CN107680610A - A kind of speech-enhancement system and method - Google Patents
A kind of speech-enhancement system and method Download PDFInfo
- Publication number
- CN107680610A CN107680610A CN201710888848.6A CN201710888848A CN107680610A CN 107680610 A CN107680610 A CN 107680610A CN 201710888848 A CN201710888848 A CN 201710888848A CN 107680610 A CN107680610 A CN 107680610A
- Authority
- CN
- China
- Prior art keywords
- spectrum
- module
- enhancing
- speech
- estimation
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Processing of the speech or voice signal 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
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Processing of the speech or voice signal 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
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/21—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being power information
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/45—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of analysis window
Abstract
The invention discloses the invention discloses a kind of speech-enhancement system and method, subtract enhancing module, Wiener filtering enhancing module, inverse Fourier transform module and voice-output device including audio input device, sound pretreatment module, endpoint detection module, signal-to-noise ratio (SNR) estimation module, more windows spectrum spectrum, signal-to-noise ratio (SNR) estimation module, including SNR estimation units and selection control unit, SNR estimation units are electrically connected with the selection control unit and endpoint detection module respectively, and selecting control unit to be composed respectively with more windows, spectrum subtracts enhancing module and Wiener filtering enhancing module is electrically connected with.The present invention, which using more windows spectrum spectrum subtracts speech enhancement technique and effectively weakens spectrum, subtracts " music noise " that voice enhancement algorithm is brought, while subtracts the drawbacks of intelligibility of speech that voice enhancement algorithm brings is damaged to judge to subtract speech enhancement technique and Wiener filtering speech enhancement technique using more windows spectrum spectrum so as to avoid in low signal-to-noise ratio spectrum by judging signal-to-noise ratio (SNR) estimation value.
Description
Technical field
The present invention relates to field of speech enhancement, and in particular to a kind of speech-enhancement system and method.
Background technology
With the fast development of VLSI technologies and DSP technologies, the realization and utilization of speech enhancement technique are also achieved, especially
It is field of speech recognition speech enhancement technique be even more receive much concern.A kind of existing Adaptive spectras of patent CN103594094A
Subtraction real-time voice Enhancement Method, this method include：Structure noisy speech has voice and the dynamic threshold differentiated without voice, proposes
Principle is updated according to the noise spectrum time-varying of dynamic threshold；The correlation extraction information of adjacent interframe is taken full advantage of, is realized pure
Net voice spectrum smoothing iterative estimate method；Actually asked for what voice signal under nonstationary noise and strong background noise was difficult to extract
Topic, gives a kind of Adaptive spectra subtraction voice enhancement algorithm；Using quick tracking noise algorithm to nonstationary noise carry out by
Frame smoothly updates, and can preferably estimate noise spectrum；Algorithm proposed by the present invention can more effectively suppress ambient noise, improve
Voice quality and intelligibility after making an uproar.
Although existing patent CN103594094A is employed, Adaptive spectra subtracts voice enhancement algorithm and quick tracking noise is calculated
Method reads noisy speech and has carried out effective denoising, but still has following drawback：(1) subtract speech enhan-cement using Adaptive spectra to calculate
The practical SNR ranges of method are more larger than narrow and to Low SNR signal intelligibility of speech damage；(2) Adaptive spectra subtracts voice
Enhancing algorithm can produce " music noise ".
The content of the invention
It is an object of the invention to provide a kind of speech-enhancement system and method, to solve what is proposed in above-mentioned background technology
Problem.
To achieve the above object, the present invention provides following technical scheme：
A kind of speech-enhancement system, including：
Audio input device, sound pretreatment module, endpoint detection module, signal-to-noise ratio (SNR) estimation module, more windows spectrum spectrum subtract increasing
Strong module, Wiener filtering enhancing module, inverse Fourier transform module and voice-output device, the signal-to-noise ratio (SNR) estimation module, bag
SNR estimation units and selection control unit are included, the SNR estimation units select control unit and end-point detection mould with described respectively
Block is electrically connected with, and the selection control unit subtracts enhancing module with more windows spectrum spectrum respectively and Wiener filtering enhancing module electrically connects
Connect, the sound pretreatment module is electrically connected with the audio input device and endpoint detection module respectively, in inverse Fu
Leaf transformation subtracts enhancing module with the voice-output device, more windows spectrum spectrum respectively and Wiener filtering enhancing module is electrically connected with.
Preferably, the pretreatment module, including pre-filtering processing, A/D modular converters and framing windowing process, it is described pre-
Filtering process uses antialiasing filter, and the framing windowing process uses Hamming window adding window.
