Two-Channel Forward Set-Membership Normalized Subband Algorithm for Acoustic Noise Reduction

This paper presents a two-channel forward set-membership normalized subband (TC-F-SM-SB-NLMS) algorithm for acoustic noise reduction. By incorporating set-membership technique, the computational complexity in updating the two cross-filters is reduced, while combining it with subband normalized adaptive filtering ensures high noise reduction performance. The proposed algorithm is compared with the two-channel forward normalized least mean square (TC-F-NLMS), its set-membership variant (TC-F-SM-NLMS), and the two-channel forward subband normalized least mean square (TC-F-SB-NLMS) algorithms. Simulation results demonstrate that the proposed algorithm outperforms these algorithms in terms of computational complexity and segmental signal-to-noise ratio (Seg-SNR) for acoustic noise reduction.