Method and system for compensating frequency response of a microphone

Abstract

The present invention relates to the technical field of audio processing, and in particular to a method and system for compensating the frequency response of a microphone. The method includes the following steps: obtaining microphone output signal data and multi-dimensional environmental data, and performing frequency response multi-level feature integration to obtain coupled frequency response data; performing frequency response-environment parameter correlation analysis based on the coupled frequency response data to obtain a frequency feature matrix; obtaining environmental noise monitoring data and performing frequency band environmental noise calibration to obtain a noise calibration matrix; performing Monte Carlo multi-dimensional parameter optimization based on the frequency feature matrix and the noise calibration matrix to obtain an optimized compensation strategy; obtaining real-time microphone output signal data and performing dynamic compensation feedback to obtain dynamic compensation feedback data; and selecting a user personalized compensation strategy according to the dynamic compensation feedback data to obtain a user microphone personalized compensation strategy. The present invention can improve the accuracy and quality of audio processing.

Description

Technical Field

[0001] The present invention relates to the technical field of audio processing, and in particular, to a method and system for compensating the frequency response of a microphone. Background Art

[0002] With the development of audio technology, the microphone, as a key device for audio acquisition, is widely used in various audio application scenarios, including speech recognition, communication, acoustic measurement, recording and other fields. However, the frequency response characteristics of the microphone directly affect the accuracy and quality of the sound it captures. In practical applications, the microphone often has the phenomenon of uneven frequency response, which will lead to audio distortion in recording or speech recognition and affect the subsequent processing results. The software compensation method realizes the compensation of the frequency response by digitally processing the audio signal output by the microphone. However, the software compensation method also has certain limitations. Software compensation often relies on prior microphone frequency response models and algorithms, which requires accurate calibration and modeling before compensation. For microphones of different brands or models, the frequency response characteristics will vary greatly. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide a method and system for compensating the frequency response of a microphone to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for compensating the frequency response of a microphone includes the following steps:

[0005] Step S1: Obtain the microphone output signal data and multi-dimensional environment data, and perform a frequency response curve conversion on the microphone output signal data to obtain the microphone frequency response curve; perform a multi-level feature integration of the frequency response according to the microphone frequency response curve and the multi-dimensional environment data to obtain the coupled frequency response data;

[0006] Step S2: Screen the main component features of the frequency band frequency based on the coupled frequency response data to obtain the main component feature data of the frequency band frequency, and perform a frequency response-environment parameter correlation analysis according to the main component feature data of the frequency band frequency to obtain the frequency feature matrix;

[0007] Step S3: Obtain the environmental noise monitoring data, map the environmental noise monitoring data to the frequency feature matrix to obtain the noise signal mapping matrix; perform a frequency band environmental noise calibration on the noise signal mapping matrix to obtain the noise calibration matrix;

[0008] Step S4: Perform an adaptive dynamic compensation strategy analysis based on the frequency feature matrix and the noise calibration matrix to obtain an initial adaptive dynamic compensation strategy; perform Monte Carlo multi-dimensional parameter optimization on the initial adaptive dynamic compensation strategy according to the coupled frequency response data to obtain an optimized compensation strategy, and transmit the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy;

[0009] Step S5: Obtain real-time microphone output signal data, and evaluate the compensation effect of the real-time microphone output signal data and the optimized compensation strategy to obtain real-time audio compensation score data; perform dynamic compensation feedback on the optimized compensation strategy based on the real-time audio compensation score data to obtain dynamic compensation feedback data;

[0010] Step S6: Build a global compensation model according to the dynamic compensation feedback data to obtain a microphone global compensation model; perform user personalized compensation strategy selection based on the microphone global compensation model to obtain a user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

[0011] Optionally, step S1 is specifically as follows:

[0012] Step S11: Obtain microphone output signal data and multi-dimensional environment data, and perform data preprocessing on the microphone output signal data and the multi-dimensional environment data respectively to obtain the microphone output signal data to be analyzed and the multi-dimensional environment data to be analyzed;