Preferably, the endpoint detection module uses the end-point detection based on frequency band variance.
Preferably, the selection control unit is ARM single-chip microcomputers, for being sentenced according to the estimate of the SNR estimation units
Disconnected selection more window spectrum spectrums subtract enhancing module and Wiener filtering speech enhan-cement module.
Preferably, more window spectrum spectrums subtract enhancing module and subtract gain calculating and enhancing amplitude spectrum including power Spectral Estimation, spectrum
Calculate, the power Spectral Estimation includes noisy speech power Spectral Estimation and power noise Power estimation.
Preferably, the Wiener filtering enhancing module includes Fourier transformation, filter coefficient calculates and frequency domain filtering.
A kind of method of speech enhan-cement, including：
Step 1：Noisy speech passes through voice-input device input system, is carried out by antialiasing filter at pre-filtering
Sampled again by A/D modular converters after reason and quantification treatment, framing and plus Hamming window finally are carried out to Noisy Speech Signal
Processing；
Step 2：Noisy Speech Signal after adding window is subjected to end points inspection using the end-point detecting method based on frequency band variance
Survey；
Step 3：Signal-to-noise ratio (SNR) estimation is done to Noisy Speech Signal, while signal-to-noise ratio (SNR) estimation value sends the selection control to
Unit, the selection control unit judge that selection more window spectrum spectrums subtract enhancing module or wiener according to default signal to noise ratio codomain
Filtering enhancing module, step 4 is carried out if selecting more window spectrum spectrums to subtract enhancing module, if selection Wiener filtering speech enhan-cement module
Why step 5 is carried out；
Step 4：Noisy Speech Signal carries out noisy speech power Spectral Estimation and noise power by the power Spectral Estimation
Power estimation, subtract the calculating of gain calculating progress power spectrum yield value by the spectrum, the enhancing amplitude spectrum calculates makes an uproar according to band
Phonetic speech power spectrum, noise power spectrum and power spectrum yield value calculate enhancing amplitude spectrum, and enhancing amplitude spectrum passes through inverse fast Fourier
Voice after synthesis enhancing after conversion；
Step 5：After Noisy Speech Signal is by Fast Fourier Transform (FFT), wave filter system is calculated according to signal-to-noise ratio (SNR) estimation value
Amplitude spectrum must be strengthened by counting and carrying out frequency domain filtering, and enhancing amplitude spectrum synthesizes enhanced language after inverse fast fourier transform
Sound.
Compared with prior art, the beneficial effects of the invention are as follows：
The present invention using more windows spectrum spectrum subtract that speech enhancement technique effectively weakens that spectrum subtracts that voice enhancement algorithm brings " music is made an uproar
Sound ", while to judge to compose using more windows, spectrum subtracts speech enhancement technique and Wiener filtering voice increases by judging signal-to-noise ratio (SNR) estimation value
Strong technology is so as to avoiding the drawbacks of spectrum in low signal-to-noise ratio subtracts the intelligibility of speech damage that voice enhancement algorithm is brought.
Brief description of the drawings
Fig. 1 is the speech-enhancement system structural representation of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other under the premise of creative work is not made
Embodiment, belong to the scope of protection of the invention.
As shown in Figure 1, a kind of speech-enhancement system, including：Audio input device, sound pretreatment module, end points inspection
Survey module, signal-to-noise ratio (SNR) estimation module, more windows spectrum spectrum subtract enhancing module, Wiener filtering enhancing module, inverse Fourier transform module and
Voice-output device, signal-to-noise ratio (SNR) estimation module, including SNR estimation units and selection control unit, SNR estimation units respectively with choosing
Select control unit and endpoint detection module is electrically connected with, select control unit to subtract enhancing module and wiener filter with more windows spectrum spectrum respectively
Ripple enhancing module is electrically connected with, and sound pretreatment module is electrically connected with audio input device and endpoint detection module respectively, inverse
Fourier transformation subtracts enhancing module with voice-output device, more windows spectrum spectrum respectively and Wiener filtering enhancing module is electrically connected with.In advance
Processing module, including pre-filtering processing, A/D modular converters and framing windowing process, pre-filtering processing use antialiasing filter,
Framing windowing process uses Hamming window adding window, and endpoint detection module uses the end-point detection based on frequency band variance.Selection control is single
Member is ARM single-chip microcomputers, for being judged to select more window spectrum spectrums to subtract enhancing module and Wiener filtering according to the estimate of SNR estimation units
Speech enhan-cement module, more windows spectrum spectrums subtract that enhancing module includes power Spectral Estimation, spectrum subtracts that gain calculates and enhancing amplitude spectrum calculates, work(
Rate Power estimation includes noisy speech power Spectral Estimation and power noise Power estimation, and Wiener filtering strengthens module and become including Fourier
Change, filter coefficient calculates and frequency domain filtering.
