CN110197670A - Audio defeat method, apparatus and electronic equipment - Google Patents
Audio defeat method, apparatus and electronic equipment Download PDFInfo
- Publication number
- CN110197670A CN110197670A CN201910480708.4A CN201910480708A CN110197670A CN 110197670 A CN110197670 A CN 110197670A CN 201910480708 A CN201910480708 A CN 201910480708A CN 110197670 A CN110197670 A CN 110197670A
- Authority
- CN
- China
- Prior art keywords
- noise
- scenarios
- type
- audio signal
- noise reduction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 56
- 230000009467 reduction Effects 0.000 claims abstract description 120
- 230000005236 sound signal Effects 0.000 claims abstract description 97
- 238000012549 training Methods 0.000 claims description 16
- 238000004378 air conditioning Methods 0.000 claims description 14
- 238000001228 spectrum Methods 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 5
- 238000000638 solvent extraction Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 description 18
- 230000008569 process Effects 0.000 description 14
- 238000011946 reduction process Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 8
- 230000007613 environmental effect Effects 0.000 description 8
- 238000001514 detection method Methods 0.000 description 7
- 230000002452 interceptive effect Effects 0.000 description 5
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 241001269238 Data Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
-
- G—PHYSICS
- 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
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/24—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being the cepstrum
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Signal Processing (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)
- Soundproofing, Sound Blocking, And Sound Damping (AREA)
Abstract
The invention discloses a kind of audio defeat method, apparatus and electronic equipments, wherein audio defeat method includes: that the type of noise scenarios is determined according to the feature of collected audio signal；Noise reduction parameters group corresponding with the type of the noise scenarios is obtained, includes at least a kind of noise reduction parameters in the noise reduction parameters group；Noise reduction is carried out to the audio signal by the noise reduction parameters in the noise reduction parameters group.The signal-to-noise ratio of audio signal can be improved in audio defeat method of the invention.
Description
Technical field
The present invention relates to audio signal processing technique field, a kind of audio defeat method, apparatus and electronic equipment are particularly related to.
Background technique
Currently, noise reduction is carried out to audio by digital signal processing algorithm, usually according to the noise situations of target scene,
It determines corresponding parameter, noise reduction process is carried out to audio using the parameter.But this noise reduction mode can only be realized to single characteristic
Noise scenarios under the audio that acquires carry out noise reduction, but ambient noise occurs the scene of dynamic change, anti-acoustic capability then compared with
It is low.
Summary of the invention
In view of this, it is an object of the invention to propose that a kind of audio defeat method, apparatus and electronic equipment, this method can
The signal-to-noise ratio of audio after improving noise reduction process.
According to the first aspect of the invention, a kind of audio defeat method is provided, comprising: believe according to collected audio
Number feature determine the types of noise scenarios；Obtain noise reduction parameters group corresponding with the type of the noise scenarios, the noise reduction
A kind of noise reduction parameters are included at least in parameter group；The audio signal is carried out by the noise reduction parameters in the noise reduction parameters group
Noise reduction.
Optionally, the method also includes: in the type for determining noise scenarios according to the feature of collected audio signal
It before, is difference by the scene partitioning according to the noise intensity of the audio signal acquired under different scenes and noise type
The noise scenarios of type；The type of each noise scenarios and preset noise parameter group are established into corresponding relationship.
Optionally, the type of the noise scenarios: environment locating for vehicle, vehicle is divided according to following at least two information
Driving status, whether vehicle window is opened, whether interior air-conditioning opens and the size of ambient noise.
Optionally, described to obtain noise reduction parameters group corresponding with the type of the noise scenarios, comprising: to extract the audio
The feature of signal；The type of noise scenarios corresponding with each frame in the audio signal is determined according to the feature；If even
The type of the corresponding noise scenarios of frame of continuous predetermined number is identical, then generates noise reduction ginseng corresponding with the type of the noise scenarios
Array.
Optionally, the method also includes: acquire the audio data under different noise scenarios, include in the audio data
Audio signal；The audio data is marked according to the noise scenarios；Use the audio data training noise after label
Scene classification model, the input of the noise scenarios disaggregated model are the feature of audio signal, the noise scenarios disaggregated model
Output be the corresponding noise scenarios of the audio signal.
Optionally, a kind of following noise reduction parameters are included at least in the noise reduction parameters group: crossing subtracting coefficient and spectrum lower limit parameter.
According to the second aspect of the invention, a kind of audio defeat device is provided, comprising: determining module is used for basis
The feature of collected audio signal determines the type of noise scenarios；Module is obtained, for obtaining and the class of the noise scenarios
Type corresponding noise reduction parameters group includes at least a kind of noise reduction parameters in the noise reduction parameters group；Noise reduction module, for by described
Noise reduction parameters in noise reduction parameters group carry out noise reduction to the audio signal.
Optionally, described device further include: division module, for making an uproar according to the determination of the feature of collected audio signal
Before the type of sound field scape, according to the noise intensity of the audio signal acquired under different scenes and noise type by the field
Scape is divided into different types of noise scenarios；Module is established, for by the type of each noise scenarios and preset noise parameter group
Establish corresponding relationship.
Optionally, the type of the noise scenarios: environment locating for vehicle, vehicle is divided according to following at least two information
Driving status, whether vehicle window is opened, whether interior air-conditioning opens and the size of ambient noise.
Optionally, the acquisition module, comprising: extraction unit, for extracting the feature of the audio signal；It determines single
Member, for determining the type of noise scenarios corresponding with each frame in the audio signal according to the feature；Generation unit is used
If identical in the type of the corresponding noise scenarios of frame of continuous predetermined number, generate corresponding with the type of the noise scenarios
Noise reduction parameters group.
Optionally, described device further include: acquisition module, it is described for acquiring the audio data under different noise scenarios
It include audio signal in audio data；Mark module, for the audio data to be marked according to the noise scenarios；Instruction
Practice module, for use label after audio data training noise scenarios disaggregated model, the noise scenarios disaggregated model it is defeated
Enter the feature for audio signal, the output of the noise scenarios disaggregated model is the corresponding noise scenarios of the audio signal.
