How good is your color vision?

[EvilScientist]

http://www.xrite.com/online-color-test-challenge

Lower the number the better. I wish computer monitors can be used to test for Tetrachromacy. To learn more about Tetrachromacy go here:

http://nymag.com/scienceofus/2015/02/what-like-see-a-hundred-million-colors.html

I seriously doubt I am a tetrachromat.
Please post your scores below. I thought this would be a fun test to do. I got a zero.

 

[zoobyshoe]

That was really hard! I got a 7, but I was certain I'd gotten them all correct!

 

[lisab]

Whew, that's a tough test! I got 69. Not good, but not surprising - my eyes perceive color differently (left eye vs right eye). My dad is colorblind, and it seems two different kinds of colorblindness run in his family. I have a cousin who is so profoundly colorblind that when he went to a town where the traffic lights are horizontal instead of vertical, he couldn't tell the "go" light from the "stop" light.

But beware - a color test online isn't accurate, due to variation between monitors.

 

[jmeps]

Nice test, 210 here. Could that be why my wife and I have "conversations" about what color that is?

 

[Evo]

I scored an 8, but towards the end everything was starting to blur together and decided to stop moving tiles although a few seemed like they were very slightly *off*.

 

[collinsmark]

Wow, I got a 0: Perfect score! (with no re-do, i.e., first time)

I attribute it in part to the fact that my monitor is calibrated. (I use a calibration tool: an older model that is sort of similar to this tool.)

On a side comment, the OP's implication is correct that computer monitors cannot be used to test for tetrachromacy. Not unless the monitor in question has more than 3 colors per pixel.

 

[DaveC426913]

Wow, I got a perfect score too!

I knew I had good colour vision (got tested for working at photo labs), but I didn't know it was that good.

(I suspect attention-span/patience is a confounding factor in scoring. I wonder how many people kept flipping tiles until they though they were perfect, as opposed to flipping tiles until they simply seemed OK.)

 

[DaveC426913]

On a side comment, the OP's implication is correct that computer monitors cannot be used to test for tetrachromacy. Not unless the monitor in question has more than 3 colors per pixel.

I'm not convinced this is true. I see the logic (screens have only RGB colours) but I'm not sure that's required. A monitor can still produce swatches with very subtle grades of blue/green, and a tetrachromat should still be able to distinguish them better than a tri.

 

[Evo]

I'm not convinced this is true. I see the logic (screens have only RGB colours) but I'm not sure that's required. A monitor can still produce swatches with very subtle grades of blue/green, and a tetrachromat should still be able to distinguish them better than a tri.

My monitor is small and I can't even guess how old it is, 10-15 years? I also have the brightness all of the way down so it won't hurt my eyes. I'd bet no one could score 0 with this thing aside from sheer luck.

 

[phion]

I scored an 11.

 

[collinsmark]

I'm not convinced this is true. I see the logic (screens have only RGB colours) but I'm not sure that's required. A monitor can still produce swatches with very subtle grades of blue/green, and a tetrachromat should still be able to distinguish them better than a tri.

Allow me to explain two sides of the same coin that demonstrate the limitation:

Case A)

Consider a particular, hypothetical RGB color monitor that generates color by combining different combinations of red, green and blue, and that the bandwidth for each fundamental pixel color (red, green or blue) is very narrow. For example, suppose that a given color is produced by using different combinations of narrow bandwidths peaking at 450 nm (blue) 510 nm (green) and 700 nm (red). The monitor would be able to generate very smooth (perceived) transitions between 450 and 700 nm by using different combinations of these three colors.

But this is partially accomplished due to limitations of the human eye. For example, to generate a typical yellow color, the monitor combines 510 nm (green) and 700 nm (red).

The (normal) human eye cannot tell the difference between that combination and an actual, single wavelength at 570 nm. If you were to take a 570 nm, yellow color filter, it would transmit a narrow 570 nm chunk of sunlight that (normal) humans couldn't distinguish from the monitor's yellow color (made by adding two individual wavelengths). And ironically, this 570 nm, yellow filter would completely block out the yellow color from this hypothetical monitor.

