Click on horizon moon to see an animated stereogram which simulates the horizon view of the binocular moon illusion experiment. Click on elevated moon to see an animated stereogram which simulates the elevated view of the binocular moon illusion experiment. These animations will delay at zero and maximum disparity to make it easier for a viewer to fuse the pair of images. In each case, the left-hand moon moves outward towards you after a slight delay, and then moves back to the distance of the right-hand moon after a similar delay. This cycle of increasing and decreasing binocular disparity is repeated to give you time to obtain good fusion. As in our actual experiment, all of the moons are, and remain, precisely the same in size. Yet the moon on the left seems to grow smaller as it draws closer. In the horizon moon example two identical pairs of moons are presented over a stereoscopic horizon scene (a Tuscan landscape photographed by Ansen Seale). In the elevated moon example the same moons are presented with the landscape is absent. In both cases the left-hand moon exhibits the illusion of growing smaller as it draws closer! The apparent change in size does not result in an apparent increase in its distance. The so-called "size-distance paradox" does not arise in this situation because two moons are present at once, and it is easy to see that the moon grows smaller (not larger) as it comes closer. In the case of the natural moon illusion one sees the apparently smaller elevated moon in isolation and, because of its smaller apparent size, observers often make the logical inference that it is farther away. This does not occur when one has the opportunity simultaneously to see an identical but more distant moon. There is no paradox. In some situations observers attend to obvious size differences, while in others they attend to differences in depth as well as size. Note that these animations simulate the experiment we conducted but do not reproduce the actual binocular experiment. The demonstrations are not identical to the experimental situation because the depth between the left-hand and right-hand moon is scaled by the distance between the viewer and the computer screen. In order for the landscape to dominate the apparent distance to the moons, the viewer would have to be located within it and the distant moon would have to be located at optical infinity (as it was positioned by the real experimental apparatus). In this animation the landscape plays only an artistic role to illustrate what the subjects observed in the real experiment. |