How a made-up color blindness infiltrated the internet

Footnotes

FOOTNOTE 1: ANOMYLIZE

While this is not a real word, it is supposed to be derived from “-anomaly”. The proper spelling should therefore be ANOMALIZE, not ANOMYLIZE. Yes, it’s a petty thing to complain about, but it’s also indicative of the lazy naming that was happening in these simulations. It’s like they are trying to see how many made up words they can screw up in one piece of code!

FOOTNOTE 2: CONDITIONAL MULTICHROMACY

Actually, all humans (and all vertebrates) with at least one active cone experience conditional multichromacy in mesopic conditions when their scotopic (rod) and photopic (cone) vision overlap. Color normal humans would therefore experience conditional tetrachromacy.

I was using cone monochromat as synonymous with blue cone monochromat. While red cone and green cone monochromats are possible, and even probable based on an intersection of tritanopia and protanopia (for example), I have never seen one described in the literature past a hypothetical. Regardless, in the case of a red-cone monochromat, this section on blue cone monochromacy would still apply, but many of the details would require revision, such as a very different luminous function.

The actual luminous function of the mesopic vision of a blue cone monochromat would be dependent on the light level, and to what degree the photopic or scotopic systems are dominant. This effect is explained indirectly by the Purkinje Effect. An arbitrary light level was chosen towards the scotopic side in this image so the rods are contributing slightly more to luminosity than the S-cones.

FOOTNOTE 3: INCOMPLETE ACHROMATOPSIA

One possibility I edited out of the video was that the Achromatomaly simulation could represent Typical Incomplete Achromatopsia (not Atypical). Complete Achromatopsia is generally synonymous with rod monochromacy, i.e. there are no functioning cones. There are several genetic causes for this, namely mutations on genes that are used by cones but not by rods.

However, some of these mutations do not completely disable the cones, rather weaken or slow their response. This may sound like a good fit for “achromatomaly”, but the effect of these partially disabled cones is incredibly complicated and actually a huge blindspot for researchers, pun intended. In fact, achromatopsia.info, an advocacy group for achromats, provides the below manually simulated image, and describes incomplete achromats as seeing in mostly black and white, with “a bit of red”, which looks  distinctly unlike the desaturated achromatomaly simulation.

The original image (center) is simulated as achromatomaly (right) and as a representative incomplete achromat may see it (left) as grays interspersed with some reds. Note, this is not how all incomplete achromats would see it.

I think this is a bit reductive – and probably false – probably indicating an individual that still has some response from his L cones (which would not look like the black/white/red “simulation”). If all cones were affected similarly, maybe the achromatomaly simulation would be somewhat accurate. However, most sources that I’ve read have indicated that incomplete achromats usually do not have all of their cones affected to the same degree, leading to wildly different manifestations of the condition in a subject’s color vision. This makes it kind of pointless to simulate.

FOOTNOTE 4: MIXED ANOMALOUS TRICHROMACY

A mixed anomalous trichromacy would be the intersection of protanomaly or deuteranomaly with tritanomaly. Together, the red-green colorblindnesses have a prevalence of 1% and 5% of males, respectively, which comes out to ~3% of the population. Numbers for tritanomaly are very wide ranging, but much rarer at 1 in 15’000, though this number represents only congenital forms of the disease. Mild acquired forms may be much more common.
So that would be 1 in 500’000. However, if we also consider that the achromatomaly simulation requires somewhat similar strengths of each of the red-green and blue-yellow components, this number gets higher. For example, a subject with strong deuteranomaly and mild tritanomaly would still express as a deutan. Assuming an even distribution of mild, moderate and severe cases for each type, then one third of subjects with mixed anomalous trichromacy would have mild-mild, moderate-moderate or severe-severe and would be well represented by the achromatomaly simulation. That is why I am using the 1 in 1’500’000 number, though this ignores acquired tritanomaly.

Transcript

So there’s this type of color blindness called Achromatomaly. If you google it, you’ll get close to two million hits but… it doesn’t actually exist. Try to find a description of it and the most authoritative source is from a dude on reddit. So why does it appear everywhere? Today on Chromophobe: the fake color blindness that just won’t die.

This story begins – as most questions in my life – on reddit when someone posted this image containing crewmates from the computer game Among Us simulated through different versions of colorblindness. Now Among Us is a terrible game for the colorblind with laughable colorblind accommodation that doesn’t even come close to addressing the actual difficulties that colorblind people have… but what struck me by this image was the simulation in the center called Achromatomaly.

While it sounded similar to Achromatopsia (a type of total color blindness), Achromatomaly specifically was not a colorblindness I’d heard about so I donned my internet armor and entered the reddit comments section. It turned out I wasn’t the only doubter raising an exception, but a group of posters confidently explained – in a I’m-right-because-my-post-is-longer-than-yours kind of fashion – that it was a blue/yellow-pink/teal color blindness. Now naturally I requested a source to which one offered:

My information about it basically comes from years of experience and talking to people on reddit.