According to the third aspect of the present invention, a kind of electronic equipment is provided, including memory, processor and being stored in is deposited
On reservoir and the computer program that can run on a processor, the processor are realized when executing described program such as the present invention first
Any one audio defeat method described in a aspect.
From the above it can be seen that audio defeat method of the invention, it can be true according to the feature of the audio signal of acquisition
Surely the noise scenarios being presently in determine the noise reduction parameters group for carrying out noise reduction process to audio signal according to noise scenarios,
It, can be continuous in ambient noise to carry out noise reduction process to audio signal based on the noise reduction parameters in the noise reduction parameters group
In the case where variation, noise reduction parameters are adaptively adjusted, with acquisition and the most matched noise reduction parameters group of current noise scene, can be directed to
The characteristic for the noise scenarios being presently in targetedly handles audio signal, thus the audio letter that can be improved that treated
Number signal-to-noise ratio.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is a kind of flow chart of audio defeat method shown according to an exemplary embodiment；
Fig. 2 is a kind of schematic diagram of audio defeat method shown according to an exemplary embodiment；
Fig. 3 is that the training process of noise scenarios disaggregated model shown according to an exemplary embodiment and use process are shown
It is intended to；
Fig. 4 is the noise reduction parameters shown according to an exemplary embodiment according in noise reduction parameters group to original signals with noise
Carry out the process flow diagram of noise reduction；
Fig. 5 is a kind of block diagram of audio defeat device shown according to an exemplary embodiment.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific embodiment, and reference
Attached drawing, the present invention is described in more detail.
It should be noted that all statements for using " first " and " second " are for differentiation two in the embodiment of the present invention
The non-equal entity of a same names or non-equal parameter, it is seen that " first " " second " only for the convenience of statement, does not answer
It is interpreted as the restriction to the embodiment of the present invention, subsequent embodiment no longer illustrates this one by one.
Fig. 1 is a kind of flow chart of audio defeat method shown according to an exemplary embodiment, as shown in Figure 1, the party
Method includes:
Step 101: the type of noise scenarios is determined according to the feature of collected audio signal；
For example, the feature of the audio signal, audio can be extracted after collecting audio signal (for after noise reduction audio signal)
The feature of signal for example can be MFCC (Mel Frequency Cepstrum Coefficient, mel-frequency cepstrum system
Number), the feature extracted is input to the noise scenarios disaggregated model that training obtains in advance, thus the class of output noise scene
Type, for example, being input to the Meier frequency spectrum cepstrum coefficient extracting after the Meier frequency spectrum cepstrum coefficient of noise reduction audio signal
Noise scenarios disaggregated model, the model are the type exported to the corresponding scene of noise reduction audio signal.
Before above-mentioned steps 101, noise scenarios can be divided previously according to the noise intensity and noise type of scene
For different type, for example, audio defeat method of the invention is applied to carry out noise reduction process to the audio obtained in vehicle
In the case where, noise scenarios can be divided into following several types:
Noise scenarios one: vehicle is in parking lot environment, and the vehicle window of idling, vehicle is opened, and the air-conditioning of vehicle is closed, environment
Noise decibel value is between 50~55db；
Noise scenarios two: vehicle is in urban road environment, low-speed running state, travel speed between 40-60km/h,
The vehicle window of vehicle is closed, and the air-conditioning of vehicle is opened, between 55~60db of environmental noise decibel value；
Noise scenarios three: vehicle is in urban road environment, low-speed running state, travel speed between 40-60km/h,
The vehicle window of vehicle is opened, and the air-conditioning of vehicle is closed, between 60~70db of environmental noise decibel value；
Noise scenarios four: vehicle is in through street environment, high-speed travel state, travel speed 80-120km/h it
Between, the vehicle window of vehicle is closed, and the air-conditioning of vehicle is opened, between 55~65db of environmental noise decibel value.
Step 102: obtaining corresponding with the type of noise scenarios noise reduction parameters group, in the noise reduction parameters group at least
Including a kind of noise reduction parameters；
For example, all types of noise scenarios, which can be preset, corresponds to different noise reduction parameters groups, all types of noises is established
Corresponding relationship between scene and each noise reduction parameters group.Wherein, a different drop is included at least in different noise reduction parameters groups
It makes an uproar parameter.The corresponding relationship between the type of noise scenarios and noise reduction parameters group can be pre-established, for example, for a certain noise field
Scape may be selected using the different noise reduction parameters group of multiple groups respectively to obtaining under the noise scenarios according to the characteristic of the noise scenarios
Audio signal carry out noise reduction, pass through many experiments, choose the optimal one group of noise reduction parameters group of noise reduction result, by the noise reduction parameters
Group is determined as noise reduction parameters group corresponding with the noise scenarios.It, can root after the type that noise scenarios have been determined in a step 102
One group of noise reduction parameters is generated according to the corresponding relationship between the type and noise reduction parameters group of preset noise scenarios.
In a kind of achievable mode, noise reduction parameters may include at least one of subtracting coefficient and spectrum lower limit parameter ginseng
Number.
Step 103: noise reduction being carried out to the audio signal by the noise reduction parameters in the noise reduction parameters group.
For example, DSP (Digital Signal can be sent to for noise reduction parameters group and audio signal to noise reduction
Processing, Digital Signal Processing) module, noise reduction process is carried out by the audio signal that DSP treats noise reduction and output is dropped
Audio signal after making an uproar.
In a kind of achievable mode, noise-reduction method of the invention can be applied in speech recognition process, to be identified
Audio signal carries out noise reduction process, to remove the noise in audio signal to be identified, to improve the precision of speech recognition.
