Logo



Something About
Our Ability to See
The World In Color

1) A Significant Discovery
2)
The Two Slides
3)
Let's Duplicate the Experiment
4)
A Bit of Analysis
5)
An Oil-Painted Version
6)
A Fascinating Viewer to Make
7)
More Images
8)
Quo Vadis?
9)
Investigating Color Deficiency
10)
Another Viewer Experiment
11)
Retinex for Dichromats
12)
3-D Shadowgraphs

(Click any image for a large view, which will
open in a new window. To continue with
the text, simply close that window)

9) Investigating Color Deficiency

b+y=?

This atmospheric shot is of some the improvised equipment I've used in an attempt to investigate the color world of those who have the most common, red-green form, of color deficiency. It started with your usual human empathy and curiosity: "you mean everything you see is just black and white? Like your doggie?" So many bits of misinformation out there, and here we all are on the net and web, one of the richest sources of misinformation yet conceived! Okay, sure, take everything I'm presenting to you here with a pinch of doubt, too. And try it out on your own. That's the most important part of modern knowledge and whatever bits of "certainty" we can find on this planet: "does it work when I duplicate the thing for myself?" I'm attempting to give you all the essentials to work it out independently, so you can draw your own conclusions. But first read about some of the experiments, and what I've learned from them, including my best guesses about what's actually going on.

monkey

We have gradually learned the causes of "color blindness" in humans and primates (BTW- many mammals see much like we do, three colors, cats have more "rods," so their color world is perhaps less intense, but not like black and white movies, either! Others see in two colors, again not strictly monochrome). Most of the time one set of cones in the eye, red, green or blue, has the wrong pigment to absorb light of the normal wavelength or color, to register the proper signals to the brain. It may be a slight shift in the pigment, almost always in the wrong direction, so one's color range is narrowed. Or it may be that it is identical with the pigment for other cones, so that both sets produce identical responses, losing discrimination over that region. Since our irises, skin and hair color differ, is this surprising? (You can see the actual retinal pigment colors at the top of the next chart: yellow, red-magenta, purple and bluish-violet -- mix them together, a reddish shade results, and that's what bounces back a photo flash, when you obtain "redeye" snapshots!) It's sex-linked, so more men than women have it, and it comes in a few varieties. The most common forms are caused by compromised red or green retinal pigments, one or the other, usually not both. There are also very rare cases of the blue pigment being affected. (Click the picture of the monkey or CLICK HERE, to read a bit more about this fascinating topic.)

plot

Historically mammals had just two color pigments, for blue and green cones. The primates of the New World are still at that earlier stage of retinal development (this is better described in the previous link, CLICK HERE). So they are like humans who have a missing red pigment (-red dichromats), the most common kind. The word, dichromats, means "two colors." Most of us are trichromats, with three colors of retinal cones. At some point our chain of mammals mutated again in our favor. The green pigment split into two kinds, one redder than the other (again note samples of the actual cone and rod pigment colors near the top -- each is the "complement" to what color it absorbs, pretty neat). I was surprised to learn how close the two actually are, that what we call "red" is more of a yellow, but it's as red as we get. The "red" does the job, even though it overlaps the green pigment pretty closely, as you can see in this plot for the normal eye. (That's why in the plot I colored the red curve more of a yellow-orange.) The small red-green pigment difference (purple vs. blue-violet) is controlled by very little DNA, which is why it's so much at risk. Our eye's blue pigment is represented more redundantly by its DNA, and is much less likely to be shifted. The rods (dim light, monochrome vision) measure similar to the green curve, moved a little to the left and with a slightly broader curve.
 
Okay, that's enough background to understand the next two images. These show the very complete color chart we reproduced for you on the last page, to print for testing the Retinex viewer. You'll see in the views below (again click each) that three charts are stacked for each image. The top chart is the reference: white light and normal vision. Below that it is the theorecital representation of an inactive green pigment, or -green. So this dichromatic eye would see with just the blue and red cones. Green would appear darker than a normal eye sees, red would be unaffected. The bottom chart is the opposite case, -red. Here the red cones are inactive, so red objects are darkened, green are unaffected.

colors
alt col
Three Color Charts,
theoretical dichromats
Three Color Charts,
with adjustments

The systematic arrangement of these charts is useful here, as it "fights" Land's Retinex processes (it's just too damn regular, not mixed up randomly, so those extra color perceptions don't kick in very much). It's a pretty tough test to reproduce well. What I've done on the left is to remove the green layer for the -green images, and pasted a copy of the red to that channel instead, which forms yellow. And for the -red below, the red is removed and replaced by green, again forming yellow. This seemed a simplistic way to go, and so the alternate version on the right was made. It's the same three charts again, but this time the missing channel is also darkened slightly. So -red has dimmer red, -green has dimmer green. It's difficult to judge what matches dichromatic vision best. And we can do a LOT better than fooling around with this sort of mathematical representation. But it's a pure way to begin, and you can study the charts yourself later, to see how each color in isolation would be affected in its brightness and hue to a "colorblind eye."
 
