CN108174328A - The auto gain control method and control device of speech processing system - Google Patents
The auto gain control method and control device of speech processing system Download PDFInfo
- Publication number
- CN108174328A CN108174328A CN201810117004.6A CN201810117004A CN108174328A CN 108174328 A CN108174328 A CN 108174328A CN 201810117004 A CN201810117004 A CN 201810117004A CN 108174328 A CN108174328 A CN 108174328A
- Authority
- CN
- China
- Prior art keywords
- signal
- operational amplifier
- processing system
- gain control
- gain
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
- H04R3/00—Circuits for transducers, loudspeakers or microphones
-
- 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
- G10L25/78—Detection of presence or absence of voice signals
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03F—AMPLIFIERS
- H03F1/00—Details of amplifiers with only discharge tubes, only semiconductor devices or only unspecified devices as amplifying elements
- H03F1/02—Modifications of amplifiers to raise the efficiency, e.g. gliding Class A stages, use of an auxiliary oscillation
- H03F1/0205—Modifications of amplifiers to raise the efficiency, e.g. gliding Class A stages, use of an auxiliary oscillation in transistor amplifiers
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Signal Processing (AREA)
- Power Engineering (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Control Of Amplification And Gain Control (AREA)
- Circuit For Audible Band Transducer (AREA)
Abstract
The auto gain control method and control device of a kind of speech processing system, the control method include：The output signal input operational amplifier of microphone is amplified；Amplified signal is converted into digital signal；Voice activity detection is carried out to the digital signal；And when it is voice signal to detect the digital signal, compare the speech signal parameter of current operation window frame and previous operation window frame, corresponding gain adjustment data is exported to the operational amplifier according to comparison result.The control device includes operational amplifier connected in sequence, analog-digital converter and microcontroller, and the Gain tuning end of the operational amplifier is connect with the microcontroller.The present invention can track voice input signal with dynamically adjusting, and the gain of operational amplifier is effectively adjusted after processing, and the voice signal of input is enable effectively to reach voltage level target zone.
Description
Technical field
The present invention relates to the auto gain control methods and control of speech processing system more particularly to a kind of speech processing system
Device processed can apply the input signal processing in systems such as compress speech, recording or speech recognitions, can effectively provide up
The input speech signal of voltage level (power level) target zone, with the follow-up relevant treatment of sharp speech processing system.
Background technology
In the speech processing systems such as compress speech, recording or speech recognition, the preposition place of analog voice input signal
Reason is it is expected processed voice signal when entering Analog-digital Converter processing, it will be able to reach using required voltage position
Then standard provides subsequent voice processing unit and carries out using processing.
For speech processing system, as front end circuit, then the input of voice is through microphone and analog circuit
Digital voice data is converted into analog-digital converter (Analog to Digital Converter, ADC).But wheat
The voice signal of gram wind input can be influenced by many reasons, and generate the voltage level of mobility.Analyzing influence voice is defeated
The reason of entering to generate different voltages level or energy, including：1) sensitivity of microphone；2) phonetic entry source (voice) and wheat
The distance of gram wind；3) the energy size that phonetic entry source makes a sound；4) factors such as energy level of ambient noise sound, this
A little voltage levels or energy that can all influence to input voice.
However, some hardware handles modes are in the front end circuit of microphone and analog circuit, using fixed gain
Method amplifies the voltage level of voice.But it is close when phonetic entry source, when and during from remote microphone, the language that is obtained
Sound signal will be suddenly big or suddenly small；Or phonetic entry source make a sound energy production size variation when, can also influence input
Voltage level.In addition, under the processing of circuit of fixed gain, ambient noise is to be amplified by synchronous, and then influence follow-up language
The correctness of sound processing.Therefore, for subsequent voice processing unit is enable to be stablized and reaches the defeated of voltage level target zone
Enter voice signal, be then required to the method according to input speech signal adjust automatically gain, to achieve the purpose that using required.