[0013] Step S12: Perform a Fourier transform on the microphone output signal data to be analyzed to obtain the microphone output signal spectrum, and perform frequency domain data standardization on the microphone output signal spectrum to obtain the microphone output signal standardized spectrum;

[0014] Step S13: Integrate the frequency response characteristics of the microphone output signal standardized spectrum to obtain the microphone frequency response characteristic data, and perform a frequency response curve conversion based on the microphone frequency response characteristic data to obtain the microphone frequency response curve;

[0015] Step S14: Perform multi-dimensional environment data fusion on the multi-dimensional environment data to be analyzed to obtain the environment fusion data to be analyzed;

[0016] Step S15: Perform frequency response multi-level feature integration according to the microphone frequency response curve and the environment fusion data to be analyzed to obtain the coupled frequency response data.

[0017] Optionally, step S15 is specifically as follows:

[0018] Step S151: Align the frequency characteristics of the microphone frequency response curve and the data of the environment to be analyzed for fusion, so as to obtain a frequency-aligned response curve and frequency-aligned environment fusion data;

[0019] Step S152: Conduct a correlation analysis on the frequency-aligned response curve and the frequency-aligned environment fusion data, so as to obtain microphone frequency-environment correlation data, and perform feature selection on the frequency-aligned environment fusion data according to the microphone frequency-environment correlation data, so as to obtain frequency-related environment data;

[0020] Step S153: Concatenate multi-dimensional feature vectors of the frequency-aligned response curve and the frequency-related environment data, so as to obtain frequency response-environment feature composite data;

[0021] Step S154: Conduct a hierarchical analysis of the environmental impact on multi-frequency responses based on the frequency response-environment feature composite data, so as to obtain frequency response hierarchical environment association data;

[0022] Step S155: Perform a mapping coupling of the frequency response hierarchical relationship on the frequency response-environment feature composite data based on the frequency response hierarchical environment association data, so as to obtain coupled frequency response data.

[0023] Optionally, step S2 is specifically as follows:

[0024] Step S21: Divide the frequency bands of the coupled frequency response data, so as to obtain frequency band divided coupled frequency response data;

[0025] Step S22: Conduct a principal component feature analysis of the frequency band divided coupled frequency response data, so as to obtain frequency band principal component feature data;

[0026] Step S23: Calculate the influence factors of the frequency response of the frequency band based on the frequency band principal component feature data, so as to obtain the influence factors of the frequency response of the frequency band;

[0027] Step S24: Quantify the influence of environmental parameters according to the influence factors of the frequency response of the frequency band, so as to obtain environmental parameter influence degree data;

[0028] Step S25: Conduct a frequency response-environment parameter association on the environmental parameter influence degree data and the frequency band principal component feature data, so as to obtain a frequency response-environment association matrix, and construct a frequency feature matrix according to the frequency response-environment association matrix.

[0029] Optionally, step S3 is specifically as follows:

[0030] Step S31: Obtain environmental noise monitoring data, and perform data preprocessing on the environmental noise monitoring data, so as to obtain environmental noise monitoring data to be analyzed;

[0031] Step S32: Perform Fourier transform on the environmental noise monitoring data to be analyzed to obtain the environmental noise signal spectrum, and perform statistical analysis on the energy distribution of the noise signal frequency bands in the environmental noise signal spectrum to obtain the energy distribution diagram of the noise signal frequency bands;

[0032] Step S33: Integrate the frequency distribution of the noise signal according to the energy distribution diagram of the noise signal frequency bands to obtain the noise signal frequency distribution matrix;

[0033] Step S34: Perform time series mapping on the noise signal frequency distribution matrix and the frequency feature matrix to obtain the initial signal mapping matrix, and perform frequency axis feature alignment on the initial signal mapping matrix to obtain the noise signal mapping matrix;

[0034] Step S35: Calibrate the environmental noise in the frequency band of the noise signal mapping matrix to obtain the noise calibration matrix.