(For the sake of completness, HERE is a set of four typical standard Dvorine-type "color blindness charts", which you've probably seen before, and are most often used by doctors, schools and employers to test for the presence of many variations of color defficiency.)

(Top of the Page)
10) Another "Viewer" Experiment

y+b lites

Which brings us to the equipment you see here. In the foreground is something that's not seen too often these days. It's a special photo darkroom safelight. The light source is a low-pressure sodium filled gas tube, not unlike mercury vapor. It takes some time to warm up to operating temperature. First the light glows dim and pinkish. After a few minutes full yellow brightness is reached. You may see them used as highway or rural lighting fixtures along the road, as some studies that indicate they provide excellent night vision for driving. They're also very "night safe" for astronomy, and cause very little light pollution, especially compared to the awful bright peach tone of "high pressure sodium vapor lamps." Gee, I hate those harsh lights! But more people seem to dislike the simpler yellow version (perverse non-astronomers...).
 
In a color darkroom it's very hard to find any light that we can see but the photo emulsions can't, except weakly. This yellow fits the bill, as its an intense very narrow light source, with a spectrum essentially of just two bright spectral lines nearly touching each other at 589 nanometers. Through a spectroscope you see the one spike, the rest is black. So everything we see with such a light source as the only one will appear to be monochrome. Why? Well, there is only the one wavelength that gets any energy. There's no blue whatsoever, and the red and green cones pick up the 589 nm. the same way, so there's no visible difference for red and green objects. These appear to be gray. Hmm... that's something interesting: a light in which we confuse red and green... could that property prove useful someday? Anyway, here's a standard minimal color chart, as seen with sodium light (taken with a normal color digital camera):

chart Na
Photo Color Chart
under sodium light only

Of course it doesn't look quite this way to our eyes. We don't notice the pure yellowness if there's nothing else to compare it to, and so it seems "whiter," more monochrome. To see what that Kodak chart really looks like, check the image to the left below, seen in white light, same digital color camera. (Since your clicks open new windows each time, you might want to hold a new image window open in your browser, then return to the text window to open another image, so you can compare them simultaneously on screen, given a large enough monitor.) The trick of low-pressure sodium merging red and green vision never completely left me. I grew up in a town which had sodium lamps in several areas as street lights, and I used to bring color photos and comic books with me when my family would go for a night drive, to study the way "everything turned to black and white" (well, yellow) when that odd light came into the car onto those images. My parents tolerated such quirks, thank goodness. Years later I was musing about color deficiency with a good friend, who was a -red dichromat, and the images from those street lights, later my darkroom safelight, jolted my memories. I wonder if the loud click of recognition was audible?
 
So -- if you view objects under sodium light only, you lose the distinctions of red and green cones, right? But you'd be missing out on blue, and Steve could easily tell what was blue, or violet or mauve, and so forth, he explained. Okay, then what we really need is to add a second light which is a pure deep blue. That plus the sodium light, and you should see approximately what my friend saw! Yes, it's true. To the right below is an example of the results. I'm surprised my digital color camera could pick it up this well. This is the same color chart, but sitting in front of those two lights, just like the photo up at the top of this page. I've color corrected the balance, as a camera is not as "smart" as our eyes, and insisted on reproducing a mild purple cast overall. Fixing that doesn't add any information (even so, I left some of the purple cast in). This is what my "eureka" moment turned up, when it simulates the color world of a red-green "colorblind" person:

chart norm
chart y+b
Photo Color Chart under
white light, normal color vision
Photo Color Chart under
sodium and deep blue lights

I find this rather astonishing, don't you? This is certainly NOT the dull, monochrome world the books usually tell us about. Those shades, while not fully normal in range and hue, are still very colorful, indeed. I note that the deep green swatch on Kodak's color chart is dark enough that here it comes out nearly black. But red is only slightly darkened. So this would be what a -green dichromatic person would perceive, fairly closely. To simulate -red vision we'd need to move the sodium light to a shorter wavelength... (right, sure; is there a simple, not too costly source for monochromatic yellow green light?)
 
