Quote - We have six; a, b, c, d, e, and turbidity. Of these, only turbidity has any sort of intuitive value. The others are used as follows:

sky pixel luminance = [1 + a * exp(b / cos theta)] * [1 + c * exp(d * gamma) + e cos^2 gamma]

where theta is the zenith angle of the sky pixel being rendered and gamma is the angle between the sky pixel and the sun.

Does this help? Not even me. And it doesn't even begin to explain how colors are produced - it only talks about luminance.

The code in the Lux file Sky.cpp has this model applied three times in SkyLight::SkyLight to calculate zenith_x, zenith_y, and zenith_z.
As the turbidity is involved along with a bunch of unexplained constants*, could these be the colour component values ?

• Which match those given in the appendix of this paper -> http://www.cs.utah.edu/~shirley/papers/sunsky/sunsky.pdf