Linguistic Effects Based Novel Filter for Hearing Aid to Deliver Natural Sound and Speech Clarity in Universal Environment

Abstract

This paper presents a novel filter for hearing aid to improve the speech clarity and to deliver natural sound in universal environment existing aids cannot achieve reconfigurability to improve the performance of hearing aid. The sound pressure level changes in universal environment with reverberation due to external noise and linguistic effect in turn affects the speech perception. The external noise such as wind, impulse and background noise reduce with existing hearing aid filter. The linguistic effect such as syllable, phonemic, long distance co articulation and vocal effects needs a filter based on the frequency and magnitude of the effects, to deliver the sound with clarity and natural sound. In this article, we propose Frequency Range based Reconfigurable (FRR) Filter bank for hearing aid to improve the speech intelligibility. The FRR filter bank consists of dyadic wavelet, RLS, ISTFT, 1-D Lifting wavelet filters. Each filter in the bank reconfigure swiftly according to sound pressure level in universal environment with reverberation. From the results of 50 samples, we obtain a better speech sound with speech clarity and natural sound of about 95 % when compared with the existing hearing aid speech sound. Experimental results are validated.