A psychophysical investigation of the perception of depth with stereoscopic television displays

Spain, Edward Huland
Journal Title
Journal ISSN
Volume Title
A series of four experiments was conducted to investigate the influence. of three video system parameters on the scaling of depth intervals with a stereoscopic (stereo) television display. A geometrical model of TV stereoscopy is presented which describes the influence of variations of the three video system parameters on retinal disparities. Experiment One investigated the independent and interactive effects of camera interaxial separation and lens magnification on depth interval scaling. Experiment Two investigated the effects of camera convergence angle on depth interval scaling. Experiments Three and Four partially replicated the video system parameters used in Experiment One under more complex scene conditions. For all experiments, ocular fatigue induced by various combinations of video system parameters was also measured. Four trained observers participated in Experiments One and Two. An additional observer participated in Experiments Three and Four. The apparatus used to produce depth intervals in the televised scene was a Howard-Dolman two-rod device. Rods were set to a pre-selected depth interval and observers were required to verbally report perceived depth interval and match the depth interval haptically by means of a sliding peg device. Stereoscopic TV imagery was provided by means of a beamsplitter camera station and a polarizer display station. All experiments included monoscopic (2-D) and direct view control conditions. Ocular fatigue tests which were administered included a questionnaire, flicker fusion threshold adjustments, and a near-far test of oculomotor adjustment time. Multifactorial repeated-measures analyses of covariance were performed on data derived from all experiments with a .05 level set for statistical significance. For Experiments One, Two, and Three stereoscopic imagery produced depth interval estimates which were superior to those found under monoscopic viewing conditions. In addition, increasing camera separation and thereby increasing disparities beyond "natural stereo" values produced improvements in depth interval estimation. Camera convergence exerted a significant effect on performance with camera convergence in front of objects to be compared in depth providing highest accuracy. Lens magnification, which affected retinal disparities, was not found to exert a significant influence on depth interval estimation. For all experiments, no evidence of ocular fatigue was found under any conditions tested. Results are discussed in light of previous studies of depth resolution with stereo TV displays, geometrical models of retinal disparities, and models to account for the discrepancies in depth perception observed between stereo TV and direct viewing conditions.
Thesis (Ph. D.)--University of Hawaii at Manoa, 1984.
Bibliography: leaves 163-170.
xvi, 189 leaves, bound ill. 29 cm
Depth perception, Stereoscopic television
Access Rights
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.