Sure, we all know that deep down the typical source for redditors is generally other redditors but most people aren’t so bold as to claim that as a sufficient source. The poster continued:

Very little was actually written on google about Achromatomaly or maybe even nothing at all. That’s how rare it is.

…and they were kind of right here, because while giving 2 million google results, Achromatomaly only had 9 academic results in google scholar. For reference, the term Achromatopsia has more than 1000x the academic references as Achromatomaly and Achromatopsia only affects 1 in 30 000 individuals.

Wait a second… you guys want to do some math? I took a list of genetic conditions and I plotted for each of them the number of google scholar hits against the conditions prevalence (about how common it is) and this provided me a waaay better power law fit than I was expecting. Now, if we take Achromatomaly’s nine google scholar hits and place it on the line of best fit, the prevalence of Achromatomaly comes out to about one in 12 billion, which is less than one person alive today… flawless logic!

…and so in very much the same way that CGP Grey spent hundreds of hours trying to find the origin of the name Tiffany, I got a little obsessed with the origin of the term Achromatomaly.

I started with the source and asked the creator of the Among Us image what simulator he used, pointing me to the COBLIS colorblind simulator: a popular applet on COLBLINDOR, which simulates for color normals how the colorblind would view a certain image. But COBLIS never makes any mention of Achromatomaly, so I had to dig deeper. When COBLIS was first released in October 2008, it was based on ColorJack’s ColorMatrix software. There is no mention of Achromatomaly in the COBLIS release notes, but in an earlier March 2007 post reviewing another ColorJack app (Sphere), the maker of COBLIS weighs in on ColorJack’s general color blindness naming convention:

The wordings Achromotopy and Achromatomaly used in the tool of ColorJack are also wrong. The correct names are either rod monochromacy or achromatopsia for complete colorblindness… and on the other side blue cone monochromacy.

So the path to Achromatomaly pointed squarely to colorjack.com: a now defunct website but still accessible in part using the Wayback Machine on archive.org: an absolutely indispensable tool during this search. Cruising through the archive, there were several mentions of Achromatomaly on colorjack.com, so I definitely picked up the scent. Locating their ColorMatrix code and looking at their attributions, we can see that they based their simulator on one by Matthew Wickline. The Wickline simulator is actually still online, but in place of an Achromatomaly simulation, it includes a simulation for Atypical Achromatopsia. Compared to Achromatomaly, that term gets far fewer hits on Google, but because it’s an actually real condition, far more on Google Scholar. This disparity is because Atypical Achromatopsia is an outdated term for Blue Cone Monochromacy.

So it seems ColorJack took Wickline’s atypical achromatopsia simulation and just changed its name to Achromatomaly, suggesting it may indeed be simulating blue cone monochromacy as COLBLINDOR suggested. Going even deeper into Wickline’s sources, we find a 1999 article by Thomas Wolfmaier at HCIRN, who wrote the original code based on… finally.. an academic paper: a 1988 paper by Meyer and Greenberg.

However, neither the original code nor the paper mentioned any simulation of achromatopsia (much less Achromatomaly). Both of them focused on simulating dichromacy, which is actually much more difficult to model compared to achromatopsia, which is essentially monochromacy and can be approximated by a standard black and white filter. Even more difficult to model is anomalous trichromacy like protanomaly. So what they did – and what pretty much every modern simulator still does – is simulate anomalous trichromacy as a weighted average between a dichromacy simulation like protanopia and the original image.

Now we can deduce what Wickline did:

  1. He took the dichromacy and anomalous trichromacy code from Wolfmaier;
  2. He made his own achromatopsia simulator which is just a black and white converter;
  3. He saw the weighted average calculation for anomalous trichromacy and made an “anomylizing” function;
  4. He read somewhere that cone monochromacy is less severe than achromatopsia which… it is… in some ways…
  5. …and finally he applied the function to achromatopsia to create this weaker “anomylized” atypical achromatopsia.

So wickline is the source of the certainly questionable Achromatomaly simulation, but what about the term? That didn’t come from Wickline. Did ColorJack just decide that atypical achromatopsia was a mouthful and change it to Achromatomaly?

I did some more digging into the name. Based on the COLBLINDOR post that criticized ColorJack’s use of the term Achromatomaly, I had a latest possible date of genesis: March 2007, which should at least make Googling a lot easier. And maybe it should be no surprise that neither Google Scholar for Achromatomaly pre-2007 nor the standard Google results showed a single result… which strongly indicates that ColorJack just made the term up… which is not a satisfying ending! Like… what the hell was he thinking?

I had one more path to go down. It was COBLIS V1 that was based on ColorJack. V2 was based on MaPePer, which was itself based on COBLIS V1 and therefore ColorJack, but before COBLIS V2 came out, MaPePer changed their source from ColorJack to something else… and they were so kind as to explain why on their github with a link to an archived conversation.

In June 2008, a developer for yet another unrelated application (Inkscape) was looking into colorblind simulators to incorporate into their app. After looking into the ColorJack code he wrote some criticisms in his dev journal and kept searching. Months later, the ColorJack developer pops up in the comments to say:

You’re right. [My] ColorMatrix version is very simplified and not accurate. I created that ColorMatrix one night, and since then it’s shown up in many places. I should probably take that page down before it spreads more.”