In addition, at present there are many voice activity detection (Voice Activity Detection, VAD) method, when
Have in domain using the input average energy of voice sound frame, root-mean-square value (Root Mean Square, RMS) or in a frequency domain
Have using characteristic values such as the spectrum peaks (Spectral Peak) of voice sound frame, to be made whether the judgement of speech activity.But
It is to detect speech activity using aforementioned operation method, belongs to higher computation complexity；Especially in the spy using frequency domain
During value indicative, need input voice making frequency spectrum conversion, can also increase processing time and computation complexity.Therefore, these voices are lived
Though dynamic detection method can effectively detect speech activity, it is not useful in the application using the microprocessor of low order.
Invention content
The object of the present invention is to provide the auto gain control method and control device of a kind of speech processing system, the controls
Method and control device can track voice input signal with dynamically being adjusted in speech processing system, and effectively adjustment is transported after processing
The gain of amplifier is calculated, the voice signal of input is enable effectively to reach voltage level target zone.
In order to achieve the above object, the technical solution adopted by the present invention is as follows：
A kind of auto gain control method of speech processing system, the auto gain control method include：
The output signal input operational amplifier of microphone is amplified；
Amplified signal is converted into digital signal；
Voice activity detection is carried out to the digital signal；And
When it is voice signal to detect the digital signal, compare the voice of current operation window frame and previous operation window frame
Signal parameter exports corresponding gain adjustment data according to comparison result to the operational amplifier.
In the auto gain control method of above-mentioned speech processing system, it is preferable that carry out language to the digital signal
The step of sound activity detection, includes：
The digital signal is read by sample frequency；
Obtain the peak-peak of digital signal in each sampling sound frame；
Compare the peak-peak, obtain the peak swing of signal in operation window frame；And
By the peak swing compared with the threshold values set, judge whether the digital signal is voice signal.
In the auto gain control method of above-mentioned speech processing system, it is preferable that each operation window frame includes four
Sound frame is sampled, the time span of each sampling sound frame is 20~40ms.
In the auto gain control method of above-mentioned speech processing system, it is preferable that gain adjustment data meets progressive
Adjust the principle of the op-amp gain.
In the auto gain control method of above-mentioned speech processing system, it is preferable that the speech signal parameter is
The peak swing of voice signal.
In the auto gain control method of above-mentioned speech processing system, it is preferable that the auto gain control method
It further includes：It is initial value by the Gain tuning of the operational amplifier when on startup or judgement is not voice signal.
In the auto gain control method of above-mentioned speech processing system, it is preferable that described according to comparison result
In the step of corresponding gain adjustment data being exported to the operational amplifier, when speech energy is continuously less than setting lower limit,
Quickly reduce the gain of the operational amplifier.
A kind of automatic gain control equipment of speech processing system, the automatic gain control equipment include from input terminal to
Output terminal operational amplifier connected in sequence, analog-digital converter and microcontroller, the Gain tuning of the operational amplifier
End is connect with the microcontroller, and the microcontroller is used for：Sample the digital signal of the analog-digital converter output；It is right
The digital signal carries out voice activity detection；And when detecting voice signal, compare current operation window frame and previous fortune
The speech signal parameter of window frame is calculated, corresponding gain adjustment data is exported to the operational amplifier according to comparison result.
In the automatic gain control equipment of above-mentioned speech processing system, it is preferable that the input of the operational amplifier
Anti aliasing purposes wave filter is provided between end and microphone.
In the automatic gain control equipment of above-mentioned speech processing system, it is preferable that the microcontroller is configured to：According to
The gain adjustment data is generated, and when speech energy is continuously described less than quickly being downgraded when setting lower limit according to principle is incrementally adjusted
The gain of operational amplifier.
Compared with prior art, the present invention at least has the advantages that：
Voice input signal can be made to reach set voltage level target zone, stable speech digit hits is provided
According to Audio Processing Unit, the operation of speech processes related application is carried out.
Directly using microphone input voice signal carry out automatic growth control, without to microphone input voltage level into
Row calibration procedure.
Voice activity detection can be carried out to microphone input signal, with sharp correct progress speech gain adjustment.
Description of the drawings
Fig. 1 is the flow chart of exemplary embodiments auto gain control method；
Fig. 2 is the flow chart of wherein voice activity detection method；
Fig. 3 is the definition schematic diagram for sampling sound frame and operation window frame；
Fig. 4 is the functional block diagram of exemplary embodiments automatic gain control equipment.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and examples.
