Abstract

Isoluminant sine wave gratings appear almost square to most observers, i.e., as bars of color separated by a fuzzy edge. A simple hypothesis can explain this phenomenon qualitatively: the observer adapts to the mean color in the grating, contrast discrimination is then most acute around the adaptation color, and discrimination falls off as a function of distance in color space from the adaptation point. An equivalent hypothesis is that a sigmoid response function with compressive nonlinearities on both sides of the mean color introduces higher harmonics of the physical sine wave into the percept. This response function has been estimated by measuring increment thresholds for brief tests and flashes at different points along lines in color space while keeping adaptation constant at the mid-white. The results can be used to predict both the apparent squaring and the change in duty cycle. Additional tests of this hypothesis can be performed by adapting to different lights and comparing the change in appearance of the grating to the change in increment threshold curves. The estimated response function along a color line can be used to correct a grating so that it appears sinusoidal.

© 1989 Optical Society of America