Application of Wavelet Transform in Speech Denoising

Jianfen Ma

DOI: https://doi.org/10.3969/j.issn.1007-9432.2001.03.008

2001-01-01

Abstract:A new speech denoising method based on the wavelet transform is presented.Filtering signal in the wavelet domain,by the frequency components acceptable for ear from the original speech,is a simple and effective speech denoising method.

What problem does this paper attempt to address?

Denoising Speech Based on Deep Learning and Wavelet Decomposition

Li Wang,Weiguang Zheng,Xiaojun Ma,Shiming Lin

DOI: https://doi.org/10.1155/2021/8677043

2021-07-16

Scientific Programming

Abstract:The work proposed a denoising speech method using deep learning. The predictor and target network signals were the amplitude spectra of the wavelet-decomposition vectors of the noisy audio signal and clean audio signal, respectively. The output of the network was the amplitude spectrum of the denoised signal. Besides, the regression network used the input of the predictor to minimize the mean square error between its output and input targets. The denoised wavelet-decomposition vector was transformed back to the time domain by the output amplitude spectrum and the phase of the wavelet-decomposition vector. Then, the denoised speech was obtained by the inverse wavelet transform. This method overcame the problem that the frequency and time resolution of the short-time Fourier transform could not be adjusted. The noise reduction effect in each frequency band was improved due to the gradual reduction of the noise energy in the wavelet-decomposition process. The experimental results showed that the method has a good denoising effect in the whole frequency band.

computer science, software engineering
Wavelet Transform Application for Signal Denoising

Hong Zhao

2001-01-01

Abstract:In the paper, the way of detection of signal based on wavelet transform is researched, it analyses and processes signal of the various frequency at different scales. The method of nonlinear wavelet threshold denoising can estimate the useful signal from the noisy so as to improve the resolving power and signal to noise ratio.
Denoising by Wavelet Soft-Thresholding in Fault Analysis

刘毅华,赵光宙

DOI: https://doi.org/10.3969/j.issn.1674-3415.2004.24.004

2004-01-01

Abstract:Denoising by soft-thresholding has been a standard algorithm in signal denoising. In fault measurement, it not only requires to reduce the noise, but also to reserve the singularity of signal. By analysing the different characteristics of signal and noise in wavelet analysis,this paper proposes an improved algorithm to enhance the denoising technique of wavelet soft-thresholding in estimating noise intensity, selecting threshold and adjusting wavelet coefficients. Simulations show that the new method solves the conflict commendably between protection of the local signals characteristics and restriction of noise, and can make proper analysis for different fault signals.
Adaptive Multiresolution Speech Enhancement Algorithm Based on Wavelet Transform

Zheng Yuanjin,Li Lemin,Wen Maosheng

DOI: https://doi.org/10.1007/s11767-999-1028-9

1999-01-01

Abstract:In this paper, an adaptive multiresolution speech enhancement algorithm based on wavelet transform is put forward. It can make adaptive filtering to noise speech both at scales and among scales. So that the noise parts during the frequency intervals which decrease hearing quality mostly are reduced efficiently. Both the SNR and subject hearing quality of denoised speech are high and good.
DENOISING METHOD BASED ON WAVELET TRANSFORM COEFFICIENT CORRELATION AND LOCAL LAPLACIAN MODEL

ZHANG Zhi-gang,ZHOU Xiao-jun,YANG Fu-chun,XIE Ming-xiang

DOI: https://doi.org/10.3969/j.issn.1000-3835.2008.11.007

2008-01-01

Abstract:In the denoising method proposed,wavelet transform coefficients on each level were separated into two sorts according to their correlation between the coefficients on the level and on its adjacent level,namely,the efficient co- efficients and the unefficient coefficients.Then the coefficients of the two sorts were processed,following corresponding rules respectively.The unefficient coefficients were directly set into zero,while local neighborhood statistical distribution of the efficient coefficients was assumed to be submit to Laplace model.Under such prior distribution,maximum a posteri- ori (MAP) estimator was used to restore the wavelet coefficients from the noisy observations,and new wavelet coefficients were used to reconstruct the denoised signal.Analyzing results show that the method can effectively denoise signal and im- prove the signal-to-noise ratio better than the traditional denoising method like hard-threshold and soft threshold.
Method and Application of Wavelet Shrinkage Denoising Based on Genetic Algorithm

