NMF-based speech enhancement using multitaper spectrum estimation

Attabi, Yazid; Chung, Hanwook; Champagne, Benoı̂t; Zhu, Wei-Ping

doi:10.1109/icsigsys.2018.8373564

Cited by 4 publications

(4 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This paper proposes an innovative approach by generating distinct particles using multitaper bases. Each particle represents the frequency domain analysis of the current windowed signal across different multitaper bases, employing orthogonal tapers for spectral smoothing and variance reduction [34].…”

Section: The Core Aim Of Instantaneous Frequency Estimation Technique...mentioning

confidence: 99%

Multicomponent collaborative time-frequency state-space model for vibration signal decomposition under nonstationary conditions

Huang,

Chen,

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

The implementation of Prognostics and Health Management (PHM) strategies is essential for enhancing the safety and maintenance of rotating equipment in chemical plants. The examination of vibration signal behaviours under variable-speed conditions and the development of signal decomposition methods in such contexts are of substantial theoretical and practical relevance. This paper proposes a novel multicomponent collaborative time-frequency state-space method for the decomposition of vibration signals. The method employs a multi-component proportional model to accurately describe the synchronisation of high-frequency components with the rotational frequency component. Based on this framework, the instantaneous frequency change curves of multiple components are input into the Vold-Kalman filter algorithm for precise decomposition of multicomponent vibration signals under variable-speed conditions. Experimental results demonstrate the effectiveness of the proposed method in achieving accurate instantaneous frequency tracking and signal decomposition, showing clear advantages over traditional methods.

show abstract

Section: The Core Aim Of Instantaneous Frequency Estimation Technique...mentioning

confidence: 99%

Multicomponent collaborative time-frequency state-space model for vibration signal decomposition under nonstationary conditions

Huang,

Chen,

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…In multitaper time frequency algorithm, a set of orthogonal tapers is applied to the short-time signal to smooth the spectral estimates for the reduction of spectral variance [31]. The Hermite function is used as window function in multitaper time frequency algorithm, which is defined as follows [32].…”

Section: Multitaper Time Frequency Distributionmentioning

confidence: 99%

Graph constrained empirical wavelet transform and its application in bearing fault diagnosis

Tan,

Zhao,

et al. 2024

Eng. Res. Express

View full text Add to dashboard Cite

The signal decomposition based on frequency domain distribution is a fundamental methodology for mechanical component fault diagnosis. However, existing methods face challenges such as susceptibility to noise interference and limited adaptability. Therefore, this paper proposes the graph constrained empirical wavelet transform (GCEWT) method. This method introduces structured information, such as the interrelationships among different parts of the frequency domain distribution of vibration signals, into the boundary detection process of empirical wavelet transform. The high-dimensional connectivity among different parts of the time-frequency distribution is utilized to construct an adjacency matrix. By constructing an adjacent graph, the proposed method encodes the adjacency relationships among frequency bands to constrain the low-dimensional spatial relationships between them. In conjunction with spectral clustering algorithms, the GCEWT method determines the boundaries for empirical wavelet transformation in the frequency domain. This approach achieves structured and adaptive decomposition of vibration signals from components of critical equipment, facilitating the structured and adaptive extraction of fault features. The effectiveness of the proposed method is validated using vibration data from both wind turbine drivetrain systems and aircraft engines. The experimental results demonstrate that the proposed method yields more reasonable signal decomposition results compared to traditional algorithms. Additionally, the proposed method proves to be more effective in extracting weak fault features of bearings in the presence of noise.

show abstract

“…Automatic bird species identification provides a suitable way to analyze the huge audio data from long-term monitoring programs [3]. However, the bird audio collected in nature is always accompanied by ambient noises, which consequently affect the accuracy of the bird species identification [4]. Therefore, audio enhancement should be carried out before identification, to improve identification accuracy.By exploiting the spatial properties of speech and noise signals, WASN techniques can significantly outperform single-channel techniques in terms of improving interference suppression and reducing speech distortion [5][6][7][8][9].…”

mentioning

confidence: 99%

mentioning

confidence: 99%

Improved Distributed Minimum Variance Distortionless Response (MVDR) Beamforming Method Based on a Local Average Consensus Algorithm for Bird Audio Enhancement in Wireless Acoustic Sensor Networks

Xie

Xing

et al. 2019

Applied Sciences

View full text Add to dashboard Cite

Featured Application: With this bird audio enhancement method, the bird audio collected through the WASN (Wireless Acoustic Sensor Network) can be processed to produce better quality audio, which is more suitable for bird species identification based on the bird audio, then a higher accuracy of identification will be achieved. Abstract: Currently, wireless acoustic sensor networks (WASN) are commonly used for wild bird monitoring. To better realize the automatic identification of birds during monitoring, the enhancement of bird audio is essential in nature. Currently, distributed beamformer is the most suitable method for bird audio enhancement of WASN. However, there are still several disadvantages of this method, such as large noise residue and slow convergence rate. To overcome these shortcomings, an improved distributed minimum variance distortionless response (IDMVDR) beamforming method for bird audio enhancement in WASN is proposed in this paper. In this method, the average metropolis weight local average consensus algorithm is first introduced to increase the consensus convergence rate, then a continuous spectrum update algorithm is proposed to estimate the noise power spectral density (PSD) to improve the noise reduction performance. Lastly, an MVDR beamformer is introduced to enhance the bird audio. Four different network topologies of the WASNs were considered, and the bird audio enhancement was performed on these WASNs to validate the effectiveness of the proposed method. Compared with two classical methods, the results show that the Segmental signal to noise ratio (SegSNR), mean square error (MSE), and perceptual evaluation of speech quality (PESQ) obtained by the proposed method are better and the consensus rate is faster, which means that the proposed method performs better in audio quality and convergence rate, and therefore it is suitable for WASN with dynamic topology. wireless acoustic sensor networks (WASN) becoming more and more popular, WASNs are commonly used for monitoring bird audio long-term [2]. Automatic bird species identification provides a suitable way to analyze the huge audio data from long-term monitoring programs [3]. However, the bird audio collected in nature is always accompanied by ambient noises, which consequently affect the accuracy of the bird species identification [4]. Therefore, audio enhancement should be carried out before identification, to improve identification accuracy.By exploiting the spatial properties of speech and noise signals, WASN techniques can significantly outperform single-channel techniques in terms of improving interference suppression and reducing speech distortion [5][6][7][8][9]. Although WASN has many advantages, it also has several challenges, such as the limited energy and calculation ability of each node. There are two kinds of audio enhancement algorithms for WASNs. The first is centralized, all the data is transferred to a so-called fusion center (FC) for further enhancement. The second is distributed, the enhancements are performed on all ...

show abstract

NMF-based speech enhancement using multitaper spectrum estimation

Cited by 4 publications

References 27 publications

Multicomponent collaborative time-frequency state-space model for vibration signal decomposition under nonstationary conditions

Multicomponent collaborative time-frequency state-space model for vibration signal decomposition under nonstationary conditions

Graph constrained empirical wavelet transform and its application in bearing fault diagnosis

Improved Distributed Minimum Variance Distortionless Response (MVDR) Beamforming Method Based on a Local Average Consensus Algorithm for Bird Audio Enhancement in Wireless Acoustic Sensor Networks

Contact Info

Product

Resources

About