When sound from audio input device after incoming call big speech-enhancement system, carried out in pretreatment module at pre-filtering
Reason, A/D analog-to-digital conversions, framing windowing process, end-point detection is being carried out by endpoint detection module, is obtaining noisy speech, SNR estimates
Meter carries out signal-to-noise ratio (SNR) estimation to noisy speech, and control unit selects Wiener filtering enhancing module and more windows according to signal-to-noise ratio (SNR) estimation value
Spectrum spectrum subtracts enhancing module, and carrying out inverse Fourier transform after speech enhan-cement obtains enhanced voice.
A kind of method of speech enhan-cement, including：
Step 1：Noisy speech passes through voice-input device input system, is carried out by antialiasing filter at pre-filtering
Sampled again by A/D modular converters after reason and quantification treatment, framing and plus Hamming window finally are carried out to Noisy Speech Signal
Processing；
Step 2：Noisy Speech Signal after adding window is subjected to end points inspection using the end-point detecting method based on frequency band variance
Survey；
Step 3：Signal-to-noise ratio (SNR) estimation is done to Noisy Speech Signal, while signal-to-noise ratio (SNR) estimation value sends selection control unit to,
Selection control unit judges that selecting more window spectrum spectrums to subtract enhancing module or Wiener filtering strengthens module according to default signal to noise ratio codomain,
Step 4 is carried out if selecting more window spectrum spectrums to subtract enhancing module, if why selection Wiener filtering enhancing module carries out step 5；
Step 4：Noisy Speech Signal carries out noisy speech power Spectral Estimation by power Spectral Estimation and noise power spectrum is estimated
Meter, subtracts the calculating of gain calculating progress power spectrum yield value by spectrum, and enhancing amplitude spectrum calculates according to noisy speech power spectrum, made an uproar
Power sound spectrum and power spectrum yield value calculate enhancing amplitude spectrum, and enhancing amplitude spectrum synthesizes language after enhancing after inverse Fourier transform
Sound；
Step 5：After Noisy Speech Signal is by Fourier transformation, filter coefficient is calculated simultaneously according to signal-to-noise ratio (SNR) estimation value
Amplitude spectrum must be strengthened by carrying out frequency domain filtering, and enhancing amplitude spectrum synthesizes enhanced voice after inverse Fourier transform.
The signal-to-noise ratio (SNR) estimation value obtained after estimation of the noisy speech by SNR estimation modules is estimated less than default signal to noise ratio
Evaluation, then Wiener filtering enhancing module is selected to ensure the intelligibility after speech enhan-cement, when signal to noise ratio evaluation is higher than default noise
Compared estimate value then subtracts enhancing module using more windows spectrum spectrum effectively reduces the influence of " musical noise " while voice is strengthened.
Although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of changes, modification can be carried out to these embodiments, replace without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (7)
- A kind of 1. speech-enhancement system, it is characterised in that including:Audio input device, sound pretreatment module, endpoint detection module, signal-to-noise ratio (SNR) estimation module, more windows spectrum spectrum subtract enhancing mould Block, Wiener filtering enhancing module, inverse Fourier transform module and voice-output device, the signal-to-noise ratio (SNR) estimation module include SNR Estimation unit and selection control unit, the SNR estimation units are electric with the selection control unit and endpoint detection module respectively Property connection, it is described selection control unit respectively with more windows spectrum spectrum subtract enhancing module and Wiener filtering enhancing module be electrically connected with, institute State sound pretreatment module to be electrically connected with the audio input device and endpoint detection module respectively, the inverse Fourier transform Subtract enhancing module with the voice-output device, more windows spectrum spectrum respectively and Wiener filtering enhancing module is electrically connected with.
- 2. a kind of speech-enhancement system according to claim 1, it is characterised in that the pretreatment module includes pre-filtering Processing, A/D modular converters and framing windowing process, the pre-filtering, which is handled, uses antialiasing filter, at the framing adding window Reason uses Hamming window adding window.
- 3. a kind of speech-enhancement system according to claim 1, it is characterised in that the endpoint detection module is used and is based on The end-point detection of frequency band variance.