Audio defeat method of the invention can determine the noise field being presently according to the feature of the audio signal of acquisition
Scape determines the noise reduction parameters group for carrying out noise reduction process to audio signal according to noise scenarios, to be based on the noise reduction parameters
Noise reduction parameters in group carry out noise reduction process to audio signal, can be adaptive in the continually changing situation of ambient noise
Noise reduction parameters should be adjusted, it, can be for the noise field being presently in acquisition and the most matched noise reduction parameters group of current noise scene
The characteristic of scape targetedly handles audio signal, thus the signal-to-noise ratio for the audio signal that can be improved that treated.
In a kind of achievable mode, the audio defeat method may also include that according to collected audio signal
Before feature determines the type of noise scenarios, according to the noise intensity of the audio signal acquired under different scenes and noise class
The scene partitioning is different types of noise scenarios by type；For example, n grades can be divided into noise intensity first, need to refer to
Noise type includes m kind, can be preset when noise intensity meets a certain shelves in n grades in scene, and simultaneously in the scene
Under in collected audio signal simultaneously comprising at least i (the 0 < i < m) kind in m kind noise type, it is determined that the scene is one
Specified noise scenarios.Wherein, audio defeat method of the invention is being applied to carry out noise reduction to the audio obtained in vehicle
In the case where processing, noise type can include: tire is made an uproar, wind is made an uproar, engine noise, noisy voice and other motor vehicle noises
Deng intensity, that is, noise size of noise, for example, the decibel value of noise.
It is described to obtain corresponding with the type of noise scenarios noise reduction parameters group and wrap in a kind of achievable mode
It includes: extracting the feature of the audio signal；Noise field corresponding with each frame in the audio signal is determined according to the feature
The type of scape；If the type of the corresponding noise scenarios of frame of continuous predetermined number is identical, generate and the noise scenarios
The corresponding noise reduction parameters group of type.For example, carrying out feature extraction to audio signal, being sent into decoder, (decoder passes through preparatory
The noise scenarios disaggregated model that training obtains determines the noise scenarios being presently according to audio frequency characteristics), by decoder to sound
Frequency signal is decoded frame by frame, when (the decoding of continuous 3-10 (for an example of above-mentioned predetermined number) corresponding decoding result of frame
As a result it is the corresponding noise scenarios of present frame) it is identical, that is, taking decoding result is target noise scene, according to the target noise scene
Noise reduction parameters corresponding with the scene (for example, crossing subtracting coefficient α and spectrum lower limit parameter β) is generated, the noise reduction parameters of generation are exported.It is based on
This, for the audio signal to noise reduction, it can be achieved that loading noise reduction parameters corresponding with noise scenarios, frame by frame so as to according to noise
The characteristic of scene treats noise reduction audio signal and carries out noise reduction, improves the signal-to-noise ratio of audio signal.
In a kind of achievable mode, the audio defeat method may also include that under the different types of noise scenarios of acquisition
Audio data, include audio signal in the audio data；For example, can be adopted respectively under ready-portioned multiple noise scenarios
Collect audio data, collected noise may include that tire is made an uproar, wind is made an uproar, engine noise, noisy voice, other motor vehicle noises etc. are more
Kind different type ambient noise.The audio data is marked according to the type of the noise scenarios；For example, will be first
The audio data acquired under noise scenarios is labeled as the first noise scenarios, the audio data mark that will be acquired under the second noise scenarios
The second noise scenarios are denoted as, and so on, the audio data acquired under third noise scenarios is labeled as third noise scenarios,
The audio data acquired under the 4th noise scenarios is labeled as the 4th noise scenarios.It is made an uproar using the audio data training after label
Sound field scape disaggregated model, the input of the noise scenarios disaggregated model are the feature of audio signal, the noise scenarios classification mould
The output of type is the corresponding noise scenarios of the audio signal.In training noise scenarios disaggregated model, using DNN (depth
Neural network) or the model trainings mode such as HMM (hidden Markov model) realize.
In a kind of achievable mode, the type of the noise scenarios: vehicle can be divided according to following at least two information
Whether the driving status of locating environment, vehicle, vehicle window are opened, whether interior air-conditioning is opened and the size of ambient noise.Example
Such as, a noise scenarios can be defined according at least two in these information.It gives one example, is made an uproar with the division of both the above information
The type of sound field scape then defines the noise field in this example, it is assumed that vehicle is in driving status and the vehicle window of vehicle is opened
Scape is the first noise scenarios；Assuming that the vehicle window that vehicle is in driving status and vehicle is closed, then defining the noise scenarios is the
Two noise scenarios；Assuming that vehicle remains static and the vehicle window of vehicle is opened, then defining noise scenarios is third noise field
Scape；Assuming that vehicle remains static and the vehicle window of vehicle is closed, then defining noise scenarios is the 4th noise scenarios.One is lifted again
A example divides the type of noise scenarios with four kinds of information in information above, stops in this example, it is assumed that current vehicle is in
The vehicle window of parking lot environment, idling, vehicle is opened, and the air-conditioning of vehicle is closed, and between 50~55db, this makes an uproar environmental noise decibel value
Sound field scape is defined as the first noise scenarios.Assuming that current vehicle is in urban road environment, low-speed running state, travel speed
Between 40-60km/h, the vehicle window of vehicle is closed, and the air-conditioning of vehicle is opened, and between 55~60db of environmental noise decibel value, this is made an uproar
Sound field scape is defined as the second noise scenarios.Assuming that current vehicle is mutually in urban road environment, low-speed running state, traveling speed
Degree between 40-60km/h, open by the vehicle window of vehicle, and the air-conditioning of vehicle is closed, should between 60~70db of environmental noise decibel value
Noise scenarios are defined as third noise scenarios；Assuming that current vehicle is in through street environment, high-speed travel state, traveling speed
Degree between 80-120km/h, close by the vehicle window of vehicle, and the air-conditioning of vehicle is opened, between 55~65db of environmental noise decibel value,
The noise scenarios are defined as the 4th noise scenarios.Wherein, the first noise scenarios, the second noise scenarios, third noise scenarios with
And the 4th noise scenarios be one of noise scenarios mark, to distinguish different noise scenarios, in addition to this, the first noise field
Scape and the second noise scenarios can be also used for indicating the intensity of the noise in noise scenarios, such as making an uproar in the first noise scenarios
The intensity of sound is greater than the intensity of the noise in the second noise scenarios.
