Subband adaptive filtering for acoustic feedback compensation in hearing aids

Abstract

The use of subband adaptive filtering for acoustic feedback compensation in hearing aids are discussed. The outputs of the filters are replicas of the subband-decomposed probe signal that are subtracted from the true feedback probe signal measured at the microphone of the hearing aid. The resulting error is used in the normalized least-mean square (NLMS) algorithm to adjust the adaptive filter coefficients to the true feedback path. Subband adaptive filters are preferred over other filters, as they offer more flexibility in dealing with disturbing speech. The performance of feedback compensation on these subband signals allows the system to achieve better tracking of changing feedback paths and offers additional flexibility in dealing with performance-stability challenges.