基于HFM和感知后滤波器的语音增强

研究了只能获得带噪信号的情况下的语音增强问题。将语音信号看作由高斯噪声激励的自回归(AR)过程,观测噪声为加性高斯白噪声,把信号转化为状态空间模型。首先用隐马尔可夫模型(HMM)估计AR参数和噪声的方差作为卡尔曼滤波器初值,估计信号作为参数估计的中间值给出,然后将估计信号通过一个感知滤波器平滑以消除残余噪声。仿真结果表明该算法有良好的性能。     

一种新的含噪混沌信号降噪算法

该文针对低信噪比、非高斯加性噪声和混沌动力学系统参数未知的含噪混沌信号降噪问题,提出了一种基于粒子滤波(Particle Filtering, PF)的降噪新算法。该算法将混沌信号和动力学系统中的未知参数作为一个多维状态矢量,利用PF方法递推计算多维状态矢量的联合后验概率分布,进而实现了对混沌信号的最优估计。对于混沌信号轨道分离过快所导致的退化问题,提出了有效的解决方法,并利用核平滑和自回归(Auto-Regression, AR)模型建模的方法分别实现了非时变以及时变参数的递推估计。仿真实验的结果表明,与现有的降噪方法相比,该文提出的新算法能够更加有效地抑制含噪混沌信号中的加性噪声。     

Speech Enhancement Using Perceptual Wavelet Packet Decomposition and Teager Energy Operator

It has been shown in the literature that the perceptual wavelet packet decomposition (PWPD) and the Teager energy operator (TEO) are useful for various speech processing systems and speech enhancement applications, respectively. By the use of the PWPD and the TEO, this paper presents an improved wavelet-based speech enhancement method. The main advantage of the proposed method is that the over thresholding of speech segments which is usually occurred in conventional wavelet-based speech enhancement schemes can be avoided. As a consequence, the enhanced speech quality of the proposed method can be increased substantially from those of conventional approaches. In addition, the proposed method does not require a complicated estimation of the noise level or any knowledge of the SNR. Using speech signals corrupted by additive and real noises, experimental results demonstrate that the speech enhancement method presented in this paper is capable of outperforming conventional noise cancellation schemes.     

一种基于自适应噪声估计的宽带语音增强算法

为解决传统算法对噪声适应性较差,残留音乐噪声较强的问题,本文提出了一种基于自适应噪声估计的宽带语音增强算法。该算法可应用于宽带语音编码器,以提升在噪声环境下的编码质量。本文所提算法利用谱熵对噪声类型进行有效的判别,将背景噪声分为白噪声和有色噪声两类,并根据噪声特性选择适当的噪声估计方法。在白噪声背景下,选择一种谱平滑的方法;在有色噪声背景下,则选择经典的最小值控制递归平均算法。在此基础上结合经典的统计模型方法,构建一种具有较强噪声鲁棒性的宽带语音增强算法。在ITU-T G.160标准下对算法进行性能测试,测试结果表明,在不同强度的背景噪声环境下,增强语音的信噪比提高都较为明显。同时,在低信噪比情况下,该算法有效的抑制了严重影响听觉质量的音乐噪声现象。      

Constrained iterative speech enhancement with application to speechrecognition

The basis of an improved form of iterative speech enhancement for single-channel inputs is sequential maximum a posteriori estimation of the speech waveform and its all-pole parameters, followed by imposition of constraints upon the sequence of speech spectra. The approaches impose intraframe and interframe constraints on the input speech signal. Properties of the line spectral pair representation of speech allow for an efficient and direct procedure for application of many of the constraint requirements. Substantial improvement over the unconstrained method is observed in a variety of domains. Informed listener quality evaluation tests and objective speech quality measures demonstrate the technique's effectiveness for additive white Gaussian noise. A consistent terminating point of the iterative technique is shown. The current systems result in substantially improved speech quality and linear predictive coding (LPC) parameter estimation with only a minor increase in computational requirements. The algorithms are evaluated with respect to improving automatic recognition of speech in the presence of additive noise and shown to outperform other enhancement methods in this application     

Smoothing approach using forward–backward Kalman filter with Markov switching parameters for speech enhancement