In a kind of achievable mode, audio defeat method of the invention can be applied to believe audio collected in vehicle
Number noise reduction process is carried out, for example, can be applied in the voice interactive system in vehicle, for collected to voice interactive system
Audio carries out noise reduction, to improve the precision of speech recognition, and then the responding ability of voice interactive system can be improved.
Fig. 2 is a kind of schematic diagram of audio defeat method shown according to an exemplary embodiment, as shown in Fig. 2, the party
Method may include handling as follows:
Collected audio signal is sent to ADC (Analog-to- by sound signal collecting device, such as microphone
Digital Converter, analog/digital converter), so that analog signal (i.e. collected audio signal) is converted to number
Word signal；
Digital signal is sent into intelligent noise scene-detection algorithms module, and (the intelligent scene-detection algorithms module can be according to upper
Type of the noise scenarios disaggregated model described in the text to the feature output noise scene of the audio signal of input) and digital signal
Handle (DSP) algoritic module；
Intelligent noise scene-detection algorithms module carries out the identification of noise scenarios to the digital signal received, and judgement is current
Locating noise scenarios generate the noise reduction parameters group being adapted to current noise scene, by noise reduction according to the noise scenarios judged
Parameter group is sent into digital signal processing algorithm module；
Digital signal processing algorithm module comes from intelligent noise scene-detection algorithms for what noise reduction parameters were updated to receive
Noise reduction parameters in the noise reduction parameters group of module carry out noise reduction process, number to audio signal using updated noise reduction parameters
By treated, audio signal is sent into rear end voice wake-up/recognition processing module to signal processing algorithm module, to treated
Audio signal carries out the operation such as speech recognition or voice wake-up.
The characteristics of above-mentioned intelligent noise scene-detection algorithms module can more accurately identify ambient noise present, for number
Signal processing algorithm module provides the noise reduction parameters group being more suitable for.
Above-mentioned digital signal processing module can get most according to the noise scenarios dynamically load noise reduction parameters being presently in
Good noise reduction effect and best voice signal-to-noise ratio, in voice interactive system, can for rear end speech recognition, voice wake up etc. after
Processing module provides best voice signal, and then the noise immunity of whole voice interactive system can be improved.
Fig. 3 is the process of trained noise scenarios disaggregated model shown according to an exemplary embodiment and using noise field
The process schematic of scape disaggregated model is illustrated the two processes below in conjunction with Fig. 3 and (wherein, flows shown in part on Fig. 3
Journey is the process of training noise scenarios disaggregated model, and process shown in Fig. 3 middle-lower part is to use noise scenarios disaggregated model
Process).
The process of training noise scenarios disaggregated model includes following processing:
It, can be respectively in above-mentioned noise scenarios one, noise scenarios two, noise scenarios three and noise under application scenarios of driving a vehicle
Acquisition noise data under this four noise scenarios of scene four, obtain each driving scene noise database.For example, in each noise field
50 hours environmental noise datas are acquired under scape respectively, collected noise data for example may include that tire is made an uproar, wind is made an uproar, engine is made an uproar
A variety of different types of noise datas such as sound, noisy voice, other motor vehicle noises.
Each audio is marked according to noise scenarios belonging to noise data each in noise database, is extracted in database
The feature of audio, wherein assuming that collecting noise data A in noise scenarios one, then described in one noise data of noise scenarios
Noise scenarios；
The feature for extracting collected audio signal under each noise scenarios in noise database, according in noise database
The feature training noise scenarios disaggregated model of collected audio signal, obtains noise scenarios disaggregated model under each noise scenarios.
Wherein, the training of noise scenarios disaggregated model can be realized using model training methods such as DNN or HMM.
Process using noise scenarios disaggregated model may include handling as follows:
To grandfather tape make an uproar audio signal (audio signal to noise reduction) carry out feature extraction, be sent into decoder (decoder
The corresponding noise scenarios of audio signal are determined using noise scenarios disaggregated model), grandfather tape audio signal of making an uproar is solved frame by frame
Code, obtains decoding result, which includes that grandfather tape is made an uproar the types of the corresponding noise scenarios of audio signal.As continuous 3-10
The decoding result of frame is identical, that is, determines that decoding result is the type of current noise scene, generate the type pair with the noise scenarios
Noise reduction parameters group is answered, which is exported to digital signal processing module, so that digital signal processing module is according to this
Noise reduction parameters form a team grandfather tape make an uproar audio signal carry out noise reduction process.
Fig. 4 is the noise reduction parameters shown according to an exemplary embodiment according in noise reduction parameters group to original signals with noise
The process flow diagram of noise reduction is carried out, as shown in figure 4, carrying out noise reduction process to grandfather tape audio signal of making an uproar according to noise reduction parameters groups
Process includes following operation:
In digital signal processing module, to grandfather tape make an uproar audio signal carry out sub-frame processing, add 20ms Hamming window, take 10
Second frame moves, and carries out FFT (Fast Fourier Transformation, fast Fourier transform) variation frame by frame, obtain frequency spectrum and
Phase information.
The noise reduction parameters in noise reduction parameters group come in using the transmission of intelligent noise scene detection module, such as cross subtracting coefficient α
Spectrum is calculated according to the following formula with spectrum lower limit parameter β to subtract:
If
Wherein, α (being more than or equal to 1) was subtracting coefficient, it mainly influences the distortion level of speech manual.β (is greater than 0 less than 1)
Spectrum lower limit parameter, can control residual noise number and music noise size, Y (ω) indicates original signals with noise, X
(ω) indicates that clean speech signal, D (ω) indicate additive noise.
