Real-time Implementation of Binary Mask Algorithm for Hearing Prosthetics

Swati R. Pawar; Hemantkumar B. Mali

doi:10.37591/jomsd.v3i1.5204

Abstract

As the current speech enhancement algorithms can give results for improved speech audibility only. So, our main motivation is to develop speech enhancement algorithm that would improve performance in speech for hearing impaired persons. In this paper, an algorithm is presented for improving speech intelligibility. Various speech enhancement algorithms were developed but only some of them can be used for real time hearing aid applications. This proposed algorithm can be used for practical hearing prosthetic devices. Implementation of the binary asking algorithm uses a bank of band-pass filters to perform mapping of signals. Also, classification is performed with a signal-to-noise (SNR) estimate and a comparator. This includes spatial filtering method, classification of signals such as original and noisy signal. After this based on SNR threshold, level signals are recombined to obtain reduced noise level in speech signal. In this, Matlab implementation of binary mask algorithm is provided which shows better results for speech intelligibility as compared to other algorithms.