A New Parametric Adaptive Source Domain Method for Sound Field Synthesis

Abstract

In this brief, we aim at developing an adaptive method that localizes and traces source positions from acoustic signals received by microphone arrays, which provides valuable source domain information for sound field synthesis. The proposed method explores a sparse representation of the sound pressure measurements that expand an overcomplete source domain. We enforce sparsity by imposing L1 regularization on the cost function. A new adaptive method is then formulated to solve the multiple time and frequency samples. Using the sparse source domain information, the sound field can be synthesized. The mean squares performance analysis of the proposed algorithm is carried out and an automatic selection criterion is introduced for the regularization parameter. Experiments are conducted to examine the effectiveness of the method. It shows that the proposed method has improved performance in terms of resolution, tracking capability, and possible frequency aliasing free effect.