Abstract
The most recent formulation of the Intrinsic Constraint (IC) Model of Domini, Caudek and Tassinari (in press) predicts that the perceived depth of a simulated 3D structure, specified by only one depth cue will be perceived as shallower than the same structure specified by two depth cues. While empirical evidence supports this hypothesis, the results may be confounded by presentation of stimuli using computer monitors. That is to say, simulated monitor viewing provides cues to flatness via pixilation, reduced blur gradient and accommodation which may contribute underestimations of depth. When depth-cues are seen in isolation, more weight may be given to flatness cues and therefore, the single-cue stimuli may appear flatter. To address this criticism textured real surfaces were employed in order to remove any contributions from flatness cues. These surfaces were slanted around the Y axis at a range of angles and viewed by observers in three experimental conditions: stereo-only, motion-only, and combined stereo and motion. In the stereo-only condition the surface was static and viewed binocularly. In the motion-only condition the surface was viewed monocularly while oscillating at a range of angular velocities about the vertical axis. Results were consistent with earlier findings by Domini (1999) where increases in angular velocity produced increases in perceived slant. In the combined-cue condition the surface was seen binocularly with oscillatory motion over the same range of velocities as in the motion-only condition. The results confirm those of Domini et al. and the predictions of the IC model in that the surface in the combined-cue condition was judged as more slanted than the surface viewed in the motion-only and stereo-only conditions.