A Study on Noise Reduction Algorithms for speech signal in Hearing Aids

Abstract

Today there is many people having hearing problems. The main complaint of people with hearing loss is the low ability to deduce speech in a noisy environment. The aim is to study various noise reduction techniques and compare them and arrive at a conclusion regarding which one of them is most suitable for enhancing speech signals.  The hearing aid is a device that can acquire, process and feedback the acoustic signal in real time. In this case, several impedance matching algorithms, various filter bank techniques, signal processing algorithm and echo cancellation are discussed. The FFT algorithm is used to make this application, which is quite easy to implement and requires less computation. The different Noise reduction methods are – Spectral subtraction, Adaptive Filter based on LMS Algorithm, Weiner Filter, Chebyshev Filter and Kalman Filter. The purpose of noise reduction algorithms is to provide a quality and clarity voice signal in background noise, interference sources and reverberation effects. Most single channel algorithms operate in the time frequency domain, while multichannel algorithms rely on spatial filtering.