Soudeh A. Khoubrouy

Criteria for estimating an FIR filter for cancelling the feedback path signal in hearing aid system

Many adaptive methods use a Finite Impulse Response (FIR) filter to estimate and cancel the effect of acoustic feedback from the loudspeaker to the microphone in a hearing aid system. In these methods the order of the filter is assumed to be fixed and known a priori, even if the feedback path characteristics change from one environment to another. Underestimation of the order of the filter and thus, the filter itself, results in inadequate feedback cancellation, while overestimation of the filter order increases computational complexity, delay, and power consumption without significantly improving cancellation of the feedback path effect. In this paper an adaptive method using short-duration noise injection with new criteria is proposed to recursively find the FIR filter of proper order for estimating the feedback path and cancelling its effect. The proposed method requires no a priori knowledge of the acoustic feedback path or its order. Using the speech and music signals as the desired signals at the input microphone, the performance of the proposed method is analyzed and compared with those based on the convergence of Misalignment (MISA) and Perceptual Evaluation of Speech Quality (PESQ) criteria.

Voice activation detection using Teager-Kaiser energy measure

Voice Activity Detection (VAD) has a wide application in speech coding, speech recognition, and saving the computation and bandwidth of a communication system which uses Voice over Internet Protocol (VoIP). In this paper, a modified Teager-Kaiser Energy (TKE) operator is adapted and used as a VAD feature. The modification includes a combination of different orders of TKE operator which shows good performance and robustness in detecting the speech signal under different Signal to Noise Ratios (SNRs). Simulation results indicate higher accuracy of voice detection using the proposed TKE feature instead of the energy feature usually applied in VAD methods.

Microphone Array Processing Strategies for Distant-Based Automatic Speech Recognition

Robust distant speech recognition (DSR) is necessary in many speech technology applications using multiple microphones but has received only limited treatment in the literature. In this paper, we work on communicating with vehicle voice-controlled system which is one of the applications of DSR. Two approaches for DSR are i) signal-level combination using beamforming followed by automatic speech recognition (ACR), and ii) word hypothesis-level combination using several speech recognition engines followed by confusion network combination or followed by recognizer output voting error reduction (ROVER). In addition to these approaches, it is possible to examine training-level combination by training the recognizer on audio signals from multiple channels (microphones). In this paper, the authors investigate how these methods can be leveraged for in-vehicle ACR using the CU-Move corpus. The authors propose various combinations of these three methods to find an optimum structure for in-vehicle ACR. The authors also investigate the effect of speaker adaptation (SA). The authors experience shows that applying SA on individual channels and merging the results with ROVER reduces the negative effects of SA reported by others in the field, and illustrates the overall improvement obtained with front-end enhancement techniques in DSR.

Howling detection in hearing aids based on generalized Teager-Kaiser operator

With the ongoing miniaturization in the hearing aid industry, acoustical coupling between the loudspeaker and the microphone(s) of the hearing aid causes a major problem to users. Howling is one of the most severe and annoying consequences of this acoustical coupling. This study presents a howling detection method using the Generalized Teager-Kaiser Operator (GTKO). Since the GTKO is both time and frequency sensitive, its resolution parameter must be assigned properly to ensure satisfactory performance of this operator in the frequency range of the input signal for the hearing aid. In order to cover the entire band of the input signal with appropriate resolution parameters, the input signal is decomposed into a filterbank (i.e., uniform and nonuniform filterbanks). GTKO is applied to the output of each band to detect the howling, and the resolution parameter of the GTKO block is selected depending on the central frequency of that particular band. Experimental results compare the performance of each proposed method with two known howling detection approaches, Peak-to-harmonic power ratio (PHPR) approach and a multiple-feature approach. The proposed method has high detection probability and short detection time. It is also shown that considering a hybrid algorithm which includes the PHPR approach with each of the proposed methods (i.e. combination of GTKO blocks with different types of filterbanks) results in lower false alarm probability.

A new delayless sub-band filtering method for cancelling the effect of feedback path in Hearing aid systems

Performance of commonly used Hearing aid systems is degraded by the presence of acoustic feedback between loudspeaker and microphone. Prediction Error Method Adaptive Feedback Canceller (PEM-AFC) has been proposed recently that could attenuate the feedback effect. In this paper, we present a new delayless frequency-based sub-band filtering method for alleviating the effect of feedback path for the Hearing aid systems. The proposed method avoids sub-band distortions and has low computational complexity making it suitable for low-power cost-effective hearing aid system designs. Performances of the two methods are compared and simulation results are presented.

