A semisoft thresholding method based on Teager energy operation on wavelet packet coefficients for enhancing noisy speech

In this paper, a speech enhancement method based on noise compensation performed on short time magnitude spectra is presented following the geometric approach. Here the noise estimate to be subtracted from the noisy speech spectrum is proposed to be determined by exploiting the low frequency regions of noisy speech of the current frame rather than depending only on the initial silence frames. We argue that this approach of noise estimation offers the capability of tracking the time variation of the non-stationary noise. A noise estimation confidence parameter is also introduced to prevent the subtraction of the overestimated noise. A procedure is thus formulated to compensate for the distortion in the magnitude spectrum of the noisy speech signal. The noise compensated magnitude spectrum is then recombined with the noisy unchanged phase spectrum to produce a modified complex spectrum prior to synthesize an enhanced frame. Extensive simulations are carried out using NOIZEUS database in order to evaluate the performance of proposed method. It is shown in terms of objective measures, spectrogram analysis and informal subjective listening test that the proposed method consistently outperforms one of the state-of-the-art methods of speech enhancement from noisy speech corrupted by babble or car noise at high as well as very low levels of SNR.

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.

Exemplar-based techniques, where the noisy speech is decomposed as a linear combination of the speech and noise exemplars stored in a dictionary, have been successfully used for speech enhancement in noisy environments. This paper extends this technique to achieve speech dereverberation in noisy environments by means of a nonnegative approximation of the noisy reverberant speech in the frequency domain. A novel approach for estimating the room impulse response (RIR) together with the speech and noise estimates using a non-negative matrix deconvolution (NMD)-based technique is proposed. In addition, we extend an existing technique based on nonnegative matrix factorisation (NMF) that performs speech derever-beration in noise-free environments to noisy scenarios. New estimators for jointly obtaining the RIR and exemplar weights for the NMD and NMF-based formulations are presented. The proposed techniques are evaluated on the noise-free and noisy reverberant speech in the CHiME-2 WSJ0 database and are shown to yield better speech enhancement in terms of signal-to-distortion ratio (SDR), perceptual evaluation of speech quality (PESQ) and cepstral distance (CD) measures.

Recently, deep neural networks (DNNs) have become the mainstream strategy for speech enhancement task because it can achieve the higher speech quality and intelligibility than the traditional methods. However, these DNN-based methods always need a large number of parallel corpus consisting of clean speech and noise to produce noisy data for the training of the DNN in order to improve the generalization of the network. As a result, this implies that many noisy speech signals that are collected in real environment cannot be used to train the DNN because of the lack of corresponding clean speech and noise. Additionally, as we know, noise varies with the time and scenario, so we cannot obtain parallel speech and noise due to infinite noise data and some limited speech data. Thus, the network training with unparallel speech and noise data is essential for the generalization of the network. To address this problem, we propose a novel parallel-data-free speech enhancement method, in which the cycle-consistent generative adversarial network (CycleGAN) and multi-objective learning are employed. Our method is also able to make best use of the benefits of multi-objective learning. On the training stage, we utilize two different encoders to encode the features of clean speech and noisy speech, respectively. Then, two forward generators are immediately used to predict the ideal time-frequency (T-F) mask and log-power spectrum (LPS) of clean speech. Two inverse generators are applied to map the magnitude spectrum (MS) and LPS of noisy speech, respectively. In addition, four discriminators are used to distinguish the real speech features from the generated features. Two encoders, four generators and four discriminators are simultaneously trained by using adversarial, identity-mapping, latent similarity and cycle-consistent loss. On the test stage, we directly utilize the forward generators and encoders to acquire the enhanced speech. The experimental results indicate that the proposed approach is able to achieve the better speech enhancement performance than the reference methods. Moreover, the proposed method is also effective to improve speech quality and intelligibility when the networks are trained under the parallel data.

