Color Vision

This page describes the physics, biology, and psychology of how we see colors.

This might seem overly geeky, but it gives you a foundation that can be used to understand color mixing.

We'll make this as simple as possible by leaving out the nastier details. We'll also skip most of the references to the original research that discovered all this stuff.

Why This Is Important
A Word About Measuring, Encoding, and Describing
Physics
Biology
Psychology
Fakery
Related Pages

Why This Is Important

To those of us with normal vision, color is a very obvious thing. Color is all around us. The apple is red. The pencil is yellow. How could so obvious a thing be hard to understand?

Unfortunately, many of the things that appear obviously true to us are actually wrong. Here are some problems:

Color is a subjective phenomena - a psychological response to brain stimulation.
The spectrum contains colors that you can not see.
You can see colors that do not exist in nature.
Sometimes two colors will look exactly identical, but the physical stimuli generating the color that you perceive are quite different.

And here's a practical application for somebody who just wants to color light or mix some colors:

You have two sources of greenish yellow light, that both look exactly the same.
You try putting a color filter in front of the lamps to change the color of their light.
One lamp looks exactly the same with and without the filter.
The other lamp turns red when you add the filter.
What's going on?

A Word About Measuring, Encoding, and Describing

Before we begin, I need to take a moment to blither about how things are measured, encoded, described.

Information can be encoded in many forms. The temperature of the water in the hot tub could be encoded as:

100.95° F on an expensive lab thermometer
101 degrees F on the cheap thermometer dangling in the water
3.4 volts on the thermister that is part of the hot tub controller
"hot enough" when the thermostat triggers, turning off the heater
"Damn hot!" when you slide into it
"It seems cooler now" 2 minutes later

All of these values encode the temperature of the water, but they do so with varying amounts of precision and objectivity.

Color vision, indeed vision as a whole, is also subject to varying amounts of precision and objectivity, depending on what part of the visual system you are looking at.

Some parts, like the physics of light, are objective and can be measured with as much precision you are willing to pay for. When you get to biology, differences between individuals make vision subjective, but clustering around an average for members of the species. But when you get into the psychology, all bets are off. It could be that you and I perceive something utterly different when we look at the same object, but we simply agree to call the thing that we are looking at "orange". Because, when you get to the end of the vision system, the result is completely subjective. We don't really see the world as it is. Our brains construct an internal model that we percieve.

Sometimes our perceptual model corresponds well to the objective measurements of the real world. Sometimes our perceptual model does not correspond well to the real world, but we don't know it, because our perception works well enough to keep us from noticing the difference. Sometimes our perceptual model does not correspond well to the real world, and we are fooled into doing the wrong thing - an illusion.

A perception is not right or wrong. You can't really "misperceive". A perception is your mind's model of the world. It is what it is. But sometimes your perception is closer to reality, and sometimes it is further away.

Physics

Photons

What we call "visible light" is just a part of the larger electromagnetic spectrum. This electomagnetic radiation is composed of streams of photons, which exhibit some of the characteristics of waves, and some of the characteristics of particles.

Envision taking a close look at a photon as it travels along like a wave.

Certain characteristics of this wave are easily measured.

The path from one peak to the next is a "cycle".

The number of cycles that the photon traces in a second is the "frequency". The unit of measure is the "Hertz" (one cycle per second), or multiples like kilohertz (one thousand cycles per second) and megahertz (one million cycles per second).

The distance between two peaks is a "wavelength". It could be measured in feet and inches, but most often uses meters, nanometers (nm), or some other metric unit.

The the maximum distance that the wave goes in either direction is the "amplitude".

The frequency and wavelength have an inverse relationship: the more cycles per second (higher frequency), the shorter wavelength.

It should also be noted that lower energy photons have longer wavelengths (low frequency), and higher energy photons have shorter wavelengths (high frequency). [For the curious, the energy of a photon with frequency v is E = h v, where h is "Planck's constant", approximately 6.626 × 10^-34 joule-seconds. That won't be on the Wolfstone Final Exam.]

