CN111797708A - Airflow noise detection method and device, terminal and storage medium - Google Patents
Airflow noise detection method and device, terminal and storage medium Download PDFInfo
- Publication number
- CN111797708A CN111797708A CN202010536859.XA CN202010536859A CN111797708A CN 111797708 A CN111797708 A CN 111797708A CN 202010536859 A CN202010536859 A CN 202010536859A CN 111797708 A CN111797708 A CN 111797708A
- Authority
- CN
- China
- Prior art keywords
- audio
- sample
- signal
- sample audio
- machine learning
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 238000003860 storage Methods 0.000 title claims abstract description 13
- 230000005236 sound signal Effects 0.000 claims abstract description 143
- 238000010801 machine learning Methods 0.000 claims abstract description 68
- 238000000034 method Methods 0.000 claims description 36
- 238000012549 training Methods 0.000 claims description 32
- 238000000605 extraction Methods 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 13
- 238000010606 normalization Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 7
- 238000012706 support-vector machine Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000009432 framing Methods 0.000 description 3
- 230000010354 integration Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000011897 real-time detection Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000005311 autocorrelation function Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Abstract
The embodiment of the invention discloses an airflow noise detection method, which comprises the steps of collecting an original audio signal in a micro loudspeaker, extracting the characteristics of the original audio signal to obtain audio characteristics, and finally detecting whether the original audio signal has airflow noise or not according to the audio characteristics by a machine learning classifier. In addition, an airflow noise detection device, a computer device and a storage medium are also provided.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to the technical field of computers, in particular to an airflow noise detection method, an airflow noise detection device, a terminal and a storage medium.
[ background of the invention ]
The micro-speaker cavity is narrow and small, and the air current that the vibrating diaphragm vibration caused during operation flows and is not smooth and easy. Under the condition of high voltage, the motion amplitude of the diaphragm is large, air in the cavity generates turbulence, and airflow noise of ' sand and ' fizz ' is caused under certain frequencies, so that the user experience is seriously influenced.
With the popularization of power amplifier applications of intelligent devices such as mobile phones and tablets, the problem of airflow noise when music is played under a large voltage is increasingly serious. To improve this problem, it is necessary to detect whether or not there is a noise in the micro-speaker. At present, the existing detection method for the noise of the airflow comprises the following steps: (1) the micro loudspeaker of the intelligent equipment plays audio signals and judges and detects the audio signals through human ears; (2) and analyzing and detecting the audio signal sent by the micro loudspeaker of the intelligent equipment by adopting an electroacoustic instrument. These detection methods have high process cost, long cycle, difficult accuracy guarantee, and limited versatility, and therefore, a new method for detecting the noise of the airflow is urgently needed.
[ summary of the invention ]
In view of the above, the present invention provides an airflow noise detection method, an airflow noise detection device, a computer device, and a storage medium, which are used to solve the problem of low efficiency in airflow noise elimination detection in the prior art.
The specific technical scheme of the embodiment of the invention is as follows:
in a first aspect, an embodiment of the present invention provides an airflow noise detection method, which is applied to a micro speaker, and the method includes:
collecting original audio signals in a micro loudspeaker;
extracting the characteristics of the original audio signal to obtain audio characteristics;
and detecting whether the original audio signal has air flow noise or not according to the audio features through a machine learning classifier.
In a second aspect, an embodiment of the present invention further provides an airflow noise detection apparatus, where the apparatus includes:
the signal acquisition module is used for acquiring original audio signals in the micro loudspeaker;
the characteristic extraction module is used for extracting the characteristics of the original audio signal to obtain audio characteristics;
and the noise detection module is used for detecting whether the original audio signal has airflow noise or not through a machine learning classifier according to the audio features.
In a third aspect, an embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the airflow noise detection method as described above when executing the computer program.
In a fourth aspect, embodiments of the present invention also provide a computer-readable storage medium, which includes computer instructions, when the computer instructions are executed on a computer, cause the computer to execute the steps of the airflow noise detection method as described above.