With reference to Fig. 1, the auto gain control method of exemplary embodiments speech processing system includes：
Step S100, the output signal input operational amplifier of microphone is amplified；
Step S200, amplified signal is converted into digital signal；
Step S300, voice activity detection is carried out to the digital signal；And
Step S400, when it is voice signal to detect the digital signal, compare current operation window frame and previous operation
The speech signal parameter of window frame exports corresponding gain adjustment data according to comparison result to the operational amplifier.
With reference to Fig. 2, the method that voice activity detection is carried out to the digital signal, i.e. step S300 includes：
Step S301, the digital signal is read by sample frequency.
More specifically include：The output valve of analog-digital converter (DAC) is read, which is expressed as PK (Peak
value)。
Since the output valve of analog-digital converter is waveform sampling value, the meeting in transfer process when for voice signal
Signed magnitude is generated, so before the comparison of sound frame sample magnitude, further comprises and PK values takes absolute value, the absolute value table
It is shown as PKCUR=| PK |.
Step S302, the peak-peak of digital signal in each sampling sound frame is obtained.
More specifically include：It is continuous to compare the sample magnitude in same sound frame, and the maximum value after comparing is stored, when complete
Hits and numerical value into 1 sound frame compare, that is, obtain the PK maximum values of the sampling sound frame (that is, the maximum peak of digital signal
Value), it is expressed as
PKMAXFn, n=0,1,2,3
Wherein, Fn, n=0,1,2,3 be 4 continuous sampling sound frame PK maximum values.It stores and adopts in agc algorithm
The buffer of sample sound frame PK maximum values is to recycle, that is, is carrying out next sampling sound frame before processing, will adjust storage
The register location parameter of next sampling sound frame PK maximum values, i.e. step S305, (n+1) mod 4.
Step S303, peak-peak described in comparison (i.e. above-mentioned PK maximum values), the maximum for obtaining signal in operation window frame are shaken
Width.
In an exemplary embodiment, operation window frame is made of continuous four sampling sound frames, is adopted by comparing continuous four
The PK maximum values of sample sound frame, select maximum value
PKMAX=max (PKMAXFn), n=0,1,2,3,
As the peak swing of signal in operation window frame, then voice signal judgement is carried out using PKMAX.
Step S304, by the peak swing compared with the threshold values set, judge whether the digital signal is voice letter
Number.
The judgement formula that voice signal judgement is carried out using PKMAX is as follows：
If PKMAX >=PKTHD, then VA=1 (judge that the window frame signal is voice signal),
if PKMAX<PKTHD, then VA=0 (judge that the window frame signal is not voice signal)；
Wherein, PKTHD is the threshold values of setting.
If judging result is not voice signal, the parameter and associated initial value of step S305, adjustment sampling sound frame are performed
Setting, the sample magnitude into next sound frame is read and numerical value compares.
If judging result is voice signal, performs step S400 and carry out Gain tuning, specifically include：Step S401, life
Into gain adjustment data；And step S402, gain adjustment data exported to the Gain tuning end of aforementioned operational amplifier.
In step S401, the method for generating gain adjustment data includes：When detecting as voice signal, relatively more current fortune
The speech signal parameter of window frame and previous operation window frame is calculated, corresponding gain adjustment data is generated according to comparison result, makes operation
The gain of amplifier is tuned up or is turned down.Wherein, the speech signal parameter uses the peak swing of voice signal in operation window frame,
I.e. above-mentioned PKMAX.As it can be seen that the foundation of Gain tuning is the PKMAX of current operation window frame and previous operation window frame
PKMAXPRE。
Further, gradual increase/reduction mode adjustment gain is taken, the voltage level for enabling input speech signal is steady
Surely increase/reduction reaches the voltage level target area of setting.
The method of generation gain adjustment data further comprises, when on startup or judgement is not voice signal, by institute
The Gain tuning for stating operational amplifier is initial value.The initial value is the experiment value that speech processing system obtains.
