Performance Analysis of Inbuilt Hearing Aid using Signal Enhancement by Deep Learning

Abstract

Audio Signal processing is a method that uses intensive algorithms that are applied to audio signals. Audio signals are in the form of both analog and digital signals and they are the typical representation of sound. The frequency of audio ranges from 20Hz to 20,000 Hz, and 20Hz is the lower limit of our ears and 20,000Hz is the upper limit of our ears. The process of audio signal processing gives the desired audio by removing the unwanted noise from the speech signal. This process balances the time and frequency range. This process also aims on commutative methods by altering sounds and removes echo, unwanted noise and over modulation. Recent literatures focus on removal of noise from the audio signal. We are dealing with enhancing the quality of speech. Speech consists of various noises such as stationaries noises and non-stationary noises. Several strategies are proposed which are based on Deep learning and Deep Neural Networks to overcome this problem. The main goal of the paper is improvement in the quality of speech signals that are corrupted by noise. This will enhance the performance of digital hearing aid using Deep Neural Networks before it delivers to the needy people and also to measure and analyze the emotion of speech.