New GGX light distribution for VRayLights with directionality

[matanov]

We all know the issue with VRayLights and directionality seen through reflection/refraction. Now we have something ready as a fix, but it comes at a price - the light beam looks different.

At the example below - plane and two disc lights with directionality set to 0.94 / 0.85 / 0.6 respectively. Left image is the old method, right one the new GGX light distribution.



I can provide a build for those who want to test it locally (V-Ray 6 license required).

Eager to hear your thoughts!

[Crayox13]

Cool if it fixes that bug. I don't mind the different shape a bit.

[flat]

Is the previous light distribution physically wrong? (Is the new GGX light distribution more physically accurate?)

Compatibility should be fine if there is a switch to revert to the previous light distribution.

[Vizioen]

Same question as flat, which is the more physical correct distribution? For me I don't mind though as long as the bug is fixed

[vlado]

Both distributions are correct (they are similar to adding an IES profile to a light), however with the GGX-like distribution, the light is not so dark when viewed from the side.

Best regards,
Vlado

[Joelaff]

Can we still get full directionality with the new method? In other words, what about a value of 1.0 ? It is already difficult to control when you want a very directional light (by difficult to control I mean we have to use many decimal places to make small tweaks once we get close to 1.0. Rarely do we ever want a true 1.0, but we often want the light o look a heck of a lot more like a spotlight, and tweaking fine values between 0.98 and 1.0 is touchy.

[GLASS-CANVAS]

The new GGX lights look great to me, perfect for washing up the sides of buildings at night etc. I presume there will be a "GGXness" spinner to set it as required? Or is it just an on/off tickbox? I'm hoping for the former!

[psanitra]

matanov I wonder how this new behavior affects the direct visibility of the lights source, if at all. Talking about the usual problem of area light being shown as black square/circle in render, if the directionality is too high and camera is outside of that solid angle. It`s annoying to deal with, adding another piece of geo under with light material and all that. So i wonder if with this update we still going to see black area. I understand the reality vs the cg approach difference of the phenomena, hence why it`s black. I think artistically, people expect to see source as something bright, even at low glazing angles. Thoughts?

[matanov]

We are still experimenting, but yes - in this implementation the camera angle is still affecting the looks of the source. GLASS-CANVAS - yes there is a GGX spinner in this variant.

[Crayox13]

So this is still not in Vray6 hotfix3? I just installed it and my LEDs are still black behind the glass.

[matanov]

So this is still not in Vray6 hotfix3? I just installed it and my LEDs are still black behind the glass.

No, it's planned for Update 1 (V-Ray 6.1).

[Crayox13]

Oh I see. Thx

[themaxxer]

it's not in update 1

[vlado]

No, not yet as we changed the original approach to something that seems to work better and there wasn't time to fully implement it for Update 1.

Best regards,
Vlado

[jackieteh]

Hi, any update on this? is it implemented in VRay 6?