The embodiment of the invention has the following beneficial effects:
after the method, the device, the terminal and the storage medium for detecting the airflow noise are adopted, the original audio signal in the micro loudspeaker is collected, then the characteristic extraction is carried out on the original audio signal to obtain the audio characteristic, finally, the machine learning classifier is used for detecting whether the airflow noise exists in the original audio signal according to the audio characteristic, and the machine learning classifier can reflect the inherent characteristics of the existence of the airflow noise or the absence of the airflow noise through training, so that the machine learning classifier is used for detecting whether the airflow noise exists in the original audio signal, the accuracy rate of the airflow noise detection is obviously improved, the cost is reduced, the real-time detection or the off-line detection can be realized, and the applicability of the detection method is improved.
[ description of the drawings ]
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Wherein:
FIG. 1 is a flow chart of a method of detecting noise in an air flow according to one embodiment;
FIG. 2 is a flow diagram of a method for audio feature extraction according to one embodiment;
FIG. 3 is a flow chart of a method of detecting airflow noise according to another embodiment;
FIG. 4 is a flow diagram of a method for machine learning classifier training in accordance with one embodiment;
FIG. 5 is a flow diagram of a sample audio signal set acquisition method according to one embodiment;
FIG. 6 is a schematic diagram of an embodiment of an apparatus for detecting noise in an airflow;
fig. 7 is a schematic diagram of an internal structure of a computer device for operating the airflow noise detection method according to an embodiment.
[ detailed description ] embodiments
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The problem of low accuracy and high cost of manual detection or instrument detection in the traditional technology is solved.
In view of the above problems, in the present embodiment, an airflow noise detection method is particularly proposed. The method may be implemented in dependence on a computer program which is executable on a computer system based on the von neumann architecture.
As shown in fig. 1, the airflow noise detection method provided in this embodiment is applied to a micro speaker, and specifically includes the following steps:
step 102: the original audio signal in the micro-speaker is captured.
Wherein the original audio signal refers to an unprocessed signal in the WeChat loudspeaker. Specifically, sound information on the micro-speaker may be collected as an original audio signal by a sound card, and video or audio data may also be collected as an original audio signal from the micro-speaker. Generally, the original audio signal is small in the cavity of the micro-speaker and precise in structure, but the amplitude of the diaphragm is large, so that the air flow easily forms turbulence in the cavity and generates flow-induced noise.
Step 104: and extracting the characteristics of the original audio signal to obtain audio characteristics.
The audio features are features for characterizing an audio signal, wherein the audio features may be audio features including time domain features, frequency domain features, cepstral domain features, or a combination of these features. Specifically, the original audio signal may be subjected to feature extraction by a toolkit such as C/C + +, Python, MATLAB, and the like, to obtain an audio feature, and more specifically, the time domain feature, the frequency domain feature, and the cepstral domain feature may be a mel-frequency cepstrum coefficient (MFCC), a linear prediction coefficient (LPCC), a component envelope, and the like. Preferably, in order to distinguish the airflow noise more accurately, a combination of several features, i.e., a time domain feature, a frequency domain feature and a cepstrum domain feature, may be used as the audio feature, so as to improve the accuracy of analyzing the airflow noise in the original audio signal.
Step 106: and detecting whether the airflow noise exists in the original audio signal according to the audio features through a machine learning classifier.
The Machine Learning (ML) classifier is a Machine Learning algorithm model with classification capability after training. The machine learning classifier of the present embodiment is configured to classify an original audio signal characterized by audio features into one of an air flow noise signal and an air flow noise signal. The Machine learning classifier may adopt an SVM (Support Vector Machine) classifier, logistic regression, an integration method, a random forest, a neural network model, and the like, wherein the integration method may be a Boosting algorithm, a Bagging algorithm, and a variation thereof. In practice, the recognition accuracy reaches more than 98% when the SVM classifier is used for detecting the airflow noise, and the recognition accuracy reaches more than 97% when the integration method is used for detecting the airflow noise.
Specifically, the classifier that can utilize at least one machine learning model for classification is used as part of the training of the machine learning classifier, the training input is the audio features of various acquired original audio signals, and a relational classifier is established that corresponds the audio features to whether or not there is an airflow noise. The machine learning classifier is enabled to have the capability of judging whether the input audio features have airflow noise or not. In this embodiment, the machine learning classifier is a two-classifier, and two classification results are obtained, that is, there is an airflow noise or there is no airflow noise. The machine learning classifier can reflect the inherent characteristics of existence or nonexistence of the airflow noise through training, so that whether the airflow noise exists in the original audio signal or not is detected by the machine learning classifier, and the accuracy of airflow noise detection is obviously improved.
