Some observations on the masking effects of two-dimensional stimuli

Gratings that differ in orientation by as much as 62.5 deg from that of a signal grating raise the signal's threshold by nearly a log unit. The spatial-frequency tuning of the masking effect reaches a maximum slightly below the spatial frequency of the maskers but far from that of any quadratic distortion product. Further, the location of the peak does not depend much on the relative orientation of the signal and maskers thus making it unlikely that the masking effect can be explained in any simple way by the presence of visual nonlinearities. This illustrates the difficulty of attempting to explain human performance in even relatively simple discrimination experiments with models based on mechanisms tuned for spatial frequency and orientation.