We still seem to be getting somewhere here. Next is the same fancy, regular color chart we've seen before, like those theoretical representations above. Remember, with so much regularity you lose most of the Retinex effect, so this is a difficult chart to reproduce, worse than the simpler Kodak chart. There are two views, one under while light to the left, and the other under our special yellow plus blue lights to the right.

full norm
full y+b
Fullstep Color Chart under
white light, normal color vision
Fullstep Color Chart under
sodium and deep blue lights

No doubt about it, this is more disappointing than the first color chart. The red tones seem darkened as well as the green. It probably also depends on what dyes or pigments are used in the printing. Oh, well, you'd have to shift the sodium lamp wavelength to the right slightly to tweak that -- out of the question for at-home experimenters. What we get with sodium light is not quite a -green simulation, yet not quite -red, either, but something in between, a merging of the properties. Cut to the chase: two red-green deficient friends observed with these lights, here in the darkened studio, when the experiment was still new. And they said that everything looked close to the colors they usually see, but that certain hues were made slightly lighter, some slightly darker. But the effect was very close, if not exactly the same to their dichromatic vision.
 
Of course that has to be the case. We're using very sharp spikes of color here, not the usual smooth spreads of spectrum given off by incandescent lights, by daylight, or even the better fluorescent lights. Some pigments and dyes on fabrics and printing inks will reflect a wide enough color band that the differences are averaged out. But it's not the case with every color example. A concern of professional color technicians is something called color metamerism. Many samples just look different when seen under fluorescent, incandescent, daylight (direct sun) or cloudy-bright conditions. They change slightly, but visibly. And these metameric shifts are what my friends were able to detect (they're very bright, precise people). My two lights didn't provide "normal" lighting conditions, causing small changes of lightness and darkness. Otherwise, the effect is close, and for some objects it's nearly an exact match. Oh, well, it's better than I expected, and I am frankly still smiling over this one. I hope you can find a way to try it for yourself, although these images should get you going.
 
It's been a while since I lugged out the lights again this week, to take these digital photos to show you. I ought mention that the source for the blue light is no longer the old slide projector I used to use. That had a very narrow projector's beam that was too bright in one spot, too dim elsewhere. And it got quite hot for color filters. This time I have one of those wonderful Ott-Lights, something recent which is just dandy for home construction projects, equipment repairs and tweaks. Great for older eyes, too -- I love these small, bright, near daylight spectrum work lamps (thanx for the tip, Carol!). It also runs very cool. With a spectroscope I can see it has many good lines and wide color regions, much better than most fluorescent lights. There's also a dandy deep blue line. I isolated that one by taping two layers of deep blue acetate color filter over the lamp, followed by a large Wratten #47 color gel, which is an even deeper blue. It doesn't photograph above anywhere near as rich a blue as it is to the eye. The effect is a very reasonably pure source of diffuse blue light, which blends well with the color safelight's sodium yellow line (forgot to say, it's a Thomas Duplex "Super Safelight"). So that's what you see in the equipment shots above!

(Top of the Page)
11) Retinex "Red-White" for Dichromats

two ideas

Now it's only logical to put together these two ideas we've been looking at: Land's Retinex, and these novel "colorblind" simulations. It so happens there's an idea test subject -- that student oil painting I made back in college. So let's fetch it and put it down in front of both the sodium and deep blue light sources. You can see a digital photo here taken in an otherwise dark room. It's the same setup used on the yellow-blue images of color charts above. But the painting is too large to illuminate evenly with those lights, at least for the camera. Let's save it anyway. We get this stylized dramatic two-toned spotlight effect. You can tell the blue light came from the leftside above, the yellow from below, "a roaring fire under the moonlight !"
 
As we examine the larger version, notice that Retinex comes to the rescue for most "color blindness." One can plainly detect sensations of red, purple, yellow, orange and blue from the original that by now has become familiar to us. The green color is not so good, but that's the fault of my painting, slightly too much black. It's the one painted mixture error that's difficult to illuminate properly under any lighting. Evenso, isn't this something?! We have clear evidence here about how two processes conspire to cancel out each other's weaknesses, and "let the light come shining through!" What we see above is definitely a color image, is it not? And it has an even wider spectral pallet than those old two-color early Technicolor (and Cinecolor, Truecolor, Royal Color, etcetera) films.
 
Yet all we're really looking at is an oil-painting made using just three tubes of pigment: red, white and black, lighted by one yellow only sodium light and one deep blue filtered fluorescent lamp, simulating a -green dichromat's limited color world rather closely. But, but, but... yes, it would appear that we have to adjust our thinking here. The "classic" RGB theory of color vision could not have been the whole story. I don't know about you, but this finding gave me great delight, that those who have color "blindness" are not really so "blind" to colors after all. You have to hand it to the redundancies of evolution that we have more than one way to perceive the sensations of light and color, and where one may not work so well, another is there to take over. Pretty kewl, no?



