CN107342093A - A kind of noise reduction process method and system of audio signal - Google Patents
A kind of noise reduction process method and system of audio signal Download PDFInfo
- Publication number
- CN107342093A CN107342093A CN201710423887.9A CN201710423887A CN107342093A CN 107342093 A CN107342093 A CN 107342093A CN 201710423887 A CN201710423887 A CN 201710423887A CN 107342093 A CN107342093 A CN 107342093A
- Authority
- CN
- China
- Prior art keywords
- signal
- audio
- sent
- radio frequency
- microphone
- 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
- 230000005236 sound signal Effects 0.000 title claims abstract description 43
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000011946 reduction process Methods 0.000 title claims abstract description 27
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 41
- 238000000926 separation method Methods 0.000 claims abstract description 22
- 230000001755 vocal effect Effects 0.000 claims abstract description 22
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000004088 simulation Methods 0.000 claims description 4
- 241000208340 Araliaceae Species 0.000 claims description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 2
- 230000008859 change Effects 0.000 claims description 2
- 235000008434 ginseng Nutrition 0.000 claims description 2
- 230000009467 reduction Effects 0.000 abstract description 7
- 238000010908 decantation Methods 0.000 abstract description 4
- 230000007613 environmental effect Effects 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000002087 whitening effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
-
- G—PHYSICS
- 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
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
-
- 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
- G10L21/0272—Voice signal separating
- G10L21/028—Voice signal separating using properties of sound source
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M1/00—Substation equipment, e.g. for use by subscribers
- H04M1/02—Constructional features of telephone sets
- H04M1/19—Arrangements of transmitters, receivers, or complete sets to prevent eavesdropping, to attenuate local noise or to prevent undesired transmission; Mouthpieces or receivers specially adapted therefor
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M1/00—Substation equipment, e.g. for use by subscribers
- H04M1/72—Mobile telephones; Cordless telephones, i.e. devices for establishing wireless links to base stations without route selection
- H04M1/724—User interfaces specially adapted for cordless or mobile telephones
- H04M1/72448—User interfaces specially adapted for cordless or mobile telephones with means for adapting the functionality of the device according to specific conditions
- H04M1/72454—User interfaces specially adapted for cordless or mobile telephones with means for adapting the functionality of the device according to specific conditions according to context-related or environment-related conditions
-
- 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
- G10L2021/02082—Noise filtering the noise being echo, reverberation of the 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
- 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
- G10L2021/02087—Noise filtering the noise being separate speech, e.g. cocktail party
-
- 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
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
- G10L2021/02161—Number of inputs available containing the signal or the noise to be suppressed
- G10L2021/02166—Microphone arrays; Beamforming
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Human Computer Interaction (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Environmental & Geological Engineering (AREA)
- Computer Networks & Wireless Communication (AREA)
- Telephone Function (AREA)
- Soundproofing, Sound Blocking, And Sound Damping (AREA)
- Circuit For Audible Band Transducer (AREA)
Abstract
The invention provides a kind of noise reduction process method and system of audio signal, voice signal is received by the microphone of terminal, and by the voice signal received by microphone array column processing after, the voice signal received is parsed using blind source separation algorithm, the only benchmark audio digital signals containing vocal sections isolated, radio frequency receiver is sent to after the benchmark audio digital signals are changed into analog signal；The radio frequency receiver sends the analog signal received by radio frequency amplifier.The method disclosed in the present and system, by microphone array and with reference to software algorithm, sound is caught and solution is separated by decantation to voice signal, not only than having higher lifting at this stage so as to bring the acoustic contrast become apparent from and after natural sound, with present noise reduction to have more qualitative leap and lifting to user in sound quality.
Description
Technical field
The present invention relates to audio defeat technical field, more particularly to a kind of noise reduction process method of audio signal and it is
System.
Background technology
With the quick popularization of handheld device, user for equipment function and it is intelligentized require more and more higher, why
Sample makes handheld device more intelligent, and function is more specialized, variation, is more efficiently used in daily life,
Through becoming our task of top priority problem to be solved.It is the standard configuration of all mobile phones on call function, but speech quality
All it is different between certain each mobile phone, the factor being directed to has a lot.The quality of network, the debugging of antenna for mobile phone with
And the device such as mobile phone receiver in itself and microphone and internal hardware design are all the factors influenceed.Current each mobile phone
Producer is all updating the performance of this major function.
