SMPTE Meeting Presentation: Your Eyes Don't do the Math: Effect of Temporal Display Protocols on Perceived brightness

摘要

It is usually assumed that in temporally multiplexed stereoscopic displays, the perceived binocular brightness is set by the total rate at which photons enter the human visual system, with the timing of these photons being immaterial on the timescales relevant to temporal multiplexing. In order to appear as bright as a 2D system, a perfect temporally multiplexed stereo display would therefore require twice the photon flux, even if it avoided all other losses. Yet this assumption has not been tested in human observers. Here, we examine how brightness perception depends on when photons are delivered to each eye. We find that stimuli appear equally bright when light is presented in alteration to each eye as when it is presented simultaneously in both eyes. Thus, there is no brightness advantage for temporally multiplexed stereo. We do find a novel non-linearity affecting brightness perception in short impulses of light, as in a Cathode Ray Tube. Light pulses presented at low frequency appear brighter than when the same number ofphotons is delivered at high frequency. However, this does not apply to the longer-duration light steps found in modern displays such as Liquid Crystal Displays. We conclude that, when evaluating such stereoscopic displays, one should measure luminance with a single photometer pointing at the screen through the same eyewear as the viewer. The recorded luminance then can be directly compared to that measured by a single photometer for 2D displays. Additionally, we point out that due to the logarithmic relation between physical luminance and perceived brightness in the human visual system, halving luminance does not halve the perceived brightness.