One more digression before we get off the topic of color deficiencies. Clearly those of us with normal color vision can't ever be quite sure of exactly what a dichromatic person sees. From all we've learned, though, it's clear that we have evolved, and also as individuals have developed, many checks and balances, so important to our survival must be an ability to see in color (don't eat the green or yellow berries, just the red or blue ones... ;^). I wouldn't be at all surprised if everyone learns while growing up how to interpret color sensations, along with all those other necessary stimulus-parsing skills needed to grasp the world around us. We compensate for what is not fully present, and adapt to the abilities we do posses.

Only under certain conditions will someone notice or comment on our differences. Do we really know if everyone else "sees" the same blue sky, the same "green" grass, the same "red" fire engine? In the way we're discussing, no, we don't exactly. But in the last decade or so the actual retinal pigments from many human eyes has been measured. We can now say that most normal eyes contain the very same 3 + 1 pigments. So we should respond to identical colors in identical fashion with our eyes, even if our brains interpret them in slightly personal ways. And we can categorize those who have compromised retinal pigments, calculate the effects and the ambiguities. Where there's an important signal or warning color that must be perceived by everyone, like a traffic light, we try to standardize on other clues, like the relative location (yes, the red one's on top). Even the original old two-light traffic signals in NYC still had the red above the green.

hp-65

At other times good old serendipity plays a role. Or perhaps I'm not giving credit where credit is due. In any event, I'll leave you with the photo here of my venerable "ancient" Hewlett-Packard HP-65 pocket calculator (click on it for a lifesized view). It was given to me by a generous friend for Xmas of 1974, and I've treasured it long before it became a collectable (it still works!). This was the first handheld "computer" produced. Had a similar power and speed to the original Eniac "electronic brain" from the late 40's. You'd write short programs, test them, and save them on little magnetic strips. Long story for another time. Why bring this up? Well look at the keys! The colors used are ivory, light gray and black, and for the ever important "function keys" yellow-gold and blue. Those are the best possible color choices for the huge majority of color deficient people, lying right along the blue-yellow color axis we've been investigating up above.
 
Long before "political correctness" and and "disabled-friendly" were even concepts, here was a small device that did not penalize those with incomplete color vision (note: most LED's were red back then). I noticed it the first time I saw one and grinned like a Cheshire. Perhaps I'm just goofy (don't answer that). But it's a relief to have a web page to pass on this observation. For important controls on computers and other equipment, please, please:

 
Choose colors along the yellow-blue axis if you want certain buttons and knobs to be instantly recognizable by most of us!
(Yellow - Gold - Ochre - Tan - Sienna - Brown - Ivory - White - Black - Gray - BlueGray - Sky Blue - Medium Blue - Cerulean - Cobalt - Indigo...)

Of course the "yellow" can be somewhat more orange or lime, the blue more aqua or light violet, with minimal compromise. And perhaps our colorful web pages, including this one, ought be considered in light of this concern. At least we should make the lightness different when using confusable pairs (I've tried to do that with the red-green link colors here). Okay, enough "walla-walla" on this fer sure. Time for something else much less serious, don't you think...?

(Top of the Page)
12) 3-D Shadowgraphs

3-d stuff

I'll just bet most of you have seen 3-D books, photographs, magazines and comics like the motley assortment shown here. These are technically called "two-color anaglyphs," and usually are printed in a red colored ink and a greenish blue one (cyan). It's very common to receive a pair of similar red-cyan spectacles of some kind with the publication, the red usually for the left eye, the cyan one for the right. They are usually inexpensive cardboard affairs, either to hand hold before you eyes, or with ear pieces to wear like normal glasses, with or without any corrective spectacles. This shot shows only a few I could put my hands on quickly.
 
During the yearlong mid century love affair with 3-D movies in 1953-54, many such printed examples appeared. Several new series of 3-D Comics and adventures were launched, which printed drawings cleverly retraced into two copies, some objects shifted according to a chart by a few millimeters left or right, others by different amounts, such that the viewer would see something that matched our spectroscopic vision's expectations. As a kid I remember finding such comics, and also seeing several motion pictures in 3-D. Perhaps some of you did, too. Usually the films were projected with the much better way of keeping the left-right image pair sorted out: "Polaroid" filters and glasses (that didn't work with printing, polaroid printing is esoteric and costly, so the older red-green idea was used instead). Yes, Edwin Land's first big invention, polarizing "J-sheet," arrived just in time to provide Hollywood an easy, cheap way to give viewers now attracted to the Tee-Vee an incentive to return to the empty movie-houses: motion pictures that leapt from the screen: Three-Dimensional Movies!
 