But a prism can tell them apart! Separate the colors using a prism or a diffraction grating, and the monitor's light would still form two, separate peaks in the rainbow: one at 510 nm and another at 700 nm. However the 570 nm yellow filter, filtering sunlight, would produce a single peak at 570 nm.

Humans (at least typical humans) cannot tell this apart, but certain other animals can. If we want an instrument to test whether a being can tell the two apart, the RGB monitor alone is not capable of producing both.

Case B)

The bandwidth of a fundamental pixel color (red, green or blue) of a typical monitor is not necessarily all that narrow. A monitor transmitting pure green will produce a spectrum peaking at around 510 nm, but with a wider rolloff, compared with the hypothetical, narrow bandwidth discussed above.

With the wider bandwidths it is conceivable that combining a wide 510 nm spectrum with a widish 700 nm spectrum could sum to a wide spectrum with a single peak at 570 nm (yellow).

But then this monitor would not be capable of producing narrow bands of any particular color; not even the primary, fundamental RGB colors. If we were to test whether a being is capable of distinguishing between narrow wavelengths, again the monitor is not capable of producing them.

Food for thought):

http://theoatmeal.com/comics/mantis_shrimp

 

[Evo]

I scored an 11.

I *LOVE* your avatar, what class, charm and grace!

 

[phion]

I *LOVE* your avatar, what class, charm and grace!

Thank you, Evo. I'm glad you like it.

 

[lisab]

(I suspect attention-span/patience is a confounding factor in scoring. I wonder how many people kept flipping tiles until they though they were perfect, as opposed to flipping tiles until they simply seemed OK.)

*wonders if DaveC was reading my mind while I was doing the test*

 

[DaveC426913]

*wonders if DaveC was reading my mind while I was doing the test*

Sort of. More like reading my own mind.

I realized after I'd taken it that I had spent as much time on it as I'd needed to ensure I got the best result possible. I often don't do that, and I imagine most people don't. I made several passes, flipping just one tile one space and checking it, until I was sure.

So I suspect my results had more to do with patience/motivation than visual prowess.

I would bet dollars to doughnuts your score would drop considerably if you tried it again and spent more time on it,

 

[DaveC426913]

Allow me to explain two sides of the same coin that demonstrate the limitation:

I think you're over-examining it though.

A tetrachromat should be able to distinguish between RGB values 0,191,193 and 0,193,191 (two shades of teal). There's only three colours being transmitted there, but the tetrachromat's extra near-green cones will see the second one as more green and less blue, so they will see two different shades, whereas someone with only one set of green cones will not have enough resolving power to distinguish them.

There's nothing special about their vision when it comes to perceiving narrow frequencies; simply put, they have a ability to distinguish finer gradients of blue-green than others - regardless of how the colours are rendered.

 

[collinsmark]

I think you're over-examining it though.

A tetrachromat should be able to distinguish between RGB values 0,191,193 and 0,193,191 (two shades of teal). There's only three colours being transmitted there, but the tetrachromat's extra near-green cones will see the second one as more green and less blue, so they will see two different shades, whereas someone with only one set of green cones will not have enough resolving power to distinguish them.

There's nothing special about their vision when it comes to perceiving narrow frequencies; simply put, they have a ability to distinguish finer gradients of blue-green than others - regardless of how the colours are rendered.

It might be best explained with a picture.

Figure 1 shows two types of "yellows" that most humans find indistinguishable. If an animal had a type of cone that could detect the 570 nm wavelength in isolation, it might perceive an entirely different color than the yellow humans see by combining red and green light. In that animal, the two "yellows" could be as different of colors as humans perceive green and magenta differently. And that doesn't even address the other possible color combinations with existing colors.

It is not possible for an RGB monitor (narrowband or otherwise) to produce the 570 nm type of yellow in isolation of the Green+Red type of yellow. With narrowband primary colors, the 570 nm type "yellow" is impossible to produce (with RGB) at all. With wider bandwidth primary colors it can be produced, but not in isolation.

