This is a post on my own colorblindness and thoughts about it on my blog, December 1, 2000:
http://www.anigami.com/jimwich/jimwich_archives/jimwich_12_2000.html
Being somewhat red/green colorblind myself (enough to flunk the tests), I’ve always been very interested in color vision and how it works.
When I was in college, taking a color theory course, I had an opportunity to compare what I saw with what others around me with normal color vision saw, using Pantone chips. Two things became apparent over the semester, regarding my color vision and how it compared to others’:
1 - A person’s color vision could be represented by a dimensional plotting. The visible color spectrum can be divided into equally spaced points (colors) for testing an individual’s color vision as chroma intensity and value level is decreased. The results can be plotted as an irregular, dropoff curve, probably varying as the testing moves along the spectrum. As the testing proceeds along the entire visble color spectrum, these individual slope profiles can be stacked to extrude a dimensional hillside that would represent a person’s color vision profile.
2 - I speculate that colors along the portions of the spectrum I have cone deficiencies for, look somewhat similar to me as they do to others, only darker. Stopsigns and Coke cans seem pretty red to me. Grass and John Deere tractors look green, green, green to me. Really bright and vivid. But I think my color vision begins lower and drops off much faster than normal people’s when the chroma (color intensity) drops toward the pastel range, or when light is lowered. For example, I simply can’t see the green in oxidized copper. And low-chroma pastels like light pink and mint green can sometimes look similar if faint enough or in low light.
The component of the red and green spectrums that I “don’t see” yield an overall darker effect to the color. i.e.: that light is actually not being picked up, so the source appears darker. This is why when I was little and was asked what color the purple crayon was, I’d answer, “dark blue.” And likewise brown looked like “dark green.” To put it another way, if there were a blue and purple that people with “normal” color vision would see as having the same value (light to dark), I would see the purple as darker, on account of its red component that I’m not picking up. And I often confuse pure green on a monitor (R0, G255, B0) with bright yellow (R255, G255, B0). It only looks slightly “darker,” sorta mustardy I guess. In fact, I never think of this pure RGB green when I think of green, since I have trouble recognizing it/distinguishing it from yellow. The green has to be enough into the blue spectrum as grass green before I start recognizing it as green.
In my color class we had to each make large posters using a 24 x 24 grid of one inch square pantone chips. I made something using randomly distributed colors, but alternating light, dark, light, dark, etc. in a subtle checkerboard pattern. Towards the middle I used brighter colors and towards the outside I used progressively more pastels. To me it looked clearly like a checkerboard that got fainter towards the outside. To everyone else in the class it looked completely random. They couldn’t see the checkerboard!
It’s surprising, but I’ve been searching for years and have found very little published on color vision beyond the old Ishihara tests that have been around for years. One would think color vision would be graded similarly to, say, nearsightedness or farsightedness. After all, we don’t simply go around saying, “he’s blind, she’s normal, she’s blind, he’s normal” and so on.
I contend that color vision could be mapped for individuals as a 3D “hillside” which represented the color spectrum horizontally and the steepness of dropoff at any point along the hillside denoting the rate of perception dropoff for that portion of the spectrum. It would be plotted using datapoints gathered from a randomly-flashed set of computer graphics. I imagine the test cards or graphics themselves could be configured very much like the Ishihara graphics, only there would be more of them, representing finer differentiations of chroma and value ranging from full intensity and value to none. My graphic above hints at what a resulting plot of test results might look like, though it doesn’t show the variance in chroma and value that I envision also being encoded in the plot. The Y-axis would represent intensity of chroma and the X-axis for each slice would represent the lightness value from full intensity to completely dark. Dips and steeper dropoffs along the Z-axis of the spectrum would show color vision deficiencies.
The test would be structured similarly to a hearing test where various tones and volumes are played. The computer would interpret the results from any one test as a point on a particular color’s sloping curve, ultimately assembling all the datapoints into a hillside graphic. It seems that the resulting map would be quite a bit more informative than the Ishihara numbers-in-the-dots tests as they are used. Case in point - my uncle is also colorblind, but quite a bit moreso than I am, with him having difficulty seeing oranges growing in a tree (I have no problem with leaf green or orange orange). Yet we’re both coarsely defined as “colorblind.” I believe there should be a much more precise measurement and rating of colorvision. That way, I could compare my color vision profile hillside to his and see how they differ, where along the visible color spectrum, and by how much.
