CN107274909A - A kind of active the machine audio removing method in speech recognition - Google Patents
A kind of active the machine audio removing method in speech recognition Download PDFInfo
- Publication number
- CN107274909A CN107274909A CN201710455665.5A CN201710455665A CN107274909A CN 107274909 A CN107274909 A CN 107274909A CN 201710455665 A CN201710455665 A CN 201710455665A CN 107274909 A CN107274909 A CN 107274909A
- Authority
- CN
- China
- Prior art keywords
- audio
- amplifier
- output end
- circuit
- swept
- 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
- 238000000034 method Methods 0.000 title claims abstract description 21
- 230000005236 sound signal Effects 0.000 claims abstract description 22
- 230000007613 environmental effect Effects 0.000 claims abstract description 18
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 108090000623 proteins and genes Proteins 0.000 description 6
- 230000004044 response Effects 0.000 description 6
- 238000011161 development Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000007634 remodeling Methods 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 238000004088 simulation Methods 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
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/20—Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech
-
- 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
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- 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
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/223—Execution procedure of a spoken command
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Signal Processing (AREA)
- Circuit For Audible Band Transducer (AREA)
- Stereophonic System (AREA)
Abstract
The invention discloses active the machine audio removing method in a kind of speech recognition, it is characterised in that comprises the following steps：Audio-frequency generator sends swept-frequency signal；Swept-frequency signal is scanned through circuit and obtains external audio characteristic；Output is modified to audio bucking circuit according to external audio characteristic.Active the machine audio that the present invention is also provided in a kind of speech recognition eliminates circuit, including：Audio-frequency generator, scanning circuit and audio bucking circuit, the audio-frequency generator are used to produce swept-frequency signal and local audio signal, and swept-frequency signal exports external audio characteristic by scanning circuit, and local audio signal is exported by audio bucking circuit.The present invention is offset to the environmental audio for being mingled in phonetic order, so as to transmit clearly user voice signal.
Description
Technical field
The present invention relates to technical field of voice recognition, and in particular to active the machine audio in a kind of speech recognition is eliminated
Method and circuit, are mainly used in the device that vehicle-mounted and intelligent toy etc. includes speech recognition.
Background technology
In the speech recognition systems such as vehicle-mounted at present and intelligent toy, can usually it be related to equipment while playing audio, while needing
Receive external voice instruction.Now microphone can pick up the audio and external voice of the machine broadcasting simultaneously.And speech recognition system
System needs as far as possible pure external sound signal, it is therefore desirable to offset the machine audio signal using technological means.It is existing
Technical approach be the pickoff signals from the audio output circuit of the machine, and the signal of microphone pickup feeds back to audio and supported together
Disappear algorithm.
The key of audio cancellation algorithms is to try to obtain real the machine play signal.And the signal and sheet of microphone pickup
Ground feedback signal is compared, and receives the influence of the acoustic frequency response characteristic of loudspeaker sound chamber, microphone and surrounding environment, it is therefore desirable to
In algorithm, local feedback signal is subjected to signal condition, to obtain and microphone is tried one's best similar signal.
For the product of determination, the frequency response curve of loudspeaker sound chamber and microphone is fixed, and the frequency of product use environment
It is uncertain to ring curve.Such a audio cancellation algorithms are difficult to have good effect under various usage scenarios.Work as production
When the use environment of product changes, phonetic recognization rate will be unstable, with not good environmental suitability.
In addition, if the frequency response curve of the different product of design, loudspeaker sound chamber and microphone also will be different.This is required for
Expend the substantial amounts of debug time of audio engineer and go the different device of adaptation and use environment, add R＆D costs.
The content of the invention
In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is that providing a kind of speech recognition
In active the machine audio removing method and circuit.
In order to achieve the above object, the present invention is achieved using following technical scheme：
Active the machine audio removing method in a kind of speech recognition, it is characterised in that comprise the following steps：
Audio-frequency generator sends swept-frequency signal；
Swept-frequency signal is scanned through circuit and obtains external audio characteristic；
Audio hair emperorship sends local audio；
Local audio is exported by the amendment of audio bucking circuit according to external audio characteristic.
