Abstract

An adaptive filtering approach is proposed in this paper for Stereophonic Acoustic Echo Cancellation (SAEC) in order to obtain convergence rate improvement and computational complexity reduction. The hybrid mono/stereo structure between Fast Affine Projection (FAP) and Frequency-domain Normalized Least Mean Square (F-NLMS) algorithms is introduced, based on subband adaptive filtering. Simulation results on speech signals suggest that enhanced performance in terms of rate of convergence can be achieved at low computational cost.