It is worth mentioning that, in this embodiment, the machine learning classifier is used, and whether the original audio signal has the airflow noise is detected according to the audio features, so that the method is applicable to an airflow noise detection scene of offline detection or real-time detection, and the applicability of the airflow noise detection method is improved.
According to the airflow noise detection method, the original audio signals in the micro loudspeaker are collected, feature extraction is carried out on the original audio signals to obtain audio features, and finally, the machine learning classifier is used for detecting whether the original audio signals have the airflow noise or not according to the audio features.
As shown in fig. 2, in an embodiment, performing feature extraction on an original audio signal to obtain an audio feature includes:
The signal segment refers to a segment of audio signal in the original audio signal, and specifically, the original audio signal may be divided into a plurality of audio frames, the division may be performed in an overlap framing manner, that is, an end portion (e.g., 100 milliseconds) of a previous frame of the original audio signal is used as a start portion of a next frame of the original audio signal, and a plurality of signal segments of the original audio signal may be obtained through overlap framing. The original audio signal may also be framed by a framing function in the MATLAB tool, such as enframe (), to obtain a plurality of signal segments of the original audio signal. And then, respectively performing feature extraction on each divided signal segment to obtain sub-audio features corresponding to each signal segment, where the sub-audio features in this embodiment are the same as the audio features in step 104, and are not described herein again. And finally, combining the plurality of sub-audio features according to the sequence corresponding to the signal segment, wherein the combination mode can be that all the sub-audio features are combined, or a specific number of sub-audio features are selected to be combined to obtain the audio features. It can be understood that the original audio signal is segmented and then subjected to feature extraction, so that the features of each signal segment are more uniform and accurate, and the subsequent analysis is more reliable.
Furthermore, the original audio signals in different time periods are acquired in an off-line mode, and after the original audio signals are segmented, the speed of feature extraction can be increased by extracting features, so that the features of the signal segments are more uniform and abundant, and the accuracy of audio features is improved.
In one embodiment, before performing feature extraction on each signal segment to obtain sub-audio features, the method further includes:
and carrying out normalization processing on the signal segment.
Wherein, each sub-audio feature distribution that each signal section corresponds can be wider, can adjust these signal sections to predetermineeing the interval through the normalization and handle for the characteristic of each signal section is obvious abundantly more, thereby has further guaranteed the unity of each signal section, is favorable to improving the degree of accuracy of the sub-audio feature that the signal section corresponds.
As shown in fig. 3, in one embodiment, the method further comprises:
step 108: acquiring sample audio signals and corresponding marks in the sample audio signal set, wherein the marks of the positive sample audio signals are free of air flow noise, and the marks of the negative sample audio signals are air flow noise;
step 110: extracting each sample audio characteristic in the sample audio signal set to obtain a sample audio characteristic set;
step 112: and training a machine learning classifier according to the sample audio feature set and the corresponding marks.
Specifically, the method includes collecting original audio signals played by a WeChat speaker, wherein a part of the original audio signals contain obvious airflow noise, and the other part of the original audio signals do not contain airflow noise, dividing the original audio signals into signal segments for labeling, then extracting features of each audio signal, wherein the features include time domain features, frequency domain features, cepstrum domain features and the like, mixing audio features of audio signals of positive samples with audio features of audio signals of negative samples, wherein random mixing can be adopted, all sample audio features and corresponding labels form a sample audio feature set, in the embodiment, the sample audio signals comprise positive samples and negative samples in a centralized manner, the negative samples comprise characteristics of the audio features with airflow noise, and a machine learning classifier trained by using the sample audio signal set can learn more accurate classification rules, thereby further improving the accuracy of the detection of the airflow noise.
In one embodiment, the audio features include at least one of time domain features, frequency domain features, and cepstral domain features.