MEMS microphone is more and more used in mobile phone at present, in view of its higher electric property, and higher sound
Frequency performance, the signal to noise ratio of voice signal when improving call, reduces distortion rate, has at present using two omni-directional microphones
And their position is designed while analyzed using calculating of the audio coprocessor to the environmental noise of entrance, and disappear
Except unwanted environmental noise, the solution of recording quality can be further lifted.
Most of at present is to be realized by beam forming technologies in the range of certain angle is set to the model
Sound in enclosing carries out the enhancing of voice signal and abated the noise, and will not be that acoustic filtering in the angle falls, but due to making an uproar
Sound and echo it is direction-free, it is also possible to the background environment noise from speaker, robustness can poor one for algorithm
A bit, because typically more Mike's noise reductions, the noise reduction Mike in addition to knowing sound Mike collects voice signal, can only also carried out
Suppress environmental noise and background noise can not be completely eliminated, another factor for influenceing algorithm performance is the position on microphone
Performance can all be had a strong impact on by putting the distance for the Mike that layout requirements are higher, between the direction of microphone aperture and two.
Therefore, prior art is improved up for further.
The content of the invention
In view of above-mentioned weak point of the prior art, it is an object of the invention to provide the user a kind of terminal microphone
The method and mobile terminal of noise reduction, overcome the defects of microphone denoising is high to microphone detailing requiments itself in the prior art.
The technical proposal for solving the technical problem of the invention is as follows：
A kind of noise reduction process method of audio signal, suitable for being performed in the intelligent terminal with microphone apparatus, wherein, including
Following steps：
Step A, the microphone of terminal receives voice signal, is sent after the voice signal received is formed into microphone array to sound
Frequency word signal processor；
Step B, described audio digital signal processor is parsed using blind source separation algorithm to the voice signal received, point
The vocal sections wherein contained are separated out, and the only benchmark audio digital signals containing vocal sections isolated are sent to base
Provided with processor；
The benchmark audio digital signals are changed into analog signal by step C, described BBP, and by the analog signal
It is sent to radio frequency receiver；
Step D, described radio frequency receiver sends the analog signal received by radio frequency amplifier.
The noise reduction process method of described audio signal, wherein, also include before the step A：
Step A01, to electric on microphone, and send initialization and notify application processor, the application processor loading prestores
Algorithm judges whether audio digital processing device initializes success to the audio digital processing device, if success, performs step A,
Otherwise step A02 is performed;
Step A02, false command is returned to show to application processor, and by display interface.
The method of described terminal microphone denoising, wherein, after the step A02, in addition to：
Step A03, BBP identifies new call, and sends hardware interrupt to application processor by I2C buses；
Step A04, application processor obtains the interrupt signal, and microphone array is initialized, and configuration audio ginseng
Number, and interrupt signal is sent to BBP；
Step 05, the BBP receive the interrupt signal by radio frequency receiver, and by radio frequency amplifier by institute
Interrupt signal is stated to send.
The method of described terminal microphone denoising, wherein, the step B also includes：
Step B1, the performance data contained in each railway digital signal is identified；
Step B2, the performance data is subjected to blind source separation algorithm parsing, isolates human voice signal and background noise.
The method of described terminal microphone denoising, wherein, the A also includes：
Step A1, the analog signal received is sent to audio by the voice signal received and/BBP by microphone
Signal converter；
The analog signal received is changed into data signal by step A2, described audio signal converter, and the numeral is believed
Number it is sent to audio digital signal processor.
A kind of noise reduction process system of audio signal, wherein, including：Microphone, audio digital signal processor, at base band
Manage device and radio frequency recipient and radio frequency amplifier；
The microphone, sent for receiving voice signal, and by the voice signal received to audio digital signal processor；
The audio digital signal processor, for being parsed using blind source separation algorithm to the voice signal received, point
The vocal sections wherein contained are separated out, and the only benchmark audio digital signals containing vocal sections isolated are sent to base
Provided with processor；
The BBP, for the benchmark audio digital signals to be changed into analog signal, and by the analog signal
It is sent to radio frequency receiver；
The radio frequency receiver, for the analog signal received to be sent by radio frequency amplifier.
The noise reduction process system of described audio signal, wherein, the application processor includes：Loading is additionally operable to prestore calculation
Method judges whether audio digital processing device initializes success, if success, is connect by microphone to the audio digital processing device
Voice signal is received, otherwise returns to false command, and show by display interface.
