Ted S. Wada

On Dealing with Sampling Rate Mismatches in Blind Source Separation and Acoustic Echo Cancellation

The lack of a common clock reference is a fundamental problem when dealing with audio streams originating from or heading to different distributed sound capture or playback devices. When implementing multichannel signal processing algorithms for such kind of audio streams it is necessary to account for the unavoidable mismatches between the actual sampling rates. There are some approaches that can help to correct these mismatches, but important problems remain to be solved, among them the accurate estimation of the mismatch factors, and achieving both accuracy and computational efficiency in their correction. In this paper we present an empirical study on the performance of blind source separation and acoustic echo cancellation algorithms in this scenario. We also analyze the degradation in performance when using an approximate but efficient method to correct the rate mismatches.

Acoustic echo cancellation based on independent component analysis and integrated residual echo enhancement

This paper examines the technique of using a memoryless noise-suppressing nonlinearity in the adaptive filter error feedback-loop of an acoustic echo canceler (AEC) based on normalized least-mean square (NLMS) when there is an additive noise at the near-end. It will be shown that introducing the nonlinearity to “enhance” the filter estimation error is well-founded in the information-theoretic sense and has a deep connection to the independent component analysis (ICA). The paradigm of AEC as a problem that can be approached by ICA leads to new algorithmic possibilities beyond the conventional LMS family of techniques. In particular, a right combination of the error enhancement procedure and a properly implemented regularization procedure enables the AEC to be performed recursively and continuously in the frequency domain when there are both ambient noise and double-talk even without the double-talk detection (DTD) or the voice activity detection (VAD) procedure.

Enhancement of Residual Echo for Robust Acoustic Echo Cancellation

This paper examines the technique of using a noise-suppressing nonlinearity in the adaptive filter error feedback-loop of an acoustic echo canceler (AEC) based on the least mean square (LMS) algorithm when there is an interference at the near end. The source of distortion may be linear, such as local speech or background noise, or nonlinear due to speech coding used in the telecommunication networks. Detailed derivation of the error recovery nonlinearity (ERN), which “enhances” the filter estimation error prior to the adaptation in order to assist the linear adaptation process, will be provided. Connections to other existing AEC and signal enhancement techniques will be revealed. In particular, the error enhancement technique is well-founded in the information-theoretic sense and has strong ties to independent component analysis (ICA), which is the basis for blind source separation (BSS) that permits unsupervised adaptation in the presence of multiple interfering signals. The single-channel AEC problem can be viewed as a special case of semi-blind source separation (SBSS) where one of the source signals is partially known, i.e., the far-end microphone signal that generates the near-end acoustic echo. The system approach to robust AEC will be motivated, where a proper integration of the LMS algorithm with the ERN into the AEC “system” allows for continuous and stable adaptation even during double talk without precise estimation of the signal statistics. The error enhancement paradigm encompasses many traditional signal enhancement techniques and opens up an entirely new avenue for solving the AEC problem in a real-world setting.

Batch-Online Semi-Blind Source Separation Applied to Multi-Channel Acoustic Echo Cancellation

Semi-blind source separation (SBSS) is a special case of the well-known blind source separation (BSS) when some partial knowledge of the source signals is available to the system. In particular, a batch adaptation in the frequency domain based on independent component analysis (ICA) can be effectively used to jointly perform source separation and multichannel acoustic echo cancellation (MCAEC) through SBSS without double-talk detection. Many issues related to the implementation of an SBSS system are discussed in this paper. After a deep analysis of the structure of the SBSS adaptation, we propose a constrained batch-online implementation that stabilizes the convergence behavior even in the worst case scenario of a single far-end talker along with the non-uniqueness condition on the far-end mixing system. Specifically, a matrix constraint is proposed to reduce the effect of the non-uniqueness problem caused by highly correlated far-end reference signals during MCAEC. Experimental results show that high echo cancellation can be achieved just as the misalignment remains relatively low without any preprocessing procedure to decorrelate the far-end signals even for the single far-end talker case.