Ma Qi-ming,Wang Xuan-yin,Du Shuan-ping

DOI: https://doi.org/10.1631/jzus.2006.a0361

2006-01-01

Abstract:Genetic algorithm (GA) based on wavelet transform threshold shrinkage (WTS) and translation-invariant threshold shrinkage (TIS) is introduced into the method of noise reduction, where parameters used in WTS and TIS, such as wavelet function, decomposition levels, hard or soft threshold and threshold can be selected automatically. This paper ends by comparing two noise reduction methods on the basis of their denoising performances, computation time, etc. The effectiveness of these methods introduced in this paper is validated by the results of analysis of the simulated and real signals.
Research on Speech Signal Denoising Algorithm Based on Wavelet Analysis

Linglong Tan,Yehui Chen,Fengzhi Wu

DOI: https://doi.org/10.1088/1742-6596/1627/1/012027

2020-08-01

Journal of Physics: Conference Series

Abstract:Abstract In this paper, wavelet analysis is used to remove noise superimposed on the speech signal to achieve denoising of the speech signal. The prepared speech signal is superimposed with Gaussian white noise with different signal-to-noise ratios, and is removed by using a forced noise removal processing method and a given threshold denoising processing method; The wavelet soft and hard threshold denoising method is used to separately superimpose different Gaussian white noise speech signals, and the hard threshold and soft threshold denoising are respectively performed, and different wavelets are used for processing; Four threshold selection methods are used to denoise the noisy signal using different wavelets. The denoising effects of these methods are simulated in MATLAB software, and combined with the signal-to-noise ratio of the signal after denoising, the denoising effects of various methods are judged. Experimental results show that soft threshold denoising is slightly better than hard threshold denoising. Among the four threshold selection methods, Rigrsure rule denoising is the best. Although other methods are effective, they are not obvious.
A New Method for Adaptive Speech Noise Canceling Based on Wavelet Transform

Tianshu Qu,Yisong Dai

DOI: https://doi.org/10.3321/j.issn:1000-6753.2001.02.017

2001-01-01

Abstract:For denoising the wide-band speech signal, this paper presents a new adaptive noise canceling method based on discrete wavelet transform (DWTANC) . This method analyzes the signal and noise into different frequency range with multiresolution analysis (MRA), and makes the signal and noise simple in frequency range, so the adaptive filter's orders would be shorted, and its self-cor-relative matrix' s orders will also be shorted. So it improves the performance of adaptive noise canceling system, shorts the adaptive time and enhances the signal noise ratio (SNR), at last, the computer simulation test validates the conclusion above.
A Denoising Method Based on New Wavelet Transforms

ZHAO Yang,SHI Yun-hui,LI Bei-li

DOI: https://doi.org/10.3969/j.issn.1672-464X.2006.04.007

2006-01-01

Abstract:The defects of the classical wavelet transforms are analyzed systemically. In the classical continuous wavelet inverse transform, the integral variables are independent. As the integral variables are dependent,the corresponding numeral integral has high resolution. The fact that the integral variable of the inverse transform appears in the denominator of integrand reduces resolution of the numeral integral and blocks the application of them. So a new type of continuous wavelet transform is proposed. No matter what we discretize integral variables, any numeral integral has a high resolution, and the integral variable does not appear in the denominator of the integrand. In addition, the corresponding numerical method is obtained. Finally, a filter based on the new CWT is present. By dealing with the signals with white noise and outlier, the white noise and outlier in the signals could be completely wiped off and the original signals could be reconstructed without prior distribution of the noise. compared with other traditional wavelet denoising methods, the new CWT can achieve higher signal-to-noise ratio.
Frequency Band Threshold Denoising Method Based on Wavelet Multi-resolution Analysis