The noise reduction process system of described audio signal, wherein, the BBP, it is additionally operable to identify new call,
And hardware interrupt is sent by I2C buses and receives the interrupt signal to application processor, and by radio frequency receiver,
And the interrupt signal is sent by radio frequency amplifier；
The application processor, it is additionally operable to obtain the interrupt signal, audio microphone array is initialized, and configuration
Audio frequency parameter, and interrupt signal is sent to BBP.
The noise reduction process system of described audio signal, wherein, the audio digital signal processor, it is additionally operable to identify
The performance data contained in each railway digital signal, and blind source separation algorithm parsing is carried out to the performance data, isolate people
Acoustical signal and background noise.
The noise reduction process system of described audio signal, wherein, in addition to audio signal converter；
The audio signal converter, for receiving the simulation letter of the wind-borne voice signal of Mike and BBP transmission
Number, the voice signal and/or analog signal are changed into data signal, and the data signal is sent to digital audio letter
Number processor.
Beneficial effect, the invention provides a kind of noise reduction process method and system of audio signal, pass through the Mike of terminal
Wind receive voice signal, and by the voice signal received by microphone array column processing after, send to audio digital signals
Manage device；The audio digital signal processor is parsed using blind source separation algorithm to the voice signal received, is isolated
The vocal sections wherein contained, and the only benchmark audio digital signals containing vocal sections isolated are sent at base band
Manage device；The benchmark audio digital signals are changed into analog signal by the BBP, and the analog signal is sent
To radio frequency receiver；The radio frequency receiver sends the analog signal received by radio frequency amplifier.It is presently disclosed
Method and system, by microphone array and with reference to software algorithm, sound is caught and solution is separated by decantation to voice letter
Number, it is not necessary to multiple auxiliary microphones ensure that environmental sound signal can meet the requirement of the parsing of algorithm enough, so calculating
The robustness of method can be very good, is not only become apparent from sound quality than there is higher lifting at this stage so as to be brought to user
And the acoustic contrast after natural sound, with present noise reduction has more qualitative leap and lifting.
Brief description of the drawings
Fig. 1 is the step flow chart of the noise reduction process method of audio signal of the present invention.
Fig. 2 is the principle schematic of the method that terminal processes initialize in the method for the invention specific embodiment.
Fig. 3 is the embodiment step schematic diagram of the noise reduction process of method sound intermediate frequency signal of the present invention.
Fig. 4 is the theory structure schematic diagram of the noise reduction process system of audio signal of the present invention.
Embodiment
To make the objects, technical solutions and advantages of the present invention clearer, clear and definite, develop simultaneously embodiment pair referring to the drawings
The present invention is further described.It should be appreciated that specific embodiment described herein is used only for explaining the present invention, and do not have to
It is of the invention in limiting.
Current speech recognition only can identify and handle the voice signal for being mixed with ambient noise with noise reduction enhancing algorithm, and
For the PMD EDM signal of voice as background noise with regard to helpless, the present invention is main to enter pedestrian using blind source separation algorithm
Separation of the part point with noise section.Because blind source separation algorithm is not by the shadow of the sound characteristics such as voice signal fundamental tone harmonic
Ring, under the conditions of the prior information of no targeted voice signal, the ear of the mankind is imitated by using microphone array, will be adopted
The aliasing voice signal collected is separated, so as to extract our target voices interested.
The invention discloses a kind of noise reduction process method of audio signal, suitable in the intelligent terminal with microphone apparatus
Middle execution, as shown in figure 1, comprising the following steps：
Step S1, terminal microphone receive voice signal, will the voice signal that received formed microphone array after send to
Audio digital signal processor.
The microphone of terminal receives voice signal, and the voice signal received is sent into audio digital signals processing
Device, wherein, microphone has 2, is arranged at the front end of terminal.Terminal can be that tablet personal computer, mobile phone or other intelligence are worn
Wear terminal.
In specific embodiment, the signal for being sent to audio digital signal processor not only includes what microphone received
Extraneous voice signal, in addition to the analog signal that radio frequency receiver receives.
In order to realize that the smoothly transmission data signal of microphone to audio digital signals processing, is also wrapped before this step
Include：
To electric on microphone, and send initialization and notify application processor, the application processor loads the algorithm that prestores to institute
Audio digital processing device is stated, judges whether audio digital processing device initializes success, if success, voice is received by microphone
Signal, and the voice signal received is formed after microphone array and sent to audio digital signal processor, otherwise return wrong
Application processor is arrived in instruction by mistake, and is shown by display interface.
