Abstract
It has long been known that the perceived aspect ratio of two perpendicular extents is not veridical – especially when one of the extents is shown in depth. Recent work conducted in high fidelity virtual reality (VR) suggested that bias in perceived aspect ratio is strongly correlated to bias in perceived slant, on which the aspect ratio stimuli is presented (Li and Durgin, 2010). In the present study, two experiments were conducted in VR to extend this previous finding. In the first experiment, the perceived aspect ratio of an L-shaped arrangement of balls on slants from 9° to 90° was measured. Two viewing distances (2 m and 8 m) and two types of slants (pitch slant and yaw slant) were used. The bias for the yaw slant was found to be systematically larger than that for the pitch slant, but this difference could be attributed to a vertical horizontal illusion. For both types of slant, the bias in the perceived aspect ratio increased with viewing distance. In the second experiment, bias of perceived aspect ratio was examined with slants from 6° to 24° (to test for a range effect), and the texture was either scaled or not scaled with viewing distance in two groups of subject. There was weak evidence of a range effect on perceived slant, but the changed scaling of texture with distance did have a reliable influence on aspect ratio judgments. Overall, the results of the present study can be quantitatively explained by assuming that the perceived slant function is a scaled sine function of the simulated slant, which is consistent with the verbal estimation data of real slants within reachable distance (Durgin, Li and Hajnal, 2010).
Meeting abstract presented at VSS 2013