Chao Wang

2008-01-01

Abstract:A wavelet frequency band threshold denoising method based on multi-resolution analysis is presented and the primary arithmetic steps is given.The soft threshold method and frequency band threshold method are used to denoise a speech signal respectively.The performances of wavelet threshold denoising method according to the objective evaluation standards: signal to noise ratio(SNR),root square mean error(RSME) are evaluated.Experimental results demonstrate that wavelet frequency band threshold method is more efficient for speech signal denoising.It can greatly improve the SNR and definition.
A New Method of Denoising of Vibration Signal and Its Application

Feng Miao,Rongzhen Zhao,Xianli Wang

DOI: https://doi.org/10.1155/2020/7587840

IF: 1.616

2020-05-30

Shock and Vibration

Abstract:In order to improve the performance of the denoising method for vibration signals of rotating machinery, a new method of signal denoising based on the improved median filter and wavelet packet technology is proposed through analysing the characteristics of noise components and relevant denoising methods. Firstly, the window width of the median filter is calculated according to the sampling frequency so that the impulse noise and part of the white noise can be effectively filtered out. Secondly, an improved self-adaptive wavelet packet denoising technique is used to remove the residual white noise. Finally, useful vibration signals are obtained after the previous processing. Simulation signals and rotor experimental vibration signals were used to verify the performance of the method. Experiment results show that the method can not only effectively eliminate the mixed complex noises but also preserve the fault character details, which demonstrates that the proposed method outperforms the method based on the wavelet-domain median filter.

engineering, mechanical,mechanics,acoustics
Image Denoising Based on Mean Filter and Wavelet Transform

Qingkun Song,Li Ma,JianKun Cao,Xiao Han,Jiankun Cao

DOI: https://doi.org/10.1109/aits.2015.17

2015-08-01

Abstract:The traditional mean filter in the processing of the image noise reduction can make the image edge information lose, the image is obscure. The wavelet transformation in the image processing can not only suppresses image noise but also has the strong ability of image enhancement. So I combine two filtering methods for image noise reduction processing. The experiments and the test in this paper show that the denoising effect by combining of wavelet transform and mean filter is superior to simple denoising method and the existing combination of the denoising method no matter from visual aspect and in terms of signal-to-noise ratio and root mean square error.
Signal Denoising Using Undecimated Second Generation Wavelet Transform

鲍文,周瑞,李宁,杨建国,于达仁

DOI: https://doi.org/10.3321/j.issn:0258-8013.2008.20.015

2008-01-01

Abstract:Signal denoising is an important problem in the field of fault diagnosis of electromechanical equipment. For the wavelet coefficients of noises are trending towards zero rapidly as the scale increasing, the denoising methods based on wavelet transform have been widely researched. The construction of undecimated second generation wavelet transform (USGWT) was studied and the calculating formulas of its predicting and updating operators in time domain were given. Furthermore, the computational complexity of both second generation wavelet transform (SGWT) and undecimated second generation wavelet transform was analyzed and a signal denoising method using undecimated second generation wavelet transform was proposed. Owing to the approximate and detail coefficients at each level after decomposition using undecimated wavelet transform have the same length as the original signal, the transform can possess time invariant property and retain more characteristics in time domain, and therefore improve the effect of denoising. The simulated signal and real-life signal were used to test this method. It was shown that the presented method is effective for signal denoising and the fault feature of original signal can be desirably retained.
Study on processing of wavelet speech denoising in speech recognition system

Xinmei Zhong,Yunzhong Dai,Yong Dai,Tao Jin

DOI: https://doi.org/10.1007/s10772-018-9516-7

2018-05-08

International Journal of Speech Technology

Abstract: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.