It is typically the variation changed from small to big since the energy of phonetic entry is when entering system；But when phonetic entry knot
Shu Shi is then rapidly by the sound change of tune into asonant state.Therefore, the method for generating gain adjustment data is further also wrapped
It includes, when speech energy is continuously less than setting lower limit, quickly reduces the gain of the operational amplifier.Exist to avoid ambient noise
Still lasting amplification, is caused in voice activity detection to the erroneous judgement of phonetic entry under high-gain numerical value.
In exemplary embodiments, each operation window frame includes four sampling sound frames, and the time span of each sampling sound frame is
20ms.The definition of sampling sound frame and operation window frame is shown in Fig. 3.With reference to Fig. 3, according to time axial adjustment input speech signal,
Each sampling sound frame can one speech peaks of output (Peak value, abbreviation PK value)；Continuous 4 sampling sounds frame forms 1 increasing
The operation window frame of benefit adjustment carries out operation using the speech peaks in window frame, judges whether voice signal and generation gain tune
Entire data.
For the sound frame sampling time length of speech processes, according to the time that sound frame is sampled in the correlative study of speech sample
Length is the more appropriate sampling time usually between 20~40ms.Therefore, the time span of sound frame is sampled in an exemplary embodiment
It is to be adjusted according to sample frequency, makes the time span of sampling sound frame about in 20ms or so；If sample frequency is 8KHz, sound frame is sampled
When interior sampling number is 160, the time span of sampling sound frame is equal to 20ms.In addition, the operation window frame of voice activity detection
The time span of time span, then equal to 4 sampling sound frames.Sampled data in each Gain tuning operation window frame has 75% to be
Be folded (overlap), it is such design be to make in Gain tuning algorithm, will not be because of the variation of noise signal in single sound frame
It influences, and then significantly adjust gain numerical value；The operation window frame of suitable length can obtain relatively stable Gain tuning number
Value.
With reference to Fig. 4, exemplary embodiments automatic gain control equipment 100 includes：It is connected in sequence from input terminal to output terminal
Operational amplifier 101, analog-digital converter 103 and microcontroller 102, the Gain tuning end of the operational amplifier 101 with
The microcontroller 102 connects, and the microcontroller 102 is used for：Sample the number letter that the analog-digital converter 103 exports
Number；Voice activity detection is carried out to the digital signal；And when detecting voice signal, compare current operation window frame with before
The speech signal parameter of one operation window frame exports corresponding Gain tuning number according to comparison result to the operational amplifier 101
According to.
Further anti aliasing purposes wave filter is provided between the input terminal of the operational amplifier 101 and microphone 200
(Anti-Aliasing Filter,AAF)300.Signal beyond speech processes range can be filtered out, and increase by the wave filter
Add antijamming capability of the input signal to noise.Further, Common Mode Rejection Ratio (Common-Mode is used after filtering
Rejection Ratio, CMRR) pattern is input to operational amplifier 101.
Further, the microcontroller 102 is configured to：The gain adjustment data is generated according to principle is incrementally adjusted, and
The gain of the operational amplifier 101 is quickly downgraded when speech energy is continuously less than setting lower limit.
Further, the microcontroller 102 is configured to：Each operation window frame includes four sampling sound frames, each sampling sound
The time span of frame is 20~40ms.
As it can be seen that voice activity detection and gradual Gain tuning model are merged in the microcontroller 102.Phonetic entry
After microphone 200 carries out pickup, the sound string wave signal of generation enters automatic increase after the filtering of anti aliasing purposes wave filter 300
Beneficial control device 100 is handled, and in automatic gain control equipment 100, the voice signal voltage level of input passes through operation amplifier
After the enhanced processing of device 101, digital voice data is converted to through analog-digital converter 103, then via microcontroller 102
Processing, obtains the gain adjustment data of next sound frame phonetic entry, and feeds back (Feedback) gain adjustment data to fortune
Amplifier 101 is calculated, the gain of next sound frame phonetic entry is provided, the energy size of phonetic entry is enable to be automatically adjusted to electricity
Press level target area.
Digital voice data after aforementioned processing is completed then is supplied to digital speech processing unit 400 to carry out required
Processing or calculation, for example, digital filtering, speech recognition, data compression storage etc., reach the application demand of speech processing system.
