CN108231084A - A kind of improvement wavelet threshold function denoising method based on Teager energy operators - Google Patents
A kind of improvement wavelet threshold function denoising method based on Teager energy operators Download PDFInfo
- Publication number
- CN108231084A CN108231084A CN201711260681.5A CN201711260681A CN108231084A CN 108231084 A CN108231084 A CN 108231084A CN 201711260681 A CN201711260681 A CN 201711260681A CN 108231084 A CN108231084 A CN 108231084A
- Authority
- CN
- China
- Prior art keywords
- wavelet
- signal
- threshold
- teager energy
- denoising
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 23
- 230000006872 improvement Effects 0.000 title claims abstract description 15
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 18
- 230000006870 function Effects 0.000 claims description 51
- 230000009466 transformation Effects 0.000 claims description 8
- 238000000205 computational method Methods 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 230000005236 sound signal Effects 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 8
- 238000001914 filtration Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000009514 concussion Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Acoustics & Sound (AREA)
- Signal Processing (AREA)
- Algebra (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Tea And Coffee (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
A kind of improvement wavelet threshold function denoising method based on Teager energy operators is claimed in the present invention; it is related to Speech Signal De-Noise field; this method passes through the calculating to the wavelet coefficient progress Teager energy operators after wavelet decomposition; the gap between noise and signal wavelet coefficient is increased, is conducive to the adaptively selected of threshold value；By being improved to soft and hard threshold denoising function, it is proposed that a kind of new threshold function table, the function can not only overcome the discontinuous caused pseudo- Gibbs' effect of hard threshold function, also overcome the constant deviation problem occurred after soft-threshold function denoising.The present invention has greatly improved to signal-to-noise ratio and with mean square error.
Description
Technical field
The invention belongs to Speech Signal De-Noise field, particularly a kind of improvement wavelet threshold based on Teager energy operators
Function denoising method.
Background technology
Voice signal usually contains noise during acquisition or propagation, and noise therein has seriously affected signal
Processing and the follow-up work of analysis.Therefore, the denoising of voice signal becomes work most basic and important in field of signal processing.
Traditional denoising method mainly includes linear filtering and nonlinear filtering, such as medium filtering, Wiener filtering and card
Kalman Filtering etc..These types of method be disadvantageous in that removal noise while, it is difficult to embody signal non-stationary property and
Correlation.In recent years, wavelet transformation was widely used in the denoising of signal with the characteristics of its multiresolution, it time domain with
Frequency domain can show good signal local feature, can efficiently extract the transient state information in signal, be suitble to believe non-stationary
Number carry out careful analysis, achieve preferable denoising effect.Wherein, wavelet threshold denoising because its realization it is simple, calculation amount it is small into
For wavelet analysis, most commonly used one kind, wavelet threshold denoising effectively can remove noise and retain original in denoising application
Voice signal, so as to preferably improve the signal-to-noise ratio and mean square error of signal.
Most important during wavelet threshold denoising is exactly selection about threshold value and threshold function table.If threshold value is too small,
It then will appear denoising deficiency, member-retaining portion noise；Threshold value selection is excessive, then will appear excessive denoising, by the weak feature in signal
Ingredient is mistakenly considered noise and eliminates.By the Donoho and Johnstone hard -thresholds proposed and Soft-threshold Denoising Method in practice
It has obtained widely applying, has also achieved preferable effect, but the discontinuous of hard threshold function causes reconstruction signal to be susceptible to
Pseudo- Gibbs' effect；Although soft-threshold function continuity is preferable, constant deviation is constantly present between estimated value and actual value.
If threshold value is too small, it will appear denoising deficiency, member-retaining portion noise；Threshold value selection is excessive, then will appear excessive denoising, signal
In weak characteristic component be mistakenly considered noise and eliminate, the loss of signal is caused to fall part useful information.Therefore, in order to obtain preferably
Denoising effect selects suitable threshold value and threshold function table particularly important.
