Microphone array post-filtering for non-stationary noise suppression

We present a multi-microphone speech enhancement approach for minimizing the log-spectral amplitude (LSA) distortion in nonstationary noise environments. An adaptive beamformer with a generalized sidelobe canceller structure is applied to the noisy observed signals. In addition to the beamformer primary output, it provides reference noise signals by projecting the input signals onto the noise-only subspace. Presumably, a desired signal component is stronger at the beamformer output than at any reference noise signal, and a noise component is strongest at one of the reference signals. Hence, the ratio between the transient power at beamformer output and the transient power at the reference signals indicates whether such a transient is desired or interfering. Based on a Gaussian statistical model, and an appropriate decision-directed a priori SNR estimate, we derive an estimator for the signal presence probability. This estimator controls the rate of recursive averaging in obtaining a noise spectrum estimate by the minima controlled recursive averaging (MCRA) approach. Subsequently, spectral enhancement of the beamformer output is achieved by applying an optimal gain function, which minimizes the MSE of the log-spectra.

A large amount of neurophysiological research is based on the study of biological neural populations. The data is gathered from extra-cellular recordings with multi-electrode arrays (MEAs). The signal is a stream containing an unknown number of neural sources superpositioned with non-stationary noise components. In order to analyze the recorded spike trains, a prior separation into its individual components is required. The increasing number of sensors on a MEA surface demand an automatic spike sorting procedure. In this article the proposed spike sorting method replaces the manual steps with artificial neural networks to enable an unsupervised separation of mixed signal components. Furthermore the artificial neural network based feature extraction allows realtime processing and ensures high flexbibility and adaptivity for unknown signals with non-stationary noise components.

Supervised single channel dual domains speech enhancement using sparse non-negative matrix factorization

A noise reduction technique to reduce background noise in noisy speech is proposed. We have proposed the noise reduction method which uses a noise reconstruction system. However, since a residual speech signal is included in the input signal of a noise reconstruction filter (NRF) used for reconstructing the background noise, the long time average value of error signal for estimating the background noise is needed not to estimate the speech signal. Therefore, the ability of tracking the non-stationary noise is decreased. In order to solve this problem, we propose the noise reconstruction system with adaptive line enhancer (ALE). Since ALE works to obtain the signal occupied by noise components, the input signal of the NRF includes only a few speech components. Therefore, we can give the high tracking ability to NRF.

Laguerre escalator lattice and feedforward/feedback active noise control

Reducing highly non-stationary transient noise, such as keyboard typing noise, remains a challenging problem for many singlechannel speech enhancement algorithms. This paper proposes two approaches based on nonnegative matrix factorization (NMF) and probabilistic latent component analysis for transient noise reduction using a pre-trained transient noise dictionary and a universal speaker-independent speech dictionary. In addition, we develop an NMF-based speech enhancement scheme to simultaneously reduce transient and non-transient background noise, in which a low-dimensional dictionary is learnt from the noisy observations to model the background noise. We exploit the temporal dependencies of speech and background noise to design and apply informative priors via a probabilistic framework, while ignoring the temporal dynamics of the transient noise. Experimental results show that the proposed algorithms can improve the perceptual evaluation of speech quality (PESQ) up to 1.2 MOS for the keyboard typing noise.

The real-time implementation of the existing multi-channel Wiener filter (MWF) algorithms suffer from performance degradation due to the lack of robustness against estimation errors of the second-order statistics. The reasons are twofold: one, the estimation of the statistics relies on real voice activity detector (VAD), which often fails in adverse environments. Second, the MWF solutions involve estimation of the second order clean speech statistics, which also exaggerates the errors. This paper presents an MWF algorithm that requires neither VAD nor clean speech statistics. Performance evaluation under real scenarios shows that the proposed method outperforms the conventional MWF solution in terms of the trade-off between noise reduction and speech distortion.