Funny that Land's name should come up so logically here. A page back we investigated a simple viewer box that anyone handy could assemble at home to explore these Retinex vision abilities we're blessed with at birth or soon afterwards. You'll remember that box had red and green filtered lights, like this:

rg box

Well, that box gives me a very distinct impression of Yogi-Berra's: "deja vu, all over again." Reason being that when the first 3-D craze came around that I remember (there were still others earlier), I had come up with a silly, useless "kid's invention." I thought it was original, but others have come up with similar ideas since then. This consisted a pair of flashlight bulbs placed into a small cardboard box, with some batteries. There was a slot in the front -- to slip a pair of cheap cardboard 3-D glasses into -- one filter positioned behind each 1" round hole that I'd cut to let the light shine out. It wasn't very different from the viewer box you see here (bet you didn't see this connection coming), although smaller and not so well made, and without any brightness controls. Anyway relax, this final suggestion isn't another "science at home experiment" -- it's time to play!
 
The idea was that you could place this on a stool in front of a wall in a darkened room. The beams would shine on the wall, overlapped. If you put your hand in front of it, holding a pencil or some scissors, or an outstretched finger, and moved it at the small light box, you would see reddish and blue-green shadows cast on the wall that followed your pantomime. If you put on a matching pair of 3-D glasses to the ones inserted in the front of the box, something truly wonderful happened (well I thought so at the time -- it seemed like magic!). The shadows lifted off the wall and floated in front of you, with an amazingly realistic solidity and depth! I don't have very good depth perception myself, as my eyes were slightly crossed at birth, but even I could tell that this really worked! (Note: Andre de Toth, the director of the finest 3-D movie, "House of Wax," was blind in one eye. True story.) I invited my friends and family to see childish "puppet shows" and mini melodramas with hand puppets and props I'd gathered together. They'd sit beside the box slightly to the front, so they couldn't see easily what I was doing, and watch the pedestrian "mella-dramma" come off the wall and into their faces. It looked something like this:

3-d shadow

This digital snapshot (don't forget to click it for new window with large view) shows my hand up to the left casting some two color 3-D Shadowgraphs (that's what I've always called them, anyway) on the white poster board propped up behind it (the light box is off screen to the far left). I'm holding a tool for removing computer chips from sockets here, no particular reason, it happened to be handy. If you look through 3-D "anaglyph" glasses like those with the books above, you can make out in this informal photo a little of the effect. But it's so much better in reality, and especially in motion (shadows don't have a lot of details). Often I placed the box on a bed, just above a pillow, shining straight up. Then I invited my vict.. I mean audience member (vidience member?) to lie down and look up at the ceiling through the glasses. Say what? Go ahead, you won't believe this... So much dropped on their faces from above no one lasted more than a few minutes, but, hey, it was a lotta fun!
 
If you want to try this yourselves (and I do recommend it, can't you tell...? ;^), you can use the same box from the previous viewer. Note how the green filter is over the right hand light source, red over the left, and they're separated not far from the spacing of our eyes? I was looking ahead (back?) to this stunt. You should make a modification, though. Carefully remove those two special deep filters, they don't let much light through. Get a regular medium red photo filter, and a greenish blue one from the same camera store. These can be nearly any reasonably pure tone, somewhere in the Wratten mid 20's for red, and the mid 60's for the cyan. Or take apart some anaglyph 3-D glasses and use the color cellophane (block out any light that may leak around them if they're too small).
 
If you'd prefer to use Polaroid filters instead, as they're less fatiguing on the eyes, sure, you can do that. It bugged me for a few months why polaroids didn't at first seem to work at all with my adolescent invention. I thought I had gotten some "inferior" polaroids from the 3-D movies. Nope. It's the screen. You need one that doesn't diffuse and "depolarize" the light. Same as most movie theaters use, I later discovered. Aluminum paint on wood or fiberboard will work dandy. So will metallic poster board sheets that good art supply stores carry, the "silvery" kind. Some slide / movie projection screens (metallized lenticular) are excellent. I guess one could repaint a room with aluminum (ha-ha), but that was as far as home rules let me go with that idea. In a small space (like a closet) I used the polaroids with a metallic screen, otherwise stuck with with the red/cyan version on walls and ceiling. Still do. Whichever way you go, it's surprising fun, that much I promise!

--Wendy Carlos


Go back to Pt. I
(Top of the Page)


Back to the Wendy Carlos Home Page

Wendy Carlos, ColorVision3
© 2001-2008 Serendip LLC. No images, text, graphics or design
may be reproduced without permission. All Rights Reserved.