Browse Prior Art Database

Tandem Operating Echo Suppressor and Echo Canceller

IP.com Disclosure Number: IPCOM000043289D
Original Publication Date: 1984-Aug-01
Included in the Prior Art Database: 2005-Feb-04
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Carsalade, D: AUTHOR [+3]

Abstract

In a voice transmission system wherein hybrid transformers are used for converting two-wire transmission paths into four-wire transmission paths and vice-versa, echoes are often generated which disturb the transmission. Two conventional means are used to solve the echo problem. The first one completely suppresses the speech signal suspected of including echo, the other cancels the echo from the return path. Echo cancelling requires the synthesis of an echo replica using an adaptive digital filter. Such a filter needs to be initialized, i.e., its coefficients have to be set at initial values to shorten the filter adapting operations as much as possible. Initialization is usually performed by sending a predetermined training sequence, and analyzing the returning sequence to derive information on the echo therefrom.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 76% of the total text.

Page 1 of 1

Tandem Operating Echo Suppressor and Echo Canceller

In a voice transmission system wherein hybrid transformers are used for converting two-wire transmission paths into four-wire transmission paths and vice-versa, echoes are often generated which disturb the transmission. Two conventional means are used to solve the echo problem. The first one completely suppresses the speech signal suspected of including echo, the other cancels the echo from the return path. Echo cancelling requires the synthesis of an echo replica using an adaptive digital filter. Such a filter needs to be initialized, i.e., its coefficients have to be set at initial values to shorten the filter adapting operations as much as possible. Initialization is usually performed by sending a predetermined training sequence, and analyzing the returning sequence to derive information on the echo therefrom. Needless to say, in most cases, such an initialization process should be performed prior to beginning any voice traffic. Also, the longer the echo path, the longer the delay between sending the training sequence and receiving the echo. Therefore, echo suppressors are better being limited to medium range echo paths (e.g., 32 msec delays). This means that echoes occurring with a delay greater than 32 msec will not be taken into account.

Such a situation may be improved by mixing echo cancelling, operating on medium range echoes, with echo suppressing controlling high delay range. The basic algorithm would...