2009.10.18 RGB color space → IPA vowels
Several years ago, on a run (I can remember exactly where I was), I considered my synesthesia in the context of the phonetic vowel space. I concluded then that mapping the color wheel around the edges of the vowel space, with red at the high front corner (for [i]), yellow at the low front corner (for [a]), green or so at the low back corner (for [ɒ]), and blue at the high back corner (for [u]), with the rest of the saturated colors in between along the edges and decreasing saturation toward the middle, was something I had been doing subconsciously. This corresponded well with my color-letter synesthesia, where "e" is orange, "a" is yellow, "o" is maybe blue, "u" is definitely blue, and "y" is pinkish. It also corresponds well to my association of front vowels with "lightness" or "warmth" and back vowels with "darkness" or "coolness," which seems due to the lower F2 of back vowels, and to my perception that [a] is the "brightest" of all, just as yellow is the most luminous of the saturated colors.
Finally, the red-to-blue segment along the high vowel edge seemed suitable, because just as the purple area of the color wheel (between blue and red) is the only area whose colors cannot be produced by a single wavelength of light — that is, they don't lie on the spectrum of pure colors, i.e. the rainbow — the high vowels between [i] and [u] are the only edge vowels that lie outside of the cardinal "spectrum" of what one could call "pure" vowels, which starts with [i], descends to [a], moves back to [ɒ], and then ascends to [u] — that is, precisely the sequence that maps to the pure color spectrum. The only part of this scheme that is slightly unsatisfying is that the letter "i" is white under my synesthesia, so the red color it's given in the scheme is strange — I almost want to replace it with "m," because that's the only letter that's red for me. However, it works so well with the rest of the mapping that I don't mind it.
The other day, for some reason (I was working on linguistics things already), it occurred to me to try mapping the entire RGB color space to the vowel space, by correlating each color dimension (red, green, and blue) to one of the three main vowel dimensions: height, backness, and roundness. I tried a few combinations before settling on one that ended up corresponding to the old two-dimensional version described above, that is, with the edge vowels mapping to mostly the same colors of the color wheel. This system is red → front, green → low, and blue → round, and it looks like this:
Again, saturation decreases toward the middle, with [ə] being gray. The only slightly off part to me is that [ɶ] is white, because [a] is "lighter" than [ɶ] and because the nagging feeling that [i] ought to be white is still present. Everything else works out very well, though, so I can't complain. In particular, [y] as magenta is quite suitable given my association of pink with "y" (whereas in the 2D scheme, [y] is much closer to red), and [ɯ] as black seems suitable because I perceive it as a quite "dark" and also less "colorful" sound than [u], although strictly speaking I think [u] ought to be a darker color than [ɯ]. As a whole, though, this scheme coordinates strikingly well with my subconscious synesthetic treatment of the vowel space.