Wherein, the time domain characteristics comprise a short-time average zero crossing rate and a short-time autocorrelation function; the frequency domain characteristics comprise extracting a short-time power spectral density function; cepstral domain feature mel-frequency cepstral coefficients and linear prediction cepstral coefficients. In this embodiment, at least one audio feature is extracted to analyze audio signals by using characteristics of different audio features, thereby improving the accuracy of detection.
It should be noted that, in this embodiment, after the time domain feature, the frequency domain feature, the cepstrum domain feature, or the combination of these features is trained by the machine learning classifier, the optimal audio feature is determined according to the machine learning classifier.
As shown in fig. 4, in one embodiment, training a machine learning classifier based on a sample audio feature set and corresponding labels includes:
step 112A: acquiring a discrete parameter value set of a machine learning classifier;
step 112B: training a machine learning classifier corresponding to each parameter value according to each parameter value in the discrete parameter value set and the sample audio feature set, and obtaining the classification prediction accuracy of the machine learning classifier corresponding to the corresponding parameter value;
step 112C: screening out the maximum classification prediction accuracy, acquiring corresponding parameter values and audio features of the sample audio feature set, and training a machine learning classifier according to the acquired parameter values and the sample audio feature set.
The discrete parameter value set is a set formed by a plurality of discrete parameter values. The parameter value is a value of a parameter required by training the machine learning classifier. Specifically, sampling can be performed in a continuous parameter value range according to a first step length to obtain a series of discrete parameter values, so as to form a discrete parameter value set. If the machine learning classifier includes a plurality of parameters to be learned, a discrete parameter value set corresponding to each parameter can be obtained. If the machine learning classifier adopts an SVM classifier, the parameters are punishment coefficients. Specifically, each parameter value in the discrete parameter value set can be traversed, the currently traversed parameter value and the audio feature corresponding to the sample audio feature set are respectively utilized to train the machine learning classifier, and the classification prediction accuracy corresponding to the machine learning classifier is obtained until all the parameter values in the discrete parameter value set and the audio feature corresponding to the sample audio feature set are traversed.
Further, dividing the sample audio feature set into a training set and a testing set, traversing each parameter value in the discrete parameter value set and the audio feature corresponding to the sample audio feature set, training a machine learning classifier by using the currently traversed parameter value and the audio feature corresponding to the sample audio feature set in the training set, predicting the testing set by using the trained machine learning classifier, obtaining the known classification result of the testing set, and comparing the predicted result with the known classification result to obtain the classification prediction accuracy of the corresponding machine learning classifier. And comparing the classification prediction accuracy rates obtained in the step 112B to find out the maximum classification prediction accuracy rate, and obtaining the parameter value used by the machine learning classifier for training the maximum classification prediction accuracy rate and the audio feature corresponding to the sample audio feature set, so as to continue training the machine learning classifier by using the obtained parameter value and the audio feature corresponding to the sample audio feature set.
In this embodiment, the sample audio feature set is used to quickly find out appropriate parameter values and audio features, so that the parameter values, the audio features and the sample audio feature set are used for training, and the efficiency of training the machine learning classifier can be improved.
As shown in fig. 5, in one embodiment, obtaining sample audio signals and corresponding labels in a sample audio signal set comprises:
Specifically, in order to ensure diversity and watchband property of the acquired original audio signals, it is ensured that one part of the original audio signals contains airflow noise and the other part of the original audio signals does not have airflow noise, so that hardware and human interference are avoided in the same environment. Simultaneously, the same audio acquisition equipment is adopted, a sample audio signal set is acquired by adjusting the gain of the micro loudspeaker, namely, the high point of the voltage is controlled by adjusting the gain, so that the vibration amplitude of the vibrating diaphragm is controlled, the richness of the acquired original audio signal is ensured, and the richness and the comprehensiveness of the sample audio signal set are improved.
The method comprises the steps of dividing sample audio signals in a sample audio signal set to obtain a plurality of sample audio segments corresponding to each sample audio signal, dividing the sample audio signals to be beneficial to improving the uniformity of the sample audio segments, determining the marks of the sample audio signals through a preset loudspeaker model according to the sample audio segments, and ensuring the accuracy of the marks of the sample audio signals through a continuous learning method. In the embodiment, the accuracy of training of the machine learning classifier is improved by acquiring a comprehensive and abundant sample audio signal set and accurately marking.
