Determining the accuracy of sound field synthesis systems in reproducing subjective source locations

Sound field synthesis systems developed and applied to study human hearing perception differ in terms of the number and arrangement of loudspeakers in rooms of different sizes, and methods used to generate virtual sound environments. Research has evaluated how well such systems physically reproduce room acoustic conditions, but limited subjective data are available. This paper seeks to develop a method for determining the accuracy of the perceived localization of a virtual sound source generated using a multichannel sound synthesis system. A test method is applied to the sound field synthesis facility at Boys Town National Research Hospital, a room (5.8 m x 5.2 m x 2.7 m) with reverberation time of 0.16 s at 125 Hz and below 0.024 s at 250 Hz and above. Short bursts of broadband speech-shaped noise are presented at a number of virtual source locations under free-field and modeled reverberant-room conditions, and listeners are asked to point to the subjective source location. Subjective localization results are compared as functions of virtual sound location, and parameters of early reflections and reverberation in the modeled sound environment. Results are intended to guide future research on subjective room acoustics relevant to children's’ communication needs. [Work supported by NIH GM109023.]