About area lighting

[Marc Chaumier]

Hi, Roy, I need your advice about the settings of the translucent plastic or glass, to reproduce the "Barrissol effect".

On the pictures, I show you the faces only tagged as area lights, 10 or 20 cm behind the faces tagged as thin plastic or glass objects. The light is given by LEDs, but is not uniform.

Thanks.

[Roy Hirshkowitz]

You'll need to try some things.  I would start by placing area lights behind the panels, and making the panels using glass with roughness.  This should work better than making the panels themselves the light sources-- this light would be uniform in this case.  Try to limit the number of light sources if possible (it may not be.)

Let me know what you get or if you want me to take a look at a model.

[Marc Chaumier]

Thanks Roy, That's what I was trying, but I need to put some "glow" on the glass. I would like that the glass glowes, just with the area lights, but I am at the beginning of the experience. In case, I send you the dwg, if you have some time.

[Marc Chaumier]

Thanks Mario, for your advices, I'll try all these way. I'll post the result.

[Roy Hirshkowitz]

Put 6 smaller light sources behind the screen.  Gave the screen some thickness for better refraction by copying and offseting the surface.  Changed the screen material a little-- IOR 1.6, Roughness 1.0, Transparency .85, -- no glow.  Until you're convinced that you can't adequately model this physically, avoid all tricks such as glow or those that Mario is suggesting.

e1, BTW.  This stuff isn't working well with e4-- I need to make some changes here.

[Jan Verzelen]

roughness is set to maximum i see, is there a way to make the material more diffuse? Maybe changing the distance between light source and screen or is there no place left? Is there any change by tagging the lights as omnidirectional or diffuse?

[Roy Hirshkowitz]

Not really, no-- but you can change the shape and configuration of the light sources behind the screen.  I did these quickly for illustration only.  I also only let this run for a minute or so-- things should smooth out a bit after a while.

[Marc Chaumier]

Thanks a lot Roy. When I offset the surface, is it to simulate a solid, so do I have to tag the both as "thin", or not? The effect that my customer is looking for, is the case of the bigs disks at the ceiling. Something like this picture, but with the better lighting effect that you obtain. 

[Roy Hirshkowitz]

No-- don't tag it as thin.  

I increased the distance between the sources and the screen for this one.  It also uses e4 plus a tiny bit of glow to simulate the sub-surface scattering of the material which we are not going to calculate (it would take forever).  You won't be able to use e4 for yours yet, I had to make some changes here since the glow isn't working.

Put the glow in a separate channel so you can easily adjust it.  Don't get carried away with it, it is "cheating".

[Marc Chaumier]

Thanks Mario, the illumination looks very good. I'm going to take a look on your file. Thanks for the time you spend on it.

[Marc Chaumier]

Thanks Mario. I see, with your help, that omni light are better for this case. And my question to Roy: Why? What is the difference in the functioning betwen area or diffuse, and omni? The "area" picture is not finalized, but its enough to see the difference...

[Roy Hirshkowitz]

The omni light distributes light equally in all directions.

The area light distributes light based on its direction vector (straight down by default-- but an editable setting).  The light is only distributed in one hemisphere and the intensity decreases at glancing angles using the cosine with respect to the direction vector.

[Marc Chaumier]

Thanks a lot Roy. I understand, now, why the light is stronger in the space between two surfaces (close enough) tagged as omni, but not in the case of the area light.