Yung-jeng Yan

Publication Date

1998

Document Type

Dissertation/Thesis

First Advisor

Kuo, Sen M. (Sen-Maw)

Degree Name

M.S. (Master of Science)

Legacy Department

Department of Electrical Engineering

LCSH

Echo suppression (Telecommunication); Speech processing systems

Abstract

In telecommunication, the echo canceler using an adaptive filter is the most effective technique for controlling echoes. Because the precision of speech detection significantly affects how the echo path is modeled, this thesis presents the structures of two echo cancelers and two speech detectors. Of the two echo cancelers, one is based on echo path loss (EPL) of 6 dB and the other uses two different echo path models. The speech detectors using a three variable-length windows technique are implemented as the far-end speech detector. Double-talk detectors based on EPL and echo return loss enhancement (ERLE) are also developed and evaluated. Different adaptation algorithms on the adaptive filter are discussed. The normalized partial block least-mean-square (NPBLMS) algorithm is discussed and compared with the normalized least-mean-square (NLMS) algorithm. The results show that the former algorithm achieves a better performance. The center clipper is also introduced and implemented to compress the residual echo. The overall programs have been implemented on the digital signal processor TMS320C54x, where the partial maximum algorithm is also developed for the echo canceler in order to save more computation on DSP program.

Comments

Includes bibliographical references (pages [93]-94)

Extent

xi, 94 pages

Language

eng

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Share

COinS