Evaluation of Sparse Sound Field Models for Compressed Sensing in Multiple Sound Zones
To implement a number of sound field reconstruction methods, it is often necessary to get a measure of room impulse responses (RIRs) of a region of interest. However, in many cases this requires a time-consuming effort due to repeated measuring processes. Compressed sensing can provide an alternative solution to obtain RIRs at any location in a domain. In this article we evaluate two different sparse sound field models and a compressed sensing algorithm for the creation of multiple sound zones. RIR estimates are obtained from the sparse models and used to derive the optimal loudspeaker filters. The experimental study indicates a significant improvement of the sound zone system performance from 300 to 3000 Hz using a reduced amount of RIRs.
Authors: Leny Vinceslas (Loughborough University London), Hyun Lim (Loughborough University London) and Ahmet Kondoz (Loughborough University London)
