Distortions of posterior visual space

View From Outside the Viewing Sphere

The ‘viewing sphere’, as defined by Euclid and explored by Gibson as the ‘optic array’, is generally thought of as wrapped around the eye. Can an observer step out of it? With currently popular photographic techniques, the spectator is forced to, because the viewing sphere is presented as a pictorial object. Then the question is whether human observers are able to use such pictorial representations in an intuitive manner. Can the spectator ‘mentally step into the interior’ of the pictorial viewing sphere? We explore this issue in a short experiment. Perhaps unsurprisingly, because the eye cannot see itself, the short answer is no.

Magic Circle

Full-horizon cylindrical projections of the optic array are in common use. One wonders whether the public actually profits from such pictorial information, since the space behind one's back does not exist in visual awareness. In an experiment, a test image included six persons located at the corners of an irregular hexagon centred at the camera. Two persons faced the camera, two turned their back to the camera and two others faced a direction at right angles to the camera. The distances to the camera were unequal and varied from 1 to 2 m. Participants were asked to draw a ground plan of the perceived configuration, including actors and camera, on the basis of viewing the picture. As with any picture there exist many possible interpretations, the ambiguity grows even more when the angular scope of the picture is unknown. Almost all naïve viewers parse this planispheric (Mercator) representation so as to have the whole scene in front of them, with the actors standing in a circle, facing each other. They take the viewpoint to be outside the circle. Only a few placed the viewpoint inside the circle, which is indeed another reasonable interpretation (in this case the actual one).

The Planispheric Optic Array

The “planispheric optic array” is a full-horizon Mercator projection of the optic array. Such pictures of the environment are coming in common use with the availability of cheap full-view cameras of reasonable quality. This introduces the question of whether the public will actually profit from such pictorial information in terms of an understanding of the spatial layout of the depicted scene. Test images include four persons located at the corners of a square centered at the camera. The persons point at each other in various combinations. Participants in the experiment judge who is pointing at whom in a number of such photographs. It is found that certain very systematic and huge errors are the rule, indicating that naïve viewers are quite unable to parse such planispheric representations.