First to electric on microphone in this step, and start to initialize application processor, whether initialization is successfully carried out
Judge, if initialization failure, judge current microphone or application processor cisco unity malfunction, therefore return to false command and arrive
Display interface, reminded for user.
In order to ensure the normal work on terminal audio frequency hardware, this step is further comprising the steps of：
BBP identifies new call, and sends hardware interrupt to application processor by I2C buses；
Application processor obtains the interrupt signal, and microphone array is initialized, and configuration audio frequency parameter, and sends
Interrupt signal is to BBP；
The BBP receives the interrupt signal by radio frequency receiver, and is believed described interrupt by radio frequency amplifier
Number send.
Step S2, described audio digital signal processor is solved using blind source separation algorithm to the voice signal received
Analysis, the vocal sections wherein contained are isolated, and the only benchmark audio digital signals containing vocal sections isolated are sent out
It is sent to BBP.
The parsing of voice signal is carried out in this step using blind source separation algorithm, microphone receives in above-mentioned steps S1
Voice signal, after microphone array column processing, a signal source is formed, signal source is sampled, identifies each way
The performance data contained in word signal, the performance data is subjected to blind source separation algorithm parsing, isolates human voice signal and the back of the body
Scape noise.
Specifically, the blind source separation algorithm used in this step is the progressive orthogonalization Fixed-Point Algorithm based on kurtosis.I
, according to available sample value z (z is the data crossed of PCA whitening pretreatments), can be calculated since some initial vector w
MakeKurtosis absolute value increase most fast direction, vectorial w is then gone into the direction.Know that wTz's is high and steep by formula 1.5
Degree absolute value derivative be：
（1）
In above formula, for the data of PCA albefactions, have, because we are on unit ballOptimize, w should be projected to after each step computing on unit ball, i.e., by w divided by its norm.Make formula（1）
The gradient of middle kurtosis is equal with w, can obtain：
（2）
After each fixed point iteration, w will divided by its norm to meet to constrain accordingly, we can count first
Calculate formula（2）The item on the right, and w is assigned it to as new value：
（3）
Last convergent vectorial w withLinear combination can give our one of independent elements.In albefaction
It is vectorial corresponding to different independent elements in spaceIt is mutually orthogonal directions, according to this characteristic, we can be to having been estimated that
Independent element carry out progressive orthogonalization, i.e., independent element is estimated one by one, can so avoid algorithm
Multiple independent elements simultaneously converge in same kurtosis maximum.
；
Then it is right againStandardization：Divided by its norm.
At the stable convergence of algorithm,WithThere should be identical direction, i.e., before and after iterationWith's
Dot product should be no better than 1, it is possible thereby to carry out the judgement of algorithmic statement.When p is equal to the independent element number for needing to estimate
The separation computing of ICA algorithm is completed, obtains independent element.
Above-mentioned algorithm principle can be from《Academic research》Disclosed in magazine the 1st phase in 2011《More voice PMD EDM signals
Blind source separation algorithm is studied》In find more detailed explanation.
Therefore vocal sections can be separated from the voice signal received using above-mentioned algorithm, the audio number
The vocal sections isolated are passed to BBP by word signal processor.
The benchmark audio digital signals are changed into analog signal by step S3, described BBP, and by the mould
Intend signal and be sent to radio frequency receiver.
BBP receives changes into analog signal by the data signal that audio digital signal processor is sent, and is sent to
Radio frequency receiver.The conversion of the BBP not only subscriber signal, it is additionally operable to control the call control of terminal call module
System, the time for being additionally operable to manage and controlling transmitting-receiving to transmit, and to being powered to radio frequency receiver and radio frequency amplifier, order
Send and receive, voice signal is sent and received, and the voice signal after parsing is carried out into digital revolving die plan and to reception
Numeral carries out analog-to word operation etc..
Step S4, described radio frequency receiver sends the analog signal received by radio frequency amplifier.
The radio frequency receiver receives the analog signal that BBP is sent, and the analog signal is sent into radio frequency and put
Big device is sent to base station, and radio frequency receiver is additionally operable to the audio signal that reception antenna receives, and by the audio signal
BBP is sent to, audio signal converter is sent to by BBP, will be simulated and believed by audio signal converter
Number data signal is changed into, so as to vocal sections therein in being separated by audio digital signal processor.
