Wavelet-Based Speech Enhancement Using Time-Frequency Adaptation

This paper proposes an improved speech enhancement algorithm based on Wiener-Filtering, which addresses the problems of speech distortion and musical noise. The proposed algorithm adopts the masking properties of human auditory system on calculating the gain of spectrum point, in order that the signal in the enhanced speech whose energy is lower than the threshold will not be decreased further and the less distortion will be brought to enhanced speech by the trade-off between the noise elimination and speech signal distortion. What’s more, in order to eliminate the “musical noise”, a spectrum-shaping technology using averaging method between adjacent frames is adopted. And to guarantee the real-time application, two-stage moving-average strategy is adopted. The computer simulation results show that the proposed algorithm is superior to the traditional Wiener method in the low CPU cost, real-time statistics, the reduction of the speech distortion and residual musical noise.

This paper presents a subband adaptive filter based on wavelet filter-bank for speech enhancement. The adaptation of Kalman filter in wavelet domain has effectively reduced the non-stationary noise. A perceptual weighting filter exploiting the masking properties of psychoacoustic model is concatenated with the Kalman filter to further improve the intelligibility of speech. The proposed method owns its merits from the successful porting of Kalman filter into the wavelet domain so that speech analysis and enhancement can be carried out in time-frequency spectrum based on the human auditory model. Experimental results show that the speech enhancement system is capable of reducing noise with little speech degradation in adverse noise environments and the overall performance is superior to several competitive methods.

Wavelet theory has already achieved huge success. For Secondary Ions Mass Spectrometry (SIMS) signals, denoising the secondary signal, which is altered by the measurement, is considered that an essential step prior to applying such a signal processing technique that aims enhance the SIMS signals.The most efficient and widely used wavelet denoising method is based on wavelet coefficient thresholding. This process involves three important steps; wavelet decomposition: the input signals are decomposed into wavelet coefficients, thresholding: the wavelet coefficients are modified according to a threshold, and reconstruction: the modified coefficients are used in an inverse transform to obtain the noise-free-signal. Several researchers have used thresholding wavelet denoising techniques. The choice of wavelet type and the level of resolution can have a significant influence; it is important to note that the choice of resolution level depends on the type of signal we are dealing with, the nature of the present noise, and our specific goals for the denoised signal. It is generally recommended to test different resolution levels and evaluate their impact on the quality of the denoised signal before making a final decision. Moreover, the results obtained in wavelet denoising can be significantly influenced by the selection of wavelet types. The chosen wavelet type plays a crucial role in the extraction of signal details. Indeed, the effectiveness of denoising the MD6 sample has been demonstrated by the results obtained with sym4, db8, Haar and coif5 wavelets? These wavelets have effectively reduced noise while preserving crucial signal information, leading to an enhancement in the quality of the denoised signal.

An improved speech enhancement scheme is proposed based on log-spectral amplitude minimum mean-square error (LSA-MMSE) method in this paper. Burg spectrum and speech absent probability (SAP) are used to improve estimation of priori signal-to-noise ratios (SNR). The estimation of priori SNR is based on "decision-directed" criterion. Burg spectrum is introduced for noisy speech power spectrum estimation. The SAP is utilized to adjust smoothing parameter. Simulation results show that the improved estimator can well balance de-noising degree, residual musical noise and voice distortion, making enhanced voice sounds more natural.

In this paper, a perceptual subspace speech enhancement method using masking properties of human auditory system is proposed. Noise was estimated at every frame using Speech Presence Probability (SPP). Variance normalization was further done to remove the abrupt changes in the noisy speech signal, removing noise. A combined masking threshold is calculated from simultaneous and temporal masking, which is used to decide the gain parameters for the algorithm. Spectral Domain Constrained (SDC) estimator was employed in determining the filter coefficients. The performance evaluation was done using wcep and WSS objective measures and subjective informal listening test. The results show a superior performance of the proposed method over some of the existing speech enhancement methods.

Classical speech enhancement algorithms often require a good estimation of the short-time power spectrum using, for instance, the periodogram methods. However, it is well known that traditional periodogram methods are prone to induce large variance, hence produces the “musical noise” after enhancement. To alleviate this problem, multitaper spectrum (MTS) estimators with wavelet denoising were proposed. In this paper, we investigate the properties of the MTS of noisy speech signals. We find that, in the log MTS domain, the variance of noise varies according to the magnitude of the underlying speech spectrum. It implies that when applying wavelet denoising to the log MTS, the constant threshold used in the traditional methods is not appropriate. Based on this observation, we further develop a wavelet denoising method with adaptive threshold for estimating power spectrum using multitaper. Simulation results show that the spectrum estimated using the proposed method is consistently more accurate than the traditional uniform thresholding methods. Hence, it further improves the current speech enhancement algorithms using the MTS approaches.

Most algorithms for speech enhancement in the spectral domain focus on the acquisition of an estimator of the clean speech parameter, such as spectrum or amplitude. Enhanced speech quality with these methods mainly depends upon the accuracy of the estimator. When the signal-to-noise ratio (SNR) becomes lower, e.g., less than 3 dB, the enhanced speech often shows unsatisfactory quality. Here, we propose a new method for speech enhancement in the spectral domain. It is based on our view that speech characteristics are perceived mainly by part of the spectral components which have higher local (or instantaneous) SNRs. These components have a special importance on speech SNR enhancement and speech understanding. Identifying these components in the noisy spectrum with a decision system and constructing enhanced speech just by the same noisy amplitude and phase, the result is similar to those popular algorithms. All that is needed is a local SNR decision for each spectral component. The threshold for the decision is derived and two popular methods for local SNR (LSNR) estimation are suggested. It shows better performance under lower SNR signals. Though the number of these kind of spectral components varies with different signal SNR, any algorithms which can pick up more than 90% of it is sufficient to match any existing methods. Primary considerations and results are shown.

