Abstract
Stereoscopic depth perception involves computing the magnitude of the spatial displacement of image features in one eye compared to the same image features in the other eye. It is generally accepted that these computations are performed in parallel in disparity-channels tuned to different ranges of spatial frequencies. One purpose of the present research was to measure the temporal characteristics of the underlying spatial-frequency-tuned disparity-channels. This was done by measuring the highest velocity at which observers could accurately report the depth of a sinusoidal stimulus. It was found that the highest velocity that could be achieved was inversely proportional to the spatial frequency of the stimulus. A second purpose of the research was to understand the interactions among the spatial-frequency-tuned disparity-channels. Again, this was done by measuring the velocity-limit, but this time using compound sinusoidal stimuli. If the underlying spatial-frequency channels were independent, the upper velocity-limit for accurate depth perception should be equal to the faster component (i.e. the lower spatial frequency component) of the compound stimulus. The results were consistent with the notion of independent channels, provided the component spatial frequencies in the compound stimulus were separated by 2 or more octaves. At smaller separations, when the component frequencies were separated by only 1 octave, the slower component (i.e. the higher spatial frequency component) influenced performance. Our results are consistent with the emerging view that binocular depth perception is accomplished in the visual system by spatial-frequency-tuned disparity-channels that are largely independent above two octaves.