Based on the same inventive concept, an embodiment of the present invention provides an airflow noise detection apparatus 600, as shown in fig. 6, including: a signal collecting module 602, configured to collect an original audio signal in the micro speaker; a feature extraction module 604, configured to perform feature extraction on the original audio signal to obtain an audio feature; and a noise detection module 606, configured to detect whether an airflow noise exists in the original audio signal according to the audio feature by using a machine learning classifier.
Specifically, the airflow noise detection apparatus 600 of the present embodiment, as shown in fig. 6, includes: a signal collecting module 602, configured to collect an original audio signal in the micro speaker; a feature extraction module 604, configured to perform feature extraction on the original audio signal to obtain an audio feature; and a noise detection module 606, configured to detect whether an airflow noise exists in the original audio signal according to the audio feature by using a machine learning classifier. The machine learning classifier can reflect the inherent characteristics of existence or nonexistence of the air flow noise through training, so that whether the air flow noise exists in the original audio signal or not is detected by the machine learning classifier, the accuracy rate of air flow noise detection is obviously improved, the cost is reduced, and the applicability of the detection method is improved.
It should be noted that, the implementation of the apparatus for detecting an airflow noise in this embodiment is consistent with the implementation concept of the method for detecting an airflow noise, and the implementation principle thereof is not described herein again, and specific reference may be made to the corresponding content in the method.
FIG. 7 is a diagram illustrating an internal structure of a computer device in one embodiment. The computer device may specifically be a server or a terminal. As shown in fig. 7, the computer device 700 includes a processor 710, a memory 720, and a network interface 730, all connected via a system bus. The memory 720 includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program that, when executed by the processor, causes the processor to implement a method of airflow noise detection. The internal memory may also have stored therein a computer program that, when executed by the processor, causes the processor to perform a method of airflow noise detection. Those skilled in the art will appreciate that the architecture shown in fig. 7 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as a particular computing device may include more or less components than those shown in fig. 7, or may combine certain components, or have a different arrangement of components.
In one embodiment, the method for detecting the noise of the air flow provided by the present application can be realized in the form of a computer program, and the computer program can be run on a computer device as shown in fig. 7. The memory of the computer device may have stored therein the various program modules that make up the means for airflow noise detection. For example, the signal acquisition module 602, the feature extraction module 604, and the noise detection module 606.
A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of: collecting original audio signals in a micro loudspeaker; extracting the characteristics of the original audio signal to obtain audio characteristics; and detecting whether the original audio signal has air flow noise or not according to the audio features through a machine learning classifier.
In one embodiment, the performing feature extraction on the original audio signal to obtain an audio feature includes: segmenting the original audio signal to obtain a plurality of signal segments; respectively extracting the characteristics of each signal segment to obtain sub-audio characteristics; and combining the sub-audio features according to the signal segments to obtain the audio features.
In one embodiment, before the performing the feature extraction on each signal segment to obtain the sub-audio feature, the method further includes: and carrying out normalization processing on the signal segment.
In one embodiment, the method further comprises: acquiring sample audio signals and corresponding marks in the sample audio signal set, wherein the marks of the positive sample audio signals are free of air flow noise, and the marks of the negative sample audio signals are air flow noise; extracting each sample audio characteristic in the sample audio signal set to obtain a sample audio characteristic set; and training a machine learning classifier according to the sample audio feature set and the corresponding marks.
In one embodiment, the audio features include at least one of time domain features, frequency domain features, and cepstral domain features.
In one embodiment, training a machine learning classifier based on the sample audio feature set and corresponding labels includes: acquiring a discrete parameter value set of a machine learning classifier; training a machine learning classifier corresponding to each parameter value according to each parameter value in the discrete parameter value set and the sample audio feature set, and obtaining the classification prediction accuracy of the machine learning classifier corresponding to the corresponding parameter value; screening out the maximum classification prediction accuracy, acquiring corresponding parameter values and audio features of the sample audio feature set, and training a machine learning classifier according to the acquired parameter values and the sample audio feature set.