Bingo… the code was a careless one-night hack job. If you look at all of the simplifications and assumptions in the code you realize this wasn’t a guy that was checking multiple sources to get accurate terminology! Not everyone has dozens of hours to waste on tracking down the origin of a single term… that’s just me! What ColorJack saw was a clear pattern of name changes when simulations got sent through the “anomylize” function. Protanopia to protanomaly; deuteranopia to deuteranomaly; tritanopia to tritanomaly… and then just applied that same name conversion to achromatopsia. Achromatomaly is an understandable extrapolation of that rule and tired programmers are really good at noticing patterns.
…but his foreboding comment that his mistakes were going to spread uncontrollably through the web is just so… prescient, considering the eventual proliferation of his made-up Achromatomaly term.

But let’s not put all the heat on ColorJack. While Wickline was not the first to “anomylize” dichromacies, he may have been the first to “anomylize” achromatopsia. While “anomylizing” is still today the popular method to convert dichromacies to anomalous trichromacies, does it even make sense on achromatopsia? Wickline himself, COLBLINDOR’s COBLIS simulator and several subsequent academic papers equate Achromatomaly to blue cone monochromacy… but is this simulation even representative of how blue cone monochromats see?

…I mean, not even close! Now bear with me, as we’re going to get a little bit technical with the rest of this video. If you want any more information on any of these terms that i’m going to use, check out the link in the description to the glossary on my website.

Wickline’s reasoning for making this simulation was probably that blue cone monochromacy (what he called atypical achromatopsia) was formally referred to (in full) as “atypical incomplete achromatopsia”. “Incomplete” suggesting a weaker form of the colorblindness. But importantly, achromatopsia is not just total color blindness; It is a whole host of other symptoms from day blindness to very poor visual acuity and it is THOSE symptoms (not the color blindness) that are considerably weaker in blue cone monochromacy.

In normal conditions, cone monochromats still have absolute black and white vision, not this partially desaturated Achromatomalous vision. The important caveat here is that in specific twilight conditions known as Mesopic Vision, cone monochromats experience “Conditional Monochromacy”, where the rods and cones (which by themselves are both monochromatic) interact to generate “some” color. However, the achromatomaly sim also doesn’t grok with this conditional multichromacy for two very important reasons.

Now there isn’t a simulation that existed to show this aspect of blue cone monochromacy, so I had to hack together one that is an approximate simulation of what the conditional dichromacy of blue cone monochromats would look like. So please don’t try to reproduce it, because I don’t want this spreading around the internet like “Achromatomaly”. However, importantly, it gives – at best – dichromatic vision, while the Achromatomaly simulation is still very much trichromatic. Not to mention, the Achromatomaly simulation doesn’t account for the huge shift in the luminosity function of cone monochromats.

Cone Monochromats are more sensitive to shorter wavelengths because their eyes are all made up of blue cones and rods. Because they lack the green and red cones, which are sensitive to longer wavelengths, red, orange and yellow wavelengths are essentially black to blue cone monochromats. They certainly don’t appear like the washed out red that is shown in the Achromatomaly simulation.
Hell, the Achromatomaly simulation is so clearly NOT atypical achromatopsia as Wickline labeled it, maybe ColorJack knew it too. Maybe he deliberately changed the name from the clearly false to something fake, yet reasonable as a kind of “placeholder”… maybe… I mean it’s kind of hard to lay blame here.

So I know what you’re thinking… if Achromatomaly isn’t an actual condition and the Achromatomaly simulation can’t represent blue cone monochromacy, does this simulation… CAN this simulation represent anything?

I made a video about combining types of CVD and one of the conclusions was that a mixture of tritanomaly and either protanomaly or deuteranomaly was possible, but would be quite rare. In this case, the tritan defect and the protan or deutan defect would manifest and exist mostly independently from each other. If the two defects were about the same intensity, then yeah… the individual would probably see something mostly similar to the Achromatomaly simulation. But calling that “mixed” anomalous trichromacy Achromatomaly would be… well it would be stupid… because:

  1. “Mixed” already perfectly describes what’s going on, and
  2. It has nothing to do with achromatopsia. They’re trichromats not achromats; they don’t have any achromatopsia symptoms; the genes involved are completely different.

Like… you know how Columbus coined the term “West Indies” because he thought Cuba and Indonesia were practically identical besides being very much NOT.

Or how Reggaeton sounds nothing like reggae?

Shut up Dalton, go listen to Beethoven.

Mixed anomalous trichromacy is nothing like achromatopsia do not call it Achromatomaly. So should we relabel the Achromatomaly simulation as “mixed anomalous trichromacy”? It would be more accurate… but you know… with an estimated prevalence of just 1 in 1.5 million people and no instance ever being described in literature, maybe it’s just better to let the simulation die!

You know what we shouldn’t let die? This channel. So if you want to see more, subscribe and maybe leave me a comment about how this video was completely self-obsessive.

This is Chromaphobe.


Posted

in

by

Tags:

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *