Speech Enhancement in Non-Stationary Noise Environments: An Efficient Approach
Year: 2009
Type: preprint
Abstract: In speech communication and particularly in mobile voice communication, noise reduction is a vital task for improvement in the quality of communication. Several speech enhancement algorithms available in the literature are mainly for the reduction of noise in stationary environments. In this paper, a two-stage noise reduction algorithm is proposed, which suits both quasi-stationary and non-stationary noise environments. The algorithm is based on two-stage noise filtering, using spectral gain and perceptually motivated weighting techniques. The proposed speech enhancement algorithm and three other algorithms are used to reduce noise in different noisy speech samples, the results of which are compared and discussed.
Source: SSRN Electronic Journal
Institution: CMR University
Language: English
Topic: Speech and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W129366757