Further, the scanning circuit includes：First DAC, the first amplifier, the first loudspeaker, the first ten
Sound device, the second amplifier, the first ADC converters and comparison circuit；Swept-frequency signal is divided into two-way, wherein turning all the way with the first DAC
Parallel operation is connected, and the output end of first DAC is connected with the first amplifier, the output end connection of first amplifier
First loudspeaker plays swept-frequency signal into environment, and environmental audio is acquired by the first sound pick-up, and described the first ten
The output end of sound device connects the second amplifier, and the output end of second amplifier connects the first ADC converters, described first
ADC converters output end connection comparison circuit, while swept-frequency signal another road output be also connected to comparison circuit, by than
Compared with circuit output external audio characteristic.
Further, the audio bucking circuit includes：Second DAC, the 3rd amplifier, the second loudspeaker,
Four high guaily unit device, the 2nd ADC converters, the second sound pick-up, the 5th amplifier, the 3rd ADC converters and acoustic characteristic inverse operation electricity
Road；The output end of second DAC connects the 3rd amplifier, and the output end of the 3rd amplifier is divided into two-way, its
In connect loudspeaker all the way, another amplifiers of Lu Yu tetra- connection, the output end of the 4th amplifier connects the 2nd ADC conversions
Device, the output end of second sound pick-up connects the 5th amplifier, and the output end of the 5th amplifier connects the 3rd ADC conversions
Device, the output end of the 2nd ADC converters and the 3rd ADC converters is connected with acoustic characteristic inverse operation circuit.
Further, the swept-frequency signal is the sampling frequency in the range of 300Hz~3KHz.
Further, the swept-frequency signal uses the music signal for including sweep check signal.
Further, the external audio characteristic includes amplitude versus frequency characte, time delay and phase characteristic.
A kind of active the machine audio in speech recognition eliminates circuit, it is characterised in that including：Audio-frequency generator, sweep
Scanning circuit and audio bucking circuit, the audio-frequency generator are used to produce swept-frequency signal and local audio signal, and swept-frequency signal is led to
Circuit output external audio characteristic is over-scanned, local audio signal is exported by audio bucking circuit.
Further, the scanning circuit includes：First DAC, the first amplifier, the first loudspeaker, the first ten
Sound device, the second amplifier, the first ADC converters and comparison circuit；Swept-frequency signal is divided into two-way, wherein turning all the way with the first DAC
Parallel operation is connected, and the output end of first DAC is connected with the first amplifier, the output end connection of first amplifier
First loudspeaker plays swept-frequency signal into environment, and environmental audio is acquired by the first sound pick-up, and described the first ten
The output end of sound device connects the second amplifier, and the output end of second amplifier connects the first ADC converters, described first
ADC converters output end connection comparison circuit, while swept-frequency signal another road output be also connected to comparison circuit, by than
Compared with circuit output external audio characteristic.
Further, the audio bucking circuit includes：Second DAC, the 3rd amplifier, the second loudspeaker,
Four high guaily unit device, the 2nd ADC converters, the second sound pick-up, the 5th amplifier, the 3rd ADC converters and acoustic characteristic inverse operation electricity
Road；The output end of second DAC connects the 3rd amplifier, and the output end of the 3rd amplifier is divided into two-way, its
In connect loudspeaker all the way, another amplifiers of Lu Yu tetra- connection, the output end of the 4th amplifier connects the 2nd ADC conversions
Device, the output end of second sound pick-up connects the 5th amplifier, and the output end of the 5th amplifier connects the 3rd ADC conversions
Device, the output end of the 2nd ADC converters and the 3rd ADC converters is connected with acoustic characteristic inverse operation circuit.
Further, the swept-frequency signal is the sampling frequency in the range of 300Hz~3KHz.
Further, the swept-frequency signal uses the music signal for including sweep check signal.