This doesn't specifically apply to human tetrachromats only, btw, but rather to color perception in general, to animals limited to certain sets of cones, and how that relates to monitors limited to RGB.

Figure 1. Humans (typical humans) find these two "yellows" indistinguishable.

The yellow used here is just one such example. (Cyan would be the next easiest example.)

There's a whole rainbow of other possible color perceptions for animals having more than three cone types.

[Edit: corrected a typo in the attached figure (originally had 470 nm for yellow rather than 570 nm). Oops. Also, added some text expanding on the ramifications of Figure 1.]

 

[DaveC426913]

I understand all that.

What does it have to do with RGB monitors making poor test equipment for tetrachromats? (Which was the original issue.)

 

[nsaspook]

Not too bad, 16 before finishing the morning coffee.

 

[zoobyshoe]

I realized after I'd taken it that I had spent as much time on it as I'd needed to ensure I got the best result possible. I often don't do that, and I imagine most people don't. I made several passes, flipping just one tile one space and checking it, until I was sure.

So I suspect my results had more to do with patience/motivation than visual prowess.

I would bet dollars to doughnuts your score would drop considerably if you tried it again and spent more time on it,

Once I sized up the task, I saw that a good score would require more than usual care, and that's what I put into it. Regardless, it is quite possible you put even more care into it than me, and that, if I did it again, I could get a better score by being yet more careful. But that would seem to defeat the purpose of the test. That is: I'm assuming a person with really good color vision wouldn't need to take extraordinary care: the differences between one tile and the next would be more obvious to that person.

What I'm saying, I guess, is that it shouldn't be a test of who can take extraordinary care or not, but about who naturally and effortlessly sees fine differences.

It could be there is no such person and that the perception of fine differences is always about training yourself to notice them. I don't know, and maybe the creators of this kind of test should sort that out first.

 

[Jonathan Scott]

"Perfect color vision", 0, admittedly using my 22" monitor rather than my laptop. Pity they can't spell "colour".

 

[collinsmark]

I understand all that.

What does it have to do with RGB monitors making poor test equipment for tetrachromats? (Which was the original issue.)

Sorry if I didn't make this clear before. In order to make a proper "experimental group" you will need a piece of test equipment (in some for or another) that can produce a particular wavelength of light (on the rainbow) that is neither red, green nor blue (nor a linear combination of red, green and blue -- a single wavelength is what we're after).

RGB monitors are great at producing linear combinations of red, green and blue. But they are limited to those.

An RGB monitor is incapable of producing 570 nm yellow in isolation, for example. If you wanted to test for some animal's tetrachomatic perception of this wavelength, an RGB monitor would not be able to produce this wavelength in isolation of red and green wavelengths, making the RGB monitor a poor choice of test equipment for the experimental group of tests.

 

[Andy Resnick]

http://www.xrite.com/online-color-test-challenge

Please post your scores below. I thought this would be a fun test to do. I got a zero.

Woo hoo! 151! My second-favorite number.

 

[zoobyshoe]

Woo hoo! 151! My second-favorite number.

And what's your third favorite?

 

[DaveC426913]

Sorry if I didn't make this clear before. In order to make a proper "experimental group" you will need a piece of test equipment (in some for or another) that can produce a particular wavelength of light (on the rainbow) that is neither red, green nor blue (nor a linear combination of red, green and blue -- a single wavelength is what we're after).

Why?

It seems you're making assumptions about how this "has" to work. I can get behind you setting how you might do it, but you seem to be using phrases like "need to" and "can't" and "this is how it works" etc.

OK, granted, a rigorous controlled experimental setting is not going to use equipment that has so many variables, but is that your premise?

 

[collinsmark]

OK, granted, a rigorous controlled experimental setting is not going to use equipment that has so many variables, but is that your premise?

The premise is that an animal or human with tetrachomatic color perception (including the cones, optic nerve and brain processing ability) will see a given wavelength of the rainbow which is neither red, blue nor green as a distinct color from its red-gree-blue superposition counterpart.

