A biased multichannel adaptive algorithm for room equalization

The use of adaptive algorithms in multichannel equalization has become essential to compensate room effects of real sound reproduction systems. Due to the high complexity and number of compensation filters that involve these multiple-input multiple-output (MIMO) systems, a compromise has to be taken to provide good equalization without increasing the complexity of the adaptive algorithm. The impulse responses of a multichannel equalizer are usually long and exhibit a high (or unknown) degree of sparsity, which results in least-mean-square (LMS) type algorithms showing slow convergence speed. Recently, proportionate adaptive schemes have been introduced to accelerate filter convergence and to exploit sparsity in echo cancellation and active noise control systems. Moreover, it is possible to reduce the error of the adaptive filters by biassing the weights, specially under low signal-to-noise ratio condition. In this paper we propose a biased proportionate adaptive algorithm for multichannel room equalization in several scenarios. Experimental results show that the proposed adaptive algorithm significantly outperforms the traditional LMS based ones.