English Else
Application of Wavelet Analysis in Denoising Seismic Data

Qing Yi Liu

DOI: https://doi.org/10.4028/www.scientific.net/amm.530-531.540

2014-02-01

Applied Mechanics and Materials

Abstract:The random noise is the kind of noise with wide frequency band in seismic data detected by the optical acceleration sensors. The noises influence and destroy the useful signal of the seismic information. There are a lot of methods to remove noise and one of the standard methods to remove the noise of the signal was the fast Fourier transform (FFT) which was the linear Fourier smoothing. In this paper, the novel denoising method based on wavelet analysis was introduced. The denoising results of seismic data with the noise with FFT method and wavelet analysis method, respectively. SNRs of the signal with noise, FFT denoisng and wavelet analysis denoising are-8.69, -1.13, and 8.27 respectively. The results show that the wavelet analysis method is prior to the traditional denoising method. The resolution of the seismic data improves.

English Else
Multiscale robust filtering technique

吕立华,宋执环,李平

DOI: https://doi.org/10.3969/j.issn.1000-7059.2004.03.004

2004-01-01

Abstract:The principle of wavelet transformation used to eliminate white noise is analyzed and it is tested that this method is not perfect in case of non-white noise.Combined with robust statistics,a robust wavelet transformation which can restrain the disturbance with contaminated distribution is studied to provide a practical multiscale robust filtering method.The result of simulation shows that this method is robust and effecient.
An Improved Wavelet De-noising Method for Time Series Analysis.

Yan-Fang Sang,Dong Wang,Ji-Chun Wu,Qing-Ping Zhu,Ling Wang

DOI: https://doi.org/10.3390/e11041123

IF: 2.738

2009-01-01

Entropy

Abstract:On the basis of discussing some key problems about wavelet de-noising as: choice of reasonable wavelet function, determination of reasonable wavelet coefficients thresholds and choice of suitable threshold processing-means, an improved wavelet denoising method has been proposed. Then by Monte-Carlo tests, the validity of this method is verified. Analyses results show that compared with traditional methods (FT, SURE and MINMAX), this improved wavelet denoising method is more accurate and reliable. Furthermore, because of based on information entropy theories to choose the reasonable wavelet coefficients thresholds, the de-noising results by the improved method are the global optimum.
Introduction to Wavelets and their applications in signal denoising

Cigdem POLAT,Mehmet Siraç ÖZERDEM

DOI: https://doi.org/10.17678/beuscitech.349020

2018-06-28

Bitlis Eren University Journal of Science and Technology

Abstract:The aim of this study is providing a comprehensive background information related to the roots of both Fourier Transform (FT) and Wavelet Transform (WT) along with an experiment related to applications of WT techniques. The paper describes several applications of WT and provides background information on FT. Fourier Transform (FT) is a concept that has a long history yet several issues related to resolution and uncertainty of time –frequency. Even though there are several adapted forms of FT such as Short Time Fourier Transform (STFT), which intend to solve the problems, certain limitations remain. Wavelet Transform (WT) is an alternative transformation technique emerged in order to fully tackle these diverse and complicated issues. In this paper, the background information related to the roots of FT and WT are given. Some of the problems that WT addresses are examined. WT is a tool that has many advantages among them is noise reduction and compression. We reviewed several studies that use the noise reduction capability of WT alone or combined with other signal processing tools. Discrete Wavelet Transform (DWT) based algorithm is also examined as a noise reduction technique and carried out in MATLAB setting. Analysis on a speech signal which contaminated with keyboard sound also a number spelling female voice containing unknown noise are performed. Different types of thresholding and mother wavelets were in consideration and it was revealed that Daubechies family along with the soft thresholding technique suited our application the most.

