[jfrancis]

I find the methods for combining the layers fairly straightforward in that they are the methods I would have used myself if, working alone, I to come up with a way of combining the separate passes, but I find myself wondering...

Is screening the light contribution (or even simply adding it) really what goes on in the renderer? I haven't tried to reverse engineer it to see, but somehow I imagined something more exotic might be happening.

How much are these blend modes an approximation to what 'really' goes on in an unbiased renderer?

[Brany]

The lights are additive, so the adding of each light channels is the final render. Blending all light channels in ps (32 bit layers) with linear dodge must produce an image like the final render in maxwell.