Skip to content

Was Van Gogh Colorblind?

A friend of mine (and amazing landscape/nature photographer) Jim Goldstein sent me a Tweet and pointed out simmering new conjecture in the art community that Vincent Van Gogh might have been color blind, specifically a protanope.  I seem to remember some discussion of this years ago, particularly given that one can rather nicely simulate both protanopia and deuteranopia in Adobe Photoshop with built in filters (View>Proof Setup>Protanopia/Deuteranopia).  However, the current speculation comes about from Kazunori Asada who wrote up a Tumbler blog entry here describing how he came about his idea and a subsequent app he wrote to simulate color vision and color blindness.  The image above shows Kazunori Asada’s approximation of Van Gogh’s Starry Night and Flowering Garden with Path with the original on the left and the Asada transform approximation on the right.

When Van Gogh’s works are viewed with Asada’s colorspace transformation, the works gain additional depth and clarity.  Contrast is increased and many of the works appear less surreal and more “normalized” or “realistic” in their use of color.  Its an interesting speculation, but it should be noted that Mr. Asada’s software does not *precisely* match the color spectrum of those that might have protanopia.  From my simulations in Photoshop, Mr. Asada’s color mappings are not accurate and they seem to be more “gentle” approximations of what a true protanope would see.


My approximation of what a true protanope would actually see Starry Night and Flowering Garden with Path is immediately above.  Again, the original is on the left and the transformed approximation is on the right.  It seemed to me that there is too much red in Asada’s transform, and what you see on the right in my approximation should be a more accurate representation of what a protanope sees.


So, was Van Gogh colorblind?  hard to say for sure, but if he was a protanope, I don’t see how he could have selected the red and green flowers in Flowering Garden with Path.  In the inset image above from the flowers (tulips?) in Flowering Garden with Path, the original zoomed in section is on the left and my approximation of how a protanope would see it is on the right.  If Van Gogh was a protanope, he would have a really difficult time seeing and determining which colors were which and the painting in this region might be very different.  Therefore, my estimation would be that Van Gogh was not in fact colorblind.

However, given that my expertise is not on color processing, I’ll leave that exploration up to the true experts.  Though I’d love to see Robert Marc and Michael Marmor weigh in on this.  Robert has a talk he gives on color that is truly stunning.  The talk brings you from the physics and psychophysics of vision through the neurobiology of color perception and is on a level that even the layman can understand.  It would be ideal for a TED talk actually…

As an aside, I worked with a software developer Bradley Grimm a couple of years ago to create an app to do real time color blindness mapping.  Brad ended up coding it for the Android rather than the iOS I had hoped for, but it was a fun concept to play with.  That said, I have some interesting speculations for a more sophisticated app, so if anyone in the iOS developer community is interested, let me know and we can talk about ideas.

Categories: Interesting, Interesting Links.

Tags: , , ,

Comment Feed

24 Responses

  1. The original and “normalized” images look the same to me, and I am a proud protanope. :)

    • Precisely. The opsins in the cones should not have any ability to discriminate the green from the red which is my point with respect to Asada’s interpretation. Asada has more red in his filter and the reality is that a protanope would not be able to see that.

  2. Watch his TED talk which will gove more idea what color deficient is;search%3AAsada

  3. In that time often false colors where used by painting artist most of them refused using black, or to darken with black and took contrast colors instead, which seams strange at first, but a true artist knows how to fool the eyes, and make often use, of how we tend see something. Colors or shapes, they work with in new ways in contrast to the daily world around us.

    • Peter, this is indeed true and in fact… one of the ways that the retina actually processes information with edge detection and center surrounds to enhance detail. The funny thing is that some artists understood more than they knew, just how to “see”. Its pretty cool to reverse engineer how various artists worked.

