[jfrancis]

I guess one of the changes for 2.0 was bright, colored emitters hold their color when they are made very bright, instead of going white. I remember a comment to the effect of 'great for neon tubes'



Do we actually want that behavior, though? It seems to me that unless an emitter is a very narrow spike of color it should change hue/saturation when it brightens.

[tom]

It's only valid for pure RGB entries such as 255,0,0. In real, such purity is not possible. It's a violet/pink which has all color guns greater than zero so the exposure turns it to white. If you take the photo with higher shutter you will notice the core color will be back. Here's a quick example:

[jfrancis]

I'm glad it works that way.

I could have sworn when I tested it I wasn't getting the white overexposure when I played with exposure and multilight emitter brightness, but maybe I was mistaken.

[tom]

As I've said, if you set pure colors it won't work like in the example above. A red made of 255,0,0 will remain red forever. You should set it something like 255,50,50 to expect it to go white. You know, multiplication (exposure) with 0 returns 0, so...

[jfrancis]

As I've said, if you set pure colors it won't work like in the example above. A red made of 255,0,0 will remain red forever. You should set it something like 255,50,50 to expect it to go white. You know, multiplication (exposure) with 0 returns 0, so...

You know what it was? I was playing with candle-like emitters and it seemed under 2000° K the emitter went from dull orange to bright yellow but never went white. I guess there is no blue in the mix at that color temperature?

[tom]

You know what it was? I was playing with candle-like emitters and it seemed under 2000° K the emitter went from dull orange to bright yellow but never went white. I guess there is no blue in the mix at that color temperature?

Yes, it looks like that's the reason. But it could be a precision problem, we will check.

[jfrancis]

Interesting.

I did a little reading. If I understand the physics correctly, it's about black body radiation curves and under 2000 K vastly most of the radiation is infrared and virtually none is blue in theory, but in practice the emissivity of materials is bluer than one would expect, and white is achievable.

[tom]

Exactly. That's why I suspect it could be a precision issue. Because, I'm so sure it should normally go white.