Abstract
Horizontal disparity plays an important role in theories of stereopsis and in the creation of stereoscopic scenery. The conventional explanation for its effectiveness is the nominally horizontal separation of the eyes, which makes horizontal the dominant disparity direction. Without denying this reason, an unconventional response might point to two qualifications: (1) Horizontal disparities do not play a special role in the perception of the stereo depth of one-dimensional stimuli. (2) The two-dimensional laboratory stimuli from which we know so much about stereo vision have almost exclusively been vertically oriented, symmetric about the vertical, or isotropic. The average effective disparity direction of these stimuli—measured perpendicular to component orientations and down-weighting horizontally oriented components—is horizontal. We tested perceived-depth predictions from horizontal disparities against those from averaged perpendicular disparities. We used oriented two-dimensional stimuli: plaids whose sinusoidal components had a 30 degree orientation difference. The plaid's orientation, given by the mean of the component orientations, was either vertical or oblique. Two plaids appeared on each trial for 176 ms. Their orientations were the same or different, as were their retinal disparity directions, which were perpendicular to the plaid's orientation or at an angle from the perpendicular. The disparity magnitude of one stimulus varied across trials; the other disparity was constant. Observers judged the relative depth of the two stimuli, without feedback. Perceived depth, measured by the point of subjective depth equality, followed the orientation-specific disparity predictions rather than horizontal disparity predictions: Perceived depth varied with relative disparity in the direction perpendicular of the average stimulus orientation. We suggest that depth from disparity is calculated along this perpendicular direction, and therefore the effective disparity direction varies with the orientation of the stimulus rather than being imposed by the horizontal separation of the eyes.
Meeting abstract presented at VSS 2018