Present invention has the advantages that：
The method that the active audio that the present invention is used is offset, one is considered as always by loudspeaker, sound pick-up and surrounding environment
The external audio feature of body, sweep check signal is actively sent by audio-frequency generator, and then the signal with microphone pickup enters
Row compares, the disposable amplitude versus frequency characte and time delay, phase characteristic for obtaining external audio response；The Dynamic gene of feedback signal is used for
Audio bucking circuit is controlled, after Dynamic gene is obtained, audio-frequency generator sends local audio, and local audio is in audio
Under the processing of bucking circuit, the audio of external environment condition is offset by Dynamic gene, so as to export pure local audio, works as user
When sending phonetic order by the equipment, the environmental audio for being mingled in phonetic order is offset, clearly used so as to transmit
Family voice signal.The parameter for the audio cancellation algorithms that the present invention fixes development, becoming can be carried out certainly according to environment
Adapt to the parameter of adjustment；It is combined using swept-frequency signal with other audio signals, such as music signal or voice signal are obtaining
The process of new parameter is taken, influence very little is used on user；Development adaptation difficulty is reduced, voice knowledge can be quickly added
Other environmental suitability, improves overall discrimination.
Brief description of the drawings
Fig. 1 is active the machine audio removing method flow chart in a kind of speech recognition proposed by the present invention；
Fig. 2 is that active the machine audio in a kind of speech recognition proposed by the present invention eliminates circuit structure diagram；
Fig. 3 is the scanning circuit structure chart proposed by the present invention；
Fig. 4 is the audio bucking circuit structure chart proposed by the present invention.
Embodiment
Below in conjunction with the accompanying drawings and embodiment describes the specific embodiment of the invention：
As shown in Figures 1 to 4, it illustrates the embodiment of the present invention,
Active the machine audio removing method in a kind of speech recognition, it is characterised in that comprise the following steps：
Step 101, audio-frequency generator sends swept-frequency signal；
Step 102, swept-frequency signal is scanned through circuit and obtains external audio characteristic；
Step 103, audio hair emperorship sends local audio；
Step 104, local audio is exported by the amendment of audio bucking circuit according to external audio characteristic.
When environment changes or speech recognition effect is not good, user can pass through Environment sound under squelch
Swept-frequency signal actively occurs for frequency corrective operations, triggering, carries out recalculating for audio frequency parameter.In addition, this method is also an option that
Select and periodically send test swept-frequency signal, actively to adapt to the change of environment and avoid environmental noise from influenceing.
The method that the active audio that the present invention is used is offset, one is considered as always by loudspeaker, sound pick-up and surrounding environment
The external audio feature of body, sweep check signal is actively sent by audio-frequency generator, and then the signal with microphone pickup enters
Row compares, the disposable amplitude versus frequency characte and time delay, phase characteristic for obtaining external audio response.Because the audio signal actively sent
Consistent with local feedback signal, thus can by external audio response characteristic inverse transformation into local feedback signal adjustment because
Son.That is, it is almost consistent with microphone pickup signal after local feedback signal is calculated by the factor, reach automatic straightening
Purpose.The Dynamic gene of feedback signal is used to be controlled audio bucking circuit, after Dynamic gene is obtained, audio-frequency generator
Local audio is sent, local audio offsets the audio of external environment condition by Dynamic gene under the processing of audio bucking circuit, from
And pure local audio is exported, and when user sends phonetic order by the equipment, the ambient sound to being mingled in phonetic order
Frequency is offset, so as to transmit clearly user voice signal.
Scanning circuit includes：First DAC 221, the first amplifier 222, the first loudspeaker 223, the first sound pick-up
224th, the second amplifier 225, the first ADC converters 226 and comparison circuit 227；Swept-frequency signal is divided into two-way, wherein all the way with
One DAC 221 is connected, and the output end of first DAC 221 is connected with the first amplifier 222, and described first puts
The output end of big device 222 connects the first loudspeaker 223 and plays swept-frequency signal into environment, by the first sound pick-up 224 to ring
Border audio is acquired, and the output end of first sound pick-up 224 connects the second amplifier 225, second amplifier 225
Output end connects the first ADC converters 226, and the output end connection comparison circuit 227 of the first ADC converters 226 is swept simultaneously
Another road output of frequency signal is also connected to comparison circuit 227, and external audio characteristic is exported by comparison circuit 227.