Invention content
Present invention seek to address that above problem of the prior art.A kind of signal-to-noise ratio that can improve signals with noise is proposed,
Reduce the improvement wavelet threshold function denoising method based on Teager energy operators of mean square error.Technical scheme of the present invention is such as
Under：
A kind of improvement wavelet threshold function denoising method based on Teager energy operators, includes the following steps：
One section of S1, acquisition voice signal, and the noise of different signal-to-noise ratio is added to collected voice signal, it obtains band and makes an uproar
Voice signal；
S2, five layer scattering wavelet decompositions are carried out to the obtained Noisy Speech Signal of step S1, obtains the small wavelength-division of each layer
Solve coefficient；
S3, the calculating that Teager energy operators are carried out to each layer coefficient of wavelet decomposition that step S2 is obtained, obtain wavelet systems
Several Teager energy operator values；
S4, by the Teager energy operators in step S3 by the Hamming window of one 32, normalizing then is carried out to its value
Change is handled, and calculates the threshold value during denoising；
S5, the threshold value in step S4 carry out denoising using improved threshold function table to Noisy Speech Signal；Change
Into being embodied in：By being improved to the common threshold continuity of a function, gradual and constant deviation problem；
S6, coefficient of wavelet decomposition is reconstructed, obtains the voice signal after denoising.
Further, the step S2 carries out discrete wavelet transformation to Noisy Speech Signal, obtains the wavelet coefficient of each layer,
The step of discrete wavelet transformation is：
Wavelet function ψ (t) is defined first, and it is translated and operation of stretching can obtain cluster wavelet function ψ a, b (t)：
A ＞ 0, b ∈ R, a represent that scale factor b represents shift factor, a0And b0
Represent extension step-length, j represents the scale of wavelet decomposition；
It is as sliding-model control respectively by a, b：
A=a0 j, b=ka0 jb0 j,k∈Z,a0≠1
After sliding-model control, we are available cluster discrete wavelet function ψj,k(t)：
Then wavelet coefficients of the signal f (t) after wavelet transform can be expressed as：
WψTable
Show orthogonal wavelet transformation, f represents voice signal, and t represents the time, and k represents number of nodes.
Further, the calculating step that the step S3 asks for Teager energy operators is：
First, the conitnuous forms for defining nonlinear energy operator are：
In formula,Continuous T eager energy operators TEO, x (t) represents a continuous speech signal, when x (n) be one from
When dissipating voice signal：
N represents the time point of discretization；
TEO values then are calculated to coefficient of wavelet decomposition：wj,m(k) the wavelet coefficient T after wavelet decomposition is representedj,m(k) it represents
The value of each layer Teager energy operators；
Further, the computational methods of threshold value are in the step S4：
Obtained TEO values are smoothed, it is allowed to obtain M=T* by the Hamming window that a length is at 32 points
H, * represent convolution, and H is Hamming window, and T represents Tj,m(k) write a Chinese character in simplified form, represents the value of each layer Teager energy operators；And M is carried out
Normalized obtains M '：
So as to which its adaptive threshold can be indicated with following expression：
THj,m(k)=λj,m(1-αjM'j,m(k))
In formula, λj,mRepresent threshold value, j, m represent m-th of subband of jth layer, α respectivelyjFor the adjustment parameter based on each layer.
Further, it is described
Wherein, Nj,mFor the length of m-th of subband of jth layer, σ j, m represent the standard deviation of Gaussian noise：
Median represents mediant estimation.
Further, in the step S5, improved threshold function table is：
In formula, λ is threshold value, and n is positive integer, wherein,A threshold value that can be adjusted automatically is equivalent to,
When | wj,kDuring | >=λ,With | wj,k| it is continuously increased,Constantly reduce, and work as | wj,k| ＜ λ
When, it does not set to 0, but a smooth transition region is formd between noise and signal.
Further, in step S6, the computational methods that the signal after denoising is obtained to signal progress wavelet reconstruction are：
Here, c is one and the unrelated constant of original signal, wherein
Represent the complex conjugate of ψ j, k (t).
