Spatial sampling versus acquisition time of room impulse responses for low-frequency sound zones

Abstract

Sound zone methods aim to create multiple listening areas within a room, allowing independent playback of different audio content. Achieving this requires a loudspeaker array, multiple microphones per sound zone, and a set of control filters, which design involves the room impulse responses (RIRs) between loudspeakers and microphones. Under certain acoustic conditions, acquiring these RIRs is possible using very short acquisition times. At lower frequencies, longer wavelengths enable spatial sampling of the sound zones with fewer points, reducing the number of microphone positions required and amount of data to process. This study exploits this characteristic to render and evaluate low-frequency sound zones in two real rooms, with few spatial sampling points. We assess the performance by objectively evaluating multiple spatial sampling configurations and RIR acquisition times. Furthermore, we evaluate the system's ability to control sound beyond the sampled regions by assessing the sound field over wider areas. For sound zones roughly the size of a human head, satisfactory performance is achieved in low reverberation conditions with at least two microphones per ear, even for RIR acquisition times as short as 150 μs. Conversely, the performance degradation in highly reflective environments cannot be compensated using additional microphones nor longer acquisition times.