It is more detailed in order to be carried out to method of the present invention below with invention specific embodiment of the present invention
Explanation.
Mainly included the following steps that during specific embodiment：
With reference to Fig. 2, before microphone collection voice signal is started, in addition to：Electricity and audio signal processor etc. on microphone
The initialization of hardware, if initializing successfully, configures voice-frequency channel and audio frequency parameter, otherwise sends false command and is handled to application
Device, and initialization mistake is reminded by display screen.
And the new call of BBP identification, and hardware interrupt is sent to application processor by bus, it is described
After application processor receives interrupt signal, system task priority is changed, and current task is first hung up, audio is carried out from new
Initialization, configure voice-frequency channel and audio frequency parameter.
Present video transformation task interrupts, BBP transmission interrupt signal to Anneta module, during Anneta module is sent
Break signal is to air base station.
With reference to shown in Fig. 3, after microphone initializes successfully, closing of the circuit, terminal starts to obtain current talking voice letter
Number, that is, recorded, and the voice signal after recording is sent to audio digital signal processor and carries out arithmetic analysis, point
Vocal sections are separated out, the voice isolated then is transferred to BBP, the vocal sections that BBP will be isolated
Data signal changes into analog signal and is sent to radio frequency receiver, and radio frequency receiver sends signal by radio frequency amplifier.
External sound signal, including voice and noise are received since 2 digital microphones（Echo, reverberation and need
The low frequency noise to be filtered）, data signal is changed into the analog signal received by audio signal processor, it is every all to enter all the way
Row individually sample and be input to audio digital signal processor combine loaded blind source separating based on kurtosis gradually
Enter orthogonalization Fixed-Point Algorithm.Due to each voice signal having time domain and spatial domain.Needed per signal all the way with algorithm
The mathematical model of offer is compared, and identifies the general direction of sound, position, loudness, background, quality is far and near, and tone etc. is big
The data of partial mixed characteristic.Because algorithm superperformance is determined in the seizure for microphone voice and positional distance
Requirement be all very little.By algorithm model presented hereinbefore by these numerals re-establish mathematical model establish it is virtual
Digital space domain and time-domain, analog controller is then turned by numeral and exported, cross the simulation that microphone array captures
Signal gets data signal by the analog-to-digital signal processing unit of high-precision high sampling rate, and delivers to digital processing unit
In, arithmetic analysis is carried out, the baseline audio signal of only voice is separated by decantation to by blind source separation algorithm solution, by two-part number
Word signal carries out hybrid algorithm parsing, always filters to isolate background noise and human voice signal, finally obtains the people become apparent from
Sound, and data signal is delivered in BBP by I2S interfaces.
The present invention discloses a kind of noise reduction process system of audio signal on the basis of above-mentioned noise reduction process method is disclosed
System, as shown in figure 4, including：Microphone, audio digital signal processor, BBP and radio frequency recipient and radio frequency amplification
Device；
The microphone, sent for receiving voice signal, and by the voice signal received to audio digital signal processor；
The audio digital signal processor, for being parsed using blind source separation algorithm to the voice signal received, point
The vocal sections wherein contained are separated out, and the only benchmark audio digital signals containing vocal sections isolated are sent to base
Provided with processor；
The BBP, for the benchmark audio digital signals to be changed into analog signal, and by the analog signal
It is sent to radio frequency receiver；
The radio frequency receiver, for the analog signal received to be sent by radio frequency amplifier.
The application processor includes：It is additionally operable to loading and prestores algorithm to the audio digital processing device, judges audio number
Whether word processing device initializes success, if success, receives voice signal by microphone, otherwise returns to false command, and lead to
Display interface is crossed to show.
The BBP, it is additionally operable to identify new call, and hardware interrupt is sent to application by I2C buses
Processor, and the interrupt signal is received by radio frequency receiver, and sent the interrupt signal by radio frequency amplifier
Go out；
The application processor, it is additionally operable to obtain the interrupt signal, audio microphone array is initialized, and configuration
Audio frequency parameter, and interrupt signal is sent to BBP.
The audio digital signal processor, it is additionally operable to identify the performance data contained in each railway digital signal, and
Blind source separation algorithm parsing is carried out to the performance data, isolates human voice signal and background noise.
The noise reduction process system of described audio signal, in addition to audio signal converter；
The audio signal converter, for receiving the simulation letter of the wind-borne voice signal of Mike and BBP transmission
Number, the voice signal and/or analog signal are changed into data signal, and the data signal is sent to digital audio letter
Number processor.