It advantages of the present invention and has the beneficial effect that：
The present invention first to after wavelet decomposition wavelet coefficient carry out Teager energy operators calculating, increase noise and
Gap between signal wavelet coefficient, conducive to the adaptively selected of threshold value；Then common soft, hard-threshold denoising function is carried out
It improves, it is proposed that a kind of new threshold function table, the function can not only overcome the discontinuous caused pseudo- gibbs of hard threshold function
Effect also overcomes the constant deviation problem occurred after soft-threshold function denoising.
The present invention does not connect for the aliasing of signal and noise wavelet coefficients, threshold function table during denoising at threshold value
There are the problems such as constant deviation for continuous, wavelet coefficient estimated value and original value, it is proposed that a kind of changing based on Teager energy operators
Into threshold function table denoising method.The calculating of Teager energy operators is carried out to wavelet coefficient first, is increased between signal and noise
Difference, conducive to the selection of threshold value；Quantification treatment is carried out to wavelet coefficient by improved threshold function table again, is efficiently avoided hard
Constant deviation problem caused by the discontinuous caused pseudo- Gibbs' effect and soft-threshold function of threshold function table.Improved denoising side
Method has greatly improved on denoising effect, and signal-to-noise ratio improves a lot, and language is also avoided while utmostly denoising
The distortion of sound signal has stronger practicability in the actual process of voice signal.
Description of the drawings
Fig. 1 is that the present invention provides improvement wavelet threshold function denoising method of the preferred embodiment based on Teager energy operators
Flow chart.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, detailed
Carefully describe.Described embodiment is only the part of the embodiment of the present invention.
The present invention solve above-mentioned technical problem technical solution be：
As shown in Figure 1, the present invention provides a kind of improvement threshold function table denoising method based on Teager energy operators,
It is characterized in that, includes the following steps：
S1 acquires one section of voice signal, and add different signal-to-noise ratio to collected voice signal with CoolEditPro
Noise obtains Noisy Speech Signal；
S2 carries out each layer coefficient of wavelet decomposition that step S2 is obtained the calculating of Teager energy operators, obtains small wavelength-division
Solve the TEO values of coefficient；(specific steps)
The step of discrete wavelet transformation is：
We define wavelet function ψ (t) first, and it are translated and operation of stretching can obtain cluster wavelet function ψ a, b
(t)：
A, b are respectively by we as sliding-model control：
A=a0j, b=ka0jb0 j,k∈Z,a0≠1
After sliding-model control, we are available cluster discrete wavelet function ψj,k(t)：
Then wavelet coefficients of the signal f (t) after wavelet transform can be expressed as：
S3 carries out each layer coefficient of wavelet decomposition that step S2 is obtained the calculating of Teager energy operators, obtains small wavelength-division
Solve the TEO values of coefficient；
First, the conitnuous forms that we define nonlinear energy operator are：
In formula,It is continuous T eager energy operators TEO, x (t) represents a continuous speech signal, then when x (n) is one
During a discrete voice signal：
TEO values then are calculated to coefficient of wavelet decomposition：
S4 calculates the threshold value during denoising；
Obtained TEO values are smoothed, it is allowed to obtain M=T* by the Hamming window that a length is at 32 points
H, * represent convolution, and H is Hamming window, and M is normalized to obtain M '：
So as to which its adaptive threshold can be indicated with following expression：
THj,m(k)=λj,m(1-αjM’j,m(k))
Wherein, Nj,mLength for m-th of subband of jth layer.σ j, m represent the standard deviation of Gaussian noise：
S5, propose improve threshold function table the step of be：
It is soft, hard threshold function for common threshold function table：
Formula (1) represents hard threshold function, and formula (2) represents soft-threshold function.Voice after hard threshold function denoising often has
Very big concussion remains much noise；Soft-threshold function treated wavelet coefficient and former wavelet coefficient there are constant deviation,
The useful high fdrequency component in part is had lost, causes voice distortion.
The shortcomings that in order to overcome two above function, the present invention propose a kind of improved threshold function table：
In formula, λ is threshold value, and n is positive integer.Wherein,It is equivalent to a threshold value that can be adjusted automatically.
