Single channel speech enhancement techniques for removal of additive noise

Abstract

In addition to the known applications of voice communication , cell phones , hands-free voice recording , automatic speech recognition , interactive voice response systems , man-machine interface , requires at least a microphone and signals are usually corrupted by noise reverberation and background noise . So , the voice signal must be " clean " by digital signal processing tool to be played before storage or transmission . This article describes an speech enhancement technology for single voice channel. Various strategies have been proposed and are described from human to simulate and compare the past in this article . But there are still some obstacles to overcome these methods . Therefore, the aim is to modify or in combination with a single voice channel from the real world , a significant noise, such as airports , cars , restaurants , train stations and other inhibiting characteristics of different algorithms voices improvement efforts improvisation to outperform traditional methods , because the failure is made to develop a new algorithm. The algorithm is based on simulation of objective and subjective evaluation.