English Else
Audio Noise Reduction from Audio Signals and Speech Signals

Nishan Singh,V. Laxmi

Abstract:Speech signal analysis is one of the important areas of research in multimedia applications. Discrete Wavelet technique is effectively reduces the unwanted higher or lower order frequency components in a speech signal. Wavelet-based algorithm for audio de-noising is worked out. We focused on audio signals corrupted with white Gaussian noise which is especially hard to remove because it is located in all frequencies. We use Discrete Wavelet transform (DWT) to transform noisy audio signal in wavelet domain. It is assumed that high amplitude DWT coefficients represent signal, and low amplitude coefficients represent noise. Using thresholding of coefficients and transforming them back to time domain it is possible to get audio signal with less noise. Our work has been modified by changing universal thresholding of coefficients which results with better audio signal. In this various parameters such as SNR, Elapsed Time, and Threshold value is analyzed on various types of wavelet techniques alike Coiflet, Daubechies, Symlet etc. In all these, best Daubechies as compared to SNR is more for Denoising and Elapsed Time is less than others for Soft thresholding. In using hard thresholding Symlet wavelet also works better than coiflet and Daubechies is best for all. Efficiency is 98.3 for de-noising audio signals which also gives us better results than various filters.

Engineering,Computer Science
Wavelet based human voice identification system

Maryam Mohammed Mubarak al Balushi,Vidhya Lavanya R,Sreedevi Koottala,Ajay Vikram Singh,R. Vidhya Lavanya

DOI: https://doi.org/10.1109/ictus.2017.8286002

2017-12-01

Abstract:This paper investigates the use of wavelet transform in order to remove noise from the signals. One of the ongoing research in multimedia applications is speech signal processing. Wavelet denoising technique is attempted to reduce and remove noise from the audio signal. It is therefore required to transform audio signal to wavelet domain by using discrete wavelet transform followed by denoising algorithm. Both soft and hard thresholding of denoising technique is used to compare the performance of human noise identification. Denoising the signal is performed in the transformation domain and improvement in the denoising will be achieved in various families of wavelet transform. There are several types of wavelets such as Haar, Symlets, BiorSplines, Discrete Meyer (Dmey) and Reverse Biothogonal. In this paper, identification of human noise is verified with Dmey and Fejer- Korovkin wavelets. Comparative analysis is done to calculate SNR after the spectral subtraction. The quality of the audio signal is determined by Mean Square Error (MSE) and Signal to Noise Ratio (SNR) of the denoised signal. The simulation results are performed in Matlab program. From the result analysis, it is shown that the Fejer- Korovkin wavelet filter has a better performance as compared to other type of wavelet transform and can be able to identify and differentiate human voices with others.

Application of Wavelet Transform in Speech Denoising

Denoising Speech Based on Deep Learning and Wavelet Decomposition

Wavelet Transform Application for Signal Denoising

Denoising by Wavelet Soft-Thresholding in Fault Analysis

Adaptive Multiresolution Speech Enhancement Algorithm Based on Wavelet Transform

DENOISING METHOD BASED ON WAVELET TRANSFORM COEFFICIENT CORRELATION AND LOCAL LAPLACIAN MODEL

Method and Application of Wavelet Shrinkage Denoising Based on Genetic Algorithm

Research on Speech Signal Denoising Algorithm Based on Wavelet Analysis

A New Method for Adaptive Speech Noise Canceling Based on Wavelet Transform

A Denoising Method Based on New Wavelet Transforms

Frequency Band Threshold Denoising Method Based on Wavelet Multi-resolution Analysis

A New Method of Denoising of Vibration Signal and Its Application

Image Denoising Based on Mean Filter and Wavelet Transform

Signal Denoising Using Undecimated Second Generation Wavelet Transform

Study on processing of wavelet speech denoising in speech recognition system

Application of Wavelet Analysis in Denoising Seismic Data

Multiscale robust filtering technique

An Improved Wavelet De-noising Method for Time Series Analysis.

Introduction to Wavelets and their applications in signal denoising

Audio Noise Reduction from Audio Signals and Speech Signals

Wavelet based human voice identification system