In this paper, a smoothing approach for enhancing speech signals degraded by statistically independent additive nonstationary noise is developed. The autoregressive hidden Markov model (ARHMM) is used for modeling the statistical characteristics of both the clean speech signal and the nonstationary noise process. In this case, the speech enhancement comprises a weighted sum of the conditional mean estimators for the composite states of the models for the speech and noise, where the weights are equal to the posterior probabilities of the composite states, given the noisy speech. The conditional mean estimators use a smoothing approach based on two Kalman filters with Markovian switching coefficients, where one of the filters propagates in the forward-time direction and the other propagates in the backward-time direction with one frame. The proposed method is tested on speech signals degraded by Gaussian colored noise or nonstationary noise at various input signal-to-noise ratios. An approximate improvement of 4.7–5.2 dB in SNR is achieved at input SNR 10 and 15 dB. Also, in comparison with conventional method (Ephraim, IEEE Trans. Signal Process. SP-41 (April 1992) 725–735), our proposed method shows improvement of about 0.3 dB in SNR.     

Parametric SNR Estimation Based on Auto-Regressive Model in AWGN Channels

Signal-to-noise ratio（SNR）estimation for signal which can be modeled by Auto-regressive（AR）process is studied in this paper.First,the conventional frequency domain method is introduced to estimate the SNR for the received signal in additive white Gauss noise（AWGN）channel.Then a parametric SNR estimation algorithm is proposed by taking advantage of the AR model information of the received signal.The simulation results show that the proposed parametric method has better performance than the conventional frequency doma in method in case of AWGN channel.     

Recursive parameter estimation for noisy autoregressive signals (Corresp.)

The problem of recursively estimating the unknown parameters of a scalar autoregressive (AR) signal observed in additive white noise, including signal power and noise variance, is considered. A state-space model in a canonical but noninnovations form is used to represent the noisy AR signal. An algorithm based on a system identification/parameter estimation technique known as the recursive prediction error method is presented for recursive parameter estimation. Two simulation examples illustrate the effectiveness of the proposed algorithm.     

基于概率耦合的双直接判决先验信噪比估计算法

直接判决（DD,Decision-Directed）算法结构简单、音乐噪声抑制能力较好,是当前语音增强领域最为常用的先验信噪比估计方法.但该算法对于滑动因子的选取数值较为敏感,且估计性能要受到时延问题的限定.本文首先采用实际的语音和噪声数据,根据音乐噪声残留及输出语音失真两方面的评测标准对DD算法中滑动因子的取值问题进行了研究,通过数据分析给出了其较为明确的上下边界值;然后基于语音及噪声信号的复高斯分布模型,采用软判决技术对两个具有不同滑动因子的DD算法进行概率耦合,提出了一种具有双DD结构的先验信噪比估计算法.该算法可以充分结合两个具有不同特性DD算法的优点,在音乐噪声抑制及限制语音失真等方面均获得了较为理想的输出效果.多种噪声背景及输入信噪比条件下的仿真结果表明,相对于目前流行的几种先验信噪比估计算法,本文提出算法具有更为优良的估计性能.     

Dual H∞ Algorithms for Signal Processing— Application to Speech Enhancement

This paper deals with the joint signal and parameter estimation for linear state-space models. An efficient solution to this problem can be obtained by using a recursive instrumental variable technique based on two dual Kalman filters. In that case, the driving process and the observation noise in the state-space representation for each filter must be white with known variances. These conditions, however, are too strong to be always satisfied in real cases. To relax them, we propose a new approach based on two dual H_infin filters. Once a new observation of the disturbed signal is available, the first H_infin algorithm uses the latest estimated parameters to estimate the signal, while the second H_infin algorithm uses the estimated signal to update the parameters. In addition, as the H_infin filter behavior depends on the choice of various weights, we present a way to recursively tune them. This approach is then studied in the following cases: (1) consistent estimation of the AR parameters from noisy observations and (2) speech enhancement, where no a priori model of the additive noise is required for the proposed approach. In each case, a comparative study with existing methods is carried out to analyze the relevance of our solution.     

