A hybrid mono/stereo acoustic echo canceler

Abstract

In many applications, such as teleconferencing, multimedia workstations, televideo gaming, etc., stereo sound is already, or will soon be, implemented to give spatial realism that mono systems cannot offer. In such hands-free systems, stereophonic acoustic echo cancelers are absolutely necessary for full-duplex communication. In this paper, we propose a new acoustic echo canceler (AEC) based on a fundamental experimental observation that the stereo effect is due mostly to sound energy below about 1000 Hz. The principle of the hybrid mono/stereo AEC is to use stereophonic sound with a stereo AEC at low frequencies (e.g., below 1000 Hz) and monophonic sound with a conventional mono AEC at higher frequencies (e.g., above 1000 Hz). This solution is a good compromise between the complexity of a full-band stereo AEC and spatial realism. For the stereo case, we borrow from a previous innovation and add a small nonlinearity into each channel in order to accurately identify the two receiving room impulse responses. Index Terms-Acoustic echo cancelation, adaptive filter, stereo. I. INTRODUCTION I N MANY applications, such as teleconferencing, multimedia workstations, televideo gaming, etc., stereo sound is already, or will soon be, implemented to give spatial realism that mono systems cannot offer. In such hands-free systems, stereophonic acoustic echo cancelers (AEC's) are absolutely necessary for full-duplex communication [1]. Stereophonic acoustic echo cancellation is, in principle, a generalization of the single-channel case [1]. Fig. 1 shows this technique for one microphone in the receiving room on the left; similar analysis will apply to the other microphone signal. As we can see from the diagram in Fig. 1, the two microphone signals in the transmission room on the right are obtained by filtering from a common source. This gives rise to a nonuniqueness problem that does not arise for the single-channel AEC [1]. As a result, the usual adaptive algorithms converge to solutions that depend on the impulse responses in the transmission room. This means that for good echo cancellation one must track not only the changes in the receiving room, but also the rapid changes in the transmission room (for example, when one person stops talking and another person starts). Note that several authors have proposed structures using only a single adaptive filter [2], [3], but these methods do not address the convergence problem discussed in [1].

A fast exact weighted subband adaptive algorithm and its application to mono and stereo acoustic echo cancellation

“…(34) (or Eq. (35)) and (14). The stereophonic case is characterized by the strong correlation between x 1 ðnÞ and x 2 ðnÞ resulting in the appearance of two additional terms at the end of Eq.…”

“…Stereophonic teleconferencing provides more natural acoustic perception than monophonic teleconferencing [7,14]. With the ongoing prevalence of high-speed processors and the introduction of efficient low-complexity adaptive algorithms, stereophonic acoustic echo cancellation has become an achievable task those days and has been implemented successfully [9].…”

“…This results in a very ill-conditioned correlation matrix, and consequently very slow convergence of gradient-based adaptive algorithms. In practical situations, the adaptive filter orders are lower than the receiving room impulse responses giving rise to an additional performance problem of large bias in the adaptive filter coefficients solution [1,7,14].…”

“…The source signal sðnÞ is zero mean white Gaussian with unity variance. The transmission room impulse responses G 1 and G 2 are each of length 4096, measured in a teleconferencing room [7,14]. The receiving room impulse responses W Ã 1M and W Ã 2M are also of length 4096, measured in a teleconferencing room, and all at 16 kHz sampling rate.…”

A new fast Newton‐type adaptive filtering algorithm for stereophonic acoustic echo cancellation (SAEC)

“…It is shown that the only solution to the non-uniqueness problem is to reduce the correlation between the stereo signals, and an efficient low-complexity method for this purpose was also given in [8]. Reference [10] presents a combination of mono and stereo echo cancellations, which has the benefit of lower complexity than a pure stereo solution. Lately, attention has been focused on the investigation of other methods that decrease the cross-correlation between the channels in order to get well-behaved estimates of the echo paths [11].…”