IFFT (FFT inverse transformation) transformation is carried out, time-domain signal is converted to by frequency-region signal and phase signal, after obtaining noise reduction
Audio signal.
Fig. 5 is a kind of block diagram of audio defeat device shown according to an exemplary embodiment, as shown in figure 5, the device
50 include following component part:
Determining module 51 (module may include above-mentioned intelligent noise scene-detection algorithms module), for according to collected
The feature of audio signal determines the type of noise scenarios；
Module 52 is obtained, for obtaining noise reduction parameters group corresponding with the type of the noise scenarios, the noise reduction parameters
A kind of noise reduction parameters are included at least in group；
Noise reduction module 53, for carrying out noise reduction to the audio signal by the noise reduction parameters in the noise reduction parameters group.
Wherein, obtaining module 52 and noise reduction module 53 can be realized by above-mentioned digital signal processing module.
In a kind of achievable mode, described device may also include that division module, for believing according to collected audio
Number feature determine noise scenarios type before, according to the noise intensity of the audio signal acquired under different scenes and make an uproar
The scene partitioning is different types of noise scenarios by sound type；Module is established, for by the type of each noise scenarios and in advance
If noise parameter group establish corresponding relationship.
In a kind of achievable mode, the acquisition module can include: extraction unit, for extracting the audio signal
Feature；Determination unit, for determining the type of noise scenarios corresponding with each frame in the audio signal according to the feature；
Generation unit generates and the noise field if the type of the corresponding noise scenarios of frame for continuous predetermined number is identical
The corresponding noise reduction parameters group of the type of scape.
In a kind of achievable mode, described device may also include that acquisition module, for acquiring under different noise scenarios
Audio data includes audio signal in the audio data；Mark module is used for according to the noise scenarios to the audio number
According to being marked；Training module, for using the audio data after label to train noise scenarios disaggregated model, the noise scenarios
The input of disaggregated model is the feature of audio signal, and the output of the noise scenarios disaggregated model is that the audio signal is corresponding
Noise scenarios.
In a kind of achievable mode, the type of the noise scenarios: vehicle can be divided according to following at least two information
Whether the driving status of locating environment, vehicle, vehicle window are opened, whether interior air-conditioning is opened and the size of ambient noise.
The device of above-described embodiment for realizing method corresponding in previous embodiment there is corresponding method to implement
The beneficial effect of example, details are not described herein.
Based on the same inventive concept, the embodiment of the invention also provides a kind of electronic equipment, including memory, processor and
The computer program that can be run on a memory and on a processor is stored, the processor is realized as above when executing described program
Audio defeat method described in any embodiment.
Based on the same inventive concept, the embodiment of the invention also provides a kind of non-transient computer readable storage medium, institutes
Non-transient computer readable storage medium storage computer instruction is stated, the computer instruction is for executing the computer such as
Audio defeat method described in upper any embodiment.
It should be understood by those ordinary skilled in the art that: the discussion of any of the above embodiment is exemplary only, not
It is intended to imply that the scope of the present disclosure (including claim) is limited to these examples；Under thinking of the invention, above embodiments
Or can also be combined between the technical characteristic in different embodiments, step can be realized with random order, and be existed such as
Many other variations of the upper different aspect of the invention, for simplicity, they are not provided in details.
In addition, to simplify explanation and discussing, and in order not to obscure the invention, it can in provided attached drawing
It is connect with showing or can not show with the well known power ground of integrated circuit (IC) chip and other components.Furthermore, it is possible to
Device is shown in block diagram form, to avoid obscuring the invention, and this has also contemplated following facts, i.e., about this
The details of the embodiment of a little block diagram arrangements be height depend on will implementing platform of the invention (that is, these details should
It is completely within the scope of the understanding of those skilled in the art).Elaborating that detail (for example, circuit) is of the invention to describe
In the case where exemplary embodiment, it will be apparent to those skilled in the art that can be in these no details
In the case where or implement the present invention in the case that these details change.Therefore, these descriptions should be considered as explanation
Property rather than it is restrictive.
Although having been incorporated with specific embodiments of the present invention, invention has been described, according to retouching for front
It states, many replacements of these embodiments, modifications and variations will be apparent for those of ordinary skills.Example
Such as, discussed embodiment can be used in other memory architectures (for example, dynamic ram (DRAM)).
The embodiment of the present invention be intended to cover fall into all such replacements within the broad range of appended claims,
Modifications and variations.Therefore, all within the spirits and principles of the present invention, any omission, modification, equivalent replacement, the improvement made
Deng should all be included in the protection scope of the present invention.
Claims (12)
1. a kind of audio defeat method characterized by comprising
The type of noise scenarios is determined according to the feature of collected audio signal；
Noise reduction parameters group corresponding with the type of the noise scenarios is obtained, includes at least a kind of noise reduction in the noise reduction parameters group
Parameter；
Noise reduction is carried out to the audio signal by the noise reduction parameters in the noise reduction parameters group.
2. the method according to claim 1, wherein the method also includes:
Before determining the type of noise scenarios according to the feature of collected audio signal, according to what is acquired under different scenes
The scene partitioning is different types of noise scenarios by the noise intensity and noise type of audio signal；
The type of each noise scenarios and preset noise parameter group are established into corresponding relationship.
3. according to the method described in claim 2, it is characterized in that, dividing the noise scenarios according to following at least two information
Type:
Whether environment locating for vehicle, the driving status of vehicle, vehicle window open, whether interior air-conditioning is opened and ambient noise
Size.
4. the method according to claim 1, wherein described obtain drop corresponding with the type of the noise scenarios
It makes an uproar parameter group, comprising:
Extract the feature of the audio signal；
The type of noise scenarios corresponding with each frame in the audio signal is determined according to the feature；
If the type of the corresponding noise scenarios of frame of continuous predetermined number is identical, the type pair with the noise scenarios is generated
The noise reduction parameters group answered.