A Novel Technique for the Identification of ARMA Systems Under Very Low Levels of SNR

In this paper, a novel technique for the identification of minimum-phase autoregressive moving average (ARMA) systems from the output observations in the presence of heavy noise is presented. First, starting from the conventional correlation estimator, a simple and accurate ARMA correlation (ARMAC) model in terms of the poles of the ARMA system is presented in a unified manner for white noise and impulse-train excitations. The AR parameters of the ARMA system are then obtained from the noisy observations by developing and using a residue-based least-squares correlation-fitting optimization technique that employs the proposed ARMAC model. As for the estimation of the MA parameters, it is preceded by the application of a new technique intended to reduce the noise present in the residual signal that is obtained by filtering the noisy ARMA signal via the estimated AR parameters. A scheme is then devised whereby the task of MA parameter estimation is transformed into a problem of correlation-fitting of the inverse autocorrelation function corresponding to the noise-compensated residual signal. In order to demonstrate the effectiveness of the proposed method, extensive simulations are performed by considering synthetic ARMA systems of different orders in the presence of additive white noise and the results are compared with those of some of the existing methods. It is shown that the proposed method is capable of estimating the ARMA parameters accurately and consistently with guaranteed stability for signal-to-noise ratio (SNR) levels as low as $-{5}~{hbox {dB}}$ . Simulation results are also provided for the identification of a human vocal-tract system using natural speech signals showing a superior performance of the proposed technique in terms of the power spectral density of the synthesized speech signal.      

Zero-crossing based spectral analysis and SVD spectral analysis forformant frequency estimation in noise

The authors discuss a method for spectral analysis of noise corrupted signals using statistical properties of the zero-crossing intervals. It is shown that an initial stage of filter-bank analysis is effective for achieving noise robustness. The technique is compared with currently popular spectral analysis techniques based on singular value decomposition and is found to provide generally better resolution and lower variance at low signal to noise ratios (SNRs). These techniques, along with three established methods and three variations of these method, are further evaluated for their effectiveness for formant frequency estimation of noise corrupted speech. The theoretical results predict and experimental results confirm that the zero-crossing method performs well for estimating low frequencies and hence for first formant frequency estimation in speech at high noise levels (~0 dB SNR). Otherwise, J.A. Cadzow's high performance method (1983) is found to be a close alternative for reliable spectral estimation. As expected the overall performance of all techniques is found to degrade for speech data. The standard autocorrelation-LPC method is found best for clean speech and all methods deteriorate roughly equally in noise     

基于听觉掩蔽效应和噪声估计的语音增强算法

谱减法在增强语音、提高信噪比的同时,残留的音乐噪声较大.在利用听觉掩蔽闻值对谱减系数进行修正的基础上,采用实时噪声估计来减少谱减法噪声估计误差,并对谱减后的语音信号进行感知滤波来进一步抑制残留音乐噪声.实验结果表明,该算法能去除噪声,增强语音,并在不影响信噪比的同时降低语音失真测度值.主观测听表明语音音质有明显提高.     

基于稀疏性的相位谱补偿语音增强算法

相位谱补偿语音增强算法通过调整相位谱对噪声进行压缩,提高重构信号的质量。针对传统的相位谱补偿(phase spectrum compensation, PSC)语音增强算法采用固定的相位补偿因子,且算法的性能易受噪声估计准确性的影响,提出了一种基于稀疏性的相位谱补偿(sparsity-based phase spectrum compensation, SPSC)语音增强算法。首先,利用噪声估计算法得到噪声幅度谱,利用基于幅度谱的语音增强算法得到目标语音幅度谱;接着,通过噪声和目标语音幅度谱之间的局部信噪比(Signal-to-Noise Ratio, SNR)来估计谱时间稀疏性;然后,利用sigmoid函数改进相位补偿因子,联合补偿因子和谱时间稀疏性,得到SPSC函数。最后,使用SPSC函数对相位谱中的谱分量进行补偿,通过短时傅里叶逆变换得到最终增强后的语音信号。仿真实验表明,在四种不同背景噪声的低信噪比下,新的相位谱补偿算法使增强语音获得了更好的LSD、PESQ和segSNR指标,说明新的算法在低信噪比下,可以有效恢复带噪语音中的语音成分,对噪声抑制效果明显,增强语音的质量和听感均有一定提升。      

一种融合相位估计的深度卷积神经网络语音增强方法

在时频域的语音增强中,幅度估计和相位估计都是影响语音增强性能的重要因素.为了在基于深度学习的语音增强方法中融合对相位的估计,本文将含噪语音短时傅里叶变换（STFT）的实部和虚部特征作为两个通道输入深度卷积神经网络,通过建立一个同步估计纯净语音STFT的实部和虚部特征的多任务学习模型,实现了对幅度和相位的同步估计.实验结果表明,相比仅考虑幅度估计的方法,本文方法具有更好的噪声抑制能力,在低信噪比条件下,显著提高了语音增强性能.     

