Page 109 -

Section 3.3 Representing Color 77

(i.e., does not conduct electricity), specularly reflected light tends to take the color
of the light source. If the surface is a conductor, the specular albedo may depend
quite strongly on wavelength, so that white light may result in colored specularities.

0.8

0.7

0.6
orange leaf
yellow leaf
0.5
Reflectance 0.4 brown dry leaf

reddish brown leaf
0.3

0.2 red leaf
brown dry leaf
0.1
black dry leaf
0
400 450 500 550 600 650 700
Wavelength in nm

FIGURE 3.7: Spectral albedoes for a variety of natural surfaces measured by Esa
Koivisto, Department of Physics, University of Kuopio, Finland, plotted against
wavelength in nanometers. These figures were plotted from data available at
http://www.it.lut.fi/ip/research/color/database/database.html.

3.3 REPRESENTING COLOR
Describing colors accurately is a matter of great commercial importance. Many
products are closely associated with specific colors—for example, the golden arches,
the color of various popular computers, and the color of photographic film boxes—
and manufacturers are willing to go to a great deal of trouble to ensure that different
batches have the same color. This requires a standard system for talking about
color. Simple names are insufficient because relatively few people know many color
names, and most people are willing to associate a large variety of colors with a
given name.

3.3.1 Linear Color Spaces
There is a natural mechanism for representing color: agree on a standard set of
primaries, and then describe any colored light by the three values of weights that
people would use to match the light using those primaries. In principle, this is easy
to use. To describe a color, we set up and perform the matching experiment and
transmit the match weights. Of course, this approach extends to give a representa-