How many primary colors?

<p>There was an article in the N Y Times today about a species of monkeys that saw only two primary colors. They are able to give them the ability to see three primaries with a transplant. This was described as seeing all the colors. My question is this. Are there more than three primaries, but we only see three? After all, after all the fuss and space travel, we are really just an advanced species of monkeys. What if there are 4, 5, or 10 primaries? </p>

<p >If you play with colored light, mixing and blending to get the most assorted colors using just three colors, you would re-discover that red, green, and blue are the three light primaries. </p>

<p > </p>

<p >If you play with water colors, mixing and blending to find three that blend and make the greatest assortment of colors, you would re-discover that the three subtractive primaries for this task are “red”, “blue”, and “yellow”. A closer look reveals the red and blue and not pure. The red is actually red + blue, a color we call magenta. The blue is blue + green, a color we call cyan. Yellow remains the classic yellow you know and love.</p>

<p > </p>

<p >If white light is beamed at a drop of yellow dye on paper, red and green light will be reflected backwards towards an observer, blue light will be seized and converted to heat. As for a drop of cyan dye, red will be absorbed and green and blue reflected. A drop of magenta dye absorbs green and reflects away red and blue. Because each absorbs one of the primary colors, cyan, magenta, and yellow are labeled subtractive primaries. </p>

<p > </p>

<p >Now when we make prints on paper we lay down cyan, magenta, and yellow. The idea is, white light consisting of red, green and blue energies, strikes the paper. The dye or pigment controls what is reflected back at the observer and based on how the colors are arranged, a color print is seen. </p>

<p > </p>

<p >Sidebar: If cyan and magenta and yellow are positioned close and adjacent we expect black to result. Due to hue inaccuracies (we have never obtained pure magenta and cyan), a de-saturated black results. Conventional photography is out of luck but ink and pigment based printing supplements using a black ink/pigment to bolster contrast.</p>

<p > </p>

<p >Prints on paper made using the ink/pigment system consist of discrete droplets of cyan, magenta, yellow, and black, on paper. Black is used to enhance contrast. Thus black ink/pigment often called the foremost or “key” tone, it kicks-off the tonal scale. The technique of using these four inks/pigments is often abbreviated as CMYK.</p>

<p > </p>

<p >Conventional silver-based color prints have the advantage being continuous tone imagery. The digital printer however must lay down droplets of CMYK varying in size and spacing. To achieve a range of colors or gray scale we require an optical illusion to fool the eye/brain into “seeing” continuous tones. This process is called “half toning”.</p>

<p > </p>

<p >Go to the library and check out Color As Seen and Photographed, a Kodak book by Ralph M. Evans of Kodak Research Lab. The book will answer all your questions. </p>

<p>"What if there are 4, 5, or 10 primaries?"<br>

====<br>

I had a dream once and I saw all types of colors, amazing colors not just the colors on the color-wheel or the sRGB gamut. hope I did not spook you out. </p>

<p><strong>There was an article in the N Y Times today about a species of monkeys that saw only two primary colors.</strong></p>

<p>This is where our ratings system breaks down<strong>.<br /> </strong></p>

<p>The reason these rumored monkeys could only see two primary colors (red and blue) was because they were <em>political </em> monkeys.</p>

<p>It's possible to expose the cones of the human retina to light of two different frequencies in varying proportions and produce the perception that light of a third frequency is seen. Red light and blue light mixed can give the perception of seeing violet light, as does a beam of high-frequency violet light. A spectrometer is going to show a spectrum of two peaks in one case and a single peak in the other, even though the eye perceives both spectra as alike. That's what it means to talk about the primary colors of red and blue--that they happen to be significant to the human visual system. <br>

When you say, "What if there are 4, 5, or 10 primaries," it might be better to say, "What if there are species with visual systems in which 4, 5, or 10 wavelengths of light are salient?" It wouldn't be too surprising when you reflect on the ultraviolet light that some insects see and we cannot, but the question would be what advantage these species would get from processing all that information.</p>

A trichromatic visual system like ours simply requires 3 primaries to, in theory, match any 4th color. In this case a

primary is a light (additive color matching) which cannot be matched by a mixture of the other two lights. For example

with red blue and green primaries a red light could not be matched by mixing blue and green etc.

 

That said, red green and blue are not the only set of 3 primaries which can fulfill the definition of primaries and

trichormatic matches - there are many possible sets of primary colors.

 

And yes there are other systems, for example many insects are tetrachromatic (require 4 primaries), as are some birds.

Even humans under very specific conditions are tetrachromatic.

 

There is no "true" system - dichromacy, trichromacy, tetrachromacy etc - depend essentially upon the number of

functional photoreceptor classes operating whenever the color match is performed. For example in very dim light

humans are monochromatic becuase only one photoreceptor class is operating - the rods. A 10 primary system would

then need 10 functioning photoreceptor classes. I don't think that has ever been discovered.

 

One of the remarkable finding reported with the dichromatic monkeys is that they only had 2 (bright light) photoreceptor

classes yet when a third class was intorduced into their retinas, somehow, appropriate retina to brain connections were

made to allow the monkeys visual systems to use the new information.

<p>There are 10 kinds of people in the world: those who understand binary and those who don't. </p>

<p>There are only 3 primary colors, but using 3 dyes, 3 pigments, 3 light waves, etc is different. It's a question of which is additive and which is subtractive. Mix 3 primary colors w/ light and you'll get white, or no color. Not the same w/ pigments. Once you understand additive and subtractive it's clear.</p>

<p>I am a painter, not a color photographer. My photography is 98% b&w. My primaries are red, blue and yellow, all as neutral as possible. Normally we use a warmer and cooler version of each, such as cadmium red and alizarin crimson. I am sure the above are very interesting, but I am thinking that we do not know that there are only three. There may be 4 primaries or more, which we do not see because of the limitations of our eyes. This is a fascinating idea if you are involved in painting. </p>

<p>Yes, the primary colors are red, green and blue. Look up close at a TV screen, that's what you'll see. If you have 3 floodlights, fitted with red, green and blue filters, if you play all 3 together on a spot you will have white light. This is an additive process, ie: each floodlight adds to the spectrum present in the lit spot. If you use only 2 of the floods, it works as follows:</p>

<p>Red + Green = Yellow<br>

Red + Blue = Magenta<br>

Blue + Green = Cyan</p>

<p>I played around with litho film and color posterizing years back. With red, green and blue gels you could mix most any color you wanted, doing multi-exposures.</p>

<blockquote>

<p>There may be 4 primaries or more, which we do not see because of the limitations of our eyes</p>

</blockquote>

<p>This really makes no sense because there are as many primary colors as there are color receptors in our eyes. No more, no less. There's nothing "fundamental" about a primary color. The physiology of the eye determines how many there are. Humans have red, green and blue sensitive cone cells, therefore by definition there are three primary colors. If we could detect UV and IR with 2 more types of cone cells (which we don't have), then there would be 5 primary colors.</p>

<p>It makes a lot of sense. The eye does not determine how many primaries there are. It determines how many we can see. What if the monkeys said there are only two, because they only see two (some monkeys-most of them see three primaries)? You have to be open to possibilities beyond your own senses. The world is not what you sense. It is what it is, and we perceive it up to the limits of our senses.</p>

<blockquote>

<p>My question is this. Are there more than three primaries, but we only see three? After all, after all the fuss and space travel, we are really just an advanced species of monkeys. What if there are 4, 5, or 10 primaries?</p>

</blockquote>

<p> Interesting. Same question could be applied to the dimensions of space and time, to our perception of reality. Brain neurologist Dr Joe Dispenza : " Our human brain processes 400 billion bits of information every second ; however ; we are only aware of about 2000 of those billions of bits of data. We therefore are not aware of all that information because we literally are not attending to those stimilu. "</p>

<p>The first responder nailed it... 3 primaries for light are RGB, 3 primaries for pigment are red yellow blue.<br>

However, our perception does something else with colors in lighting conditions that get towards either end of the scale (too bright, too dark). As it gets darker, our eyes shift to a different mode (light/dark), and color perception changes. Because our eyes function as a continually adjusting sensor, that figures into color perception. I suspect (based on personal observation) that bright tropical sun tends to shut down the light/dark sensing and causes the visual experience to rely more on color.</p>

<p>Bruce, why would you jump to the hypothesis that we're "advanced" over monkeys? What does South Carolina's voting history suggest?</p>

<p>I knew it... Phylo is waiting for the Kwisatz Haderach</p>

<p>there are 2. Black & White</p>

<p>Well, speaking at the Kiwsatz Haderach...</p>

<p>The whole concept of their being as many primary colors as there are types of photoreceptors in the eye is pure bunk. Ever see a CIE chromaticity diagram? That's the range of colors your human trichromic eyes can see. Now, pick any three points in the space bounded by the diagram, and connect then with a triangle. That's the range of color that can be presented to the eye by mixing those three colors. Here's a chromaticity diagram where the three primaries chosen really suck.</p><div></div>

<p>The largest triangle you can draw in the curvy part of the diagram still leaves "wings" between the straight sides of the triangle and the curved part of the visual range. So, no matter what three "addative primaries" you choose, they don't cover the range of the eye. That is why Sharp has a new line of LCD displays coming out that use <strong>five</strong> additive primaries.</p>

<p>Paint and dyes are an even worse problem. Alan gave a fascinating explanation that totally confused the opaque pigment (aka "paint") primaries red, blue, and yellow, with the subtractive (aka "dye") primaries cyan, magenta, and yellow. The subtractive primaries are almost never as good as additive primaries, so any combination of three dyes (such as cyan, magenta, and yellow) bounds a triangle smaller than a typical additive primary triangle. Additive primaries can touch the perimeter of the curve, subtractive primaries can't. That's why we had so many "hi-fi color" systems (and yes, that unfortunately is the term used in the industry) such as the Pantone Hexachrome (CMY+orange+green) or the CMY+red+blue that Canon likes for inkjets.</p>

<p>Bruce talked more about the paint primaries. In color lectures, I refer to the paint primaries as the "kindergarten primaries". Aristotle, the first real color scientist, came to the conclusion that we chose those three primaries because we simply have an attraction to the number three, and that the minimal functional set of paint primaries is six, red, blue, yellow, green, white, and black. He didn't have modern measuring equipment, so he didn't know that red, blue, yellow, no matter how "pure" can't cover more than 50% of the human visual range. Why? To understand this, just look at what pigments are, opaque chunks of colored substance. (I like the term "interstitial primaries" for that). So, if you take a bunch of blue chunks and an equal number of red chunks, and pack them together onto canvas, until there's no space between them, any given ray of light is either going to reflect from a blue chunk or from a red chunk. So, you get four possibilities:</p>

<ul>

<li>reddish rays (a whole range of wavelengths that could be considered red, from deep 700nm red to bright 660nm red, to orangish 620nm) reflecting from red chunks</li>

<li>bluish rays reflecting from blue chunks</li>

<li>reddish rays absorbed by blue chunks</li>

<li>bluish rays absorbed by red chunks</li>

</ul>

<p>Basically, about half the light gets absorbed, half reflected, so you've got a 50% luminosity purple. That's how the three paint primaries work, a range of luminosity from around 100% near the primaries (yellow, blue, red) to 50% between the primaries (green, orange, purple, flesh, gray). No matter what minerals you pick for the pigments, you can't get around this. Adding the green primary, and finding a more "orangy" yellow primary lets you sustain a range of near 100% luminosity from red, orange, yellow, green, and blue. Violet suffers, as does purple, but in the organic world those are uncommon, so you really only need to be able to paint them when depicting fantasy, or trying to record the colors of mineral samples.</p>

<blockquote>

<p>there are as many primary colors as there are color receptors in our eyes. No more, no less.</p>

</blockquote>

<p>Aristotle knew better. ;)</p>

<p>There are *always* more real primaries, whether additive, subtractive, or interstitial, than there are photoreceptors. The only way to cover the range of trichromic vision with just three primaries is by using negative numbers, as in the CIE XYZ primaries. Negative light is rare in the real world, even in the photo.net rating system.</p>