When | wj,kDuring | >=λ,With | wj,k| it is continuously increased,Constantly reduce, avoid soft-threshold
Constant deviation problem present in processing procedure.And work as | wj,k| be not simply to set to 0 during ＜ λ, but noise and signal it
Between form a smooth transition region, avoid and issuable concussion directly blocked in hard -threshold.
S6 carries out wavelet reconstruction, obtains the voice signal after denoising.
Here, c is one and the unrelated constant of original signal, wherein
The above embodiment is interpreted as being merely to illustrate the present invention rather than limit the scope of the invention.
After the content for having read the record of the present invention, technical staff can make various changes or modifications the present invention, these equivalent changes
Change and modification equally falls into the scope of the claims in the present invention.
Claims (7)
1. a kind of improvement wavelet threshold function denoising method based on Teager energy operators, which is characterized in that including following step
Suddenly：
One section of S1, acquisition voice signal, and the noise of different signal-to-noise ratio is added to collected voice signal, obtain noisy speech
Signal；
S2, five layer scattering wavelet decompositions are carried out to the obtained Noisy Speech Signal of step S1, obtains the wavelet decomposition system of each layer
Number；
S3, the calculating that Teager energy operators are carried out to each layer coefficient of wavelet decomposition that step S2 is obtained, obtain wavelet coefficient
Teager energy operator values；
S4, by the Teager energy operators in step S3 by the Hamming window of one 32, place then is normalized to its value
Reason calculates the threshold value during denoising；
S5, the threshold value in step S4 carry out denoising using improved threshold function table to Noisy Speech Signal；Improve body
Now：By being improved to the common threshold continuity of a function, gradual and constant deviation problem；
S6, coefficient of wavelet decomposition is reconstructed, obtains the voice signal after denoising.
2. the improvement wavelet threshold function denoising method according to claim 1 based on Teager energy operators, feature
It is, the step S2 carries out discrete wavelet transformation to Noisy Speech Signal, obtains the wavelet coefficient of each layer, discrete wavelet transformation
The step of be：
Wavelet function ψ (t) is defined first, and it is translated and operation of stretching can obtain cluster wavelet function ψa,b(t)：
A represents that scale factor b represents shift factor, a0And b0Represent extension step-length, j represents the scale of wavelet decomposition；
It is as sliding-model control respectively by a, b：
A=a0 j, b=ka0 jb0 j,k∈Z,a0≠1
After sliding-model control, we are available cluster discrete wavelet function ψj,k(t)：
Then wavelet coefficients of the signal f (t) after wavelet transform can be expressed as：
WψRepresent orthogonal wavelet transformation, f represents voice signal, and t represents the time, and k represents number of nodes.
3. the improvement wavelet threshold function denoising method according to claim 2 based on Teager energy operators, feature
It is, the calculating step that the step S3 asks for Teager energy operators is：
First, the conitnuous forms for defining nonlinear energy operator are：
In formula,It is continuous T eager energy operators TEO, x (t) represents a continuous speech signal, when x (n) is a discrete language
During sound signal：
N represents the time point of discretization；
TEO values then are calculated to coefficient of wavelet decomposition：wj,m(k) the wavelet coefficient T after wavelet decomposition is representedj,m(k) each layer is represented
The value of Teager energy operators；
4. the improvement wavelet threshold function denoising method according to claim 3 based on Teager energy operators, feature
It is, the computational methods of threshold value are in the step S4：
Obtained TEO values are smoothed, it is allowed to obtain M=T*H, * generations by the Hamming window that a length is at 32 points
Table convolution, H are Hamming window, and T represents Tj,m(k) write a Chinese character in simplified form, represents the value of each layer Teager energy operators；And M is normalized
Processing obtains M '：
So as to which its adaptive threshold can be indicated with following expression：
THj,m(k)=λj,m(1-αjM'j,m(k))
In formula, λj,mRepresent threshold value, j, m represent m-th of subband of jth layer, α respectivelyjFor the adjustment parameter based on each layer.
5. the improvement wavelet threshold function denoising method according to claim 4 based on Teager energy operators, feature
It is, it is described
Wherein, Nj,mFor the length of m-th of subband of jth layer, σj,mRepresent the standard deviation of Gaussian noise：
Median represents mediant estimation.