A single photon isn't much. A beam of light is likely to have billions and billions of photons, and they don't necessarily have the same characteristics.

Light can contain a mixture of wavelengths, some long, some short.

When all the photons in a light beam have the same wavelength, the beam is said to be "monochromatic", which means "all the same color".

The Electromagnetic Spectrum

A "spectrum" is a collection of photons with different wavelengths. The full electromagnetic spectrum is roughly laid out like this:

the trend on this "blue" side of the spectrum is:
- shorter wavelength
- higher frequency
- higher energy
gamma ray
X-ray
ultraviolet
visible light
- violet
- indigo
- blue
- green
- yellow
- orange
- red
infrared
microwave
radio
the trend on this "red" side of the spectrum is:
- longer wavelength
- lower frequency
- lower energy

Don't let the explicit list of visible colors fool you: visible light makes up only a tiny part of the electromagnetic spectrum, roughly 400 nanometers (violet) to 700 nanometers (red). Natural sunlight reaching the surface of the earth spans from near infrared through UV-A and some UV-B. Wavelengths below 285 nm and above 2,500 nm are normally absorbed by the earth's atmosphere.

Let's take a closer look at the visible spectrum.

The smooth transition is a simplification. Actually, the colors go up in teenie tiny steps, too fine for us to see.

[source: EFG2's "Spectra Lab Report" application.]

Continuous Spectrum

If you project a sharply focused beam of sunlight through a prism, multiple beams emerge. This is not a case of scattering, where the whole beam is merely spread out. The prism "sorts" the light by wavelength, a principle known as dispersion.

You see that the "white" light is actually composed of numerous different colors, smoothly transitioning from violet to red.

Newton referred to these as "pure spectral colors", because if you catch one of these beams and direct it into another prism, you can't break it up any smaller.

We are actually lucky to get such a smooth spectrum out of sunlight. That's because the sun comes close to the behavior of an ideal "black body radiator", and serves as an "incandescent" light source. If you were to project the luminous plasma of an "arc lamp", into a prism, you would see just a few bright, distinct color lines.

Emission Spectra

Consider this experiment: Elecricity is passed through a sample of Hydrogen gas. It glows with a blue light. But projecting a sharply focused beam of this light through a prism shows that the "blue" light is actually composed of distinct color lines of red, blue-green, blue-violet, and violet. In fact, hydrogen also emits ultraviolet, infrared, and radio lines that we can't see.

[source: http://chemed.chem.purdue.edu/genchem/topicreview/bp/ch6/bohr.html]

Every element has a different spectral signature of light that it emits when stimulated. Here are some simple "emission spectra" that are emitted when material is subject to an electric arc:

Monatomic gasses at low pressure tend to have spectra composed of sharp lines. Because of the dark gaps, light from such sources are considered "broken spectrum".

Putting the gas under pressure causes the spectrum to broaden into bands. Because of their complexity, molecules also exhibit bands.

Absorption Spectra

In addition to an emission spectrum, each element has an "absorption spectrum". The absorption spectrum is what you get when you pass a full spectrum beam of light through a cloud of cold gas, which soaks up some colors and lets other colors through.

This is the hydrogen absorption spectrum. The black vertical lines are wavelengths of light that hydrogen absorbs.

Compare that to the hydrogen emission spectrum, where colored vertical lines are wavelengths that hydrogen emits.

[adapted from EFG2's "Spectra Lab Report" application.]

Notice that the absorption and emission spectra are complementary: the colors of light that a cold cloud of gas will absorb are the same colors that the cloud will emit when it is energized.

The reason for this is simple: the same mechanism that emits light from an atom operates in reverse to absorb light. That mechanism uses the position of electrons within the atom to store the light energy. And it is important to note that the electrons have a limited number of positions that they are able to occupy. So, if a hydrogen atom gets hit by a photon, and that energy of that photon is not exactly the right amount necessary to shift an electron from its current position to another legal position, nothing happens. But if the photon energy is just right, the atom accepts the energy and the electron moves up. Later, when the electron drops back to its original position, it gives back exactly the same amount of energy that it previously took in, which causes the emission of a photon of exactly the same wavelength.