<blockquote>

<p>As it gets darker, our eyes shift to a different mode (light/dark), and color perception changes.</p>

</blockquote>

<p>Tom, it has to get pretty dark before that happens. It takes surprisingly little light, just 5 cd/m2 to put the eye into pure photopic (or "day") vision. That's 20-40x the light from the monitor you're looking at right now. Even if you had the room lights off, things several feet from teh monitor, illuminated only by the monitor, would still be in the consistant pure photopic range and seeing "daylight" colors. Exact same colors you'd see under 10,000 cd/m2 of direct sun.</p>

<p>There is a wonderful region from about 0.05cd/m2 to 5cd/m2 where both the cones (photopic "day vision") and rods (scotopic "night vision") are active at the same time, producing "mesopic" vision. Since the rods have a spectral respons that's nothing like any of the three cone responses, during this time, a healthy trychromic human gets to experience tetrachromic vision, and flowers take on colors that they never had before. The light is only right for a few minutes a day, so you have to be lurking about a really decent garden at just the right time to experience the magic. And, since the shift from cones to rods is also causing your vision to "go blue" at this time (Purkinje effect), it takes a little practice to see the tetrachromic colors. You might have to spend a few evenings in the garden.</p>

<p>To Joseph W.: thanks for taking the trouble to post that information. I suspected that we painters were still in kindergarten. To John Kelly: I have concluded that we are more advanced than monkeys because we have the ability to destroy our planet and they do not.</p>

<p>Joseph,</p>

<p>Context of the OP included other species, the possibility of more than the three primary colors definable to human "<em>tri</em> chromic" (your own term) eyes, and some shape other than the approximate <em>tri</em> angle of possible colors defined in your CIE chromaticity diagram. Yes, the primary colors in any reproduction system are indeed arbitrarily chosen within the gamut of visible colors, but we aren't talking about specific reproduction systems, but about what it's possible to see. </p>

<p>Think outside the envelope.</p>

<p>You can be Kwisatch Haderach if you want to, but the question is, "What is the shape of the CIE envelope in bichromic monkeys? In tetrachromic insects? In hypothetical polychromic aliens perceiving doubtless-large chunks of the electromagnetic spectrum? <strong>And is it going to be defined by the response curves of the photoreceptors in that organism?</strong> " From your comment about the five primaries Sharp uses to better fit the CIE envelope in humans, I suspect your answer to the last question is probably going to be "Yes." If so, your discussion doesn't dismiss the relationship between photoreceptors and the CIE envelope as bunk--it supports it.</p>

<p><strong>A multispectral vision experiment you can try</strong> <br /> <strong>Disclaimer</strong> <br /> This is something I devised a couple of years ago, and have performed it on both myself and several volunteers. As far as I can tell, there are no permanent ill effects, but I am not a medical practitioner, and am not fully qualified to comment in that area. There are a variety of temporary alterations, not only to your sense of color, but to your sense of the nature of objects (their material composition, purpose, and overall place in the great scheme of life). There are temporary effects on your depth perception. Reported results have also included headache and vertigo. Operating motor vehicles, power equipment, a yo-yo, or an MP3 player while experiencing these effects is not advised.<br /> <strong>Background</strong> <br /> Roy Berns and Lawrence Taplin, a couple of really bright color scientists over at RIT, Munsell, and ARTSI (they got around) came up with the idea of performing multispectral (more than three additive primary) imaging using an ordinary trichromic (three color) digital camera and some fairly broad band (affecting a lot of wavelengths of light) color filters. They determined that a camera that is built like an eye, with three spectral responses that overlap greatly, in combination with fairly simple and common filters, can perform 6 color multispectral imaging. Take two shots, with two different broad filters, and perform a bunch of math, and you get hexachromic vision.</p>

