taranvh/taranassets

Various other stuff from Taran

taranassets's Issues

Photo showing violet

Just finished watching your (great!) video on magenta not being pink, but I don't think you'll get an actual photo of a spectrum from a prism that shows violet. You already mentioned there are no violets in this reference image on github because it's in RGB, which can't represent violet, but cameras are the same - they have R, G and B colour filters in front of their sensors, so they can't capture true violet. Look for photographers trying to capture photos of violets or chicory - they can't, and the common suggestion is to adjust the colour in post-production. I suspect that is what happened with the photo you showed of a rainbow on someone's hand - the violet was undoubtedly adjusted after shooting.

Also of interest: I can't find any good graphs of the frequency response of the L cone (the red one). The "good" graph you cited from Wikipedia seems accurate, but the one with two humps also has some truth to it. The L cone does have a hump in the violet part of the spectrum, but it's much smaller than the graphs that show it indicate. Importantly, it's enough that in the violet range, although the red cone only has about 10% of the response of the blue cone, the response of the yellow-green cone is lower than the red cone here, which might explain why our brains want to make violet and purple the same. I'm not a specialist in this field though, so this is just a random layperson's idea!