Abstract
The perception of convex-concave ambiguous pictures is known to depend on the orientation of those stimuli in the frontal plane, because it is subject to the light-from-above (and convexity) biases. In Koenderink et al. (Pointing out of the picture, Perception 2004:33) it was found that perception of an obliquely and frontally viewed picture of a globally convex relief was very similar in that the pictorial relief simply was indicated by the proximal stimulus. So, they found no evidence for any ‘correction’ mechanisms under oblique viewing. We tested whether this also holds for convex-concave ambiguous stimuli. We tested frontal and oblique viewing of a real example of a convex-concave-ambiguous ‘stimulus’, a picture of a sculpted memory stone, for vertically and horizontally placed photographs in fully diffuse lighting. We found idiosyncratic differences between frontal and oblique viewing. For frontal viewing the pictorial relief was mostly perceived to exist of convex elements and by one observer as just concave elements. For oblique viewing the perceived relief changed for all observers: between 1 and 5 elements flattened or reversed depth; two observers reported that all relief elements looked flatter in the oblique than in the upright condition. This suggest that in this specific case ‘correction’ mechanisms might play a role, but it is not yet clear which and whether those relate to the picture’s spatial attitude or foreshortening. To address this issue we will extend this study testing more viewing, positioning and lighting conditions for photographs and for simple computer rendered stimuli.