- 4. a kind of speech-enhancement system according to claim 1, it is characterised in that the selection control unit is mono- for ARM Piece machine, for judging that selection more window spectrum spectrums subtract enhancing module and Wiener filtering according to the estimate of the SNR estimation units Speech enhan-cement module.
- 5. a kind of speech-enhancement system according to claim 1, it is characterised in that more window spectrum spectrums subtract enhancing module bag Include power Spectral Estimation, spectrum subtracts gain calculating and enhancing amplitude spectrum calculating, the power Spectral Estimation are estimated including noisy speech power spectrum Meter and power noise Power estimation.
- 6. a kind of speech-enhancement system according to claim 1, it is characterised in that the Wiener filtering enhancing module includes Fourier transformation, filter coefficient calculates and frequency domain filtering.
- A kind of a kind of 7. method of speech enhan-cement described in claim 1, it is characterised in that including：Step 1：Noisy speech passes through voice-input device input system, after antialiasing filter carries out pre-filtering processing Sampled again by A/D modular converters and quantification treatment, framing and plus Hamming window processing are finally carried out to Noisy Speech Signal；Step 2：Noisy Speech Signal after adding window is subjected to end-point detection using the end-point detecting method based on frequency band variance；Step 3：Signal-to-noise ratio (SNR) estimation is done to Noisy Speech Signal, while signal-to-noise ratio (SNR) estimation value sends the selection control unit to, The selection control unit judges that selection more window spectrum spectrums subtract enhancing module or Wiener filtering according to default signal to noise ratio codomain Strengthen module, step 4 is carried out if selecting more window spectrum spectrums to subtract enhancing module, if why selection Wiener filtering speech enhan-cement module is entered Row step 5；Step 4：Noisy Speech Signal carries out noisy speech power Spectral Estimation by the power Spectral Estimation and noise power spectrum is estimated Meter, subtracts the calculating of gain calculating progress power spectrum yield value by the spectrum, and the enhancing amplitude spectrum is calculated according to noisy speech Power spectrum, noise power spectrum and power spectrum yield value calculate enhancing amplitude spectrum, and enhancing amplitude spectrum passes through inverse fast fourier transform Voice after synthesis enhancing afterwards；Step 5：After Noisy Speech Signal is by Fast Fourier Transform (FFT), filter coefficient is calculated simultaneously according to signal-to-noise ratio (SNR) estimation value Amplitude spectrum must be strengthened by carrying out frequency domain filtering, and enhancing amplitude spectrum synthesizes enhanced voice after inverse fast fourier transform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710888848.6A CN107680610A (en) | 2017-09-27 | 2017-09-27 | A kind of speech-enhancement system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710888848.6A CN107680610A (en) | 2017-09-27 | 2017-09-27 | A kind of speech-enhancement system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107680610A true CN107680610A (en) | 2018-02-09 |
Family
ID=61136041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710888848.6A Pending CN107680610A (en) | 2017-09-27 | 2017-09-27 | A kind of speech-enhancement system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107680610A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110047470A (en) * | 2019-04-11 | 2019-07-23 | 深圳市壹鸽科技有限公司 | A kind of sound end detecting method |
| WO2019227590A1 (en) * | 2018-05-29 | 2019-12-05 | 平安科技（深圳）有限公司 | Voice enhancement method, apparatus, computer device, and storage medium |
| CN111986686A (en) * | 2020-07-09 | 2020-11-24 | 厦门快商通科技股份有限公司 | Short-time speech signal-to-noise ratio estimation method, device, equipment and storage medium |
| CN111986660A (en) * | 2020-08-26 | 2020-11-24 | 深圳信息职业技术学院 | Single-channel speech enhancement method, system and storage medium for neural network sub-band modeling |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030033139A1 (en) * | 2001-07-31 | 2003-02-13 | Alcatel | Method and circuit arrangement for reducing noise during voice communication in communications systems |
| CN101976566A (en) * | 2010-07-09 | 2011-02-16 | 瑞声声学科技(深圳)有限公司 | Voice enhancement method and device using same |
| CN103824563A (en) * | 2014-02-21 | 2014-05-28 | 深圳市微纳集成电路与系统应用研究院 | Hearing aid denoising device and method based on module multiplexing |
