I blogged about why red+blue = purple: xanthir.com/b4XK0 This bugged the *crap* out of young me, and no one talks about it!
4
5
3
@tabatkins Doesn't it just come down to that a 3 dimensional space can be spanned by any 3 orthogonal vectors?
1
@davidbaron Your vision is not well-described as a 3-vector space.
1
@tabatkins Why not? Each of our 3 response curves is a vector in the ∞-dimensional vector space of possible response curves.
2
@davidbaron They're not independent, is why. They are ordered linearly, so the appearance of a loop is surprising.
1
@tabatkins Do you mean overlapping ranges (still independent) or that one response curve changes when another responds?
2
@davidbaron When using a single wavelength of light, there are colors you can't produce.
1
@davidbaron Some colors are impossible with any amount of light, but you can trick your eyes into seeing them (super green, for example).
1
@davidbaron The fact that spectral yellow and non-spectral yellow look the same isn't an accident, but purple vs red+blue *is*.
1
Replying to @tabatkins
@tabatkins No, I don't buy that distinction. They're both the result of a 3-D space.

Jul 28, 2014 · 7:05 PM UTC

3
@tabatkins Functions from {1, 2, 3} → ℝ form a 3-D vector space. Functions from ℝ → ℝ form an ∞-D vector space.
1
@davidbaron Yes, you're arguing at a *much* more abstract level. Your statements are non sequiturs in the context I'm talking in.
Replying to @davidbaron
@davidbaron No. No they're not. You're implying that spectral maroon exists.
Replying to @davidbaron
@davidbaron Or you're arguing at the wrong level of discussion, more abstract than the physical basis that I'm talking about.
2
@davidbaron Abstractly, purple is just a blue and red response, with low green response. But that's impossible to do with single light color
2
more replies