用于自动语音识别系统的切换语音功率谱估计算法

针对语音识别系统在噪声环境下不能保持很好鲁棒性的问题,提出了一种切换语音功率谱估计算法。该算法假设语音的幅度谱服从Chi分布,提出了一种改进的基于最小均方误差（MMSE）的语音功率谱估计算法。然后,结合语音存在的概率（SPP）,推导出改进的基于语音存在概率的MMSE估计器。接下来,将改进的MSME估计器与传统的维纳滤波器结合。在噪声干扰比较大时,使用改进的MMSE估计器来估计纯净语音的功率谱,当噪声干扰较小时,改用传统的维纳滤波器以减少计算量,最终得到用于识别系统的切换语音功率谱估计算法。实验结果表明,所提算法相比传统的瑞利分布下的MMSE估计器在各种噪声的情况下识别率平均提高在8个百分点左右,在去除噪声干扰、提高识别系统鲁棒性的同时,减小了语音识别系统的功耗。     

联合优化深度神经网络和约束维纳滤波的单通道语音增强方法

深度神经网络（Deep neural networks,DNNs）依靠其良好的特征提取能力,在语音增强任务中得到了广泛应用。为进一步提高深度神经网络的语音增强效果,提出一种将深度神经网络和约束维纳滤波联合训练优化的新型网络结构。该网络首先对带噪语音幅度谱进行训练并分别得到纯净语音和噪声的幅度谱估计,然后利用语音和噪声的幅度谱估计计算得到一个约束维纳增益函数,最后利用约束维纳增益函数从带噪语音幅度谱中估计出增强语音幅度谱作为网络的训练输出。对不同信噪比下的20种噪声进行的仿真实验表明,无论噪声类型是否在网络的训练集中出现,本文方法都能够在有效去除噪声的同时保持较小的语音失真,增强效果明显优于DNN及NMF增强方法。     

基于子带谱熵的仿生小波语音增强

前端噪声处理直接关系着语音识别的准确性和稳定性,针对小波去噪算法所分离出的信号不是原始信号的最佳估计,提出一种基于子带谱熵的仿生小波变换(BWT)去噪算法。充分利用子带谱熵端点检测的精确性,区分含噪语音部分和噪声部分,实时更新仿生小波变换中的阈值,精确地区分出噪声信号小波系数,达到语音增强目的。实验结果表明,提出的基于子带谱熵的仿生小波语音增强方法与维纳滤波方法相比,信噪比(SNR)平均提高约8%,所提方法对噪声环境下语音信号有显著的增强效果。     

基于小波熵自适应阈值的语音信号去噪新方法

针对语音信号去噪问题, 提出小波熵自适应阈值去噪法。首先利用小波变换分解带噪语音信号, 计算小波分解后信号子带区间的小波熵, 然后将小波熵和自适应阈值相结合确定各层高频系数的阈值门限, 采用折中指数阈值函数对各层高频系数进行去噪处理, 重构降噪后的语音信号, 最后对比小波熵自适应阈值、极大极小阈值、固定阈值和无偏风险阈值去噪方法的性能。实验结果表明, 当输入信噪比为5 dB时, 小波熵自适应阈值去噪法的输出信噪比是最大的, 且其输入输出信噪比曲线高于其他三种阈值去噪法的输入输出信噪比曲线, 从而证实该算法具有更好的去噪性能。     

A new a priori SNR estimator based on multiple linear regression technique for speech enhancement

We propose a new approach to estimate the a priori signal-to-noise ratio (SNR) based on a multiple linear regression (MLR) technique. In contrast to estimation of the a priori SNR employing the decision-directed (DD) method, which uses the estimated speech spectrum in previous frame, we propose to find the a priori SNR based on the MLR technique by incorporating regression parameters such as the ratio between the local energy of the noisy speech and its derived minimum along with the a posteriori SNR. In the experimental step, regression coefficients obtained using the MLR are assigned according to various noise types, for which we employ a real-time noise classification scheme based on a Gaussian mixture model (GMM). Evaluations using both objective speech quality measures and subjective listening tests under various ambient noise environments show that the performance of the proposed algorithm is better than that of the conventional methods.     