In this paper, a hidden Markov model (HMM)-based cue parameters estimation method for single-channel speech enhancement is proposed, in which the cue parameters of binaural cue coding (BCC) are applied to single-channel speech enhancement system successfully. First, the clean speech and noise signals are considered as the left and right channels of stereo signal, respectively; and the noisy speech is treated as the down-mixed mono signal of BCC method. According to the clean speech and noise data set and the corresponding noisy speech data set, the clean cue parameters and pre-enhanced cue parameters are extracted, respectively. Then the cue HMM is trained offline, which exploits the a priori information about the clean cue parameters and the pre-enhanced cue parameters for speech enhancement. Next, using the trained cue HMM, the clean cue parameters are estimated from noisy speech online. Finally, following the synthesis principle of BCC cue parameters, the speech estimator is constructed for enhancing noisy speech. The test results demonstrate that, for the segmental signal-noise-ratio (SNR), the log spectral distortion and PESQ measures, the proposed method performs better than the reference methods.

Noisy speech enhancement based on an adaptive threshold and a modified hard thresholding function in wavelet packet domain

This paper introduces a new speech enhancement algorithm based on the adaptive threshold of intrinsic mode functions (IMFs) of noisy signal frames extracted by empirical mode decomposition. Adaptive threshold values are estimated by using the gamma statistical model of Teager energy operated IMFs of noisy speech and estimated noise based on symmetric Kullback–Leibler divergence. The enhanced speech signal is obtained by a semisoft thresholding function, which is utilized by threshold IMF coefficients of noisy speech. The method is tested on the NOIZEUS speech database and the proposed method is compared with wavelet-shrinkage and EMD-shrinkage methods in terms of segmental SNR improvement (SegSNR), weighted spectral slope (WSS), and perceptual evaluation of speech quality (PESQ). Experimental results show that the proposed method provides a higher SegSNR improvement in dB, lower WSS distance, and higher PESQ scores than wavelet-shrinkage and EMD-shrinkage methods. The proposed method shows better performance than traditional threshold-based speech enhancement approaches from high to low SNR levels.

Speech is easily interfered by external environment in reality, which results in the loss of important features. Deep learning has become a popular speech enhancement method because of its superior potential in solving nonlinear mapping problems for complex features. However, the deficiency of traditional deep learning methods is the weak learning capability of important information from previous time steps and long-term event dependencies between the time-series data. To overcome this problem, we propose a novel speech enhancement method based on the fused features of deep neural networks (DNNs) and gated recurrent unit (GRU). The proposed method uses GRU to reduce the number of parameters of DNNs and acquire the context information of the speech, which improves the enhanced speech quality and intelligibility. Firstly, DNN with multiple hidden layers is used to learn the mapping relationship between the logarithmic power spectrum (LPS) features of noisy speech and clean speech. Secondly, the LPS feature of the deep neural network is fused with the noisy speech as the input of GRU network to compensate the missing context information. Finally, GRU network is performed to learn the mapping relationship between LPS features and log power spectrum features of clean speech spectrum. The proposed model is experimentally compared with traditional speech enhancement models, including DNN, CNN, LSTM and GRU. Experimental results demonstrate that the PESQ, SSNR and STOI of the proposed algorithm are improved by 30.72%, 39.84% and 5.53%, respectively, compared with the noise signal under the condition of matched noise. Under the condition of unmatched noise, the PESQ and STOI of the algorithm are improved by 23.8% and 37.36%, respectively. The advantage of the proposed method is that it uses the key information of features to suppress noise in both matched and unmatched noise cases and the proposed method outperforms other common methods in speech enhancement.

In this paper a cumulant based Teager energy operator (TEO) method is proposed to enhance the morphological features of noisy stress ECG signals which are corrupted with Gaussian noise, and other artifacts like baseline wander. It uses higher order statistical (HOS) cumulants in combination with nonlinear TEO required for the analysis of nonlinear and non stationary signals. Higher order cumulants suppress the Gaussian noise and TEO derives an energy from nonlinear components in time and frequency domains. It exhibits an attractive feature of retrieving a clean ECG from noisy recordings with extremely poor SNR of up to 10 dB. The mean of Teager energy (TE) in time as well as frequency domains and improvement in signal to noise ratio (SNRI) are the statistical measures used for classification. Results reveal the efficacy of the proposed method in improving the signal to noise ratio and computing the energy in time and frequency domains.