In one embodiment, the obtaining of sample audio signals and corresponding labels in a sample audio signal set comprises: under the same environment, the same audio acquisition equipment is adopted, and the sample audio signal set is obtained by adjusting the gain acquisition of the micro loudspeaker; dividing the sample audio signals in the sample audio signal set to obtain the plurality of sample audio segments corresponding to each sample audio signal; and determining the mark of the sample audio signal through a preset loudspeaker model according to the sample audio segment.
A computer-readable storage medium storing a computer program, the computer program when executed by a processor implementing the steps of: collecting original audio signals in a micro loudspeaker; extracting the characteristics of the original audio signal to obtain audio characteristics; and detecting whether the original audio signal has air flow noise or not according to the audio features through a machine learning classifier.
In one embodiment, the performing feature extraction on the original audio signal to obtain an audio feature includes: segmenting the original audio signal to obtain a plurality of signal segments; respectively extracting the characteristics of each signal segment to obtain sub-audio characteristics; and combining the sub-audio features according to the signal segments to obtain the audio features.
In one embodiment, before the performing the feature extraction on each signal segment to obtain the sub-audio feature, the method further includes: and carrying out normalization processing on the signal segment.
In one embodiment, the method further comprises: acquiring sample audio signals and corresponding marks in the sample audio signal set, wherein the marks of the positive sample audio signals are free of air flow noise, and the marks of the negative sample audio signals are air flow noise; extracting each sample audio characteristic in the sample audio signal set to obtain a sample audio characteristic set; and training a machine learning classifier according to the sample audio feature set and the corresponding marks.
In one embodiment, the audio features include at least one of time domain features, frequency domain features, and cepstral domain features.
In one embodiment, training a machine learning classifier based on the sample audio feature set and corresponding labels includes: acquiring a discrete parameter value set of a machine learning classifier; training a machine learning classifier corresponding to each parameter value according to each parameter value in the discrete parameter value set and the sample audio feature set, and obtaining the classification prediction accuracy of the machine learning classifier corresponding to the corresponding parameter value; screening out the maximum classification prediction accuracy, acquiring corresponding parameter values and audio features of the sample audio feature set, and training a machine learning classifier according to the acquired parameter values and the sample audio feature set.
In one embodiment, the obtaining of sample audio signals and corresponding labels in a sample audio signal set comprises: under the same environment, the same audio acquisition equipment is adopted, and the sample audio signal set is obtained by adjusting the gain acquisition of the micro loudspeaker; dividing the sample audio signals in the sample audio signal set to obtain the plurality of sample audio segments corresponding to each sample audio signal; and determining the mark of the sample audio signal through a preset loudspeaker model according to the sample audio segment.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a non-volatile computer-readable storage medium, and can include the processes of the embodiments of the methods described above when the program is executed. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.
Claims (10)
1. An airflow noise detection method applied to a micro-speaker, the method comprising:
collecting original audio signals in a micro loudspeaker;
extracting the characteristics of the original audio signal to obtain audio characteristics;
and detecting whether the original audio signal has air flow noise or not according to the audio features through a machine learning classifier.
2. The method for detecting airflow noise according to claim 1, wherein said extracting the characteristics of the original audio signal to obtain the audio characteristics comprises:
segmenting the original audio signal to obtain a plurality of signal segments;
respectively extracting the characteristics of each signal segment to obtain sub-audio characteristics;
and combining the sub-audio features according to the signal segments to obtain the audio features.
3. The method according to claim 2, wherein before said separately performing feature extraction on each of said signal segments to obtain sub-audio features, the method further comprises:
and carrying out normalization processing on the signal segment.
4. The airflow noise detection method of claim 1, further comprising:
acquiring sample audio signals and corresponding marks in the sample audio signal set, wherein the marks of the positive sample audio signals are free of air flow noise, and the marks of the negative sample audio signals are air flow noise;
extracting each sample audio characteristic in the sample audio signal set to obtain a sample audio characteristic set;
and training a machine learning classifier according to the sample audio feature set and the corresponding marks.
5. The airflow noise detection method of claim 1, wherein the audio features include at least one of time domain features, frequency domain features, and cepstral domain features.
