The chromatic Cornsweet effect: Relationships between spatial and chromatic variables

Date of Completion

January 1989


Psychology, Experimental




A parametric investigation of the chromatic Cornsweet effect (CCE) was conducted using isoluminant stimuli presented on a color-calibrated CRT display. Five locations by three orientations (including constant L/M- and constant S-cone lines) in CIE 1976 Uniform Chromaticity Space (UCS) were sampled. Observers adjusted the chromaticities of two chromatically homogeneous rectangles spatially adjacent to the flanks of the chromatic Cornsweet stimulus so as to make a color match. Observers also made magnitude estimations of the uniformity of color to either side of the central stimulus border.^ As a result of these data, several conclusions can be drawn regarding the CCE. First, the effect is generalizable and reliable. The results are consistent with, and extend, previous work (e.g., Ware & Cowan, 1983). The CCE was obtained in all locations and all orientations sampled. The effect occurred with variations in width of the central chromatic perturbation (1.2 to 4.8 deg visual angle), in chromatic contrast (defined as linear distance between stimulus border chromaticities:.02 to.07 UCS units), in chromatic content (i.e., at different color space locations), and in the nature of the change in chromaticity across space (i.e., with constant S-cone lines, constant L/M-cone lines, and lines of intermediate orientation). However, the magnitude of the effect varied systematically with these parameters and was generally best for perturbation widths of 2.4 deg and chromatic contrasts of.05-.06.^ A significant color space location by orientation effect was obtained such that constant S-cone stimuli in the red region of UCS resulted in a stronger effect than constant L/M-cone stimuli. The opposite pattern occurred in the blue region. Possible reasons for this interaction are discussed, as are implications of the effect for color theory. ^