In the embodiment of the present invention, the first amplifier and the second amplifier use analogue audio frequency amplifier, work as audio-frequency generator
When producing swept-frequency signal, swept-frequency signal converts digital signals into analog signal by the first DAC, and the simulation is believed
Number it is amplified and is output to through loudspeaker in environment by the first analogue audio frequency amplifier, environmental audio is carried out by sound pick-up
Audio collection, and output is amplified by the second analogue audio frequency amplifier, now the audio signal after the amplification passes through again
One ADC converters convert analog signals into data signal, now send out audio the frequency sweep letter that emperorship is produced by comparison circuit
Number and the data signal that is exported by the first ADC converters be compared, will swept-frequency signal in data signal offset after obtain
Environmental audio signal, then obtains environmental audio characteristic by audio inverse transform method, and the environmental audio characteristic is rung including audio
Amplitude versus frequency characte, time delay and the phase characteristic answered.
In the embodiment of the present invention, the swept-frequency signal that audio-frequency generator is produced can use manual triggering mode, can also adopt
With automatic triggering；When being triggered using manual triggering mode, by setting trigger button, when ambient noise is larger or voice is known
When not good, audio-frequency generator can be triggered manually and produces swept-frequency signal；During using automatic triggering, swept-frequency signal can be set to send out
Go out the cycle, swept-frequency signal is sent by periodicity, automatic detection is carried out to external environment condition, to adapt to the change of environment automatically and keep away
Exempt from the influence of noise.
Audio bucking circuit includes：Second DAC 231, the 3rd amplifier 232, the second loudspeaker the 233, the 4th fortune
Put device 234, the 2nd ADC converters 235, the second sound pick-up 236, the 5th amplifier 237, the 3rd ADC converters 238 and audio special
Property inverse operation circuit 239；The output end of second DAC 231 connects the 3rd amplifier 232, the 3rd amplifier
232 output end is divided into two-way, wherein connecting loudspeaker 233 all the way, another amplifiers 234 of Lu Yu tetra- are connected, and the described 4th
The output end of amplifier 234 connects the 2nd ADC converters 235, and the output end of second sound pick-up 236 connects the 5th amplifier
237, the output end of the 5th amplifier 237 connects the 3rd ADC converters 238, the 2nd ADC converters 235 and the 3rd
The output end of ADC converters 238 is connected with acoustic characteristic inverse operation circuit 239.
In the embodiment of the present invention, when scanning circuit obtains environmental audio characteristic, that is, the audio frequency parameter of environment is obtained, is led to
Cross acoustic characteristic inverse operation circuit and output is adjusted to audio signal, so that pure audio signal is obtained, specifically, locally
The audio signal of audio output converts digital signals into analog signal by the second DAC, and passes through the 3rd amplifier
The analog signal is amplified and exported through loudspeaker, the audio signal that the second sound pick-up is exported to loudspeaker is picked up,
The audio signal of this in pick process is to be doped with environmental audio signal, and elementary audio all the way is also drawn in the 3rd amplifier out
Signal, exports fundamental digital audio signal, now the fundamental digital audio signal by the 4th amplifier and the 2nd ADC converters
Counteracting is compared by acoustic characteristic inverse operation circuit with environmental audio data signal, sent for user in environment to equipment
Phonetic order be that can obtain, then pass through sound identification module carry out phonetic order identification.
It is preferred that, swept-frequency signal is the sampling frequency in the range of 300Hz~3KHz.Wherein swept-frequency signal uses and includes frequency sweep
The music signal of test signal.In this case discomfort is brought to user in order to reduce test signal, selects suitable music letter
Number, swept-frequency signal is fused in music signal.What user heard is one section of music signal, and needed for being contained in the signal
Sweep check signal.Outside acoustic characteristic includes amplitude versus frequency characte, time delay and phase characteristic in the present invention.
Active the machine audio that the present invention is also provided in a kind of speech recognition eliminates circuit, it is characterised in that including：Sound
Frequency generator 201, scanning circuit 202 and audio bucking circuit 203, the audio-frequency generator 201 be used for produce swept-frequency signal and
Local audio signal, swept-frequency signal exports external audio characteristic, the output end and this rock noise of scanning circuit by scanning circuit 202
Frequency signal is connected with audio bucking circuit 203.
