The distance used for scaling disparities is the same as the one used for scaling retinal size.

To determine the physical size and global three-dimensional (3-D) shape of an object, retinal size and retinal disparity have to be scaled in accordance with the object's distance. We examined whether the distance used for scaling retinal disparity is the same as the distance used for scaling retinal size. Subjects adjusted the 3-D shape (size and depth) of a computer-simulated ellipsoid to match a tennis ball. Analysis of the errors when only the ellipsoid was visible in an otherwise completely dark room suggests that the distance used for scaling retinal disparity is indeed the same as that used for scaling retinal size. This was confirmed by showing that the correspondence between the distance used for scaling retinal disparity and that used for scaling retinal size does not improve when more information about distance is available (room lights on), although both distances are then much closer to the simulated distance. Finally, we show that this correspondence is not due to the use of distance-invariant higher order binocular information.