Echo Cancellation for Desktop Conferencing

July–September 1998 Copyright 1998. Lucent Technologies Inc. All rights reserved. Introduction At present, most teleconferencing systems use a single full-duplex audio channel for voice communication. These systems usually employ an acoustic echo canceler (AEC) to remove undesired echoes that result from the coupling between a loudspeaker and a microphone. As these systems evolve to an ever more lifelike and transparent audio/video medium, the need for enhanced sound realism becomes more important. This situation leads to the consideration of multichannel audio, which consists of at least two channels— that is, stereophonic sound. However, before full-duplex stereophonic teleconferencing can be deployed, the AEC problem must first be solved. In this paper, we focus on one particular application: multiparticipant stereo desktop conferencing. With single-channel sound, simultaneous talkers are overlaid and it is difficult to concentrate on one particular voice. On the other hand, by using our binaural auditory system together with multichannel presentation, we can concentrate on one source to the exclusion of others (the so-called cocktail party effect). Moreover, localization helps us identify which person is actually talking. This is a very difficult task in a monophonic presentation. Communication using stereo (or multichannel) sound likely will grow rapidly in the near future, especially over the Internet. The general scenario is as follows. Several persons in different locations would like to communicate with each other, and each one of them has a workstation. Each participant would like to see on the screen pictures of the other participants arranged in a reasonable fashion and to hear them in perceptual space in a way that facilitates identification and understanding. For example, the voice of a participant whose picture is located on the left of the screen should appear to come from the left. We assume that we are located in a hands-free environment, where the composite acoustic signal is presented over loudspeakers. This study will be limited to two channels, so we assume that each workstation is equipped with two loudspeakers (one on each side of the screen) and one microphone (somewhere on ♦ Synthesized Stereo Combined with Acoustic Echo Cancellation for Desktop Conferencing Jacob Benesty, Dennis R. Morgan, Joseph L. Hall, and M. Mohan Sondhi