6. The airflow noise detection method of claim 5, wherein training a machine learning classifier based on the sample audio feature set and corresponding labels comprises:
acquiring a discrete parameter value set of a machine learning classifier;
training a machine learning classifier corresponding to each parameter value according to each parameter value in the discrete parameter value set and the sample audio feature set, and obtaining the classification prediction accuracy of the machine learning classifier corresponding to the corresponding parameter value;
screening out the maximum classification prediction accuracy, acquiring corresponding parameter values and audio features of the sample audio feature set, and training a machine learning classifier according to the acquired parameter values and the sample audio feature set.
7. The airflow noise detection method of claim 4, wherein said obtaining sample audio signals and corresponding labels in a sample audio signal set comprises:
under the same environment, the same audio acquisition equipment is adopted, and the sample audio signal set is obtained by adjusting the gain acquisition of the micro loudspeaker;
dividing the sample audio signals in the sample audio signal set to obtain a plurality of sample audio segments corresponding to each sample audio signal;
and determining the mark of the sample audio signal through a preset loudspeaker model according to the sample audio segment.
8. An airflow noise detection apparatus, the apparatus comprising:
the signal acquisition module is used for acquiring original audio signals in the micro loudspeaker;
the characteristic extraction module is used for extracting the characteristics of the original audio signal to obtain audio characteristics;
and the noise detection module is used for detecting whether the original audio signal has airflow noise or not through a machine learning classifier according to the audio features.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the computer program performs the steps of the airflow noise detection method of any one of claims 1 to 7.
10. A computer readable storage medium comprising computer instructions which, when run on a computer, cause the computer to perform the steps of the airflow noise detection method of any of claims 1 to 7.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010536859.XA CN111797708A (en) | 2020-06-12 | 2020-06-12 | Airflow noise detection method and device, terminal and storage medium |
| PCT/CN2020/096685 WO2021248522A1 (en) | 2020-06-12 | 2020-06-18 | Current noise detection method and apparatus, terminal, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010536859.XA CN111797708A (en) | 2020-06-12 | 2020-06-12 | Airflow noise detection method and device, terminal and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111797708A true CN111797708A (en) | 2020-10-20 |
Family
ID=72803904
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010536859.XA Pending CN111797708A (en) | 2020-06-12 | 2020-06-12 | Airflow noise detection method and device, terminal and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111797708A (en) |
| WO (1) | WO2021248522A1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112487865A (en) * | 2020-11-02 | 2021-03-12 | 杭州兆华电子有限公司 | Automatic loudspeaker classification method based on machine learning |
| CN112863548A (en) * | 2021-01-22 | 2021-05-28 | 北京百度网讯科技有限公司 | Method for training audio detection model, audio detection method and device thereof |
| CN114627891A (en) * | 2022-05-16 | 2022-06-14 | 山东捷瑞信息技术产业研究院有限公司 | Moving coil loudspeaker quality detection method and device |
| CN116038076A (en) * | 2023-03-28 | 2023-05-02 | 苏芯物联技术(南京)有限公司 | Welding shielding gas control method based on scene matching |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103824557A (en) * | 2014-02-19 | 2014-05-28 | 清华大学 | Audio detecting and classifying method with customization function |
| WO2014160678A2 (en) * | 2013-03-26 | 2014-10-02 | Dolby Laboratories Licensing Corporation | 1apparatuses and methods for audio classifying and processing |
| CN108287816A (en) * | 2017-01-10 | 2018-07-17 | 腾讯科技（深圳）有限公司 | Point of interest on-line checking, Machine learning classifiers training method and device |
| CN109785850A (en) * | 2019-01-18 | 2019-05-21 | 腾讯音乐娱乐科技（深圳）有限公司 | A kind of noise detecting method, device and storage medium |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106202952A (en) * | 2016-07-19 | 2016-12-07 | 南京邮电大学 | A kind of Parkinson disease diagnostic method based on machine learning |