Wherein scanning circuit includes：First DAC, the first amplifier, the first loudspeaker, the first sound pick-up, second
Amplifier, the first ADC converters and comparison circuit；Swept-frequency signal is divided into two-way, wherein be connected all the way with the first DAC,
The output end of first DAC is connected with the first amplifier, and the output end connection first of first amplifier is raised one's voice
Device plays swept-frequency signal into environment, and environmental audio is acquired by the first sound pick-up, first sound pick-up it is defeated
Go out the second amplifier of end connection, the output end of second amplifier connects the first ADC converters, the first ADC converters
Output end connection comparison circuit, while swept-frequency signal another road output be also connected to comparison circuit, it is defeated by comparison circuit
Go out outside acoustic characteristic.
Audio bucking circuit includes：Second DAC, the 3rd amplifier, the second loudspeaker, four high guaily unit device, second
ADC converters, the second sound pick-up, the 5th amplifier, the 3rd ADC converters and acoustic characteristic inverse operation circuit；2nd DAC
The output end of converter connects the 3rd amplifier, and the output end of the 3rd amplifier is divided into two-way, wherein connection is raised one's voice all the way
Device, another amplifiers of Lu Yu tetra- connection, the output end of the 4th amplifier connects the 2nd ADC converters, and described the second ten
The output end of sound device connects the 5th amplifier, and the output end of the 5th amplifier connects the 3rd ADC converters, described second
The output end of ADC converters and the 3rd ADC converters is connected with acoustic characteristic inverse operation circuit.
It is preferred that, swept-frequency signal is the sampling frequency in the range of 300Hz~3KHz.
Swept-frequency signal uses the music signal for including sweep check signal.In this case in order to reduce test signal to using
Person brings discomfort, selects suitable music signal, swept-frequency signal is fused in music signal.What user heard is one section of music
Signal, and required sweep check signal is contained in the signal.
In the embodiment of the present invention, by actively sending swept-frequency signal, according to the feedback of ambient signal, the sound of the environment is obtained
It is used for audio after frequency indicatrix, inverse transformation and offsets parameter.So as to carry out denoising Processing to local audio so that the audio of transmission
Signal improves audio output quality, while swept-frequency signal is merged with music signal, provided for user without ambient external noise
One comfortable audition environment, is especially fused in the voice signal of report, when transmitting voice messaging, while frequency sweep is believed
Number send.
The parameter for the audio cancellation algorithms that the present invention fixes development, becoming can be carried out adaptively according to environment
The parameter of adjustment；It is combined using swept-frequency signal with other audio signals, such as music signal or voice signal, that is, is obtaining new
The process of parameter, influence very little is used on user；Development adaptation difficulty is reduced, speech recognition can be quickly added
Environmental suitability, improves overall discrimination.
The preferred embodiment for the present invention is explained in detail above in conjunction with accompanying drawing, but the invention is not restricted to above-mentioned implementation
Mode, can also be on the premise of present inventive concept not be departed from the knowledge that those of ordinary skill in the art possess
Make a variety of changes, these changes are related to correlation technique well-known to those skilled in the art, and these both fall within patent of the present invention
Protection domain.
Many other changes and remodeling can be made by not departing from the spirit and scope of the present invention.It should be appreciated that the present invention is not
It is limited to specific embodiment, the scope of the present invention is defined by the following claims.
Claims (10)
1. active the machine audio removing method in a kind of speech recognition, it is characterised in that comprise the following steps：
Audio-frequency generator sends swept-frequency signal；
Swept-frequency signal is scanned through circuit and obtains external audio characteristic；
Audio hair emperorship sends local audio；
Local audio is exported by the amendment of audio bucking circuit according to external audio characteristic.
2. active the machine audio removing method in a kind of speech recognition according to claim 1, it is characterised in that institute
Stating scanning circuit includes：First DAC, the first amplifier, the first loudspeaker, the first sound pick-up, the second amplifier, first
ADC converters and comparison circuit；Swept-frequency signal is divided into two-way, wherein being connected all the way with the first DAC, the first DAC
The output end of converter is connected with the first amplifier, and the output end of first amplifier connects the first loudspeaker by swept-frequency signal
Play into environment, environmental audio is acquired by the first sound pick-up, the output end connection second of first sound pick-up
Amplifier, the output end of second amplifier connects the first ADC converters, the output end connection of the first ADC converters
Comparison circuit, while another road output of swept-frequency signal is also connected to comparison circuit, it is special to export external audio by comparison circuit
Property.