Four powder diffraction patterns taken under different experimental conditions were denoised by a new method, i.e., thresholding of wavelet coefficients. The patterns were transformed by discrete wavelet transform applying Coiflet4 wavelet function. WLS refinements of peaks’ positions, FWHM, and intensity showed that wavelet denoising, in contrast to previously used polynomial smoothing, did not shift the maxima and preserved peak and integrated intensities. This method may therefore represent an useful alternative to polynomial filters or filters based on Fourier transform.

Reduction of musical residual noise for speech enhancement using masking properties and optimal smoothing

This study proposes a bi-directional recurrent neural network (Bi-RNN) post-processing method for speech enhancement (SE) at low signal-to noise ratios (SNR). Current speech enhancement solutions performed badly under low SNR situations. Loizou and Kim proposed a solution to reduce speech distortion errors in time-frequency (T-F) domain but it requires the knowledge of ground truth. As ground truth is unknown in real-life applications, the current study proposes to use a Bi-RNN to implement Loizou and Kim’s solution as a post-processing method for SE engines. Our solutions do not require prior knowledge of ground truth. The effectiveness of the proposed method is investigated with a spectral subtraction (SS) SE engine, a non-negative matrix factorization (NMF) SE engine, and a deep neural network ideal ratio mask (DNN-IRM) SE engine, under matched/mis-matched noise and different SNR conditions. Experimental results demonstrate that the proposed post-processing method effectively improved both perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI) for all of these SE engines, especially at low SNR conditions.

This paper proposes a new speech enhancement (SE) algorithm utilizing constraints to the Wiener gain function which is capable of working at 10 dB and lower signal-to-noise ratios (SNRs). The wavelet thresholded multitaper spectrum was taken as the clean spectrum for the constraints. The proposed algorithm was evaluated under eight types of noises and seven SNR levels in NOIZEUS database and was predicted by the composite measures and the SNRLOSS measure to improve subjective quality and speech intelligibility in various noisy environments. Comparisons with two other algorithms (KLT and wavelet thresholding (WT)) demonstrate that in terms of signal distortion, overall quality, and the SNRLOSS measure, our proposed constrained SE algorithm outperforms the KLT and WT schemes for most conditions considered.

This paper compared different wavelet de-noising results when different wavelet bases and thresholds were used. Three kinds of wavelet bases, daubechies, symlets, coiflets, and four thresholds, rigrsure, heursure, sqtwolog and minimaxi, are respectively used in the process of wavelet de-noising. Simulation results show that different selection of the wavelet bases and the thresholds have significant impact on the de-noising results. The evaluation indexes include signal to noise ratio (SNR), mean square error (MSE) and smoothness evaluation functions.

In this paper we propose an efficient single microphone (single channel) speech enhancement (SE) method for Quasi-Stationary noise environment. The proposed method estimates the noise by exploiting its quasi-periodic nature, followed by a statistical model based method to enhance the speech. An efficient reduced-order linear predictive error filtering is introduced to increase the signal to noise ratio (SNR) of the noisy speech. The proposed method is evaluated experimentally by considering the actual recorded Functional Magnetic Resonance Imaging (fMRI) machinery noise which is quasi periodic in nature, added in clean speech. Objective evaluation of our method shows improvement in both quality and intelligibility measures when tested with the sentences chosen from IEEE corpus added in broadband quasi periodic fMRI noise. The proposed method outperforms the standard statistical model based SE technique.

A reliable estimator for speech presence probability (SPP) can significantly improve the performance of many speech enhancement algorithms. Previous work showed that a good SPP estimator can be obtained by using a smooth a-posteriori signal to noise ratio (SNR) function, which can be achieved by reducing the noise variance when estimating the speech power spectrum. In this paper, a wavelet based denoising algorithm is proposed for such purpose. We first apply the wavelet transform to the periodogram of a noisy speech signal to generate an oracle for indicating the locations of the noise floor in the periodogram. We then make use of that oracle to selectively remove the wavelet coefficients of the noise floor in the log multitaper spectrum (MTS) of the noisy speech. The remaining wavelet coefficients are then used to reconstruct a denoised MTS and in turn generate a smooth a-posteriori SNR function. Simulation results show that the new SPP estimator outperforms the traditional approaches and enables a significantly improvement in the quality and intelligibility of the enhanced speeches. © 2012 IEEE.

An effective noise reduction algorithm based on signal analysis in sub-bands is proposed. Speech signal is analyzed in sub-bands using analytic decomposition in hearing critical bands where the envelopes and phases of the analytic signals are used. Noise reduction uses minimum mean square error log-spectral amplitude estimation (MMSE LSAE) noise suppression rule which preserves speech components while producing less musical residual noise. Experiments show that the nature of the band analysis in this work results in less musical noise in comparison with Fourier transform based methods. Moreover noise estimation algorithm in this work is a minima tracking algorithm which significantly reduces the level of musical noise when used in combination with MMSE LSAE.