So there are two things you need:
1) An object capable of producing the required wavelength in isolation of its RGB counterpart.
2) An object who's color is composed of red, green, blue (and possibly the wavelength in question, which is optional), such that it is of indistinguishable color from the first object when seen by typical members of the same species.

[Edit: or alternatively, if the red, green and blue constituent colors each have sufficiently wide bandwidths such that there is sufficient overlap with the wavelength in question:
1) An object having a color consisting of a linear combination of RGB, but the specific wavelength in question blocked out of the spectrum.
2) An object having the equivalent RGB combination but with the specific wavelength in question not blocked out.
The colors are balanced such that typical members of the species find the color of the two objects indistinguishable.

Either way you need both a control and experimental color set. An RGB monitor alone is not capable of producing both.]

Then a comparison is made. If a test subject easily distinguishes the two objects (where typical members of the same species cannot), it is evidence of that subject being a tetrachomat.

An RGB computer monitor is incapable of producing object (1). Some other method, object, or equipment is necessary to produce the experimental color (with the specific non-red-gree-blue wavelength).

How you you propose an experiment using an RGB monitor? As far as I can figure, an RGB monitor is only useful for information about how a subject sees different combinations of red, green and blue. That has nothing to do with being a tetrachomat, since tetrachomatic perception transcends red, green and blue (or a linear combination thereof) perception.

 

[collinsmark]

Here is a related question.

How would you go about testing a typical human's color perception of green (and their ability to use their green color cones, optic nerve and brain processing ability for the color green) if the only tool you had was a malfunctioning computer and RGB monitor, such that somewhere along the way, the green channel is broken (suppose the video card is broken and no longer can produce the green channel).

The computer can generate all shades of red, blue and magenta, but nothing with green in it. How you go about using that computer (and that computer alone) to test one's ability to see and distinguish the color green from some shade of magenta?

 

[DaveC426913]

tetrachomatic perception transcends red, green and blue (or a linear combination thereof) perception.

Not true.

"Typical" red cones are stimulated by a wide range of freqs, peaking in the red.
Typical green cones are stimulated by a wide range of freqs, peaking in the green.
Typical blue cones are stimulated by a wide range of freqs, peaking in the blue.

Tetrachromat's 4th type of cones are stimulated by a wide range of freqs, peaking in the green, but not exactly where the typical green cones peak. They don't transcend anything, they merely peak at a slightly different place.

Again, it does not matter how the colours are generated (all cones are stimulated to some degree by all frequencies), they simply have a finer-grained ability to discriminate between them.

Take a look at this, to see how tetrachromats perceive versus typicals:

https://www.physicsforums.com/attachments/84279

 

[collinsmark]

Your diagrams are incorrect, when applied to an RGB monitor. You have put two green wavelength peaks (A and B) in your diagrams. And RGB monitor is not capable of producing color peaks so close together. An RGB monitor produces one, and only one, peak of green wavelength. (It also has two other capable peaks in red and blue). (This is neglecting any possible slight overlap in the bandwidth of the RGB monitor's primary color bandwidths. I think that's reasonable, since the primary color bandwidths of RGB monitor are typically much narrower than the bandwidths of our receptor color cones in our eyes).

When adjusting colors, the RGB monitor simply changes the relative amplitudes of these three peaks. It does not and cannot change the actual wavelengths of the individual peaks.

And RGB monitor is capable of producing A, but instead of producing color B, it actually produces color A plus a touch of Blue, not shown in the diagram (the resulting power spectral density function has two peaks -- one at A and other small one in the blue part of the spectrum). But the true wavelength peaks of A and B are both (together) incapable of being produced by the same RGB monitor.

-----------------------------

That said, of course our three color receptor cones have wide bandwidths, and even some significant overlap. That's why when we see a rainbow, or sunlight separated by a prism, we can see all the colors, not just narrow patches at red, green and blue.

But those same eyes can still be fooled into thinking we are seeing any particular color of the rainbow by a linear combination of very narrow bandwidth red, green a blue light. That's enough to display a rainbow on a computer screen, and for tricromats, it's indistinguishable to an actual rainbow. It's enough to fool most people into thinking that A + a pinch of Blue, is the same as color B.

