Customized Spatial Sound for Human/Computer Interaction

University Of California-Davis, Davis CA

Investigators

Abstract

Current systems that generate spatial sound for computer-based applications employ Head-Related Transfer functions (HRTFs) and simple models of room reflections to provide the acoustic localization cues. However, the abilities of these systems to generate well-controlled spatial sound streams are quite limited. Although it is possible to position virtual sound sources to the left or right rather accurately, front/back confusion is common, localization in elevation is problematic, and localization in range is unreliable. This proposal describes a comprehensive program of research directed at solving the major problems that are the cause of these limitations. These problems are identified as (a) a mismatch between the HRTF used by the system and the listener's actual HRTF, (b) a failure to provide the correct dynamic cues that occur when the listener moves relative to the source, (c) a mismatch between synthesized room reflections and the listener's experience or expectations, and (d) a failure to render the correct spectral cues for familiar sounds, such as human speech.