The present invention is described in detail above by specific embodiment, these detailed description are only limited to help
Those skilled in the art understand that present disclosure, can not be interpreted as limiting the scope of the invention.Art technology
Various retouchings that personnel carry out said program under present inventive concept, equivalent transformation etc. should be included in the protection model of the present invention
In enclosing.
Claims (11)
1. a kind of auto gain control method of speech processing system, which is characterized in that the auto gain control method includes：
The output signal input operational amplifier of microphone is amplified；
Amplified signal is converted into digital signal；
Voice activity detection is carried out to the digital signal；And
When it is voice signal to detect the digital signal, compare the voice signal of current operation window frame and previous operation window frame
Parameter exports corresponding gain adjustment data according to comparison result to the operational amplifier.
2. the auto gain control method of speech processing system according to claim 1, which is characterized in that the number
The step of signal progress voice activity detection, includes：
The digital signal is read by sample frequency；
Obtain the peak-peak of digital signal in each sampling sound frame；
Compare the peak-peak, obtain the peak swing of signal in operation window frame；And
By the peak swing compared with the threshold values set, judge whether the digital signal is voice signal.
3. the auto gain control method of speech processing system according to claim 2, which is characterized in that each operation window
Frame includes four sampling sound frames, and the time span of each sampling sound frame is 20~40ms.
4. the auto gain control method of speech processing system according to claim 1, which is characterized in that in the root
According to comparison result into the step of operational amplifier output corresponding gain adjustment data, gain adjustment data meets progressive
Adjust the principle of the op-amp gain.
5. the auto gain control method of speech processing system according to claim 1, which is characterized in that the voice
Signal parameter is the peak swing of voice signal.
6. the auto gain control method of speech processing system according to claim 1, which is characterized in that the automatic increasing
Beneficial control method further includes：When on startup or judgement is not voice signal, the Gain tuning by the operational amplifier is first
Initial value.
7. the auto gain control method of speech processing system according to claim 1, which is characterized in that in the root
According to comparison result into the step of operational amplifier output corresponding gain adjustment data, set when speech energy is continuously less than
It fixs in limited time, quickly reduces the gain of the operational amplifier.
8. a kind of automatic gain control equipment of speech processing system, which is characterized in that the automatic gain control equipment includes
Operational amplifier connected in sequence, analog-digital converter and microcontroller, the operational amplifier from input terminal to output terminal
Gain tuning end connect with the microcontroller, the microcontroller is used for：Sample the analog-digital converter output
Digital signal；Voice activity detection is carried out to the digital signal；And when detecting voice signal, compare current operation window
The speech signal parameter of frame and previous operation window frame exports corresponding Gain tuning according to comparison result to the operational amplifier
Data.
9. the automatic gain control equipment of speech processing system according to claim 8, which is characterized in that the operation is put
Anti aliasing purposes wave filter is provided between the input terminal and microphone of big device.
10. the automatic gain control equipment of speech processing system according to claim 8, which is characterized in that the micro-control
Device processed is configured to：The gain adjustment data is generated, and when speech energy is continuously less than setting lower limit according to principle is incrementally adjusted
When quickly downgrade the gain of the operational amplifier.