基于小波包与自适应维纳滤波的语音增强算法

语音增强主要用来提高受噪声污染的语音可懂度和语音质量,它的主要应用与在嘈杂环境中提高移动通信质量有关。传统的语音增强方法有谱减法、维纳滤波、小波系数法等。针对复杂噪声环境下传统语音增强算法增强后的语音质量不佳且存在音乐噪声的问题,提出了一种结合小波包变换和自适应维纳滤波的语音增强算法。分析小波包多分辨率在信号频谱划分中的作用,通过小波包对含噪信号作多尺度分解,对不同尺度的小波包系数进行自适应维纳滤波,使用滤波后的小波包系数重构进而获取增强的语音信号。仿真实验结果表明,与传统增强算法相比,该算法在低信噪比的非平稳噪声环境下不仅可以更有效地提高含噪语音的信噪比,而且能较好地保存语音的谱特征,提高了含噪语音的质量。     

小波包分解下的多窗谱估计语音增强算法

传统谱减法是基于短时傅里叶变换的单一分辨率算法,具有较大方差。为此,提出一种基于小波包分解下的多窗谱估计语音增强算法。将含噪语音在小波包下分解成不同频段,在不同频段下进行多窗谱谱减运算,并逐一进行小波包重构,以得到去噪后的语音信号。仿真结果表明,该算法能提高含噪语音的信噪比,降低语言失真度。     

一种新阈值函数的小波包语音增强算法

针对传统软、硬阈值函数去噪方法增强的语音存在失真的问题,提出一种新阈值函数的小波包语音增强算法,同时给出了新阈值函数和新的Bark尺度小波包分解结构。新阈值函数在小波包系数绝对值大于给定阈值的区间内,灵活地结合了软、硬阈值函数;在小波包系数绝对值小于给定阈值的区间内,用一种非线性函数代替传统阈值函数中的简单置零,实现了阈值函数的平缓过渡;新的60个频带Bark尺度小波包分解结构能更好地模拟人耳的听觉感知特性。仿真实验结果表明,在高斯白噪声和有色噪声背景下,与传统软、硬阈值函数去噪方法相比,新算法有效提高了增强语音信噪比和分段信噪比,减少了语音失真,具有更好的去噪效果。     

A hybrid speech enhancement system with DNN based speech reconstruction and Kalman filtering

In this paper, we propose a hybrid speech enhancement system that exploits deep neural network (DNN) for speech reconstruction and Kalman filtering for further denoising, with the aim to improve performance under unseen noise conditions. Firstly, two separate DNNs are trained to learn the mapping from noisy acoustic features to the clean speech magnitudes and line spectrum frequencies (LSFs), respectively. Then the estimated clean magnitudes are combined with the phase of the noisy speech to reconstruct the estimated clean speech, while the LSFs are converted to linear prediction coefficients (LPCs) to implement Kalman filtering. Finally, the reconstructed speech is Kalman-filtered for further removing the residual noises. The proposed hybrid system takes advantage of both the DNN based reconstruction and traditional Kalman filtering, and can work reliably in either matched or unmatched acoustic environments. Computer based experiments are conducted to evaluate the proposed hybrid system with comparison to traditional iterative Kalman filtering and several state-of-the-art DNN based methods under both seen and unseen noises. It is shown that compared to the DNN based methods, the hybrid system achieves similar performance under seen noise, but notably better performance under unseen noise, in terms of both speech quality and intelligibility.

     

Study on processing of wavelet speech denoising in speech recognition system

The development of society promotes the continuous progress of science and technology, and speech processing technology gradually occupies an increasingly important position in people’s life and work, which puts forward higher requirements on the speech processing technology, especially in noisy environment. Due to the complexity of the real environment, denoising processing has great practical significance. In order to improve the level of speech denoising and increase the accuracy of the speech recognition system, wavelet denoising technology was used to analyze the de-noising requirements and hard and soft threshold functions in the speech recognition system, and an improved wavelet threshold denoising algorithm was put forward. Firstly, the signals were processed by wavelet decomposition according to primary function; then denoising was performed using the improved function; finally the denoised signals were reconstructed using inverse operation. The denoising effect of the algorithm was verified. The results showed that it was effective in denoising conventional speech signals. Besides, it was applied to the speech recognition system to denoise the noisy speech collected in the real environment, and finally high system self-assessment parameters were obtained. Thus it is concluded that wavelet denoising is effective in the speech denoising of the speech recognition system and can be put into practice.     

