Abstract

Reported frequency shifts between the fovea and periphery¹ are quantitatively modeled here using data for spatially localized stimuli; test stimuli were presented in either the superior or inferior visual fields. The model assumes a set of frequency channels in which a nonlinear contrast transfer function (CTF) follows the output of each channel and the labeled channel outputs are combined in a weighted average. Frequency tuning, overall sensitivity, and nonlinearity of CTFs for a set of channels are assumed to change as a function of visual field eccentricity. Parameter values for these factors were based on published reports² or on measurements obtained for observers tested previously in the frequency shift paradigm. Theoretical predictions using the latter parameter estimates produce closer fits to the data. Furthermore, although frequency shifts were slightly larger for test stimuli presented in the superior than inferior visual field, these differences are not significant. Parameter estimates for superior vs inferior visual fields of these observers also do not differ significantly. However, an important question remains: Why don’t observers perform like ideal observers and perhaps alter the perceptual frequency labels of peripheral mechanisms to eliminate frequency shifts between fovea and periphery?

© 1987 Optical Society of America