Abstract
Purpose: In dynamic occlusion displays with multiple apertures, observers accurately connect sequential, spatially separated parts to recover the complete form of an object (e.g., Palmer et al., ARVO, 2000). In studying this effect with simpler displays containing only two misaligned, non-overlapping apertures, we discovered a strong, apparently unknown, perceptual distortion effect. Methods: Subjects viewed a horizontally translating vertical bar, whose top and bottom appeared sequentially through two vertically separated and horizontally offset apertures. Subjects made a 7-AFC among static displays showing the two parts of the bar in various relative spatial positions. We systematically explored variations in aperture width, spacing and translation velocity. Results: Subjects consistently misperceived the spatial relations of the parts, such that the leading part was displaced opposite to the direction of motion (counter to the Froelich Effect). Perceptual distortion increased in magnitude with velocity. Distortion could be eliminated by increasing the leading window width, proportional to velocity. Size of the second aperture had little effect. Conclusions: Misperceived position appears to result from misperceived motion. Velocity may be underperceived when the leading window is too narrow to encompass both receptors of bilocal movement detectors (or encompasses too few detectors for adequate summation). Previous research has shown that receptor separation of Reichardt-type detectors increases with their optimal velocity (van der Grind, Koenderink & van Doorn, 1986). These findings are consistent with a model of spatiotemporal interpolation in which a moving fragment and its velocity continue to be represented for a short time after occlusion, and the fragment (with updated position) is perceptually integrated with visible image fragments in accord with contour relatability (Kellman & Shipley, 1991).