Inter-channel decorrelation by sub-band resampling in frequency domain

This paper presents a novel decorrelation procedure by frequency-domain resampling in sub-bands. The new procedure expands on the idea of resampling in the frequency domain that efficiently and effectively alleviates the non-uniqueness problem for a multi-channel acoustic echo cancellation system while introducing minimal distortion to the signal. We show in theory and verify experimentally that the amount of decorrelation in each sub-band, measured in terms of the coherence, can be controlled arbitrarily by varying the resampling ratio per frequency bin. For perceptual evaluation, we adjust the sub-band resampling ratios to match the coherence given by other decorrelation procedures. The speech quality (PESQ) score from the proposed decorrelation procedure remains high at around 4.5, which is about the highest possible PESQ score after signal modification.

Coherent spectral estimation for a robust solution of the permutation problem

In this paper, we propose a new method based on a coherent source spectral estimation for solving the permutation problem of frequency-domain blind source separation (BSS). It combines the robustness of the State Coherence Transform (SCT) to recursively estimate a smooth phase spectrum associated with each source and the precision of the inter-frequency correlation to solve for a correct permutation. Namely, the TDOAs estimated by the SCT are used to constrain the permutation correction process in order to force the resulting filters to be coherent across frequency. This intrinsic interconnection between the TDOA information and the spectral correlation makes the new approach robust even when the signal is short in duration and spatial aliasing is substantial. Experimental results show that the proposed method is able to drastically reduce the number of permutation errors for three sources recorded in a short time block using microphones with large spacing.

Decorrelation by resampling in frequency domain for multi-channel acoustic echo cancellation based on residual echo enhancement

An inter-channel decorrelation procedure via resampling in the frequency domain for multi-channel acoustic echo cancellation (MCAEC) based on residual echo enhancement is proposed. The objective is to efficiently alleviate the non-uniqueness problem while introducing minimal distortion to the audio quality and the signal statistics. The effectiveness is illustrated with respect to the standard approach of using a memoryless nonlinearity or additive noise. A combination of the new decorrelation procedure and the residual echo enhancement technique points towards a computationally feasible yet very robust frequency-domain MCAEC system.

Towards robust acoustic echo cancellation during double-talk and near-end background noise via enhancement of residual echo

This paper examines the technique of using a noise suppressing nonlinearity in the adaptive filter error feedback loop of the acoustic echo canceler (AEC) based on the least mean square (LMS) algorithm when there are both double-talk and white background noise at the near-end. By combining the previously introduced noise suppressing technique with a compressive nonlinearity derived from the theory of robust statistics, consistently better results are obtained during double-talk as well as during single-talk when compared to the traditional approach of using only the compressive nonlinearity. It is shown that a compressive form of noise reducing nonlinearity can be derived also from the signal enhancement point of view when the noise probability density (pdf) is tailed more heavily and has a higher kurtosis than the Gaussian pdf. A combination of such a noise compressing nonlinearity and a noise suppressing nonlinearity is capable of producing results that are similar to that of the robust statistics approach during double-talk along with an added benefit of increased robustness during single-talk when there is only the background noise.

A system approach to residual echo suppression in robust hands-free teleconferencing

This paper presents a system approach to the residual echo suppression (RES) problem in a noisy acoustic environment. We propose a method that takes advantage of our existing robust acoustic echo cancellation system in order to obtain a residual echo estimate that closely resembles the true, noise-free residual echo. To achieve improved RES during strong near-end interference (e.g., double talk), a psychoacoustic postfilter is also used. The simulation results show that our RES based on the system approach outperforms a conventional estimation method. Comparing the postfiltered output to the unprocessed one indicates that our proposed RES approach can raise the PESQ score by more than half a point.

Enhancement of Residual Echo for Improved Frequency-Domain Acoustic Echo Cancellation

This paper explores the technique of integrating a noise suppressing nonlinearity to the adaptive filter error feedback loop of the frequency-domain acoustic echo canceler (AEC) when there is a continuous additive noise at the near-end. It was shown in a previous study that in the time-domain AEC case, both the echo return loss enhancement (ERLE) and the misalignment from using normalized least-mean-squares (NLMS) can be improved through the error enhancement procedure. By applying the same technique to the frequency-domain AEC, namely the generalized multidelay filter (GMDF), the misalignment can be decreased by around 5 dB on average in a numerical simulation, whereas the ERLE can be increased by 2 dB on average and by over 5 dB at some instances in a real acoustic environment. New results further support the idea that reducing the effects of distortion to the cancellation error helps to improve the performance of an adaptive filter.