Beneficial effect, the invention provides a kind of noise reduction process method and system of audio signal, pass through the Mike of terminal
Wind receive voice signal, and by the voice signal received by microphone array column processing after, send to audio digital signals
Manage device；The audio digital signal processor is parsed using blind source separation algorithm to the voice signal received, is isolated
The vocal sections wherein contained, and the only benchmark audio digital signals containing vocal sections isolated are sent at base band
Manage device；The benchmark audio digital signals are changed into analog signal by the BBP, and the analog signal is sent
To radio frequency receiver；The radio frequency receiver sends the analog signal received by radio frequency amplifier.It is presently disclosed
Method and system, by microphone array and with reference to software algorithm, sound is caught and solution is separated by decantation to voice letter
Number, it is not necessary to multiple auxiliary microphones ensure that environmental sound signal can meet the requirement of the parsing of algorithm enough, so calculating
The robustness of method can be very good, is not only become apparent from sound quality than there is higher lifting at this stage so as to be brought to user
And the acoustic contrast after natural sound, with present noise reduction has more qualitative leap and lifting.It is understood that to this
For the those of ordinary skill of field, equivalent substitution or change can be subject to technique according to the invention scheme and its inventive concept,
And all these changes or replacement should all belong to the protection domain of appended claims of the invention.
Claims (10)
1. a kind of noise reduction process method of audio signal, suitable for being performed in the intelligent terminal with microphone apparatus, its feature
It is, comprises the following steps：
Step A, the microphone of terminal receives voice signal, is sent after the voice signal received is formed into microphone array to sound
Frequency word signal processor；
Step B, described audio digital signal processor is parsed using blind source separation algorithm to the voice signal received, point
The vocal sections wherein contained are separated out, and the only benchmark audio digital signals containing vocal sections isolated are sent to base
Provided with processor；
The benchmark audio digital signals are changed into analog signal by step C, described BBP, and by the analog signal
It is sent to radio frequency receiver；
Step D, described radio frequency receiver sends the analog signal received by radio frequency amplifier.
2. the noise reduction process method of audio signal according to claim 1, it is characterised in that also wrapped before the step A
Include：
Step A01, to electric on microphone, and send initialization and notify application processor, the application processor loading prestores
Algorithm judges whether audio digital processing device initializes success to the audio digital processing device, if success, performs step A,
Otherwise step A02 is performed;
Step A02, false command is returned to show to application processor, and by display interface.
3. the method for terminal microphone denoising according to claim 2, it is characterised in that after the step A02, also wrap
Include：
Step A03, BBP identifies new call, and sends hardware interrupt to application processor by I2C buses；
Step A04, application processor obtains the interrupt signal, and microphone array is initialized, and configuration audio ginseng
Number, and interrupt signal is sent to BBP；
Step 05, the BBP receive the interrupt signal by radio frequency receiver, and by radio frequency amplifier by institute
Interrupt signal is stated to send.
4. the method for terminal microphone denoising according to claim 3, it is characterised in that the step B also includes：
Step B1, the performance data contained in each railway digital signal is identified；
Step B2, the performance data is subjected to blind source separation algorithm parsing, isolates human voice signal and background noise.
5. the method for the terminal microphone denoising according to claim 3 or 4, it is characterised in that the A also includes：
Step A1, the analog signal received is sent to audio by the voice signal received and/BBP by microphone
Signal converter；
The analog signal received is changed into data signal by step A2, described audio signal converter, and the numeral is believed
Number it is sent to audio digital signal processor.
A kind of 6. noise reduction process system of audio signal, it is characterised in that including：Microphone, audio digital signal processor, base
Provided with processor and radio frequency recipient and radio frequency amplifier；
The microphone, sent for receiving voice signal, and by the voice signal received to audio digital signal processor；
The audio digital signal processor, for being parsed using blind source separation algorithm to the voice signal received, point
The vocal sections wherein contained are separated out, and the only benchmark audio digital signals containing vocal sections isolated are sent to base
Provided with processor；
The BBP, for the benchmark audio digital signals to be changed into analog signal, and by the analog signal
It is sent to radio frequency receiver；
The radio frequency receiver, for the analog signal received to be sent by radio frequency amplifier.
