RGB Anomaloscope Color Blindness Test

The anomaloscope is the most accurate tool to classify your color blindness. Since it was developed by a German ophthalmologist just over 100 years ago it is used all over the world to check the severity of ones color vision deficiency and its specific subtype.

The classical anomaloscope is used to identify red-green color blindness. It consists of two different light sources which have to be matched. One source is a yellow lamp which can be adjusted in its brightness. The other one is made by a mixture of a red and a green lamp, whereas the mixture between those two colors can be calibrated.

According to the match the tool can tell you details about the following facts of your color blindness:

  • Red- vs. green-blindness,
  • dichromacy vs. anomalous trichromacy,
  • and the severity of your color vision deficiency,

As this tool is only available as a quite expensive equipment I tried to simulate it online. I knew that there is only a little chance to get the same results as the real anomaloscope. But I tried it anyway and designed the RGB Anomaloscope. As all computer displays are based on the colors red, green, and blue, it will always be a simple approximation of the real test.

The diagram below shows the test results as matching-lines of some red-green colorblind persons. Each point means that this person matched a certain red-green combination to some shade of yellow. Non-colorblind people would only match the point 50/50 with some variances.

RGB Anomaloscope Test Results Diagram
RGB Anomaloscope Test Results Diagram

Unfortunately the results didn’t turn out as I thought. I couldn’t find any split between red- and green-blind persons. Also it isn’t that easy to judge the severity of a red-green color blindness.

Anyway, the test is online and you are welcome to try it out. Maybe there will be some improvements in future releases of the RGB Anomaloscope Color Blindness Test, so stay tuned.

A big thanks to everybody who helped me to develop the test.

14 responses on “RGB Anomaloscope Color Blindness Test

  1. Milo XU

    I have done this before when I went to hosipital to check my color blindness problem. It is something like microscope. The doctor continuely adjusted the 2 colors in it, and asked me if they were matched at the same time.

    it was quite annoying to do this…the doctor was very impatient on doing the test for me. actually few people asked for this test. (I am from China)

  2. Betsy

    I am a red-green colorblind female. Both GF colorblind, mother is carrier. My sister is also colorblind and one brother. Other brother normal color vision.
    I also have simian lines (straight) on both palms of hands (like Tony Blair).

  3. Alano

    Hi there – fascinating test – thanks for working on it. I’ve reviewed your graphical results analysis and would be interested to hear more explanation of your phrase “Each point means that this person matched a certain red-green combination to some shade of yellow. Non-colorblind people would only match the point 50/50 with some variances”. I am a scientist and am not afraid of statistical analysis – but would appreciate help in exploring your interpretation. I am wanting (of course) to explore the interpretation of my own result. Best wishes

    Alano.

  4. Daniel Flück Post author

    Alano, what I mean with this sentence is, that non-colorblind people only match the point with the mixture of 50% green and 50% red to some medium (50%) yellow. Medium yellow in the viewpoint of the test. If you are colorblind you will also be able to match some unequal combinations of red and green to some shade of yellow.

  5. Dabamash

    It said that i was moderate to strong red/green colourblind. But Due to various test, I know that that is not true, Except for short sight, i have perfect vision. Sorry, but you test doesn’t work. try again.

  6. Wolfgang

    I’m normal sighted.
    If I do it without a filter, I get a narrow cluster in the very center with a few points,.
    If I do it with a red filter, I get a line from the center to the lower mid-left.
    If I do it with an orange filter, I get the same result as without filter.
    If I do it with a yellow filter, I get the same result as without filter.
    If I do it with a green filter, I get a horizonal line from the left to the center and a line from the center to the lower mid-right.
    If I do it with a blue filter, I get a line from the center to the lower mid-right and a maybe sparse horizonal line from the left to the center.

    Here one must strongly consider equal brightness too for a match or one
    will get a few matches ! If one has a few matches here, its IMO not severe
    if they cluster in a very small area in the very center of the result plot.

    As this test doesn’t consider changes in the blue color, its not sensitive
    to changes in the blue hue.

  7. Wolfgang

    Are you sure that the split between red- and green-blind persons is not the one in the results that go from the center towards the 50..60 vs. the ones from the center towards the 80..100 at the buttom right ?

  8. Dr. T

    The real anomaloscope should have TWO dials, one for red/green ratio and another for the intensity of the yellow light (see the website I listed). Since the screen has no real yellow light, the RGB-anomaloscope -emulator can not reproduce this 1:1 (as Daniel mentioned in the original description already). But I did some screenshots to have a detailed look at the colors, and there seems to be an additional, unnecessary, restriction:
    ALL colors used are, in RGB color space, mixtures of R and G only. This is the case not only on the mixing side, but on the “yellow” side, too. Indeed, all the cases I matched turned out to have very similar values for R and G (with B=0) on BOTH sides, so the match is hardly surprising. Using some B to adjust the “yellow” side would allow to play with a much larger range of luminescence and saturation, probably fixing the problem.

  9. Daniel Flück Post author

    Dr. T, first I implemented two dials. But it seemed to complicated to me that everybody could do the test easily. So I removed one and inserted the “no match” button.

    The problem really lies in the absence of a yellow color source in RGB. Because of that you can’t really construct metameric lights (yellow based either on yellow or on a mixture of green and red). If you could, a red-blind would see the red-green mixture darker than the “only yellow”. In RGB both is the same.

    Looking at the colors in RGB both sides will always consist only of R and G. If you look at the CIE color space the mixture is like walking a straight line from R to G and the yellow side is like staying on the spot. On both sides yellow would change the hue and/or saturation which isn’t desired.

  10. Dr. T

    Well, it seems I was hasty. I redid the test, doing screenshots of every match and a number of no-matches and played around with it (@Daniel: I can send this to you, if you are interested). My first idea with the blue does not work out at all. But in addition to two matches, were both colors were really near identical, I had to more in which I compensated a small difference in green by a huge difference in red. In numbers: RG 234,234 matched by 31,255 and 208,208 matched by 255,205. I also was able from a close non-match 144,144 with 255,116 to fabricate a match 220,130. But if the green difference is >20, there is no match any more. I know from other tests that I am highly red-weak (protan), but green about normal. This fits nicely with the numbers I get on your test. Maybe the traditional plot is not appropriate for this emulation?

  11. Vinod

    Are you from United states….
    Its a major process for health insurance in US… the medical coding is related to Health insurance cycle useful for quoting numerical codes for surgical procedures.