3. active the machine audio removing method in a kind of speech recognition according to claim 1, it is characterised in that institute
Stating audio bucking circuit includes：Second DAC, the 3rd amplifier, the second loudspeaker, four high guaily unit device, the 2nd ADC conversions
Device, the second sound pick-up, the 5th amplifier, the 3rd ADC converters and acoustic characteristic inverse operation circuit；Second DAC
Output end connect the 3rd amplifier, the output end of the 3rd amplifier is divided into two-way, wherein connect loudspeaker all the way, another
The amplifiers of Lu Yu tetra- are connected, and the output end of the 4th amplifier connects the 2nd ADC converters, second sound pick-up it is defeated
Go out the 5th amplifier of end connection, the output end of the 5th amplifier connects the 3rd ADC converters, the 2nd ADC converters
It is connected with the output end of the 3rd ADC converters with acoustic characteristic inverse operation circuit.
4. active the machine audio removing method in a kind of speech recognition according to claim 1, it is characterised in that institute
It is the sampling frequency in the range of 300Hz~3KHz to state swept-frequency signal.
5. active the machine audio removing method in a kind of speech recognition according to claim 1, it is characterised in that institute
State swept-frequency signal and use the music signal for including sweep check signal.
6. active the machine audio in a kind of speech recognition eliminates circuit, it is characterised in that including：Audio-frequency generator, scanning
Circuit and audio bucking circuit, the audio-frequency generator are used to produce swept-frequency signal and local audio signal, and swept-frequency signal passes through
Scanning circuit exports external audio characteristic, and local audio signal is exported by audio bucking circuit.
7. active the machine audio in a kind of speech recognition according to claim 6 eliminates circuit, it is characterised in that institute
Stating scanning circuit includes：First DAC, the first amplifier, the first loudspeaker, the first sound pick-up, the second amplifier, first
ADC converters and comparison circuit；Swept-frequency signal is divided into two-way, wherein being connected all the way with the first DAC, the first DAC
The output end of converter is connected with the first amplifier, and the output end of first amplifier connects the first loudspeaker by swept-frequency signal
Play into environment, environmental audio is acquired by the first sound pick-up, the output end connection second of first sound pick-up
Amplifier, the output end of second amplifier connects the first ADC converters, the output end connection of the first ADC converters
Comparison circuit, while another road output of swept-frequency signal is also connected to comparison circuit, it is special to export external audio by comparison circuit
Property.
8. active the machine audio in a kind of speech recognition according to claim 6 eliminates circuit, it is characterised in that institute
Stating audio bucking circuit includes：Second DAC, the 3rd amplifier, the second loudspeaker, four high guaily unit device, the 2nd ADC conversions
Device, the second sound pick-up, the 5th amplifier, the 3rd ADC converters and acoustic characteristic inverse operation circuit；Second DAC
Output end connect the 3rd amplifier, the output end of the 3rd amplifier is divided into two-way, wherein connect loudspeaker all the way, another
The amplifiers of Lu Yu tetra- are connected, and the output end of the 4th amplifier connects the 2nd ADC converters, second sound pick-up it is defeated
Go out the 5th amplifier of end connection, the output end of the 5th amplifier connects the 3rd ADC converters, the 2nd ADC converters
It is connected with the output end of the 3rd ADC converters with acoustic characteristic inverse operation circuit.
9. active the machine audio in a kind of speech recognition according to claim 6 eliminates circuit, it is characterised in that institute
It is the sampling frequency in the range of 300Hz~3KHz to state swept-frequency signal.