An Efficient Delayless Sub-band Filtering for Adaptive Feedback Compensation in Hearing Aid

Acoustic feedback, or signal leakage between the loudspeaker and microphone of hearing aids, creates some irritations to the users of these devices and degrades the performance of the hearing aids. Among various methods proposed to cancel the negative effect of acoustic feedback, Instrumental Variable Method Adaptive Feedback Canceller (IVM-AFC) has shown superb performance. IVM-AFC exploits two kinds of adaptive filters to pre-filter the signals and to estimate the feedback path transfer function, respectively. This AFC method typically uses Partitioned Block Frequency Domain Normalized Least Mean Square (PBFD- NLMS) algorithm for the feedback path estimation. In this paper, two alternative algorithms are introduced in addition to PBFD-NLMS. One is Discrete Fourier Transform Multi-Delay block Frequency domain NLMS (DFT-MDF-NLMS) algorithm not used for this application before. The other one is a new delayless sub-band filtering algorithm. The algorithms are evaluated using speech as the input of hearing aid. Based on the experimental results, the new sub-band filtering method possesses low computational complexity and high capability of tracking the changes in the feedback path.

Howling detection in hearing aids using Discrete Energy Separation Algorithm-2 and Generalized Teager-Kaiser Operator

Howling is one of the most annoying consequences of the acoustical coupling in hearing aids. In this paper, a method of howling detection is proposed applying Discrete Energy Separation Algorithm-2 (DESA-2) and Generalized Teager-Kaiser Operator (GTKO). Several GTKOs with various resolution parameters monitor the input signal to recognize the howling occurrence while DESA-2 is used to estimate the howling frequency. Performance of the proposed method is compared with two known howling detection approaches, i.e. Peak-to-Harmonic Power Ratio (PHPR) and Adaptive Notch Filter (ANF). Simulation results show the advantage of the proposed method in terms of having lower false detection and shorter detection time.

Evaluation of a new approach for speech enhancement algorithms in hearing aids

Several studies on hearing impaired people who use hearing aid reveal that speech enhancement algorithms implemented in hearing aids improve listening comfort. However, these algorithms do not improve speech intelligibility too much and in many cases they decrease the speech intelligibility, both in hearing-impaired and in normally hearing people. In fact, current approaches for development of the speech enhancement algorithms (e.g. minimum mean square error (MMSE)) are not optimal for intelligibility improvement. Some recent studies investigated the effect of different distortions on the enhanced speech and realized that by controlling the amplification distortion, the intelligibility improves dramatically. In this paper, we examined, subjectively and objectively, the effects of amplification distortion on the speech enhanced by two algorithms in three background noises at different SNR levels.

Order selection of the hearing aid Feedback Canceller filter based on its impulse response energy

Numerous methods have been proposed to cancel the unpleasant effects of acoustic feedback between the loudspeaker and microphone in hearing aid systems. Adaptive Feedback Cancellation (AFC) methods are often used to estimate an FIR filter for cancelling the feedback path effect. In estimating the AFC FIR filter, it is important to select the order of the filter properly; especially when the feedback path changes from one environment to another and no knowledge about it is available. Choosing improper filter order causes deficient system performance or excessive computations and power usage in the system. We present tracking of the energy of AFC FIR filters and its convergence behavior as a new criterion for determining the proper order for AFC FIR filter. Experimental results show validity of the proposed criterion.

A method of howling detection in presence of speech signal

Hearing aid users suffer from howling sound caused by acoustic coupling between the loudspeaker and the microphone(s) of this device. It is crucial to detect and eliminate the howling before it causes serious irritation to the hearing aid user. This study presents a multiple-feature method which uses voice activity detection (VAD) algorithm to reduce false alarm probability. Experimental results compare the performance of the proposed method with three conventional howling detection techniques in terms of detection probability, false alarm probability, and computational complexity. The proposed method possesses lower false alarm probability and less computational complexity compared to the other methods. This study proposes a frame-based howling detection method.The method uses a VAD algorithm in the pre-processing step.Applying VAD reduces computational complexity and false alarm probability.Multiple features of the howling are checked to increase the accuracy.