6. the improvement wavelet threshold function denoising method according to claim 4 based on Teager energy operators, feature
It is, in the step S5, improved threshold function table is：
In formula, λ is threshold value, and n is positive integer, wherein,A threshold value that can be adjusted automatically is equivalent to, when |
wj,kDuring | >=λ,With | wj,k| it is continuously increased,Constantly reduce, and work as | wj,k| during ＜ λ,
It does not set to 0, but a smooth transition region is formd between noise and signal.
7. the improvement wavelet threshold function denoising method according to claim 6 based on Teager energy operators, step
In S6, the computational methods that the signal after denoising is obtained to signal progress wavelet reconstruction are：
Here, c is one and the unrelated constant of original signal, whereinIt representsψj,k(t)Complex conjugate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711260681.5A CN108231084B (en) | 2017-12-04 | 2017-12-04 | Improved wavelet threshold function denoising method based on Teager energy operator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711260681.5A CN108231084B (en) | 2017-12-04 | 2017-12-04 | Improved wavelet threshold function denoising method based on Teager energy operator |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108231084A true CN108231084A (en) | 2018-06-29 |
| CN108231084B CN108231084B (en) | 2021-09-10 |
Family
ID=62653137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711260681.5A Active CN108231084B (en) | 2017-12-04 | 2017-12-04 | Improved wavelet threshold function denoising method based on Teager energy operator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108231084B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109611703A (en) * | 2018-10-19 | 2019-04-12 | 宁波市鄞州利帆灯饰有限公司 | A kind of LED light being easily installed |
| CN109697702A (en) * | 2018-11-02 | 2019-04-30 | 浙江工业大学 | Medical ultrasound image denoising method based on bending wave conversion |
| CN110808059A (en) * | 2019-10-10 | 2020-02-18 | 天津大学 | Speech noise reduction method based on spectral subtraction and wavelet transform |
| CN113541729A (en) * | 2021-07-12 | 2021-10-22 | 电子科技大学 | Time-frequency graph denoising method based on time-frequency matrix |
| CN114091983A (en) * | 2022-01-21 | 2022-02-25 | 网思科技股份有限公司 | Intelligent management system for engineering vehicle |
| CN116202770A (en) * | 2023-03-21 | 2023-06-02 | 广东海洋大学 | Bearing fault diagnosis simulation experiment device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003081578A1 (en) * | 2002-03-21 | 2003-10-02 | U.S. Army Medical Research And Materiel Command | Methods and systems for detecting, measuring, and monitoring stress in speech |
| CN101625869A (en) * | 2009-08-11 | 2010-01-13 | 中国人民解放军第四军医大学 | Non-air conduction speech enhancement method based on wavelet-packet energy |
| CN103575807A (en) * | 2013-10-24 | 2014-02-12 | 河海大学 | Method for detecting structural damage of beam based on Teager energy operator-wavelet transformation curvature mode |
| CN106030707A (en) * | 2014-02-14 | 2016-10-12 | 唐纳德·詹姆士·德里克 | System for audio analysis and perception enhancement |
| CN107274908A (en) * | 2017-06-13 | 2017-10-20 | 南京邮电大学 | Small echo speech de-noising method based on new threshold function table |
-
2017
- 2017-12-04 CN CN201711260681.5A patent/CN108231084B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003081578A1 (en) * | 2002-03-21 | 2003-10-02 | U.S. Army Medical Research And Materiel Command | Methods and systems for detecting, measuring, and monitoring stress in speech |
| CN101625869A (en) * | 2009-08-11 | 2010-01-13 | 中国人民解放军第四军医大学 | Non-air conduction speech enhancement method based on wavelet-packet energy |
| CN103575807A (en) * | 2013-10-24 | 2014-02-12 | 河海大学 | Method for detecting structural damage of beam based on Teager energy operator-wavelet transformation curvature mode |
| CN106030707A (en) * | 2014-02-14 | 2016-10-12 | 唐纳德·詹姆士·德里克 | System for audio analysis and perception enhancement |
| CN107274908A (en) * | 2017-06-13 | 2017-10-20 | 南京邮电大学 | Small echo speech de-noising method based on new threshold function table |