| CN105679330A (en) * | 2016-03-16 | 2016-06-15 | 南京工程学院 | Digital hearing aid noise reduction method based on improved sub-band signal-to-noise ratio estimation |
-
2017
- 2017-09-27 CN CN201710888848.6A patent/CN107680610A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030033139A1 (en) * | 2001-07-31 | 2003-02-13 | Alcatel | Method and circuit arrangement for reducing noise during voice communication in communications systems |
| CN101976566A (en) * | 2010-07-09 | 2011-02-16 | 瑞声声学科技(深圳)有限公司 | Voice enhancement method and device using same |
| CN103824563A (en) * | 2014-02-21 | 2014-05-28 | 深圳市微纳集成电路与系统应用研究院 | Hearing aid denoising device and method based on module multiplexing |
| CN105679330A (en) * | 2016-03-16 | 2016-06-15 | 南京工程学院 | Digital hearing aid noise reduction method based on improved sub-band signal-to-noise ratio estimation |
Non-Patent Citations (2)
| Title |
|---|
| 武鹏鹏等: "基于多窗谱估计的改进谱减法", 《现代电子技术》 * |
| 韦高梧: "基于单信道的语音增强算法的研究与改进", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019227590A1 (en) * | 2018-05-29 | 2019-12-05 | 平安科技（深圳）有限公司 | Voice enhancement method, apparatus, computer device, and storage medium |
| CN110047470A (en) * | 2019-04-11 | 2019-07-23 | 深圳市壹鸽科技有限公司 | A kind of sound end detecting method |
| CN111986686A (en) * | 2020-07-09 | 2020-11-24 | 厦门快商通科技股份有限公司 | Short-time speech signal-to-noise ratio estimation method, device, equipment and storage medium |
| CN111986686B (en) * | 2020-07-09 | 2023-01-03 | 厦门快商通科技股份有限公司 | Short-time speech signal-to-noise ratio estimation method, device, equipment and storage medium |
| CN111986660A (en) * | 2020-08-26 | 2020-11-24 | 深圳信息职业技术学院 | Single-channel speech enhancement method, system and storage medium for neural network sub-band modeling |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108735213B (en) | Voice enhancement method and system based on phase compensation | |
| CN109767783B (en) | Voice enhancement method, device, equipment and storage medium | |
| CN107680610A (en) | A kind of speech-enhancement system and method | |
| US8010355B2 (en) | Low complexity noise reduction method | |
| CN107610712B (en) | Voice enhancement method combining MMSE and spectral subtraction | |
| CN103594094B (en) | Adaptive spectra subtraction real-time voice strengthens | |
| CN105390142B (en) | A kind of digital deaf-aid voice noise removing method | |
| CA2458428A1 (en) | System for suppressing wind noise | |
| CN104157295A (en) | Method used for detecting and suppressing transient noise | |
| CN107316648A (en) | A kind of sound enhancement method based on coloured noise | |
| CN103117066A (en) | Low signal to noise ratio voice endpoint detection method based on time-frequency instaneous energy spectrum | |
| CN103730126B (en) | Noise suppressing method and noise silencer | |
| CN103578477B (en) | Denoising method and device based on noise estimation | |
| CN111091833A (en) | Endpoint detection method for reducing noise influence | |
| CN105702262A (en) | Headset double-microphone voice enhancement method | |
| CN110808059A (en) | Speech noise reduction method based on spectral subtraction and wavelet transform | |
| CN107731242B (en) | Gain function speech enhancement method for generalized maximum posterior spectral amplitude estimation | |
| CN108053842A (en) | Shortwave sound end detecting method based on image identification | |
| CN103400578B (en) | Anti-noise voiceprint recognition device with joint treatment of spectral subtraction and dynamic time warping algorithm | |
| CN106328160B (en) | Noise reduction method based on double microphones | |
| CN106060717A (en) | High-definition dynamic noise-reduction pickup | |
| CN104810023B (en) | A kind of spectrum-subtraction for voice signals enhancement | |
| Goel et al. | Developments in spectral subtraction for speech enhancement | |
| CN102300014A (en) | Double-talk detection method applied to acoustic echo cancellation system in noise environment | |
| Meher et al. | Dynamic spectral subtraction on AWGN speech |
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 | ||
| CB02 | Change of applicant information |
Address after: Room 803, room F1, two, innovation industrial park, No. 2800, new avenue of innovation, Hefei high tech Zone, Anhui Applicant after: Anhui Shuo Wei Intelligent Technology Co., Ltd. Address before: 230088, H2, building 374, two innovation industrial park, 2800 innovation Avenue, Hefei hi tech Zone, Anhui Applicant before: Anhui Shuo Wei Intelligent Technology Co., Ltd. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180209 |
|
| RJ01 | Rejection of invention patent application after publication |