  4. Van Gogh at least read the labels on the pigments he was using: “red”… “green”. To achieve whatever look he was going for, he knew he was using a few different saturated colors rather then one dull color. He’d go to great lengths to combine those colors to give an some impression. Why would he go to all this trouble, with multiple and sometimes very expensive bright pigments, if he was color blind? It makes no sense.

  5. I’ve read in a few Van Gogh biographies that foxglove plant, which was used as a ‘treatment’ for epilepsy and an antidepressant. One of the well-known side effects of digoxin toxicity is color distortion and light halos.

    Adam OrbitAugust 30, 2012 @ 5:12 pmReply
  6. A major problem with the suggestion of color-blindness is that it ignores van Gogh’s incredibly specific written description of his paintings.

  7. In regard to Van Gogh’s color vision, I have this to say. If he were a dichromat with only two cone types, he would only distinguish yellow and blue, white and black and brown and gray. Reds and greens could not be distinguished as colors. If he were intent on capturing reality, and this is not certain, he would use his palette of perceived colors accordingly but he would make occasional glaring errors like painting red objects, like a rose, as brown or yellow or even black. He would similarly fail with greens. Since I have not seen such errors I suspect he had trichromatic color vision. If he were an anomalous trichromat, it would be much harder to detect. Peter Gouras

    Helga KolbSeptember 2, 2012 @ 10:33 amReply
  8. Is there a way to learn more about Robert Marc’s lecture on color?

    Brayden LundquistSeptember 17, 2012 @ 8:42 amReply
    • Hey Brayden,

      Thanks for the link. There is currently no online resource for Dr. Marc’s color talk, but the next time he gives it, I’ll try and record it and post it here. Its a wonderful talk. Perhaps I should propose it for a TED talk?….

  9. Bryan, I’ve been an appreciator of webvision for many years.
    One comment on this: Asada’s blog indicates his filter was geared toward protanomal simulation rather than protanopia simulation…

    • Fair enough. This of course was Dr. Kolb’s point in describing an anomalous trichromat.

      By the way, thanks for coming back to Webvision over the years. Your presence is much appreciated.

  10. Am both a protanope and a painter. All throughout my career, I have been complimented on my colour sense. There is absolutely no doubt in my mind that Van Gogh is a protanope, and assume Matisse could be as well. I use red in my paintings quite a bit– but much like Van Gogh here. Some brush strokes on top, some dots, or some area. I have one early experimentation in impressionism, where I painted red hued dots first (could see them) and then painted the others (making red disappear).

    What I would not do typically is have a very large area of bright red with hue changes or have a mossy forest scene with brownish treetrunks and greenish moss. And am naturally attracted to blue/orange, blue/yellow combinations like the starry sky— or van gogh´s sunflowers in a vase. Another very typically protanope painting.

    • “Absolutely no doubt in my mind.”

      You’re claiming 95% confidence on diagnosing someone with a condition affecting 1% of their demographic by their paintings?

      Can you really be so sure? Can you really say these are in the 99.95th percentile of most protanopic- looking paintings? Are they 3.3 standard deviations more protanopic than the mean? How do you know they’re not deuteronopic?

  11. I’m a strong protanope, and I can very clearly tell the difference between the bottom two pictures.

  12. To me all shown versions of the paintings look exactly the same. I am protanope, but i don’t think one can state anything about van gogh being color blind. Color blind people like me will make many obvious and strange mistakes with colors that will be immediately apparent to people with normal color vision. That would be the thing to look for.

  13. Let’s also keep in mind that just because colorblind people see these images the same as each other does not mean that they see them with the same palette as color-sighted people (or indeed that color- sighted people see it the same as each other). The colorblind vision filters filter camera RGB data, not interpreted color or even light frequency distributions. colorblind people are often much more sensitive to certain types of subtle differences in images simulating colorblind vision than color-sighted people are, because their internal mapping of color has a different white balance and normalization behavior and they are much better at telling red from green than color-sighted people are when given colorblind images.

Some HTML is OK

or, reply to this post via trackback.

Continuing the Discussion