SUBBAND ADAPTIVE FILTERING FOR ACOUSTIC ECHO CONTROL USING ALLPASS POLYPHASE IIR FILTERBANKS

Adaptive filtering in subbands is an attractive alternative to full-ba nd schemes in many applications because of the potential for faster co nvergence and lower computational cost, However, the analysis of a sig nal into a subband representation and the synthesis back into its orig inal full-band form carries three main penalties, These are that 1) th e subsampling process often introduces aliasing, 2) the subband analys is and synthesis processes carry a computational overhead, thereby red ucing the gain in efficiency, and 3) the subband analysis and synthesi s processes introduce delay into the signal path, In this paper, a sub band scheme is presented that;aims to minimize these penalties, thereb y allowing the potential advantages of the subband approach to be more fully realized, The scheme is based on infinite impulse response (IIR ) filterbanks, formed from allpass polyphase filters, which exhibit ve ry high quality filtering compared to typical finite impulse response (FIR) implementations, have relatively low complexity, introduce a lim ited degree of phase distortion and have low delay, The scheme, in con junction with normalized least mean squares (NLMS) adaptive filters, i s tested in an acoustic echo control application and shown to give bet ter convergence, lower delay, and lower computational cost than a comp arable FIR subband scheme.