Color Space

The three essential parameters hue, saturation, and brightness can be thought of as defining a color space.

Using these three parameters or attributes to define a color, it is convenient to think of these parameters as forming a "color space" in analogy with three spatial dimensions. Three color "coordinates" would specify a color.

Space can be described by different coordinate systems, and the three most widely used color systems, Munsell, Ostwald, and CIE, describe the color space with different parameters. The Munsell system uses hue, value, and chroma and the Ostwald system uses dominant wavelength, purity, and luminance. The more precise CIE system uses a parameter Y to measure brightness and parameters x and y to specify the chromaticity which covers the properties hue and saturation on a two dimensional chromaticity diagram.

It is also a useful practice to use three primary colors, like blue, green, and red as vectors with which to map color space. If unit amounts of B, G, and R produce white light, then they can be used like unit vectors to define the color space.

Primary Colors

Any set of three colors which when added in appropriate combination will yield white can be considered to be primary colors. It is useful to map color space with a set of primaries like blue, green, and red. If unit amounts of B,G, and R produce white light, then they can be used like unit vectors to define the color space. The coordinates of a point in space defines a color C by

The coordinates B, G, and R are called the tristimulus values with respect to that set of primaries. Early research was done with a specific set of RGB primaries. The same color can be quantitatively matched to the CIE primaries and expressed in terms of the CIE chromaticity coordinates.

RGB Primaries

Color matching studies carried out in the 1920s by British researchers John Guild, W. D. Wright, and others, showed that colored samples could be matched by combinations of monochromatic primary colors Red (700 nm), Green (546.1 nm) and Blue (435.8 nm). The average responses of a large group of observers could be reproduced by a set of three matching functions. While purely additive combination of the three primaries could match only the range of hues in the triangle shown below, all the colors could be matched by adding a certain amount of red to the color being compared. This corresponds to negative values for the red matching function as shown. From the matching functions, one can derive tristimulus values which specify the chromaticity.

The RGB Color Matching Functions

These are the historical color matching functions worked out by painstaking experiments. They can be used to reliably predict color perception, but most applications now use the CIE functions.

The CIE functions are derived from these RGB functions in such a way as to give all positive values. The RGB primaries must include some subtraction to cover the entire range of color perception.