11. the automatic gain control equipment of speech processing system according to claim 8, which is characterized in that the micro-control
Device processed is configured to：Each operation window frame includes four sampling sound frames, and the time span of each sampling sound frame is 20~40ms.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810117004.6A CN108174328A (en) | 2018-02-06 | 2018-02-06 | The auto gain control method and control device of speech processing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810117004.6A CN108174328A (en) | 2018-02-06 | 2018-02-06 | The auto gain control method and control device of speech processing system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108174328A true CN108174328A (en) | 2018-06-15 |
Family
ID=62513311
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810117004.6A Pending CN108174328A (en) | 2018-02-06 | 2018-02-06 | The auto gain control method and control device of speech processing system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108174328A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112397085A (en) * | 2019-08-16 | 2021-02-23 | 骅讯电子企业股份有限公司 | System and method for processing voice and information |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5983183A (en) * | 1997-07-07 | 1999-11-09 | General Data Comm, Inc. | Audio automatic gain control system |
| CN101009099A (en) * | 2007-01-26 | 2007-08-01 | 北京中星微电子有限公司 | Digital auto gain control method and device |
| CN101499276A (en) * | 2008-12-25 | 2009-08-05 | 爱唱数码科技（上海）有限公司 | Voice primitive period marking method |
| CN204857179U (en) * | 2011-01-10 | 2015-12-09 | 艾利佛公司 | Pronunciation activity detector |
| CN105355197A (en) * | 2015-10-30 | 2016-02-24 | 百度在线网络技术（北京）有限公司 | Gain processing method and device for speech recognition system |
| CN105490654A (en) * | 2016-01-20 | 2016-04-13 | 山东大学 | Automatic gain controller control method of voice acquisition system and circuit thereof |
-
2018
- 2018-02-06 CN CN201810117004.6A patent/CN108174328A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5983183A (en) * | 1997-07-07 | 1999-11-09 | General Data Comm, Inc. | Audio automatic gain control system |
| CN101009099A (en) * | 2007-01-26 | 2007-08-01 | 北京中星微电子有限公司 | Digital auto gain control method and device |
| CN101499276A (en) * | 2008-12-25 | 2009-08-05 | 爱唱数码科技（上海）有限公司 | Voice primitive period marking method |
| CN204857179U (en) * | 2011-01-10 | 2015-12-09 | 艾利佛公司 | Pronunciation activity detector |
| CN105355197A (en) * | 2015-10-30 | 2016-02-24 | 百度在线网络技术（北京）有限公司 | Gain processing method and device for speech recognition system |
| CN105490654A (en) * | 2016-01-20 | 2016-04-13 | 山东大学 | Automatic gain controller control method of voice acquisition system and circuit thereof |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112397085A (en) * | 2019-08-16 | 2021-02-23 | 骅讯电子企业股份有限公司 | System and method for processing voice and information |
| CN112397085B (en) * | 2019-08-16 | 2024-03-01 | 骅讯电子企业股份有限公司 | Sound message processing system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US6910011B1 (en) | Noisy acoustic signal enhancement | |
| CN101110217A (en) | Automatic gain control method for audio signal and apparatus thereof | |
| CN111465981A (en) | Analog voice activity detector system and method | |
| JP2780676B2 (en) | Voice recognition device and voice recognition method | |
| CN111091833A (en) | Endpoint detection method for reducing noise influence | |
| JP2012088404A (en) | Noise power estimation device and noise power estimation method, and voice recognition device and voice recognition method | |
| CN112951259A (en) | Audio noise reduction method and device, electronic equipment and computer readable storage medium | |
| CN108231088B (en) | Fast automatic gain control method and device applied to voice digital signals | |
| CN108174328A (en) | The auto gain control method and control device of speech processing system | |
| WO2005020212A1 (en) | Signal analysis device, signal processing device, speech recognition device, signal analysis program, signal processing program, speech recognition program, recording medium, and electronic device | |
| WO2017128910A1 (en) | Method, apparatus and electronic device for determining speech presence probability | |
| US7646912B2 (en) | Method and device for ascertaining feature vectors from a signal | |
| US20020173957A1 (en) | Speech recognizer, method for recognizing speech and speech recognition program | |
| JP2002366192A (en) | Method and device for recognizing voice | |
| CN116884405A (en) | Speech instruction recognition method, device and readable storage medium | |
| US6785648B2 (en) | System and method for performing speech recognition in cyclostationary noise environments | |
| JP2010102129A (en) | Fundamental frequency extracting method, fundamental frequency extracting device, and program | |
| US20110234438A1 (en) | Signal Processing Apparatus | |
| CN115050380A (en) | Audio signal processing method, earphone, and computer-readable storage medium | |
| US20140023208A1 (en) | Digital audio amplification device using harmonics and method thereof | |
| US20230253010A1 (en) | Voice activity detection (vad) based on multiple indicia | |
| CN109862463A (en) | Earphone audio playback method, earphone and its computer readable storage medium | |
| JP2007206154A (en) | Voice section detection under real environment noise | |
| TWI756817B (en) | Voice activity detection device and method | |
| JP2975808B2 (en) | Voice recognition device |
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 |