Enhancement of speech signal in adverse environments by exploiting harmonicity
Date of Issue2016
School of Electrical and Electronic Engineering
Enhancement of speech signal has been an important research field over the past years because of its wide range of applications in the field of medicine, security, aerospace and communication. Numerous methods have been proposed to achieve better speech signal quality. One of the spectral enhancement methods proposed by I. Cohen gives prominent performance under non – stationary noise environments. However, it was observed to degrade in performance when there is sudden increase in noise energy. This short duration noise pulse becomes undesirable for perception and detection of speech signal. The dissertation aims to exploit the spectral feature harmonicity, to remove this transient noise and achieve better speech enhancement. Analysis on simulation results shows removal of undesired noise energy retaining better speech quality for speech estimation.
DRNTU::Engineering::Electrical and electronic engineering