Perceptual thresholds of spatial audio update latency in virtual auditory and audiovisual environments

Abstract

When observers in a virtual sound environment are in motion relative to a source, the virtual-auditory display must rapidly track the users head position and update the location-cueing acoustic filters—known as head-related transfer functions (HRTFs)—in order to accurately reflect the source’s location relative to the current head orientation and position. The end-to-end spatial audio system latency (SASL) is the time elapsed between the listener assuming a position and sound being delivered to the ears with an HRTF corresponding to that same position. To maintain the stability and plausibility of a virtual sound “object,” the SASL must lie below a perceptual temporal threshold. The current study sought to rigorously probe the threshold of SASL detectability in human observers at three different rotational head velocities and at various source locations. These conditions were repeated in an audiovisual experiment in which the sound stimulus was now accompanied by a zero-latency visual stimulus. Thus, we characterized SASL thresholds in environments where cross-modal interactions occur due to visual information accompanying sound as is often the case in virtual reality (VR) applications. Initial results reveal a rich interaction between the contextual factors of the virtual environment and listeners’ sensitivity to SASL.