基于自适应共轭梯度参数估计的顽健卡尔曼滤波语音增强算法

对于加性噪声影响下的语音信号，利用双通道输入建立起来的增广卡尔曼滤波器模型，采用自适应共轭梯度方法对纯净语音和有色噪声干扰模型分别进行参数估计，提出了一种有效的语音增强算法。由于该方法对模型参数的估计精确性较高，而且估计速度快，同卡尔曼滤波类的其它语音增强方法相比，其语音增强效果良好，且具有一定的顽健性。仿真实验表明在环境噪声很复杂的情况下，该方法仍然有效。     

一种基于最小统计噪声估计的改进谱减法

针对传统最小统计噪声估计谱减法在低信噪非平稳噪声环境下存在的噪声估计误差较大及搜索延迟较大的问题,对带噪语音进行频域平滑然后进行全局平滑,并且提出使用双向搜索的方法,减小了搜索延迟.实验结果表明,文中提出的方法有效地实现了语音增强,具有一定的使用价值.     

Inverse filtering based method for estimation of noisy autoregressive signals

In this paper we present a new method for estimating the parameters of an autoregressive (AR) signal from observations corrupted with white noise. The least-squares (LS) estimate of the AR parameters is biased when the observation noise is added to the AR signal. This bias is related to observation noise variance. The proposed method uses inverse filtering technique and Yule-Walker equations for estimating observation noise variance to yield unbiased LS estimate of the AR parameters. The performance of the proposed unbiased algorithm is illustrated by simulation results and they show that the performance of the proposed method is better than the other estimation methods.     

基于无语音概率改进的对数谱幅度估计增强算法

针对谱减语音增强法中一直存在的去噪度、残留的音乐噪声和语音畸变度三者间均衡这一关键问题,本文提出一种基于无语音概率改进的对数谱估计增强算法.该算法结合无语音概率的思想,按照纯噪声帧和带噪语音帧两种状态.有区别地实时更新语音最小均方误差的对数谱增益,并利用无语音概率参数(SAP)自适应地调节平滑系数,以求随着噪声环境的变化,在去噪度、残留"音乐噪声"和语音畸变度之间自适应地折中.实验表明,该算法在相同去噪程度下,语音畸变和音乐噪声相对其他谱减法都同时地减弱,特别在低信噪比环境下优势更明显,而且平滑参数利用SAP参数,无需多余计算,便于实时处理.     

Feature classification criterion for missing features mask estimation in robust speaker recognition

Currently, many speaker recognition applications must handle speech corrupted by environmental additive noise without having a priori knowledge about the characteristics of noise. Some previous works in speaker recognition have used the missing feature (MF) approach to compensate for noise. In most of those applications, the spectral reliability decision step is performed using the signal to noise ratio (SNR) criterion, which attempts to directly measure the relative signal to noise energy at each frequency. An alternative approach to spectral data reliability has been used with some success in the MF approach to speech recognition. Here, we compare the use of this new criterion with the SNR criterion for MF mask estimation in speaker recognition. The new reliability decision is based on the extraction and analysis of several spectro-temporal features from across the entire speech frame, but not across the time, which highlight the differences between spectral regions dominated by speech and by noise. We call it the feature classification (FC) criterion. It uses several spectral features to establish spectrogram reliability unlike SNR criterion that relies only in one feature: SNR. We evaluated our proposal through speaker verification experiments, in Ahumada speech database corrupted by different types of noise at various SNR levels. Experiments demonstrated that the FC criterion achieves considerably better recognition accuracy than the SNR criterion in the speaker verification tasks tested.