5. the method according to claim 1, wherein the method also includes:
The audio data under different types of noise scenarios is acquired, includes audio signal in the audio data；
The audio data is marked according to the type of the noise scenarios；
Using the audio data training noise scenarios disaggregated model after label, the input of the noise scenarios disaggregated model is audio
The feature of signal, the output of the noise scenarios disaggregated model are the corresponding noise scenarios of the audio signal.
6. method according to any one of claims 1 to 5, which is characterized in that in the noise reduction parameters group include at least with
A kind of lower noise reduction parameters:
Cross subtracting coefficient and spectrum lower limit parameter.
7. a kind of audio defeat device characterized by comprising
Determining module, for determining the type of noise scenarios according to the feature of collected audio signal；
Module is obtained, for obtaining noise reduction parameters group corresponding with the type of the noise scenarios, in the noise reduction parameters group extremely
It less include a kind of noise reduction parameters；
Noise reduction module, for carrying out noise reduction to the audio signal by the noise reduction parameters in the noise reduction parameters group.
8. device according to claim 7, which is characterized in that described device further include:
Division module, for before determining the type of noise scenarios according to the feature of collected audio signal, according to not
With the audio signal acquired under scene noise intensity and noise type by the scene partitioning be different types of noise field
Scape；
Module is established, for the type of each noise scenarios and preset noise parameter group to be established corresponding relationship.
9. device according to claim 8, which is characterized in that divide the noise scenarios according to following at least two information
Type:
Whether environment locating for vehicle, the driving status of vehicle, vehicle window open, whether interior air-conditioning is opened and ambient noise
Size.
10. device according to claim 7, which is characterized in that the acquisition module, comprising:
Extraction unit, for extracting the feature of the audio signal；
Determination unit, for determining the type of noise scenarios corresponding with each frame in the audio signal according to the feature；
Generation unit, if the type of the corresponding noise scenarios of frame for continuous predetermined number is identical, generation is made an uproar with described
The corresponding noise reduction parameters group of the type of sound field scape.
11. according to the described in any item devices of claim 7 to 10, which is characterized in that described device further include:
Acquisition module includes audio signal in the audio data for acquiring the audio data under different noise scenarios；
Mark module, for the audio data to be marked according to the noise scenarios；
Training module, for using the audio data after label to train noise scenarios disaggregated model, the noise scenarios classification mould
The input of type is the feature of audio signal, and the output of the noise scenarios disaggregated model is the corresponding noise field of the audio signal
Scape.
12. a kind of electronic equipment including memory, processor and stores the calculating that can be run on a memory and on a processor
Machine program, which is characterized in that the processor realizes the sound as described in claim 1 to 6 any one when executing described program
Frequency noise-reduction method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910480708.4A CN110197670B (en) | 2019-06-04 | 2019-06-04 | Audio noise reduction method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910480708.4A CN110197670B (en) | 2019-06-04 | 2019-06-04 | Audio noise reduction method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110197670A true CN110197670A (en) | 2019-09-03 |
| CN110197670B CN110197670B (en) | 2022-06-07 |
Family
ID=67753906
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910480708.4A Active CN110197670B (en) | 2019-06-04 | 2019-06-04 | Audio noise reduction method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110197670B (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111083591A (en) * | 2019-12-13 | 2020-04-28 | 恒玄科技(北京)有限公司 | Configuration method and device of noise reduction earphone and noise reduction earphone |
| CN111678994A (en) * | 2020-07-01 | 2020-09-18 | 山西宏安翔科技股份有限公司 | Acoustic emission sensing system and method |
| CN111724805A (en) * | 2020-06-29 | 2020-09-29 | 北京百度网讯科技有限公司 | Method and apparatus for processing information |
| CN111787149A (en) * | 2020-06-23 | 2020-10-16 | 上海传英信息技术有限公司 | Noise reduction processing method, system and computer storage medium |
| CN112185409A (en) * | 2020-10-15 | 2021-01-05 | 福建瑞恒信息科技股份有限公司 | Double-microphone noise reduction method and storage device |
| CN112185408A (en) * | 2020-10-10 | 2021-01-05 | Oppo广东移动通信有限公司 | Audio noise reduction method and device, electronic equipment and storage medium |
| CN112309359A (en) * | 2020-07-14 | 2021-02-02 | 深圳市逸音科技有限公司 | Method for intelligent scene switching active noise reduction of high-speed audio codec and earphone |
| CN112382284A (en) * | 2020-10-22 | 2021-02-19 | 珠海格力电器股份有限公司 | Pickup noise reduction method and device and server |
| CN112543292A (en) * | 2020-11-30 | 2021-03-23 | 维沃移动通信有限公司 | Image noise reduction method and device, electronic equipment and readable storage medium |
| CN112735457A (en) * | 2020-12-11 | 2021-04-30 | 中国汽车技术研究中心有限公司 | Voice denoising method and system |
| CN112992153A (en) * | 2021-04-27 | 2021-06-18 | 太平金融科技服务（上海）有限公司 | Audio processing method, voiceprint recognition device and computer equipment |
