Abstract
Researches pointed out the common occurrence of a phenomenon of severe undershooting of depth extents and accuracy or slighter undershooting of frontoparallel extents, namely spatial anisotropy. Despite of this strong undershooting of depth extents, many other studies were concerned with a source of information for depth perception that may work for more accurate judgments, known as motion parallax. The present study is concerned with the effects of motion parallax on spatial anisotropy of exocentric distances in distal layouts. We introduced four types of head movements, free, restrained (control groups), translational, and rotational head movements, in order to investigate that issue, and two viewing conditions, monocular and binocular, for scrutinize interactions between motion parallax and binocular disparity. Observers verbally estimated exocentric distances from a layout composed of two dimensions, saggital and frontoparallel, whose center of expansion was located 49.213 ft (15 m) away the observation point. Repeated measures ANOVA over verbal estimates showed that depth extents were strongly undershot and frontoparallel extents were slightly undershot, F(1, 36)=109.240, p=.000 (control groups), even with induced motion parallax, F(1, 36)=35.312, p=.000 (translational and rotational). No effects were found for binocular disparity, F>1. We also submitted data on linear regression analysis (D′ = b + aD), matching depth and frontoparallel estimates. Perfect fit was only obtained for translational head movement under monocular viewing, t(9)=.541, p=.602. Our results indicated that motion parallax produced by translational head movement is a powerful cue for depth in large outdoors spaces, even though we found that spatial anisotropy remains a strong perceptual phenomenon. Our data also indicated that motion parallax is a strong cue for frontoparallel extents when produced by rotational head movement. Supported by Grants from CNPq (460045/00-8 and 523572/94-8) and FAPERJ (E-26/150.699/99).