Psych 30 Lecture Notes: Color

• How Do We See Colors
• Red, Green and Blue Cones
• Color Blindness: More Prevalent Among Males
• Judging a Color

Physics of Color/Wavelength

Newton prism apparatus

Electromagnetic spectrum

Color vision begins with the physics of ligtht. Issac Newton discovered fundamental decomposition of light into separate wavelength components. Pass light through a prism, get colors of the rainbow. Visible light corresponds to a small range of the electromagetic spectrum roughly from 400nm (appears blue) to 700nm (appears red).
 
 

Spectral power distribution (SPD) is a plot of energy versus wavelength. Any light can be characterized by its spectral power distribution, that can be measured using a spectro-radiometer. Diagram of spectro-radiometer shows light source, prism that splits the light into its separate components, slit that passes narrow band of wavelengths (ideally passes only one wavelength), photodetector that measures energy at that wavelength. By moving the slit and detector, can measure the amount of energy at each wavelength. Most lights contain energy at many wavelengths. A light that contains only one wavelength is called a monochromatic light.

Measure SPD separately for each of two light sources. Then turn both lights on together. SPD of the mixture (sum of 2 lights) equals the sum of the two SPDs. That is, lights obey the additivity rule. Lights also obey the scalar rule. Doubling the intensity of a light doubles the SPD at each wavelength.

Color Matching and Trichromacy

Classic psychophysical matching experiment. Box split into two chambers, one chamber has test light, other chamber has three primary lights (the 3 primaries can be almost 3 light sources as long as they are different from one another). Small hole in the box allows subject to see the colors from the 2 chambers right next to one another. Subjects adjusts 3 knobs that set the intensities of the 3 primaries to match the test light.

Three results:

1. Lights that are physically different can look identical, are called metamers or metameric lights. Test light SPD typically is different from the SPD of the primaries. Using a spectro-radiometer, you can tell that the lights in the two chambers are different. Using your eye, they look identical.
2. 3 primaries are always enough to match any test light. With 3 primaries there is only one way to set the knobs to get the match. 2 primaries not enough, no way to achieve a match for many light sources. 4 too many, infinite number of different settings would achieve a match.
3. People behave like linear systems in the color matching experiment. Scalar rule: double the intensity of the test light, double the settings of the 3 knobs. Additivity rule: add any two test lights, 3 knobs set to sum of settings. Might expect peoples' behavior to be more complicated than this, considering all the neural activity that most go into observing the lights, making a decision, adjusting the knobs, etc.  Incredibly simple behavior in this experiment calls out for a simple and basic explanation (see below).

SPDs of a pair of metamers

Example showing the SPDs of a pair of metamers: two lights that are physically different, yet look identical.

Illustration of additivity in the color matching experiment.  Settings to (test light 1 + test light 2) = (settings to test light 1) + (settings to test light 2).

The color matching experiment is the basis for the design of color TV.  Three types of phosphors on the CRT screen glow red, green, and blue.  Yet can produce the appearance of many colors (yellow, purple, orange, etc.).  The designers of color TV took advantage of all of these results: (1) lights behave like linear systems (SPDs add), (2) people behave like linear systems in the color matching experiment, and (3) 3 primaries are all you need.

Physiological Basis of Trichromacy

Denis Baylor, here at Stanford, measured spectral sensitivities of macaque monkey cones. Chop of retina. Get a single cone into the suction pipette. Shine a light on it and measure the current. Repeat for many different wavelengths.  Repeat for each of the three cone classes.

Plots of relative response versus wavelength from Baylor's measurements, for each of the 3 cone types. The height of the curve at a certain wavelength corresponds to the probability that a photopigment molecule will absorb (and isomerize) a photon of light with that wavelength. The greater the probability of isomerization, the greater the response from the cone. S cones most sensitive to short wavelengths. L cones most sensitive to long wavelengths. M cones peak sensitivity to middle wavelengths. Amazing measurements, precise to 6 orders of magnitude.

Mulitply test light SPD times spectral sensitivities to compute cone responses. Changing the wavelength on a monochromatic light changes the relative responses of the three cone types. This is the basis of your ability to discriminate the colors of the rainbow (wavelength discrimination). Each wavelength evokes a unique ratio of cone responses.

Summary of trichromacy theory: There are three cone types that differ in their photopigments. The three photopigments are each selective for a different range of wavelengths. If two lights evoke the same responses in the three cone types, then the two lights will look the same. All that matters is the excitation in the three cone types. There are lots of lights out there that are physically different, but result in the same cone excitations (called metamers). Trichromacy is basis of color technology in the print industry and color TV.