7. the noise reduction process system of audio signal according to claim 6, it is characterised in that the application processor bag
Include：It is additionally operable to loading and prestores algorithm to the audio digital processing device, judges whether audio digital processing device initializes success, if
Success, then voice signal is received by microphone, otherwise return to false command, and show by display interface.
8. the noise reduction process system of audio signal according to claim 7, it is characterised in that the BBP, also
Hardware interrupt is sent to application processor for identifying new call, and by I2C buses, and passes through radio frequency receiver
The interrupt signal is received, and is sent the interrupt signal by radio frequency amplifier；
The application processor, it is additionally operable to obtain the interrupt signal, audio microphone array is initialized, and configuration
Audio frequency parameter, and interrupt signal is sent to BBP.
9. the noise reduction process system of audio signal according to claim 7, it is characterised in that at the audio digital signals
Device is managed, is additionally operable to identify the performance data contained in each railway digital signal, and blind source separating is carried out to the performance data
Arithmetic analysis, isolate human voice signal and background noise.
10. the noise reduction process system of audio signal according to claim 8, it is characterised in that also turn including audio signal
Change device；
The audio signal converter, for receiving the simulation letter of the wind-borne voice signal of Mike and BBP transmission
Number, the voice signal and/or analog signal are changed into data signal, and the data signal is sent to digital audio letter
Number processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710423887.9A CN107342093A (en) | 2017-06-07 | 2017-06-07 | A kind of noise reduction process method and system of audio signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710423887.9A CN107342093A (en) | 2017-06-07 | 2017-06-07 | A kind of noise reduction process method and system of audio signal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107342093A true CN107342093A (en) | 2017-11-10 |
Family
ID=60220542
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710423887.9A Pending CN107342093A (en) | 2017-06-07 | 2017-06-07 | A kind of noise reduction process method and system of audio signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107342093A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108806677A (en) * | 2018-02-14 | 2018-11-13 | 美律电子(深圳)有限公司 | Audio processing device and audio processing method |
| CN109215648A (en) * | 2018-09-03 | 2019-01-15 | 惠州市德赛西威智能交通技术研究院有限公司 | Vehicle-mounted voice identifying system and method |
| CN112185412A (en) * | 2020-09-08 | 2021-01-05 | 中国人民解放军国防科技大学 | Recognition method under multi-target acoustic signal aliasing |
| CN114390505A (en) * | 2022-01-26 | 2022-04-22 | 惠州华阳通用电子有限公司 | Emergency call control method and device based on Internet of vehicles |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1509065A1 (en) * | 2003-08-21 | 2005-02-23 | Bernafon Ag | Method for processing audio-signals |
| CN1648995A (en) * | 2004-10-12 | 2005-08-03 | 上海大学 | Voice enhancing device based on blind signal separation |
| CN202889458U (en) * | 2012-11-02 | 2013-04-17 | 姚西 | Automatic call volume regulation mobile phone based on environmental noise |
| CN103873977A (en) * | 2014-03-19 | 2014-06-18 | 惠州Tcl移动通信有限公司 | Recording system and method based on multi-microphone array beam forming |
| CN105100470A (en) * | 2015-07-22 | 2015-11-25 | 惠州Tcl移动通信有限公司 | Method and system for switching microphone function of mobile terminal |
| CN105262516A (en) * | 2015-09-07 | 2016-01-20 | 惠州Tcl移动通信有限公司 | Slave device, and system and method for automatically pairing master and slave devices |
| CN105306692A (en) * | 2015-09-29 | 2016-02-03 | 北京元心科技有限公司 | Mobile terminal and startup method thereof |
| CN205232299U (en) * | 2015-11-26 | 2016-05-11 | 惠州Tcl移动通信有限公司 | Intelligent terminal based on recording of optical sensor harmony microphone |
| CN205754811U (en) * | 2016-05-12 | 2016-11-30 | 惠州Tcl移动通信有限公司 | Mobile terminal and audio frequency processing system thereof |
-
2017
- 2017-06-07 CN CN201710423887.9A patent/CN107342093A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1509065A1 (en) * | 2003-08-21 | 2005-02-23 | Bernafon Ag | Method for processing audio-signals |