| CN113345399A (en) * | 2021-04-30 | 2021-09-03 | 桂林理工大学 | Method for monitoring sound of machine equipment in strong noise environment |
| CN113496694A (en) * | 2020-03-19 | 2021-10-12 | 上汽通用汽车有限公司 | Vehicle acoustic system, vehicle seat and vehicle |
| CN113531844A (en) * | 2021-07-06 | 2021-10-22 | 青岛海尔空调器有限总公司 | Control method and system for noise reduction of air conditioner, electronic equipment and storage medium |
| CN113556654A (en) * | 2021-07-16 | 2021-10-26 | RealMe重庆移动通信有限公司 | Audio data processing method and device and electronic equipment |
| CN113903323A (en) * | 2021-12-08 | 2022-01-07 | 比亚迪股份有限公司 | Vehicle noise reduction system and car |
| WO2022116825A1 (en) * | 2020-12-03 | 2022-06-09 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based audio processing method and apparatus, electronic device, computer readable storage medium, and computer program product |
| CN114694666A (en) * | 2020-12-28 | 2022-07-01 | 北京小米移动软件有限公司 | Noise reduction processing method and device, terminal and storage medium |
| WO2022247244A1 (en) * | 2021-05-24 | 2022-12-01 | 青岛海尔空调器有限总公司 | Voice control method for air conditioner, and air conditioner |
| WO2024139730A1 (en) * | 2022-12-30 | 2024-07-04 | 腾讯科技（深圳）有限公司 | Audio data processing method and apparatus, and device, computer-readable storage medium and computer program product |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020042712A1 (en) * | 2000-09-29 | 2002-04-11 | Pioneer Corporation | Voice recognition system |
| US20040236570A1 (en) * | 2003-03-28 | 2004-11-25 | Raquel Tato | Method for pre-processing speech |
| US20050171772A1 (en) * | 1997-10-31 | 2005-08-04 | Sony Corporation | Feature extraction apparatus and method and pattern recognition apparatus and method |
| CN1965218A (en) * | 2004-06-04 | 2007-05-16 | 皇家飞利浦电子股份有限公司 | Performance prediction for an interactive speech recognition system |
| US20120303367A1 (en) * | 2006-05-12 | 2012-11-29 | Qnx Software Systems Limited | Robust Noise Estimation |
| CN102945670A (en) * | 2012-11-26 | 2013-02-27 | 河海大学 | Multi-environment characteristic compensation method for voice recognition system |
| CN103219011A (en) * | 2012-01-18 | 2013-07-24 | 联想移动通信科技有限公司 | Noise reduction method, noise reduction device and communication terminal |
| US20140172421A1 (en) * | 2011-08-10 | 2014-06-19 | Goertek Inc. | Speech enhancing method, device for communication earphone and noise reducing communication earphone |
| CN104157294A (en) * | 2014-08-27 | 2014-11-19 | 中国农业科学院农业信息研究所 | Robustness speech recognition method for agricultural product market element information collection |
| CN105228054A (en) * | 2015-10-15 | 2016-01-06 | 深圳市大疆创新科技有限公司 | Flight instruments, filming apparatus and recording denoising device thereof and method |
| CN105427859A (en) * | 2016-01-07 | 2016-03-23 | 深圳市音加密科技有限公司 | Front voice enhancement method for identifying speaker |
| CN107197388A (en) * | 2017-06-29 | 2017-09-22 | 广州华多网络科技有限公司 | A kind of method and system of live noise reduction |
| CN107564538A (en) * | 2017-09-18 | 2018-01-09 | 武汉大学 | The definition enhancing method and system of a kind of real-time speech communicating |
| CN108766454A (en) * | 2018-06-28 | 2018-11-06 | 浙江飞歌电子科技有限公司 | A kind of voice noise suppressing method and device |
-
2019
- 2019-06-04 CN CN201910480708.4A patent/CN110197670B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050171772A1 (en) * | 1997-10-31 | 2005-08-04 | Sony Corporation | Feature extraction apparatus and method and pattern recognition apparatus and method |
| US20020042712A1 (en) * | 2000-09-29 | 2002-04-11 | Pioneer Corporation | Voice recognition system |
| US20040236570A1 (en) * | 2003-03-28 | 2004-11-25 | Raquel Tato | Method for pre-processing speech |
| CN1965218A (en) * | 2004-06-04 | 2007-05-16 | 皇家飞利浦电子股份有限公司 | Performance prediction for an interactive speech recognition system |
| US20120303367A1 (en) * | 2006-05-12 | 2012-11-29 | Qnx Software Systems Limited | Robust Noise Estimation |
| US20140172421A1 (en) * | 2011-08-10 | 2014-06-19 | Goertek Inc. | Speech enhancing method, device for communication earphone and noise reducing communication earphone |
| CN103219011A (en) * | 2012-01-18 | 2013-07-24 | 联想移动通信科技有限公司 | Noise reduction method, noise reduction device and communication terminal |
| CN102945670A (en) * | 2012-11-26 | 2013-02-27 | 河海大学 | Multi-environment characteristic compensation method for voice recognition system |
| CN104157294A (en) * | 2014-08-27 | 2014-11-19 | 中国农业科学院农业信息研究所 | Robustness speech recognition method for agricultural product market element information collection |
| CN105228054A (en) * | 2015-10-15 | 2016-01-06 | 深圳市大疆创新科技有限公司 | Flight instruments, filming apparatus and recording denoising device thereof and method |
| CN105427859A (en) * | 2016-01-07 | 2016-03-23 | 深圳市音加密科技有限公司 | Front voice enhancement method for identifying speaker |
| CN107197388A (en) * | 2017-06-29 | 2017-09-22 | 广州华多网络科技有限公司 | A kind of method and system of live noise reduction |
| CN107564538A (en) * | 2017-09-18 | 2018-01-09 | 武汉大学 | The definition enhancing method and system of a kind of real-time speech communicating |
| CN108766454A (en) * | 2018-06-28 | 2018-11-06 | 浙江飞歌电子科技有限公司 | A kind of voice noise suppressing method and device |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111083591A (en) * | 2019-12-13 | 2020-04-28 | 恒玄科技(北京)有限公司 | Configuration method and device of noise reduction earphone and noise reduction earphone |
| CN113496694A (en) * | 2020-03-19 | 2021-10-12 | 上汽通用汽车有限公司 | Vehicle acoustic system, vehicle seat and vehicle |
| CN111787149A (en) * | 2020-06-23 | 2020-10-16 | 上海传英信息技术有限公司 | Noise reduction processing method, system and computer storage medium |
