixels of the sphere lights are about 1000 -> so the curve converts those values to values below zero.

[fixeighted]

You have the disc lights set to directional I think, so there's an issue with this, as discussed here https://forums.chaos.com/forum/v-ray...ss#post1100871

[mcquarrie1]

Thank you very much for the help. Yes, you are right, the light was set to directional 0.8. Putting it back to 0.0 "solved" the issue, except I don't get the same light effect I wanted. The link you posted was interesting. I did a forum search on this topic before posting and I did not find anything, so thanks for pointing it out. Some members on that thread theorize this behavior may be physically accurate (possibly due to a combination of light honeycomb and viewing angle it seems). Whether that's true or not, not many users are looking for a black floating disc when they use a disc light. The issue looks like it was placed on a V-Ray to-do list in 2018, but it appears to not have been addressed yet. So I'll humbly re-iterate what others have said on that thread to V-Ray: a toggle to preserve the disc's lightness in reflection/refraction when using the directional function would be great. Thanks!

[cheerioboy]

What I'll do - set the disc lights how I want them visually from a lighting perspective - then make them invisible and only affect diffuse - next, duplicate all of them and give them 0 directionality, which only affects reflection/refraction. I might have to offset them vertically slightly.

Or do the same as above, but instead of duplicating the lights, use geometry of a recessed light with a light material on the lens/bulb.

[^Lele^]

We have a solution to this issue being worked on, sadly without an ETA for the time being.

[mcquarrie1]

Thanks cheerioboy for the workaround suggestion and ^Lele^ for the update.

[OPEN_RANGE]

Thank you all for your helpful replies and info. Now I know it isn't something on my end. For this I will likely just paint in Pshop.

[matanov]

https://forums.chaos.com/forum/v-ray...directionality

[JD3D_CGI]

Did this ever get fixed, as I am having a horrendous issue with it right now. These are just sphere lights.

This is Max 2023 and Vray 6... I'm pretty sure it didn't happen with Vray 5.

[matanov]

Did this ever get fixed, as I am having a horrendous issue with it right now. These are just sphere lights.

This is Max 2023 and Vray 6... I'm pretty sure it didn't happen with Vray 5.

Sphere lights don't have directionality - probably a different issue?

[JD3D_CGI]

Yeah I guess it's a different issue. It didn't happen in Vray 5

This is infuriating. A whole night of renders against a deadline, and they all throw up this disgusting mess. What am I supposed to do?

[matanov]

Without the scene it will be very difficult to guess.

Does the blackness appear in the clear RGB or only in the effectsresult channel (after denoising/lens effects)?

[Joelaff]

Almost looks like some kind of invalid (#inf or #nan) floating point pixel values.

You could try clamping any fog maps if you have them in a transparent material or on the light itself (and use intensity to bring the values back up if needed). Using a gamma curve can sometimes do this if the map is unclamped.

[^Lele^]

The only thing in v-ray that could produce NaNs of this kind would be negative component filters, like Lanczos, in rendering.
It's normally a pixel-sized ring around bright objects, however (it's only the negative component, after all), so the center is highly suspicious to me.
A scene would be great, indeed.

[JD3D_CGI]

I found it what the issue was - somehow - the curves filter in the VFB was smashing out high values to black.