Night (rod) vision: How about the rod system? You may have noticed that at under low light conditions (when your eye is dark adapted), you don't see colors. Rather, everything appears as some shade of gray.
 

This graph shows Baylor's measurements of the spectral sensitivity curve for rods.  With only one spectral sensitivity, there's no way to discriminate wavelength. Wavelength is totally confounded with intensity.

Color blindness: Either 1 (monochromats) or 2 (dichromats) missing photopigments/cone types.

• Dichromats can match with 2 pimaries.
• Dichromat will accept trichromat's match, but trichromat will not accept dichromat's match. In other words, some stimuli that look different to trichromats are metamers for dichromats.
• Rod monochromats missing all 3 cone types, only have rods. Don't see color at all, only different shades of gray. Have to wear dark sunglasses during the daytime - otherwise their photorecptors are fully bleached and they are effectively blind. Note that rod spectral sensitivity curve is more sensitive to shorter wavelengths. Because of this, the brightness of a blue object compared to a red one increases during dark adaptation, called Purkinje shift.

For the rest of you, try an online color blindness simulator to "see" what it would be like to be color blind.

Color Opponency

Trichromacy falls out from the fact that you have 3 cone types with different spectral sensitivities. In the retina, the cone signals get recombined into opponent mechanisms:

1. White/black: adds signals from all 3 cones types, L+M+S.
2. Red/green: L-M
3. Yellow/blue: L+M-S

Color appears redish when the red/green mechanism gives postive response, greenish when the red/green mechanism gives a negative response. Likewise for yellow/blue.

Color opponency in the retina:  Color opponency requires very specific wiring in the retina. Blue-yellow mechanism, for example, must receive complementary inputs from specific cone types (e.g., inhibition from S cones, excitation from L and M cones). Anatomists have identified a special subclass of ganglion cells, called bistratified cells, that do just that. Anatomical substrate for red/green opponency is still unknown.

S cones filled with flourescent dye. Turns out to be easy to stain the S cones, because their photopigment is very different from the other two types. Most of the cones are L and M cones, only a few S cones.  Because its easy to find the S cones, anatomists have been able to identify the retinal circuitry for blue-yellow.

Color Constancy and Chromatic Adaptation

Color checker overhead. Cover one square with a blue-ish filter and its color appearance is dramatically altered, e.g., the white square becomes blue-ish. But when we cover the entire thing with a blue-ish filter the color appearance doesn't change much at all.

Take photo under flourescent light, versus the same picture under daylight. The colors come out totally differently - greenish under the flourescent light and redish under the daylight - unless you do some "color correction" while developing the film. But you wouldn't see it that way if you were in the room. To you the colors would look pretty much the same under both illuminants. Phenomenon called color constancy, analogous to brightness constancy that we discussed earlier. Eye does not act like a camera, simply recording the image. Rather, the eye adapts to compensate for the color (SPD) of the light source.

Each cone type adapts independently. For example, a given L cone adapts according to local average L cone excitation. Likewise for the M cones. So that the retinal image adjusts to compensate not only to the overall intensity of the light source, but also to compensate for the color of the light source.

Color aftereffect demo (CD track 3.8.2)

Adapt to green, black, and yellow flag for 60 secs, then look at a white field and you see an afterimage of a red, white, and blue flag. Red/green, blue/yellow, black/white are complementary colors. Normally, when you look at a white field, L and M cones give about the same response so the red/green opponnent colors mechanism does not respond at all. Adapt to green, M cone sensitivity is reduced. Then when look at white field, L:M cones are out-of-balance, L cones more sensitive that M cones so red/green mechanism gives a positive response and you see red instead of white. Only lasts for a couple secs because the M cone sensitivity starts to readjust right away.

The visual system is designed to try to achieve a perceptual constancy. But as with the various brightness illusions I showed earlier, color adaptation also results in some misperceptions. Afterimage is an undesirable consequence of chromatic adaptation coupled with color opponency. Usually chromatic adaptation does the right thing - compensates for the color of the illuminant.

Simultaneous color contrast (analogous to simultaneous brightness contrast). X on left surrounded by yellow. X on right surrounded by gray. Paint/pigment of two X's is identical, yet color appearance is quite different because surrounding context is different. Color perception, like brightness perception, depends on contrast/surrounding context.

Summary