Abstract

We have previously presented estimates of the effect of different parts of the surround on the appearance of a central test. This effect was measured by surrounding a central disk with a radially symmetric spatial sinewave that varied along a color line around midwhite. As the phase of the sinewave was changed uniformly in time, so that it appeared to drift towards the center, the appearance of the center changed cyclically in time. The induced modulation was nulled by adding real temporal modulation to the disk. The amplitude of the nulling modulation was measured for spatial frequencies of 0.05-2.0 c/deg at a drift rate of 1 Hz. The function relating the amplitude of nulling modulation to the spatial frequency of the surround was low-pass. If the amplitude of induced contrast were assumed to be a weighted sum of the effect of each pixel of the surround and the effect of each pixel proportional to the contrast amplitude at that pixel, then the spatial weighting function could be estimated by the Fourier transform of the measured nulling function divided by the extent of the surround. We now present tests of the first assumption. The surround consisted of pairs of radial sinewaves of different spatial frequencies. The two sinewaves were set to be inphase at the first surround pixel, and both were drifted at 1 Hz. The amplitude of induced contrast was measured as described above. The assumption of spatial additivity was tested by comparing the functions that relate the nulling amplitude to the spatial frequency of the paired surround sinewaves for each spatial frequency. The above measurements, combined with measurements of induced amplitude as a function of inducing amplitude, will enable us to characterize the spatial properties of color induction.

© 1990 Optical Society of America