| CN1648995A (en) * | 2004-10-12 | 2005-08-03 | 上海大学 | Voice enhancing device based on blind signal separation |
| CN202889458U (en) * | 2012-11-02 | 2013-04-17 | 姚西 | Automatic call volume regulation mobile phone based on environmental noise |
| CN103873977A (en) * | 2014-03-19 | 2014-06-18 | 惠州Tcl移动通信有限公司 | Recording system and method based on multi-microphone array beam forming |
| CN105100470A (en) * | 2015-07-22 | 2015-11-25 | 惠州Tcl移动通信有限公司 | Method and system for switching microphone function of mobile terminal |
| CN105262516A (en) * | 2015-09-07 | 2016-01-20 | 惠州Tcl移动通信有限公司 | Slave device, and system and method for automatically pairing master and slave devices |
| CN105306692A (en) * | 2015-09-29 | 2016-02-03 | 北京元心科技有限公司 | Mobile terminal and startup method thereof |
| CN205232299U (en) * | 2015-11-26 | 2016-05-11 | 惠州Tcl移动通信有限公司 | Intelligent terminal based on recording of optical sensor harmony microphone |
| CN205754811U (en) * | 2016-05-12 | 2016-11-30 | 惠州Tcl移动通信有限公司 | Mobile terminal and audio frequency processing system thereof |
Non-Patent Citations (2)
| Title |
|---|
| 刘红梅: "基于FastICA盲源分离算法的语音增强系统", 《计算机与数字工程》 * |
| 栗科峰 等: "多人声混叠语音信号的盲源分离算法研究", 《信息通信》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108806677A (en) * | 2018-02-14 | 2018-11-13 | 美律电子(深圳)有限公司 | Audio processing device and audio processing method |
| CN108806677B (en) * | 2018-02-14 | 2020-10-13 | 美律电子(深圳)有限公司 | Audio processing device and audio processing method |
| CN109215648A (en) * | 2018-09-03 | 2019-01-15 | 惠州市德赛西威智能交通技术研究院有限公司 | Vehicle-mounted voice identifying system and method |
| CN112185412A (en) * | 2020-09-08 | 2021-01-05 | 中国人民解放军国防科技大学 | Recognition method under multi-target acoustic signal aliasing |
| CN114390505A (en) * | 2022-01-26 | 2022-04-22 | 惠州华阳通用电子有限公司 | Emergency call control method and device based on Internet of vehicles |
| CN114390505B (en) * | 2022-01-26 | 2023-08-18 | 惠州华阳通用电子有限公司 | Emergency call control method and device based on Internet of vehicles |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11450337B2 (en) | Multi-person speech separation method and apparatus using a generative adversarial network model | |
| CN107342093A (en) | A kind of noise reduction process method and system of audio signal | |
| CN106782589B (en) | Mobile terminal and voice input method and device thereof | |
| US20190272842A1 (en) | Speech enhancement for an electronic device | |
| US20050071158A1 (en) | Apparatus and method for detecting user speech | |
| CN106663445A (en) | Voice processing device, voice processing method, and program | |
| CN108511002B (en) | Method for recognizing sound signal of dangerous event, terminal and computer readable storage medium | |
| CN108076226A (en) | A kind of method, mobile terminal and the storage medium of speech quality adjustment | |
| CN110364156A (en) | Voice interactive method, system, terminal and readable storage medium storing program for executing | |
| CN112532266A (en) | Intelligent helmet and voice interaction control method of intelligent helmet | |
| CN110473568B (en) | Scene recognition method and device, storage medium and electronic equipment | |
| WO2019015503A1 (en) | Bluetooth headphones | |
| US11638092B2 (en) | Advanced speech encoding dual microphone configuration (DMC) | |
| CN103618835A (en) | Method and system for automatically switching scene mode according to magnitude of peripheral noise | |
| CN111863020B (en) | Voice signal processing method, device, equipment and storage medium | |
| CN109874096A (en) | A kind of ears microphone hearing aid noise reduction algorithm based on intelligent terminal selection output | |
| CN113194372B (en) | Earphone control method and device and related components | |
| CN107293305A (en) | It is a kind of to improve the method and its device of recording quality based on blind source separation algorithm | |
| WO2016056683A1 (en) | Electronic device and reverberation removal method therefor | |
| CN114141230A (en) | Electronic device, and voice recognition method and medium thereof | |
| CN110992967A (en) | Voice signal processing method and device, hearing aid and storage medium | |
| CN112289325A (en) | Voiceprint recognition method and device | |
| CN108600489A (en) | Receiver, the calibration method of loud speaker, mobile terminal and readable storage medium storing program for executing | |
| CN108235165A (en) | A kind of microphone neck ring earphone | |
| CN107659864A (en) | A kind of noise-reduction method and terminal |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171110 |