Human Systems Integration Division @ NASA Ames

Observations of effects such as Ramachandran and Cavanagh's motion capture (Vis. Res. 21:1 97-106) and Adelson's demonstration of the paradoxical apparent motion of a fluted square wave (ARVO, 1982; JOSA A 2:2 284-299) suggest that a spatial frequency analysis is required to understand a variety of motion phenomena. In this paper we examine the hypothesis that the low spatial frequencies in a stimulus dominate the motion percept in smooth, as well as apparent motion, and suggest that such effects might be considered a form of masking, as opposed to a distinct "capture" mechanism.

Our stimuli consisted of pairs of drifting vertical gratings presented within a two-dimensional gaussian spatial window. The contrast and velocity of each grating were under experimental control. When the motions of the two components are in opposite directions the resulting directional percept can be biased by changing the relative contrasts of the two components. When neither component clearly dominates we say the contrasts have been "motion equilibrated." The percept associated with a motion equilibrated stimulus commonly takes one of two forms, depending on the particular parameter values: the stimulus may appear to flicker with no clear direction of motion, or the two gratings may be perceptually segregated into two "transparent" planes.

Our results show that the contrast ratio needed for motion equilibration is equal to the threshold ratio when the individual contrasts are near threshold, but that as the contrasts are increased the ratio deviates, requiring less low frequency contrast to balance a contrast increment of the high frequency component. When 2 and 4 c/d gratings are opposed at moderate contrasts, the null ratio differs from the threshold ratio by about a factor of 2; since the threshold of the 2 c/d component is lower than that of the 4 c/d component, the results cannot be explained by assuming that subjects are matching the components for apparent contrast.