A novel wavelet seismic denoising method using type II fuzzy

Wavelet based denoising of the observed non stationary time series earthquake loading has become an important process in seismic analysis. The process of denoising ensures a noise free seismic data, which is essential to extract features accurately (max acceleration, max velocity, max displacement, etc.). However, the efficiency of wavelet denoising is decided by the identification of a crucial factor called threshold. But, identification of optimal threshold is not a straight forward process as the signal involved is non-stationary. i.e. The information which separates the wavelet coefficients that correspond to the region of interest from the noisy wavelet coefficients is vague and fuzzy. Existing works discount this fact. In this article, we have presented an effective denoising procedure that uses fuzzy tool. The proposal uses type II fuzzy concept in setting the threshold. The need for type II fuzzy instead of fuzzy is discussed in this article. The proposed algorithm is compared with four current popular wavelet based procedures adopted in seismic denoising (normal shrink, Shannon entropy shrink, Tsallis entropy shrink and visu shrink).It was first applied on the synthetic accelerogram signal (gaussian waves with noise) to determine the efficiency in denoising. For a gaussian noise of sigma = 0.075, the proposed type II fuzzy based denoising algorithm generated 0.0537 root mean square error (RMSE) and 16.465 signal to noise ratio (SNR), visu shrink and normal shrink could be able to give 0.0682 RMSE with 14.38 SNR and 0.068 RMSE with 14.2 SNR, respectively. Also, Shannon and Tsallis generated 0.0602 RMSE with 15.47 SNR and 0.0610 RMSE with 15.35 SNR, respectively. The proposed method is then applied to real recorded time series accelerograms. It is found that the proposal has shown remarkable improvement in smoothening the highly noisy accelerograms. This aided in detecting the occurrence of ‘P’ and ‘S’ waves with lot more accuracy. Interestingly, we have opened a new research field by hybriding fuzzy with wavelet in seismic denoising.     

A novel fast nonstationary noise tracking approach based on MMSE spectral power estimator

Estimating the noise power spectral density (PSD) from the corrupted speech signal is an essential component for speech enhancement algorithms. In this paper, a novel noise PSD estimation algorithm based on minimum mean-square error (MMSE) is proposed. The noise PSD estimate is obtained by recursively smoothing the MMSE estimation of the current noise spectral power. For the noise spectral power estimation, a spectral weighting function is derived, which depends on the a priori signal-to-noise ratio (SNR). Since the speech spectral power is highly important for the a priori SNR estimate, this paper proposes an MMSE spectral power estimator incorporating speech presence uncertainty (SPU) for speech spectral power estimate to improve the a priori SNR estimate. Moreover, a bias correction factor is derived for speech spectral power estimation bias. Then, the estimated speech spectral power is used in “decision-directed” (DD) estimator of the a priori SNR to achieve fast noise tracking. Compared to three state-of-the-art approaches, i.e., minimum statistics (MS), MMSE-based approach, and speech presence probability (SPP)-based approach, it is clear from experimental results that the proposed algorithm exhibits more excellent noise tracking capability under various nonstationary noise environments and SNR conditions. When employed in a speech enhancement system, improved speech enhancement performances in terms of segmental SNR improvements (SSNR+) and perceptual evaluation of speech quality (PESQ) can be observed.     

基于小波变换的语音增强方法研究

分析了小波去噪原理,根据随机噪声的小波变换系数在不同尺度上的传递特性和噪声信号奇异性与小波模极大值的关系,同时考虑到语音中浊音和清音的特点,提出了一种改进阈值的小波域语音增强方法。在阈值函数中引入参数,通过调整参数以获得最佳的小波系数的阈值估计,使得改进阈值介于硬阈值与软阈值之间。利用改进阈值对染噪语音信号的小波系数进行阈值处理,既抑制了噪声,又减少了语音段信息的损失。仿真结果表明,这是一种有效的语音增强方法。     

改进相位谱补偿的语音增强方法

针对传统单通道语音增强方法中用带噪语音相位代替纯净语音相位重建时域信号，使得语音主观感知质量改善受限的情况，提出了一种改进相位谱补偿的语音增强算法。该算法提出了基于每帧语音输入信噪比的Sigmoid型相位谱补偿函数，能够根据噪声的变化来灵活地对带噪语音的相位谱进行补偿；结合改进DD的先验信噪比估计与语音存在概率算法（SPP）来估计噪声功率谱；在维纳滤波中结合新的语音存在概率噪声功率谱估计与相位谱补偿来提高语音的增强效果。相比传统相位谱补偿（PSC）算法而言，改进算法可以有效抑制音频信号中的各类噪声，同时增强语音信号感知质量，提升语音的可懂度。     

An adaptive signal-to-noise ratio estimator in mobile communication channels

Representing the reception condition directly, signal-to-noise ratio (SNR) is an important parameter in mobile propagation channels, and therefore is widely used in system performance evaluations and adaptive applications. Hence, this paper puts forward a frequency domain SNR estimator in mobile communications, where we exploit the signal model with the band-limited fading channel and the additive white Gaussian noise. With the above model, the noise power spectrum density can be estimated from the periodogram of channel-plus-noise signals, subsequently leading to our SNR estimation. Moreover, in order to degrade the intrinsic spectrum leakage of fast Fourier transform in periodogram calculation, the leakage expression is derived analytically and then an adaptive process is proposed to make a tradeoff between leakage reduction and noise smoothing. We verify our algorithm by simulations and observe high accuracy in a wide range of velocities and SNRs. Additionally, unlike the conventional work, the proposed estimator is not strictly based on the assumption of specific Doppler spectral shapes except for the requirement of the band-limited channel, hence it is robust to general mobile channels.     