Physics Summary

Here's the important part to remember about the physics of color vision:

When light is pure (monochromatic), the color of the light can be uniquely described by its wavelength.
Some light sources produce a broad, smooth spectrum, with an assortment of wavelengths in it.
Some light sources contain just a few wavelengths of light, and are considered "broken spectrum" .

We also have some more advanced lighting concepts .

Biology

The human eye contains a large number of photoreceptors, sensitive spots that react to light. There are two types of light-sensing structures, called rods and cones.

All rod light sensors are sensitive to the same broad band of the visible spectrum. Your eye can't extract color information from rods; they just provide luminance (brightness) information. The flip side is that rods are quite sensitive. Rods are useful for seeing in dim light, and provide our peripheral vision. Since this page is devoted to color, we won't bother with any more detail about rods.

Tristimulus Model

There are three kinds of cone light sensors in the human eye. They are alike in physical structure, but differ in the particular light-sensitive chemical they use. Thus the three different kinds of cones respond to different parts of the visible light spectrum.

Greek letter	Roman letter	color band	peak sensitivity
Beta	S	blue	440 nm
Gamma	M	green	544 nm
Rho	L	red	580 nm

[I have seen slightly different peak sensitivity numbers from various sources. I won't bother tracking down the "best" numbers. The important point is that there are different curves with different peaks for each type of cone.]

Each type of cone is sensitive to a different range of light, and has a peak sensitivity at a different point in the spectrum.

Source: 1964 CIE Chromatacity project, as displayed by EFG2's "Chromaticity Lab Report" application.]

Light anywhere in the visible range from 400 to 700 nm will be detected by one or more of these sensors - remember, they overlap. The brain perceives a color depending on which of the three sensors are stimulated, and the relative strength of that stimulation.

Processing

This red/green/blue "tristimulus" model is true for the basic cone light receptors. Once the light receptors are activated, extensive processing is done in the retina, with the transformed data being sent to the brain.

[Some sticklers, and all the textbooks, will claim that the retina is actually a part of the brain.]

There are over 100,000,000 light receptors in the retina. The optic nerve bundle contains approximately 800,000 nerve fibers. Clearly, the information from many light receptors is somehow boiled down and re-encoded, to be transmitted along fewer nerve fibers. To put it more plainly, the tristimulus model doesn't go very far past the actual light receptors.

When visual information reaches the brain, even more processing is done in the brain in order to form our subjective perception.

But that's OK. We don't really need to know what's inside the magic black box. We just need to know the relationship between what goes in (light) and what comes out (sensation of color). We can skip all that processing and the form that optical input takes after the light receptors.

Gamut

Since the three tristimulus light receptors constitute the interface between us and the world of color, any device capable of recording and reproducing tristimulus values can faithfully reproduce any color that we can see. But what if our color reproduction system isn't quite capable of reproducing the tristimulus values?

The "gamut" of a color reproduction system is a collection of all possible distinct colors that the system can reproduce.

Color television uses three colors (RGB), like our eyes.
Some printing systems use three colors of ink. [If the ink is not opaque, the color of the paper is involved, too.]
Others use four, adding black (CMYK).
The paint desk at the hardware store uses numerous tint colors.

Different color reproduction systems have different ranges of colors that can be faithfully reproduced. Even when two gadgets use the same color reproduction system, such as the RGB system used in CRT-based video monitors, the exact color of the primary colors can enlarge the gamut.

Color Blindness

For the morbidly curious, color blindness happens when a flaw in the genetic code causes cones to be manufactured with incorrect color sensitive chemicals, or no color sensitive chemical at all.