10. active the machine audio in a kind of speech recognition according to claim 6 eliminates circuit, it is characterised in that
The swept-frequency signal uses the music signal for including sweep check signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710455665.5A CN107274909A (en) | 2017-06-16 | 2017-06-16 | A kind of active the machine audio removing method in speech recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710455665.5A CN107274909A (en) | 2017-06-16 | 2017-06-16 | A kind of active the machine audio removing method in speech recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107274909A true CN107274909A (en) | 2017-10-20 |
Family
ID=60067249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710455665.5A Pending CN107274909A (en) | 2017-06-16 | 2017-06-16 | A kind of active the machine audio removing method in speech recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107274909A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1103956A2 (en) * | 1999-11-27 | 2001-05-30 | Alcatel | Exponential reduction of echo and noise during speech pauses |
| CN201805538U (en) * | 2010-09-10 | 2011-04-20 | 康佳集团股份有限公司 | Circuit for improving play sound articulation according to ambient sound, and device thereof |
| CN204836533U (en) * | 2015-08-18 | 2015-12-02 | 聚熵信息技术（上海）有限公司 | Acoustic control intelligence audio amplifier with echo elimination function |
| CN106297815A (en) * | 2016-07-27 | 2017-01-04 | 武汉诚迈科技有限公司 | A kind of method of echo cancellation in speech recognition scene |
| CN106653040A (en) * | 2016-12-22 | 2017-05-10 | 上海百芝龙网络科技有限公司 | Voice audio signal sampling processing method |
-
2017
- 2017-06-16 CN CN201710455665.5A patent/CN107274909A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1103956A2 (en) * | 1999-11-27 | 2001-05-30 | Alcatel | Exponential reduction of echo and noise during speech pauses |
| CN201805538U (en) * | 2010-09-10 | 2011-04-20 | 康佳集团股份有限公司 | Circuit for improving play sound articulation according to ambient sound, and device thereof |
| CN204836533U (en) * | 2015-08-18 | 2015-12-02 | 聚熵信息技术（上海）有限公司 | Acoustic control intelligence audio amplifier with echo elimination function |
| CN106297815A (en) * | 2016-07-27 | 2017-01-04 | 武汉诚迈科技有限公司 | A kind of method of echo cancellation in speech recognition scene |
| CN106653040A (en) * | 2016-12-22 | 2017-05-10 | 上海百芝龙网络科技有限公司 | Voice audio signal sampling processing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101260131B1 (en) | Audio source proximity estimation using sensor array for noise reduction | |
| CN103986995B (en) | The method for reducing the uncorrelated noise in apparatus for processing audio | |
| CN101155438B (en) | Frequency response adaptive equalization method for audio device | |
| CN107071654A (en) | A kind of digital noise reduction method, device and a kind of audio-frequence player device | |
| CN107396158A (en) | A kind of acoustic control interactive device, acoustic control exchange method and television set | |
| CN108401204A (en) | A kind of novel active noise reduction earphone | |
| CN109769060A (en) | A kind of mobile phone active noise reducing device and method | |
| CN102223428A (en) | Noise reducing method and mobile terminal | |
| US20050207596A1 (en) | Packaged digital microphone device with auxiliary line-in function | |
| CN106878642A (en) | Audio automatic equalization system and method | |
| CN109104683A (en) | A kind of method and correction system of dual microphone phase measurement correction | |
| CN105472507A (en) | Method and system for audio optimization | |
| CN107274909A (en) | A kind of active the machine audio removing method in speech recognition | |
| CN112349291A (en) | Sound amplification system and method based on AI noise reduction model | |
| CN111815877B (en) | White noise warning method and device | |
| CN110099328B (en) | Intelligent sound box | |
| KR101600426B1 (en) | Method for reducing noise of hearing aid and apparatus using thereof | |
| CN109688503A (en) | Psychological response condition detecting system and method | |
| CN115494739A (en) | Intelligent sound and intelligent home linkage method and linkage system | |
| CN109102820A (en) | The processing method of voice signal, the processing system of voice signal and electric appliance | |
| TWI738409B (en) | Directivity hearing-aid device and method thereof | |
| CN113808566A (en) | Vibration noise processing method, device, electronic device and storage medium | |
| CN114245284B (en) | Microphone performance detection method and system | |
| CN109862458A (en) | Psychological response condition detecting system and method | |
| TWI777265B (en) | Directivity sound source capturing device and method thereof |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20180918 Address after: 518000 Building 12, Block B, North District, Beijing-based Binhe Times Square, 9289 Binhe Avenue, Shatou Street, Futian District, Shenzhen City, Guangdong Province Applicant after: Shenzhen PCTX Technology Development Co., Ltd. Address before: 518057 Korabensi, 25 west east building, 5 West Road, Nanshan District, Shenzhen, Guangdong. Applicant before: Shenzhen Huayu Wireless Technology Co., Ltd. |
|
| TA01 | Transfer of patent application right | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171020 |
|
| WD01 | Invention patent application deemed withdrawn after publication |