| CN108073856A (en) * | 2016-11-14 | 2018-05-25 | 华为技术有限公司 | The recognition methods of noise signal and device |
| EP3547308B1 (en) * | 2018-03-26 | 2024-01-24 | Sony Group Corporation | Apparatuses and methods for acoustic noise cancelling |
| CN109903752B (en) * | 2018-05-28 | 2021-04-20 | 华为技术有限公司 | Method and device for aligning voice |
-
2020
- 2020-06-12 CN CN202010536859.XA patent/CN111797708A/en active Pending
- 2020-06-18 WO PCT/CN2020/096685 patent/WO2021248522A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014160678A2 (en) * | 2013-03-26 | 2014-10-02 | Dolby Laboratories Licensing Corporation | 1apparatuses and methods for audio classifying and processing |
| CN103824557A (en) * | 2014-02-19 | 2014-05-28 | 清华大学 | Audio detecting and classifying method with customization function |
| CN108287816A (en) * | 2017-01-10 | 2018-07-17 | 腾讯科技（深圳）有限公司 | Point of interest on-line checking, Machine learning classifiers training method and device |
| CN109785850A (en) * | 2019-01-18 | 2019-05-21 | 腾讯音乐娱乐科技（深圳）有限公司 | A kind of noise detecting method, device and storage medium |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112487865A (en) * | 2020-11-02 | 2021-03-12 | 杭州兆华电子有限公司 | Automatic loudspeaker classification method based on machine learning |
| CN112863548A (en) * | 2021-01-22 | 2021-05-28 | 北京百度网讯科技有限公司 | Method for training audio detection model, audio detection method and device thereof |
| CN114627891A (en) * | 2022-05-16 | 2022-06-14 | 山东捷瑞信息技术产业研究院有限公司 | Moving coil loudspeaker quality detection method and device |
| CN116038076A (en) * | 2023-03-28 | 2023-05-02 | 苏芯物联技术(南京)有限公司 | Welding shielding gas control method based on scene matching |
| CN116038076B (en) * | 2023-03-28 | 2023-10-13 | 苏芯物联技术(南京)有限公司 | Welding shielding gas control method based on scene matching |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021248522A1 (en) | 2021-12-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111797708A (en) | Airflow noise detection method and device, terminal and storage medium | |
| CN108877775B (en) | Voice data processing method and device, computer equipment and storage medium | |
| CN108198547B (en) | Voice endpoint detection method and device, computer equipment and storage medium | |
| WO2021128741A1 (en) | Voice emotion fluctuation analysis method and apparatus, and computer device and storage medium | |
| CN109034046B (en) | Method for automatically identifying foreign matters in electric energy meter based on acoustic detection | |
| CN111816218A (en) | Voice endpoint detection method, device, equipment and storage medium | |
| CN111145782B (en) | Overlapped speech recognition method, device, computer equipment and storage medium | |
| US11386916B2 (en) | Segmentation-based feature extraction for acoustic scene classification | |
| Huang et al. | Intelligent feature extraction and classification of anuran vocalizations | |
| CN108899033B (en) | Method and device for determining speaker characteristics | |
| CN108091340B (en) | Voiceprint recognition method, voiceprint recognition system, and computer-readable storage medium | |
| CN111341333B (en) | Noise detection method, noise detection device, medium, and electronic apparatus | |
| CN116486834A (en) | Rolling sound classification method based on feature fusion and improved convolutional neural network | |
| CN114639155A (en) | Emotion recognition method, emotion recognition device, storage medium and processor | |
| CN111261145A (en) | Voice processing device, equipment and training method thereof | |
| Jena et al. | Gender recognition of speech signal using knn and svm | |
| CN113223492B (en) | Voice broadcasting system | |
| CN112786054A (en) | Intelligent interview evaluation method, device and equipment based on voice and storage medium | |
| CN112802498A (en) | Voice detection method and device, computer equipment and storage medium | |
| CN111785302A (en) | Speaker separation method and device and electronic equipment | |
| CN112216285B (en) | Multi-user session detection method, system, mobile terminal and storage medium | |
| CN111091816B (en) | Data processing system and method based on voice evaluation | |
| CN114121025A (en) | Voiceprint fault intelligent detection method and device for substation equipment | |
| JP2002062892A (en) | Acoustic classifying device | |
| CN111782860A (en) | Audio detection method and device and storage medium |
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 |