Abstract
Slanted surfaces are often seen against backgrounds or through apertures. In addition to providing a binocular context, such conditions result in the presence of monocular regions (sidebands) at the side of the binocularly slanted surface whose locations and/or relative widths could provide information about slant independent of azimuth, distance to the slanted surface, distance between surface and background/aperture or width of the slanted surface. Sidebands can be either temporal on each eye (background condition), nasal on each eye (aperture condition), or temporal and nasal on the same eye (equivalent to a surface intersecting an aperture). Using random dot stereograms Gillam & Blackburn (Perception, 1998) showed that vertical axis slant is considerably enhanced when appropriate monocular sidebands of uncorrelated texture are added to the nasal and temporal sides of the same eye with no binocular background. Here we expand these findings and also explore for the first time the pure background and aperture cases where information is given by the angular ratio of the monocular sidebands. Further investigation of the same eye case showed that for absolute slant the information provided by sidebands interacted with the disparity gradient rather than summing with it. The slant advantage with same eye sidebands also increased with the width of the binocular surface, although their absolute enhancing effect declined. It was found that background sidebands had little enhancing effect on perceived slant whereas aperture sidebands had a strong effect. Only aperture sidebands give unambiguous information about relative depth at the sides of the slanted surface. A phantom aperture is seen that “accounts for” the monocular sidebands. It was also found that under aperture conditions the slant of the surface could normalise so the aperture appeared slanted. The perceptual effects of sidebands are considered in relation to geometric constraints and to Panum's Limiting Case.
ARC Grant A00106364 To BJ Gillam