The detectability of headtracker latency in virtual audio displays

Abstract

A critical engineering parameter in the design of interactive virtual audio displays is the maximum amount of latency that can be tolerated between the movement of the listener's head and the corresponding change in the spatial audio signal presented to the listener's ears. In this study, subjects using a virtual audio display were asked to detect the difference between a control stimulus that had the lowest possible latency value for the display system (11.7 ms) and a test stimulus that had an artificially increased headtracker latency ranging from 36 to 203 ms. In a standard listening configuration with only a single virtual sound source, the results show that typical listeners are unable to reliably detect the presence of headtracker latencies smaller than 80 ms, and that even the best listeners are unable to detect changes smaller than 60 ms. However, the addition of low-latency reference tone at the same location of the target signal decreases the minimum threshold for latency detection by about 25 ms. This result suggests that augmented reality systems may require headtracker latencies smaller than 30 ms to ensure the delays are undetectable to all users in all listening environments.