Non-Patent Citations (2)
| Title |
|---|
| TAHSINA FARAH SANAM等: ""A semisoft thresholding method based on Teager energy operation on wavelet packet coefficients for enhancing noisy speech"", 《EURASIP JOURNAL ON AUDIO, SPEECH, ANDMUSIC PROCESSING》 * |
| 高亚召 等: ""基于Teager能量算子的自适应小波语音增强"", 《语音技术》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109611703A (en) * | 2018-10-19 | 2019-04-12 | 宁波市鄞州利帆灯饰有限公司 | A kind of LED light being easily installed |
| CN109697702A (en) * | 2018-11-02 | 2019-04-30 | 浙江工业大学 | Medical ultrasound image denoising method based on bending wave conversion |
| CN110808059A (en) * | 2019-10-10 | 2020-02-18 | 天津大学 | Speech noise reduction method based on spectral subtraction and wavelet transform |
| CN113541729A (en) * | 2021-07-12 | 2021-10-22 | 电子科技大学 | Time-frequency graph denoising method based on time-frequency matrix |
| CN113541729B (en) * | 2021-07-12 | 2022-06-07 | 电子科技大学 | Time-frequency graph denoising method based on time-frequency matrix |
| CN114091983A (en) * | 2022-01-21 | 2022-02-25 | 网思科技股份有限公司 | Intelligent management system for engineering vehicle |
| CN116202770A (en) * | 2023-03-21 | 2023-06-02 | 广东海洋大学 | Bearing fault diagnosis simulation experiment device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108231084B (en) | 2021-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108231084A (en) | A kind of improvement wavelet threshold function denoising method based on Teager energy operators | |
| CN107274908B (en) | Wavelet voice denoising method based on new threshold function | |
| CN109643554B (en) | Adaptive voice enhancement method and electronic equipment | |
| US11056130B2 (en) | Speech enhancement method and apparatus, device and storage medium | |
| Abd El-Fattah et al. | Speech enhancement with an adaptive Wiener filter | |
| CN109446928B (en) | Signal noise reduction method based on variational modal decomposition and minimum mean square error adaptive filter | |
| CN110808059A (en) | Speech noise reduction method based on spectral subtraction and wavelet transform | |
| CN109409206A (en) | A kind of thunder denoising method based on one-dimensional wavelet transform | |
| CN110808057A (en) | Voice enhancement method for generating confrontation network based on constraint naive | |
| Sanam et al. | Enhancement of noisy speech based on a custom thresholding function with a statistically determined threshold | |
| CN114581674A (en) | Wavelet denoising method based on improved threshold function | |
| CN109724693A (en) | A kind of fusion spectrum denoising method based on stationary wavelet | |
| CN110755069B (en) | Dynamic electrocardiosignal baseline drift correction method for jump mutation noise | |
| Batina et al. | Noise power spectrum estimation for speech enhancement using an autoregressive model for speech power spectrum dynamics | |
| Esch et al. | Model-based speech enhancement using SNR dependent MMSE estimation | |
| Sun et al. | Speech enhancement via two-stage dual tree complex wavelet packet transform with a speech presence probability estimator | |
| CN106997766B (en) | Homomorphic filtering speech enhancement method based on broadband noise | |
| Feng et al. | Voice activity detection based on ensemble empirical mode decomposition and teager kurtosis | |
| Vishwakarma et al. | De-noising of Audio Signal using Heavy Tailed Distribution and comparison of wavelets and thresholding techniques | |
| Shen et al. | Speech Noise Reduction by EMD-LMS | |
| Verma et al. | A comparative performance analysis of wavelets in denoising of speech signals | |
| Joshi et al. | Comparative analysis of thresholding techniques using discrete wavelet transform | |
| CN109271971B (en) | Noise reduction method for time sequence financial data | |
| Seyedin et al. | New features using robust MVDR spectrum of filtered autocorrelation sequence for robust speech recognition | |
| Islam et al. | Enhancement of noisy speech based on decision-directed Wiener approach in perceptual wavelet packet domain |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |