Comparison of color and luminance contrast: apples versus oranges?

Using a spatial, forced-choice, matching protocol, we have measured observers' ability to equate the contrasts of sinusoidal gratings which vary along differing directions in a 3-dimensional color space. In a given experiment, the observer obtained a perceptual match between the contrasts of two gratings whose chromaticities or luminances varied along differing chromatic directions which were selected from among five axes: an achromatic luminance axis (lum), an isoluminant axis where only S-cone activation varied (S-axis), an isoluminant axis where L- and M-cone activation varied in a complementary manner (LM-axis), an axis where only L-cone activation varied (L-axis), and an axis where only M-cone activation varied (M-axis). Even though these chromatic axes were chosen to activate independent mechanisms involved in the early stages of spatiochromatic visual processing, and despite the distinctly differing appearance of patterns from variations along differing directions, we find that observers can reliably make such pairwise contrast matches. Furthermore there is reasonable consistency of matching contrasts among observers and the pairwise contrast matches exhibit the properties of homogeneity and transitivity. This observed homogeneity and transitivity allows, for each color direction, the specification of a single scaling factor which relates perceptual contrast to physical contrast. (C) 1999 Elsevier Science Ltd. All rights reserved.