D65 Spectral Power Distribution & Replication

[thebeals]

I am digging into light theory a bit and how it correlates to material theory and am hoping some of you bright folks can give clarity (pun intended).

Two real world lights that have the same visual color could have a different spectral power distribution. This could make one material react to the two lights differently and thus appear to have slightly different looks under each light. If I have a d65 light in the real world, and I want to match that in Vray with the correct power distribution so materials react correctly, how would I do this?

Since Vray is not a spectral render engine, is this even possible? If it is not possible, is the closest we can get simply visually matching the color temperature/color and is this "close enough?"

My first theory was maybe apply individual RGB curves, similar to how one would create a manual measured metal reflection, to the light color, but I have not tested this theory and don't even know if it would have the same desired outcome.


Thanks for any help!

[Mihail.Djurev]

You have three factors affecting the apparent color. Light spectrum, material reflectance spectrum and the human eye R/G/B response curves.
You can get the spectrum of the reflected light by multiplying the light and reflectance spectra.
Then, if you take the area (integral) of the eye's R response curve multiplied by the reflected spectrum, you'll get the amount of "red". It's the same fo green and blue.

Incandescent light has a very broad spectrum, which will look consistent on all materials. However CCFLs and LEDs may have peaks and valleys in the spectrum, which may coincide or not with the valleys and peaks with the material reflectance spectrum.

V-Ray has only three spectral points, R,G and B. In theory, you could do multiple renders, as if there are more, but it would be very difficult to adjust the colors and combine the result.

You would be better off using light linking and having a light with different color shining only on the objects with an irregular reflectance spectrum. Or, you could use color corrections in post

[thebeals]

That is incredibly fascinating information, thank you so much! In regards to materials with irregular reflectance spectrum, is this fairly common? Do all materials display a little bit of it and some a lot, or just a few display it? I've been playing with LEDs and have noticed the irregular peaks and valleys using a spectrophotometer. I think a standard bulb for the print industry is Solux so I was going to try them next..

If most of the product I render is in a controlled studio type of environment, I would imagine the look of materials would stay fairly consistent and worrying about the spectral response of lights is over the top and will eventually have very little consequence in the resulting image... What are your thoughts? My end goal is to create a material and have it respond correctly in any lighting situation without any tweaks (assuming of course the material and lighting is correct in the first place).


Thanks again for your time!

[Mihail.Djurev]

I wouldn't say it's common, but I happen to have one of these materials at home.
I have two light bulbs in the same fixture, one (cheap) fluorescent and one incandescent. My window blinds appear green with the first and yellow with the second, even though both bulbs have roughly the same luminance and color temperature.

Anyway, I guess this shouldn't matter for studio lighting. It's supposed to be high quality after all.