Spectral subtraction based on two-stage spectral estimation and modified cepstrum thresholding

A novel feedforward hybrid active sound quality control algorithm for both narrowband and broadband sound-profiling

In this paper, a multi-channel post-filtering approach in reverberant environment based on detection and estimation scheme is presented. A modified Signal presence probability (SPP), which is in consideration of reverberation, is proposed with a novel estimator Direct-toreverberate ratio (DRR) to adapt to distant-talking scene. SPP is known a key estimator to instruct the updating of transient noise or residual directional interference and form gain function in the time-frequency domain, consequently a new desired signal detection scheme is proposed to improve its accuracy. Appropriate spectral enhancement technique is applied to the noisy speech signal taking advantage of the modified SPP estimator. The proposed multi-channel post-filtering is tested in different nonstationary noisy and reverberant environments. Experimental results show that it achieves considerable improvement on signal preservation of the desired speech with more noise reduction over the comparative algorithms.

Random walk noise is an important component in the output phase noise of a voltage-controlled oscillator (VCO). The relation between this non-stationary noise component and the induced timing jitter is critical for VCO performance characterization. Rigorous derivation of the relation is provided in this paper, which corrects the error in the relation obtained under wide-sense stationary assumption. Besides, the correct expression of power spectral density of the output phase noise is obtained. Additionally, discussion is provided in this paper on the measurement principle to obtain power spectral density of non-stationary phase noise.

Harmonics of power system include steady and transient components, extraction and analysis of them can improve the power quality. Tunable Q-factor wavelet transform (TQWT), for which the Q-factor can be easily specified, is adopted to accomplish the extraction of non-stationary harmonic component from the noisy signal. The proposed method decomposes a signal into noise component and non-stationary harmonic component based on suitable Q-factor. Feasibility and effectiveness of the proposed method are verified by the simulation study.

<p/> This paper introduces two short-time spectral amplitude estimators for speech enhancement with multiple microphones. Based on joint Gaussian models of speech and noise Fourier coefficients, the clean speech amplitudes are estimated with respect to the MMSE or the MAP criterion. The estimators outperform single microphone minimum mean square amplitude estimators when the speech components are highly correlated and the noise components are sufficiently uncorrelated. Whereas the first MMSE estimator also requires knowledge of the direction of arrival, the second MAP estimator performs a direction-independent noise reduction. The estimators are generalizations of the well-known single channel MMSE estimator derived by Ephraim and Malah (1984) and the MAP estimator derived by Wolfe and Godsill (2001), respectively.

Many approaches of Gene Expression Pro ling (GEP) measurements were suggested. However, mostmethods exhibit signi cant gross errors. There is a need to study and improve precision of measure-ment methods. This work did it for a speci c method, i.e., the multiplex PCR with color-taggedmodule-shuing primers [3, 1, 2]. It was designed for the precise measurement of small di erences ofgene expression. However, this method performs the single measurement of each gene in each sample.Such data cannot be used to assess precision. The original method used three dyes each measuringdi erent sample. We used this three dyes for the same sample to assess measurement quality. Thisrevealed many outliers (wrong measurements), i.e., measurements di ering more than twice from themeasured value. This paper analyses this three-sample GEP measuring system from a statisticalviewpoint. The aim is (1) to determine the precision of measurement and the sources of errors, (2) todevelop the method which identi es outliers, and (3) suggests modi cations to improve precision andreproducibility. The data processing was optimized for the given technology. However, the principleslike the transposition of uorescent dyes for outlier detection, data projection for its normalization,and clustering (EM algorithm, Gaussian statistical model) can be applied to other technologies too.

The maximum correntropy Kalman filter (MCKF) is an effective algorithm that was proposed to solve the non-Gaussian filtering problem for linear systems. Compared with the original Kalman filter (KF), the MCKF is a sub-optimal filter with Gaussian correntropy objective function, which has been demonstrated to have excellent robustness to non-Gaussian noise. However, the performance of MCKF is affected by its kernel bandwidth parameter, and a constant kernel bandwidth may lead to severe accuracy degradation in non-stationary noises. In order to solve this problem, the mixture correntropy method is further explored in this work, and an improved maximum mixture correntropy KF (IMMCKF) is proposed. By derivation, the random variables that obey Beta-Bernoulli distribution are taken as intermediate parameters, and a new hierarchical Gaussian state-space model was established. Finally, the unknown mixing probability and state estimation vector at each moment are inferred via a variational Bayesian approach, which provides an effective solution to improve the applicability of MCKFs in non-stationary noises. Performance evaluations demonstrate that the proposed filter significantly improves the existing MCKFs in non-stationary noises.