August Kaelin

Transpose properties in the stability and performance of the classic adaptive algorithms for blind source separation and deconvolution

Abstract This paper presents a tutorial review of the problem of Blind Source Separation (BSS) and the properties of the classic adaptive algorithms when either the score-function or a general (nonscore) nonlinearity is employed in the algorithm. In new findings it is shown that the separating solution for both sub- and super-Gaussian signals can be stabilized by an algorithm employing any given nonlinearity. For these separating solutions the steady-state error levels are also given in terms of the nonlinearity and the pdfs of the source signals. These results show that a transpose symmetry exists between the nonlinear algorithms for sub- and super-Gaussian signals. The behavior of the algorithm is then detailed when the ideal score-function nonlinearity is replaced by a general (hard saturation or u3) nonlinearity. The phases of convergence to decorrelated output signals and then to recovery of the source signals are explained. The results are then extended to single- and multi-channel deconvolution and shown by analysis and extensive simulation to hold for mixed and convolved source signals. The results allow the design of stable algorithms for multichannel blind deconvolution with a general nonlinearity when sub- and super-Gaussian source signals are present.

Subband signal processing for hearing aids

In this paper a hearing aid concept working in subbands is presented. Consideration is given to the compensator for recruitment of loudness and the acoustic feedback cancellation. In addition, a step-size control for fast tracking and a constant error level with nonstationary input signals is presented. The feedback canceler is adapted to the feedback path in the transform domain using a power-normalized least mean square (LMS) algorithm. The transformation into subbands is based on the modulated lapped transform (MLT). Together with the feedback canceler the proposed hearing-loss compensating forward filter is computationally very efficient.

Adaptive feedback cancelling in subbands for hearing aids

In this paper a hearing aid concept with recruitment of loudness compensation and acoustic feedback cancellation is presented. Special consideration is given to the acoustic feedback canceler which uses only the available (e.g. speech) input signal for adaptation. In principle, the feedback canceler is adapted to the feedback path in the transform domain using a power-normalized least mean square (LMS) algorithm. The transformation into uniform subbands is based on an augmentation of the modulated lapped transform (MLT). Together with the hearing-loss compensating forward filter the proposed feedback canceler is computationally very efficient.

A DSP implementation of a digital hearing aid with recruitment of loudness compensation and acoustic echo cancellation

This paper describes a DSP implementation of a digital hearing aid realized in the frequency domain that compensates for recruitment of loudness and cancels acoustic echos. In contrast to conventional systems which are based on a noise-probe signal, our echo canceler is adapted using only the available (e.g. speech) input signal. The main problems caused by a nonlinear feedforward filter are discussed using analytical results of the steady state behavior of the closed-loop hearing-aid system. The implemented DSP system is tested with a dummy behind-the-ear (bte) hearing-aid device on a KEMAR-head and results are presented.

Performance comparison of PBFDAF algorithms

This paper compares the feedback cancellation performance of different partitioned block frequency-domain adaptive filter (PBFDAF) algorithms. The comparison is made for the error level of the echo canceler of an open and closed loop system of a hearing instrument setup. Special consideration is given to the different transformations and the number of partitions. The results enable a computationally efficient design of a hearing instrument for any combination of feedback reduction performance of the predictor and band spacing requirements for the hearing-loss compensator.