| CN111724805A (en) * | 2020-06-29 | 2020-09-29 | 北京百度网讯科技有限公司 | Method and apparatus for processing information |
| CN111678994A (en) * | 2020-07-01 | 2020-09-18 | 山西宏安翔科技股份有限公司 | Acoustic emission sensing system and method |
| CN112309359A (en) * | 2020-07-14 | 2021-02-02 | 深圳市逸音科技有限公司 | Method for intelligent scene switching active noise reduction of high-speed audio codec and earphone |
| CN112309359B (en) * | 2020-07-14 | 2024-04-30 | 深圳市逸音科技有限公司 | Intelligent scene switching active noise reduction method of high-speed audio coder and earphone |
| CN112185408A (en) * | 2020-10-10 | 2021-01-05 | Oppo广东移动通信有限公司 | Audio noise reduction method and device, electronic equipment and storage medium |
| CN112185408B (en) * | 2020-10-10 | 2024-05-03 | Oppo广东移动通信有限公司 | Audio noise reduction method and device, electronic equipment and storage medium |
| CN112185409A (en) * | 2020-10-15 | 2021-01-05 | 福建瑞恒信息科技股份有限公司 | Double-microphone noise reduction method and storage device |
| CN112382284A (en) * | 2020-10-22 | 2021-02-19 | 珠海格力电器股份有限公司 | Pickup noise reduction method and device and server |
| CN112543292A (en) * | 2020-11-30 | 2021-03-23 | 维沃移动通信有限公司 | Image noise reduction method and device, electronic equipment and readable storage medium |
| WO2022116825A1 (en) * | 2020-12-03 | 2022-06-09 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based audio processing method and apparatus, electronic device, computer readable storage medium, and computer program product |
| CN112735457B (en) * | 2020-12-11 | 2022-05-17 | 中国汽车技术研究中心有限公司 | Voice denoising method and system |
| CN112735457A (en) * | 2020-12-11 | 2021-04-30 | 中国汽车技术研究中心有限公司 | Voice denoising method and system |
| CN114694666A (en) * | 2020-12-28 | 2022-07-01 | 北京小米移动软件有限公司 | Noise reduction processing method and device, terminal and storage medium |
| CN112992153A (en) * | 2021-04-27 | 2021-06-18 | 太平金融科技服务（上海）有限公司 | Audio processing method, voiceprint recognition device and computer equipment |
| CN113345399A (en) * | 2021-04-30 | 2021-09-03 | 桂林理工大学 | Method for monitoring sound of machine equipment in strong noise environment |
| WO2022247244A1 (en) * | 2021-05-24 | 2022-12-01 | 青岛海尔空调器有限总公司 | Voice control method for air conditioner, and air conditioner |
| CN113531844A (en) * | 2021-07-06 | 2021-10-22 | 青岛海尔空调器有限总公司 | Control method and system for noise reduction of air conditioner, electronic equipment and storage medium |
| CN113556654A (en) * | 2021-07-16 | 2021-10-26 | RealMe重庆移动通信有限公司 | Audio data processing method and device and electronic equipment |
| WO2023284438A1 (en) * | 2021-07-16 | 2023-01-19 | RealMe重庆移动通信有限公司 | Audio data processing method and apparatus, and electronic device |
| CN113903323B (en) * | 2021-12-08 | 2022-04-15 | 比亚迪股份有限公司 | Vehicle noise reduction system and car |
| CN113903323A (en) * | 2021-12-08 | 2022-01-07 | 比亚迪股份有限公司 | Vehicle noise reduction system and car |
| WO2024139730A1 (en) * | 2022-12-30 | 2024-07-04 | 腾讯科技（深圳）有限公司 | Audio data processing method and apparatus, and device, computer-readable storage medium and computer program product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110197670B (en) | 2022-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110197670A (en) | Audio defeat method, apparatus and electronic equipment | |
| CN108122556B (en) | Method and device for reducing false triggering of voice wake-up instruction words of driver | |
| CN101149928B (en) | Sound signal processing method, sound signal processing apparatus and computer program | |
| CN106782504A (en) | Audio recognition method and device | |
| US10224053B2 (en) | Audio signal quality enhancement based on quantitative SNR analysis and adaptive Wiener filtering | |
| US9934780B2 (en) | Method and system for using sound related vehicle information to enhance spoken dialogue by modifying dialogue's prompt pitch | |
| US9263040B2 (en) | Method and system for using sound related vehicle information to enhance speech recognition | |
| CN101154384B (en) | Sound signal correcting method, sound signal correcting apparatus and computer program | |
| CN109192200B (en) | Speech recognition method | |
| US20130185066A1 (en) | Method and system for using vehicle sound information to enhance audio prompting | |
| CN104916289A (en) | Quick acoustic event detection method under vehicle-driving noise environment | |
| CN107910013A (en) | The output processing method and device of a kind of voice signal | |
| CN110348002A (en) | The system and method for realizing voice request | |
| CN106992002A (en) | Dynamic acoustic models switching for improving noisy speech identification | |
| CN110942766A (en) | Audio event detection method, system, mobile terminal and storage medium | |
| CN111667838A (en) | Low-power-consumption analog domain feature vector extraction method for voiceprint recognition | |
| CN114155876A (en) | Traffic flow identification method and device based on audio signal and storage medium | |
| CN112017682A (en) | Single-channel voice simultaneous noise reduction and reverberation removal system | |
| CN113763965A (en) | Speaker identification method with multiple attention characteristics fused | |
| CN101866652A (en) | Voice de-noising method | |
| CN115641861A (en) | Vehicle-mounted voice enhancement method and device, storage medium and equipment | |
| CN108806725A (en) | Speech differentiation method, apparatus, computer equipment and storage medium | |
| Lee et al. | Statistical model‐based noise reduction approach for car interior applications to speech recognition | |
| CN115331656A (en) | Non-instruction voice rejection method, vehicle-mounted voice recognition system and automobile | |
| Loh et al. | Speech recognition interactive system for vehicle |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |