Color Perception

posted on 8 May 2013 by guy
last changed 8 Nov 2018

4.142855
Average: 4.1 (7 votes)
Your vote (click to rate)

ages: 12 to 99 yrs
budget: $0.00 to $20.00
prep time: 5 to 15 min
class time: 15 to 60 min

This lesson provides a tutorial in color perception and the physiology of color sensitive cells in the eye. It includes three classroom activites for exploring color vision:

  1. a game testing peripheral color vision
  2. a demonstration of mixing colored light using smart phones
  3. a test for defective color vision

Copies of the images, with captions, are included in the attached files at the bottom of the lesson.

required equipment: crayons, smart phones
subjects: Biology, Psychology
keywords: color, perception, blindness, light, photoreceptor

file attachment(s): 


Color is an illusion! There is nothing in the external world that "has" color in a scientific sense. Color is generated in our brains, in response to light striking our eyes, subconciously interpreted in a variety of contexts. Color can even be produced in our sleep, when our eyes are closed. Understanding how our eyes and brain produce a sense of color helps explain many interesting features of our vision.

This lesson discusses the structures in the eyes that process color information and the general way in which that information is used by the brain to interpret color.


Fig. 1: The visible color spectrum, with wavelengths indicated in nanometers.

the color spectrum

Humans perceive different frequencies of light as different colors.

Light is generated by vibrating electric charges (usually by electrons in atoms or in conducting materials), which give off different frequencies of electromagnetic radiation. For example, rapidly moving electrons in a hot light bulb filament are responsible for the glow of the filament. A certain frequency range of that radiation is perceived as visible light.

In air, a particular frequency of light corresponds to a particular wavelength of light, so when talking about the color of light, one often identifies it with the corresponding wavelength. Figure 1 shows the connection between wavelength (in nanometers — $10^{-9}$ m) and perceived color for the entire visible spectrum, from 380 nanometers (deep violet) to 740 nanometers (deep red). Radiation of slightly longer wavelength is classified as "infrared", which we perceive as heat. Radiation of slightly shorter wavelength is classified as "ultraviolet" and is responsible for tanning and burning skin.


Fig. 2: Schematic of layers in the retina. The back of the retina is at the top of the diagram. Adapted from work made available by Peter Hartmann and Marc Gabriel Schmid at Wikimedia Commons.

 


Fig. 3: The sensitivity (absorbance) of the different types of photoreceptors as a function of wavelength. Each curve is normalized to a maximum of 100. Short (S), medium (M) and long (L) wavelength cones are shown in solid blue, green and red, respectively. Rod absorbance is shown as the dashed black line.


Fig. 4: A scene from a rose garden as viewed by someone with working M and L cones (top) compared to the same scene viewed by someone with only one working type of cone. With only one cone, only brightness is recorded; there is no color information.


Fig. 5: Three iPhones displaying three colors, on the floor facing up.



Fig. 6: The same three iPhones stood up on edge, leaning against each other in a triangle. Red, green and blue light from the phones combine to form white light.


Fig. 7: Soap film as displayed on an iPad. (top) Full picture. (middle) Close up. This image starts to show the individual pixels in the screen. (bottom) Microscope view at 40X. The individual pixels are shown as blue-red-green triplets of LCDs.


Fig. 8: A plate from the Ishihara color perception test. The number "74" should be clearly visible to viewers with normal color vision. Viewers with one missing or altered cone type may read it as "21".

structure of the retina

As light enters the eye, it is absorbed in the retina by special photoreceptor cells called "rods" and "cones". In response, these cells produce chemical signals that jump to nearby neurons, which in turn send a chain of signals to the brain. When the brain receives these signals, it interprets the information as light shining on the corresponding region of the retina. As the light absorbed by the photoreceptors increases, more signals are sent to the brain, which in turn perceives a brighter image. Figure 2 shows a schematic of the cell structure in the retina that absorbs light and transmits the signals to the brain. The photoreceptor cells, at the back of the retina, are shown in blue and red at the top of the figure.

Rods and cones are not equally sensitive to light. Rods are much more sensitive and can be triggered by only a few photons of light, while cones require much higher intensities. Rods provide the mechanism for good night vision, but they are so sensitive that they saturate in daylight, and are useless for day vision.

Rods and cones are also not uniformly distributed over the surface of the retina. In the very center of the retina is a small region packed tightly with cones, the "fovea", which provides our highest resolution vision, good for reading and other fine work. Over most of the retina there is a mix of rods and cones, but in the very center there are no rods, and outside the fovea there are far fewer cones. For that reason, hunters and astronomers who are trying to see a dim object at night will get better results if they look a little to the side of the object in order to best utilize their rods.

human color vision

The mechanism for human color vision was postulated by Thomas Young in a paper to the Royal Society in 18021. Young correctly surmised that there are three types of photoreceptors in the eye, which are sensitive to different frequency ranges of visible light. Comparison of the signals from these different types of photoreceptors allow the brain to interpret the different light frequencies as different colors. Modern science has since discovered three different types of cone cells that behave in just this way.

One type of cone cell is mostly sensitive to short wavelength light (S cone). Another type of cone is mostly sensitive to long wavelength light (L cone), and a third type of cone is sensitive to the middle wavelengths (M cone). Figure 3 shows the relative sensitivity of the three types of cones as a function of wavelength, along with a graph of rod sensitivity (dashed line).2

Color vision most likely evolved to better discriminate objects from a background of similar brightness. An object reflecting light at 670 nanometers stimulates mostly long wavelength cones, and hardly any of the middle or short wavelength cones, and the object will be interpreted as "red". A background reflecting light at 540 nanometers may stimulate the same number of cones, but they will be mostly middle wavelength cones, with slightly fewer long wavelength cones, and consequently the background will be interpreted as "green". Figure 4 shows a rose garden as seen by someone with working M and L cones (top), and the same scene as viewed by someone with only one type of cone (bottom). A single type of cone only gives information about the brightness of the scene, without any way of identifying color. Clearly, color vision would have been very useful to our simian ancestors who wanted to find ripe fruit in the jungle, or maybe pick a rose. 

All three types of cones are scattered over most of the retina. By comparing the signals from the three types of cones in a certain region, the brain can figure out the color of light hitting that region, and thereby tell the difference between regions that are equally bright, but of different color. At night, when only one type of photoreceptor is active (the rods), you can't see any color.

activity - peripheral color vision
Since the periphery of the retina is populated mostly by rods, peripheral vision does not provide much useful information for the brain to interpret color. Sit in a dim room and have a friend hold up some similarly-shaped colored objects at your side while you look straight ahead. Crayons work very well for this exercise. Make sure you look at the objects only out of the corner of your eye. You may be able to determine the shape of the object being displayed, but unless you cheat and move your eye, you'll have a hard time determining the color.

perceived color and color mixing

By stimulating the different cone types at different levels, we can produce all the colors of the rainbow, and then some. When only our S cones are stimulated, we see blue. When only our L cones are stimulated, we see red. When our M cones are stimulated at their maximum and our L cones are stimulated at about 80% of their maximum (this ratio corresponds to a wavelength of about 540 nanometers, see figure 3), we see green. When our L cones are stimulated at about 95% of their maximum and M cones are stimulated at about 55% of their maximum (this ratio corresponds to a wavelength of about 580 nanometers), we see yellow.  By stimulating all three types of cones more or less equally, we get something on the grey scale (anywhere from black to white, depending on brightness).

We can even generate colors that are not contained in the rainbow (or the spectrum of figure 1) by using mixtures of light. If we stimulate the S and L cones, but not the M cones, we  produce a sense of red and blue, "magenta", which doesn't exist as a single frequency. There is no frequency on the response plot of figure 3 where it is possible to stimulate the S and L cones without also stimulating the M cones. Magenta may be somewhat similar in sense to spectral violet around 400 nanometers, but it has more L cone activity, and appears redder.

