Generalizeability of Latency Detection in a Variety of Virtual Environments

Abstract

User perceptual sensitivity to changes of system latency was tested in three simple virtual environments: one with only a foreground object, a second with only a background object, and a third that combined both of these elements. Prior psychophysical measurements of sensitivity, Just Noticeable Difference; and bias, Points of Subjective Equality, from our laboratory are confirmed with measurements in 13 subjects. Our measurements indicate that perceptual stability across a variety of virtual environments will require latencies less than 16 ms. We discount a possible explanation that the differences between our results and those from a study by Allison et al. could be related to a visual capture effect initially reported by L. Matin. Instead, the differences may be due to the type of psychophysical judgment rendered by the subjects and the degree to which subjects were instructed and practiced.