If your gamma sensors were made with the rho chemical, those two sets of cones would have the same spectral responses, and you couldn't tell the difference between red and green.

This can be a complex topic. We have more information on color blindness.

Biology Summary

The important thing in the biology section is that there are three types of color receptors, because these three color receptors are the interface between you and the colorful world.

Psychology

Colors are a subjective perception, the brain's internal interpretation of light. Most of the time, this perception corresponds to reality well enough that we can function well. But the perception is not exactly the same as reality.

When 400 nm light enters your eye; the cones that trigger are only the beta; and your brain perceives it as violet.
When 425 nm light enters your eye; the cones that trigger are a lot of beta and a touch of rho; and the brain perceives blue.
When 490 nm light enters your eye; the cones that trigger are about equal amounts of beta and rho, with a third as much gamma; the brain perceives a sickly greenish blue.
When 625 nm light enters your eye; the cones that trigger are mostly rho, with one quarter as much gamma; your brain perceives orange.

Psychology Summary

The important lesson is this: The light is not itself colored. It merely generates the sensation of color in your brain.

Why is this distinction important? Because the brain can be tricked.

Fakery

We now know that colors are a subjective perception, an internal interpretation that generally corresponds to reality. When the brain receives familiar stimulation from the eye, it gives us the sensation of color. But what if the brain receives unfamiliar stimulation?

Consider the color pink. Look at the rays of color cast from a prism illuminated by the sun, and you won't find pink.
Look at the spectrum and you won't find pink there, either.

Pink is not a "pure spectral color". There is no wavelength of light that corresponds to pink. But we can see it anyway. Where does it come from?

It turns out that pink is a pale shade of purple - which doesn't really exist either. Purple is what your brain perceives when light enters your eye that stimulates about equal amounts of beta and rho, and no gamma at all.
Beta and rho overlap from roughly 405 nm - 540 nm, and for most of that range, gamma will also produce a signal. The only place that we can get beta and rho without gamma is 405 nm - 425 nm. And in that range, beta is much more sensitive than rho.

There is no single wavelength in the visible spectrum that will stimulate equal signals from both beta and rho, without gamma.

But what if you mixed two beams of light - one around 405 nm (which would stimulate only beta), and one from 675 nm (which would stimulate only rho)? With this setup, you can jam into the brain combinations of signals that do not occur with pure spectral colors. And in this case, the brain perceives the combination as pink (if it's a light shade) or purple (if dark).

In addition to generating colors that are not found in the spectrum, you can fake colors that do exist in the spectrum. Shine 625 nm light into the eye, the photoreceptors in the eye get "mostly rho, with one quarter as much gamma", which the brain perceives as the color orange. But the brain will perceive the same color orange with a mixed beam of yellowish-green 550 nm (which stimulates more gamma than rho), sweetened up with a shot of 675 nm red (pure rho). [When two spectra are different, but look alike to the observer, they are called "metamers" or "monomers".]

When all of the cones are stimulated, the brain provides the sensation of "white". You also see white when presented with an extremely bright light of any color. This may be because, with the overlap of color sensors, a very bright light source will produce strong signals in all receptors.

You might remember back to your primary school art class, when Miss Arglebargle said "yellow and red make orange". Well, they don't make real orange. A prism can show the difference between Arglebargle Orange and the real thing. But they make something that fools your brain into perceiving orange - and that's good enough.

This is also the origin of the three "primary colors" that can be mixed to produce other colors. The colors work that way because the eye has light receptors tuned to the three primary wavelengths, and the brain perceives mixtures of these three stimulus wavelengths as a single different color.

Fakery Summary

The important bits:

For pure spectral colors, the human optical system can generate the same sensation of color when presented with a monochromatic light, as it senses with certain mixes of different wavelengths that stimulate the same tristimulus color receptors.
The human optical system can generate sensations of color when presented with tristimulus combinations that can not occur with monochromatic light.

You may be interested in these related pages:

Thank you for visiting. Your comments are welcome.
. . . . . .