一种改进的维纳滤波语音增强算法

提出了一种改进的语音增强算法,该算法以基于先验信噪比估计的维纳滤波法为基础。首先通过计算无声段的统计平均得到初始噪声功率谱;其次,计算语音段间带噪语音功率谱,并平滑处理初始噪声功率谱和带噪语音功率谱,更新了噪声功率谱;最后,考虑了某频率点处噪声急剧增大的情况,通过计算带噪语音功率谱与噪声功率谱的比值,自适应地调整噪声功率谱。将该算法与其他基于短时谱估计的语音增强算法进行了对比实验,实验结果表明：该算法能有效地减少残留噪声和语音畸变,提高语音可懂度。     

基于双树复小波二元统计模型的图像去噪方法

为了更有效地进行图像去噪,提出了一种基于双树复小波二元统计模型的图像去噪方法,该方法先用带参数的二元广义高斯分布(GGD)来模拟原图双树复小波系数的统计分布;然后结合最大似然估计(MLE)得到优化的参数估计;最后在此先验分布的基础上,运用最大后验概率(MAP)来估计从噪声图的小波系数中恢复原图的系数,从而达到去噪的目的。实验表明该新方法不仅可以干净地去除图像的噪声,还可以有效地保留图像细节,取得了良好的去噪效果,尤其是去噪图像的视觉效果要明显优于目前的很多算法。     

Reducing over- and under-estimation of the a priori SNR in speech enhancement techniques

Most speech enhancement methods based on short-time spectral modification are generally expressed as a spectral gain depending on the estimate of the local signal-to-noise ratio (SNR) on each frequency bin. Several studies have analyzed the performance of a priori SNR estimation algorithms to improve speech quality and to reduce speech distortions. In this paper, we concentrate on the analysis of over- and under estimation of the a priori SNR in speech enhancement and noise reduction systems. We first show that conventional approaches such as the decision-directed approach proposed by Ephraïm and Malah lead to a biased estimator for the a priori SNR. To reduce this bias, our strategy relies on the introduction of a correction term in the a priori SNR estimate depending on the current state of both the available a posteriori SNR and the estimated a priori one. The proposed solution leads to a bias-compensated a priori SNR estimate, and allows to finely estimating the output speech signal to be very close to the original one on each frequency bin. Such refinement procedure in the a priori SNR estimate can be inserted in any type of spectral gain function to improve the output speech quality. Objective tests under various environments in terms of the Normalized Covariance Metric (NCM) criterion, the Coherence Speech Intelligibility Index (CSII) criterion, the segmental SNR criterion and the Perceptual Evaluation of Speech Quality (PESQ) measure are presented showing the superiority of the proposed method compared to competitive algorithms.     

基于MAP估计双树复小波的电能质量扰动信号去噪方法

针对电能质量信号的去噪,提出了一种基于MAP估计的双树复小波电能质量扰动信号的去噪方法。首先对带噪信号进行相关性预处理,然后通过MAP方法对双树复小波分解不同层次的细节系数估计噪声方差和信号方差,并计算各层阀值从而得到去噪方案,针对带噪的电压跌落等扰动信号进行仿真,并与传统实小波去噪进行了信噪比和突变点信息保留能力的比较。仿真结果表明,所提算法速度快,去噪效果理想,且易于实现,实用性强,有良好的发展前景。     

A perceptually motivated stationary wavelet packet filterbank using improved spectral over-subtraction for enhancement of speech in various noise environments

In this paper, we propose a speech enhancement method where the front-end decomposition of the input speech is performed by temporally processing using a filterbank. The proposed method incorporates a perceptually motivated stationary wavelet packet filterbank (PM-SWPFB) and an improved spectral over-subtraction (I-SOS) algorithm for the enhancement of speech in various noise environments. The stationary wavelet packet transform (SWPT) is a shift invariant transform. The PM-SWPFB is obtained by selecting the stationary wavelet packet tree in such a manner that it matches closely the non-linear resolution of the critical band structure of the psychoacoustic model. After the decomposition of the input speech, the I-SOS algorithm is applied in each subband, separately for the estimation of speech. The I-SOS uses a continuous noise estimation approach and estimate noise power from each subband without the need of explicit speech silence detection. The subband noise power is estimated and updated by adaptively smoothing the noisy signal power. The smoothing parameter in each subband is controlled by a function of the estimated signal-to-noise ratio (SNR). The performance of the proposed speech enhancement method is tested on speech signals degraded by various real-world noises. Using objective speech quality measures (SNR, segmental SNR (SegSNR), perceptual evaluation of speech quality (PESQ) score), and spectrograms with informal listening tests, we show that the proposed speech enhancement method outperforms than the spectral subtractive-type algorithms and improves quality and intelligibility of the enhanced speech.