The vibration signals of diesel include excess noise that must be eliminated before extraction of characteristic parameters. Firstly, the effects of vibration-signal de-noising among Fourier transform, wavelet decomposition and wavelet packet decomposition are compared. Secondly, singular value decomposition is applied to de-noising vibration signals. Finally, a new de-noise method integrated with wavelet packet and singular value is presented. In this method, vibration signals are decomposed by wavelet packet, and the wavelet packet coefficient is de-noised by singular value decomposition again. The results indicate that the new de-noising method is the best. The SNR (signal-to-noise ratio) of the vibration signals of a diesel cylinder lid is the highest. The diesel vibration waveforms of combustion and valve become clear and the extracted characteristic parameters become more precise.

Speech enhancement using a wavelet thresholding method based on symmetric Kullback–Leibler divergence

After speech enhancement, some noisy speech will lose some effective speech information and cause speech signal distortion while improving its signal-to-noise ratio (SNR). According to the auditory masking effect, a method for judging the noise tolerance of noisy speech is proposed in this paper. Firstly, the necessity of judging noise tolerance is verified and analyzed. Then, based on the minimum priori SNR, the noise tolerance of noisy speech is studied, and its judgment threshold is given. Before speech enhancement, the voice segment is prejudged based on this threshold. If the minimum priori SNR ratio is less than this value at each frequency point, it is considered that the noise is intolerable at this time, and the speech enhancement is necessary. If the contrary is the case, it is determined that the noise is tolerable, and it is not necessary for speech enhancement. Finally, experiments are carried out to verify the effectiveness of the proposed method in this paper so that speech with tolerable noise can avoid speech distortion caused by enhancement.

In this paper, a novel method for stator interturn short circuit fault diagnosis in induction motors using the wavelet packet analysis is presented. Wavelet packet analysis is the extension of wavelet analysis and decompose signal into the equal bandwidth, it can achieve better decomposition result than wavelet analysis does. In the operating induction motor, electromagnetic torque is sensitive to the asymmetry due to the fault. Research indicate that the interturn short circuit fault will give rise to the sidebands at some given harmonic components of the electromagnetic torque signal and some certain frequency components change evidently, the distribution of the energy in the certain frequency bands alter as well. In the proposed method, electromagnetic torque signal is analyzed via wavelet packet decomposition and the wavelet packet coefficients are extracted. Based on each coefficient's energy eigenvalue, an eigenvector is set up. The presence of the short circuit fault makes the eigenvector change by a special rule, and the fluctuation of the load or supply unbalance make it change in another way, so it is immune to the variety of load and supply. Experimental results demonstrate the validity and effectiveness of this method.

In order to extract the week fault features contained in the vibration signal of the mechanical equipment, a new de-nosing method based on dual-tree complex wavelet packet transform (DTCWPT) and principal manifold was proposed. Firstly, the vibration signals were decomposed into several sub-frequency bands by DTCWPT, the Shannon entropy was used to seek the best basis of DTCWPT, and a new adaptive threshold function was employed to denoise the wavelet packet coefficients on the best basis of the real part and imaginary part of DTCWPT via the de-noising criterion, then the wavelet packet coefficients were reconstructed into a high dimensional space. Secondly, t-distributed stochastic neighbor embedding (t-SNE) was performed to extract a low dimensional manifold, the proposed threshold function was further applied to process the low dimensional manifold, aiming at separate the signal and noise, the principal manifold was reconstructed by the method of spectral regression analysis, thus, the signals were reconstructed back into one dimensional time series after eliminating the noise. Finally, a simulated signal and a real bearing fault signal were used to validate the proposed method, the results have demonstrated that the proposed method has a good performance of nonlinear noise reduction, and can extract fault features of rolling bearing effectively.

ASR converts speech signal into corresponding text form. The performance of an ASR decreases under noisy environment. To overcome this problem a speech enhancement need to be performed on noisy speech before being fed to an ASR system. Speech enhancement techniques have been developed over past several decades, some of these techniques introduce musical noise. To achieve further improvement in recognition accuracy, a generalized distillation framework is used in which machines learns machines. In this paper, an ASR is implemented for noisy kannada language speech using generalized distillation framework. In this framework, a teacher machine is trained with clean speech and student machine with 4 different noise speech and teacher machine help student machine to learn by providing additional information needed. During test phase, a student machine is tested with 4 different noise speech other than used in training. A DNN acoustic model is build using a 39 dimension MFSC features and bi-gram language model is created using Kaldi Speech Recognition Toolkit. Experimental results shows that generalized distillation framework for kannada noisy speech achieved a reduction in WER compared to an HMM-GMM approach.