activity - colored light mixing
We can get a feeling for how colored light mixing works by using smart phones displaying different colors. Figures 5 and 6 show how this can be done. We used three iPhones to display three different colors: red, green and blue, figure 5. (A Google image search for red, green and blue gave us what we wanted.) Then we stood the phones on edge on a white piece of cardboard and leaned them against each other to prop them up. In the region in between the phones, the carboard reflects white, figure 6. If we remove the green phone, the blue and red phones produce magenta; if we remove the red phone, the blue and green phones produce cyan; and if we remove the blue phone, the red and green phones produce yellow.

color displays

Modern electronic displays in televisions and computers use the trick of color mixing to generate a wide range of colors from just three primary color components. Typical displays use pixels constructed from a triplet of red, green and blue LCD's. Adjusting the relative brightness of the three LCD's can mimic almost any hue in the spectrum. Figure 7 shows a magnified view of an Apple iPad display, demonstrating the LCD arrangement. At the upper left of the bottom image, the green LCD's are off and the red and blue LCD's are bright. This region of the image is showing magenta. From the bottom of the image to the upper right is a swath of pixels with the blue LCD's turned off. This region is showing yellow.

other color vision

Mammals all have either 2 or 3 types of cones, leading to dichromatic or trichromatic vision. Dogs and cats are dichromatic, while apes and men are trichromatic. Some nocturnal mammals, as well as marine mammals, have no color vision.

Some other animals have different arrangements of cones. Many birds, reptiles and amphibians are tretrachromatic, while pigeons and some turtles are pentachromatic (the latter presumably evolving out of a need to distinguish various subtle shades of mud). A certain type of shrimp has been found with at least 10 different photoreceptor types.3

There is even significant variation in color vision among human individuals. There is a wide range in the numbers of L and M cones in human eyes,4 with a small fraction of people completely missing one type. In other individuals, one cone type may have a shifted spectral sensitivity from normal, leading to somewhat impaired color discrimination. Defects in the L or M cone types lead to red-green color blindness, which is the most common type of colorblindness. It is rare in females, but several percent of males suffer from some form of it. (The genes that produce photoreceptor pigments are carried on the X chromosome, and a single normal gene is dominant.)

activity — color blindness test
In 1917, Dr. Shinobu Ishihara published5 a series of visual tests to identify and diagnose color vision defects. One of his plates is shown in figure 8. Normal trichromats should clearly see the number "74" in the image. People with missing or spectrum-shifted L or M cones may read it as "21". Achromats may see nothing. Show this plate to a large class and you should discover a small fraction of males in the room (maybe 4% to 8%) will have trouble with it.

questions to ponder

  • "Brown" doesn't appear in the visible spectrum. Where does it come from?

Brown can only be produced from a color mixture. It comes from mostly L and M cone stimulations, with some S cone stimulus thrown in. On the RGB scale, values around 45%R, 30%G, 20%B produce a fairly good chocolate brown, depending on the display device.

  • How is the color "silver" made?

Silver objects reflect almost all the visible light that shines on them, and in that sense, silver might be considered part of the grey scale. What makes silver objects appear different from grey or white objects is not the frequencies of light reflected, but the way in which the light is reflected. Light rays reflected off of a matte grey surface scatter in all directions. Therefore, light of many different colors from all over the room may reflect off of a particular region of the surface and enter your eye. In a silver object, light reflects off the surface in such a way that the light ray leaves the surface at the same angle as the incoming ray.6 The rays that enter your eye from that part of the surface come from only one direction in the room, and represent the color of whatever object is in that direction. Different parts of the surface appear to be different colors, reflecting the image of whatever is in the room.

  • How do we know different people sense the same colors?

We don't. In principle, what you describe as "burgundy red" could be what I would describe as "sky blue". As long as the correlation between wavelength and experience is consistent for each of us, we would learn to use the same words for the same frequencies of light. There is some highly circumstantial evidence that suggests there could be dramatic differences between individuals in the perception of color:

Different individuals can have large variations in the density of L and M cones in their retinas.4

Some common forms of synesthesia confuse color signals in the brain with other perceptions, such as number or letter recognition ("grapheme-color synesthesia")7, or sounds ("chromesthesia")8.

Married couples can never agree on paint colors.

Michael at Vsauce offers an entertaining riff on this question at https://www.youtube.com/watch?v=evQsOFQju08.

  • Color mixtures can show us colors (like magenta) that do not appear in the color spectrum. Are there any colors that cannot even be produced by mixtures of light?

Yes, but presumably none of us have seen them. The neurologist Oliver Sacks once described an hallucination in which he claimed to see a color that no one else had ever seen. He called the color "indigo". (Of course the word "indigo" in common use refers to a color somewhere between blue and violet, derived from the Indigofera tinctoria plant, but Sacks presumably had something else in mind.) In principle, it is possible to generate the sense of a color that has never been produce by light. Suppose the M cones are stimulated, but not the L or S cones. Figure 3 shows there is no frequency of light for which it is possible stimulate the M cones without also stimulating either the L or S cones. It is also not possible to do this with any combination of frequencies (a mixture). Such a color has never been seen in the natural world, but we might expect it to look like some kind of "ultra green". Aside from direct electrical stimulation of the retina or a drug-induced coma, we are unlikely to experience it. (Neither of these experiments is recommended.)

  • Why are violet and purple so similar, given that they correspond to different sources of light?

Violet is a pure spectral color of short wavelength, around 400 nanometers. Purple, like magenta, is a mixture of short and long wavelength stimulation, with a little more short and a little less long than magenta. Figure 3 hints at the reason for this similarity. Although the response of L cones to very short wavelengths may not be very well measured yet (figure 3 does not plot the response below 370 nanometers), some researchers believe the L cone response rises somewhat at low wavelengths. In that case, spectral violet does stimulate both S cones (strongly) and L cones (a little bit), much like a mixture of blue and red light.

further resources

Bruce MacEvoy at handprint.com provides a clear and detailed introduction to light and color vision at http://www.handprint.com/HP/WCL/color1.html.

Peter Gouras at webvision gives a very detailed (though somewhat technical) description of many features of the physiology of color vision at http://webvision.med.utah.edu/book/part-vii-color-vision/color-vision/.

Steve Mould at The Royal Institution gives a lucid discussion of magenta at https://www.youtube.com/watch?v=iPPYGJjKVco.

A collection of Ishihara color test plates can be found at http://www.colour-blindness.com/colour-blindness-tests/ishihara-colour-t....

Leonard Vance has developed a classroom demonstration using a 3-bladed painted fan that is illuminated by a strobed ultraviolet LED to produce interesting color combinations at  http://www.youtube.com/watch?v=qm9d9qV39Ec.

  • 1. Young, T. "Bakerian Lecture: On the Theory of Light and Colours". Phil. Trans. R. Soc. Lond. 92 (1802): 12–48. http://rstl.royalsocietypublishing.org/content/92/12
  • 2. This plot should be taken as representative. Different researchers disagree about the precise details of the graphs. For instance, all researchers agree there is a photoreceptor sensitive to mostly short wavelengths, but claim that the sensitivity peaks anywhere from 420 to 440 nanometers.
  • 3. Cronin TW, Marshall NJ (1989). "A retina with at least ten spectral types of photoreceptors in a mantis shrimp". Nature 339 (6220): 137–40. http://www.nature.com/nature/journal/v339/n6220/abs/339137a0.html
  • 4. a. b. Hofer H, Carroll J, Neitz J, Neitz M, Williams DR. Organization of the human trichromatic cone mosaic. J Neurosci. 2005;25:9669–9679. http://www.ncbi.nlm.nih.gov/pubmed/16237171.
  • 5. S. Ishihara, Tests for color-blindness (Handaya, Tokyo, Hongo Harukicho, 1917).
  • 6. See http://hyperphysics.phy-astr.gsu.edu/hbase/phyopt/reflectcon.html
  • 7. Jäncke, L., Beeli, G., Eulig, C., & Hänggi, J. (2009). The neuroanatomy of grapheme–color synesthesia. European Journal Of Neuroscience, 29(6), 1287–1293. doi:10.1111/j.1460-9568.2009.06673.x
  • 8. Cytowic, Richard E; Eagleman, David M "Wednesday is Indigo Blue: Discovering the Brain of Synesthesia" Cambridge: MIT Press. 2009. ISBN 0-262-01279-0.


login or register to post comments