<p>I figured that if you took a beam splitter, two RGB camera sensors, and the appropriate filters, you could make an industrial camera that did this in a single shot. It worked pretty well. But for fun, I also realized that if you already had two organic cameras (eyes) hooked up to one big processor (brain) that all you needed to do was to place the appropriate filters over the eyes, and you would experience hexachromic vision directly. Now, there's drawbacks to this. The processor (brain) that runs human vision has an annoying tendency to deemphasize information from whichever eye it considers to not be producing as "high quality" information at a given time. This can even extend to "shutting down" the processing of one eye, entirely, a condition called amblyopia. I figured that rather than apply filters to both eyes, I'd take the filter responsible for the larger separation of data and apply just that filter to the dominant eye (an eye is "dominant" because the brain considers that eye to be producing "better quality" information).<br /> <strong>The experiment</strong> <br /> The X1 filter (lime green) common in the days of B&W film photography turns out to be a quite effective filter for this experiment. If you put this filter over your dominant eye, and look at the world with both eyes open, after some minutes (or hours, depending on the individual) the brain will begin to integrate the two different spectral responses, and you will begin to experience multispectral imaging. Give it time, everyone is different. Discontinue if feeling vertigo, nausea, or headaches.<br /> <strong>Common observations</strong> <br /> It will affect you in many ways, scenes with complex colors will produce more information and be harder to process. Scenes with subtle colors will yield more information. The ratio of vision to sound, touch, kinesthetic sense, smell, etc. in your world-view will alter. Your sense of the nature of objects will also alter. One subject reported that after becoming used to the multispectral effect, they were able to tell naturally occuring (live plant, etc) colors from man made colors almost immediately, causing the world to "divide" into "natural" and "artificial" in a most disconcerting way.<br /> <strong>Variations</strong> <br /> The neodymium enhancing filter can be used as the dominant eye filter instead of the X0. The X0 will cause teh hexachromic effect to operate in six relatively regular spectral bands. The much more irregular spectrum of a neodymium filter, with multiple abrupt peaks and valleys will causes a very different effect, described b one subject as "shimmery".</p>

<blockquote>

<p>Think outside the envelope.</p>

</blockquote>

<p>Charles, look at what I was posting while you wrote that, an "article length" post on how to mess up your eyes, brain, and mind by putting different color filters in front of your eyes to achieve hexachromic vision...</p>

<blockquote>

<p>You can be Kwisatch Haderach if you want to,</p>

</blockquote>

<p>The issue is not open to debate. And one's personal desires usually have little influence on whether or not one is tasked with being the Kwisatch Haderach ;)</p>

<blockquote>

<p>but the question is, "What is the shape of the CIE envelope in bichromic monkeys?</p>

</blockquote>

<p>A line.</p>

<blockquote>

<p>In tetrachromic insects?</p>

</blockquote>

<p>A flattened sphere, because they emphasize the four signals differently.</p>

<blockquote>

<p>In hypothetical polychromic aliens perceiving doubtless-large chunks of the electromagnetic spectrum? <strong>And is it going to be defined by the response curves of the photoreceptors in that organism?</strong> "</p>

</blockquote>

<p>Likely not. Unless they're incredibly big brained polychromic aliens, they probably handle large numbers of informational channels the way Earth life does, by "quantizing". By boiling the data down into concepts before it reaches the highest order parts of our though process. Look at hearing. The total opposite of vision, the spatial resolution is low (just two ears, instead of millions of cone "pixels" in the eye) but the spectral resolution is enormous (30,000 hair cells in the cochlea). But we don't view sound as a 30,000 dimensional hypersphere, we process the sound to note the relations between the frequencies. So we can note a pitch, a timber based on the relationships of the harmonics of the pitch, whether the sound has content between the harmonics (either subharmonics or an "unvoiced" noise component). We also look at its dynamics. So, we may "see" 200^3 (200 to the third power) colors, I'd bet the aliens, with 2000 spectral colors wouldn't see 200^2000 colors, they'd "see" differently, literally "seeing" materials. They'd perform (subconciously) a chemical analysis on everything the look at, and see your face as collagen fiber, mellanin pigment, and blood hemaglobin, the same way we "hear" a band and recognize a sax, a bass, a guitar, and drums.</p>