Tetrachromats on the other hand, might realize that the rainbow on the computer screen looks a bit bland compared to an actual rainbow. For certain colors, they are not "fooled" by this RGB equivalent of the single color wavelength.

But in order to test for tetrachromatic perception, you need some method to produce both for comparison. An RGB monitor cannot produce both.

-------------

If I'm still missing something, it might help if you explain how one could test for a trichromat's green color perception, and how a subject might distinguish green from magenta, if the only instrument around is a defective computer with a broken green channel. (No other instrumentation or tools allowed.)

 

[zoobyshoe]

Snopes maintains an online test for tetrachromacy is impossible:

http://www.snopes.com/politics/medical/tetrachromacy.asp

 

[collinsmark]

It might be useful to recall that a couple of decades ago, there was a lot of research and development that went into laser televisions. These were big screen TVs where lasers were used instead of the more conventional filtered lights, for rearscreen projection. Of course this was all canceled when plasma and LCD flatscreen TVs took off as they did.

But laser TVs illustrate the point. You can't get narrower bandwidth of the primary colors any more than using lasers. Lasers have extremely narrow bandwidths. And as you may know, you can't just simply change the peak wavelength of a given laser -- lasers operate at particular wavelengths.

In order to approximate the "near green" as shown in color B, the laser TV would shine the green laser together with a little bit of the blue laser, and that would fool the trichomatic human eye into thinking that color B was being shown, even though it wasn't really, in reality. It was formed by a very narrow peak at wavelength A plus a touch of the very narrow blue wavelength. And that's enough to fool the silly human into thinking that B was being shown. Most humans can't tell the difference.

 

[Algr]

What does it have to do with RGB monitors making poor test equipment for tetrachromats? (Which was the original issue.)

Monitors are RGB. Tetrachroma is RGBX by definition. Even if X is inadvertently included in one of the primaries, you would not be able to control it independently from whatever other channel it was lumped in with. A black and white monitor technically outputs RGB, but without independent control of the colors, you can't tell anything about a person's color vision using one.

 

[DaveC426913]

Your diagrams are incorrect, when applied to an RGB monitor. You have put two green wavelength peaks (A and B) in your diagrams. And RGB monitor is not capable of producing color peaks so close together. An RGB monitor produces one, and only one, peak of green wavelength.

You're misinterpreting the diagram. I debated how much detail to put in before it got confusing.

Those vertical lines are not narrow light frequencies, they are combined display colours. They are (for argument's sake) RGB 0,193,191 and 0, 191, 193. A monitor is certainly capable of displaying them.

 

[DaveC426913]

I don't know how useful this conversation is anymore. We are talking past each other. You are certain of your setup and results, and I find you keep missing the mark. Whereas you feel the same way about mine.

But I am in a stronger position because your claim (all X's cannot be done) is harder to back up. You must show all claims are invalid, including mine. (It is pointless for you to put forth a argument then turn around and demonstrate that it is false.) All I have to do is demonstrate that a single argument (Y could be done) can't be invalidated. I have the luxury of putting forth an argument that suits my premise; the onus is on you to falsify the argument of my choosing.

 

[collinsmark]

You're misinterpreting the diagram. I debated how much detail to put in before it got confusing.

Those vertical lines are not narrow light frequencies, they are combined display colours. They are (for argument's sake) RGB 0,193,191 and 0, 191, 193. A monitor is certainly capable of displaying them.

"Combined dispaly" colours? I'm not sure what that means. Perhaps we are talking past each other due to a misunderstanding of how an RGB monitor works.

Below is a more qualitative (not to scale) way of how an RGB monitor produces the colors RGB [0,193, 191] and [0, 191, 193]. Does it make sense now how this is not an adequate tool to use for a tetrachromacy test?

Notice that after combining (neglecting any slight overlap, which I think is a valid approach, given the narrower, primary color bandwidths produced by the monitor), the individual peaks do not change frequency (reciprocal of wavelength), at least not significantly. They only change their relative intensities.