(Looks like this is about to be outdated but here it is anyway)
Here’s something fun I’ve always wanted to do. And also the self-multiplication for telling apart metamers at least somewhat in action:
When multiplying the spectrum of the sun* with itself**, you can see, that the sun is actually green!
Moreover, you can nicely see that there also is a regime where it looks turquoise.
Note, this isn’t necessarily accurate: Pretty early on, the red portion of the produced colors already takes on negative values, suggesting this is quite a bit out of gamut.
Also, sun light is actually rather white. Surprising, I know.
But what I mean by that is, that it actually takes insane amounts of absorption to achieve the darkness to the right:
What you’re looking at is a 1BU radius rod where the density varies from 0 to 999999 exponentially from left to right. The rod is 10 BU long, and each BU the absorption goes up by a factor of 10 (minus 1 so it starts at 0)
In the center it therefore has a density of 999.
Of course, saying the sun is green in that sense is kind of shady. It actually kinda works with that particular spectrum, but compare, for instance, 6500K in the same vain:
It clearly goes through at least three different colors! It starts cyan, then goes blue and eventually ends up purple. In some sense any of those colors would be reasonable to assign.
By comparison, here is a supposed RGB approximation of 5777K***
This (RGB 1 0.96 .92) goes up to absorption 99999 at the right end. VERY different from what sunlight does. It looks closer to something like a 3750K spectrum (going to absorption 99 because even as a baseline this is far darker than the other ones I did above.)
It looks like that’s too yellow (judging by how long it takes to reach the red region). However, going for something more reddish ends up being way darker than the RGB version.
* approximated as a 5777K blackbody radiator.
** this is, at least after normalizing, equivalent to looking through a material with that absorption spectrum with increasing thickness or, for practical purposes, increasing density.
*** I’m sure there is a better way to do this. What I did was to simply use what this tells me https://www.wolframalpha.com/input/?i=5777K - I am not sure how they decide white point and what not so there’s a lot of caveats here, obviously