[0035] Optionally, step S35 is specifically:

[0036] Step S351: Divide the frequency data of the noise signal mapping matrix into frequency bands to obtain the noise signal band spectrum;

[0037] Step S352: Perform time series correlation of the environmental condition - band noise signal based on the noise signal frequency distribution matrix and the noise signal band spectrum to obtain the environmental condition - band noise signal change data;

[0038] Step S353: Calculate the environmental noise impact factor based on the environmental condition - band noise signal change data to obtain the environmental noise impact factor;

[0039] Step S354: Calculate the band calibration coefficient for the noise signal band spectrum according to the environmental noise impact factor to obtain the band calibration coefficient;

[0040] Step S355: Calibrate the noise signal band of the noise signal mapping matrix according to the band calibration coefficient to obtain the noise calibration matrix.

[0041] Optionally, step S4 is specifically:

[0042] Step S41: Compare the features of the frequency feature matrix and the noise calibration matrix to obtain the frequency response difference data, and integrate the noise impact factor according to the frequency response difference data to obtain the noise impact factor;

[0043] Step S42: Construct an initial noise compensation model based on the frequency response difference data and the noise impact factor;

[0044] Step S43: Design a compensation strategy by means of the initial noise compensation model, in combination with the microphone frequency response characteristic data, the frequency band calibration coefficient, and the data of the environment to be analyzed for fusion, so as to obtain an initial adaptive dynamic compensation strategy;

[0045] Step S44: Set compensation parameters according to the coupled frequency response data, so as to obtain a set of compensation parameters, and perform Monte Carlo random sampling on the set of compensation parameters, so as to obtain a set of random compensation parameters;

[0046] Step S45: Perform signal compensation simulation on the set of random compensation parameters according to the initial adaptive dynamic compensation strategy, so as to obtain signal compensation simulation data, and evaluate the compensation signal error of the signal compensation simulation data, so as to obtain compensation signal error data;

[0047] Step S46: Select a set of compensation parameter groups for the set of random compensation parameters based on the compensation signal error data, so as to obtain an optimal set of compensation parameters;

[0048] Step S47: Optimize the multi-dimensional parameters of the compensation strategy for the initial adaptive dynamic compensation strategy according to the optimal set of compensation parameters, so as to obtain an optimized compensation strategy, and transmit the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy.

[0049] Optionally, step S5 is specifically as follows:

[0050] Step S51: Obtain real-time microphone output signal data, and perform data preprocessing on the real-time microphone output signal data, so as to obtain real-time microphone output signal data to be analyzed;

[0051] Step S52: Extract the frequency characteristics of the audio signal from the real-time microphone output signal data to be analyzed, so as to obtain the real-time audio signal spectrum;

[0052] Step S53: Perform compensation strategy backtracking on the real-time audio signal spectrum according to the optimized compensation strategy, so as to obtain the real-time uncompensated audio signal spectrum;

[0053] Step S54: Evaluate the compensation effect based on the real-time audio signal spectrum and the real-time uncompensated audio signal spectrum, so as to obtain audio compensation effect evaluation data, and weight the evaluation results of the audio compensation effect evaluation data, so as to obtain real-time audio compensation score data;

[0054] Step S55: Classify the real-time audio compensation scoring data according to a preset scoring threshold. If the real-time audio compensation scoring data is greater than or equal to the scoring threshold, send an instruction to the microphone device controller to continue executing the optimization compensation strategy; if the real-time audio compensation scoring data is less than the scoring threshold, mark the corresponding optimization compensation strategy as an abnormal compensation strategy and send an instruction to the microphone device controller to stop executing the optimization compensation strategy;

[0055] Step S56: Adjust the real-time compensation strategy for the abnormal compensation strategy according to the real-time audio compensation scoring data, so as to obtain dynamic compensation feedback data.

[0056] Optionally, step S6 is specifically as follows:

[0057] Step S61: Select the principal component features of the dynamic compensation feedback data to obtain dynamic compensation feature data;

[0058] Step S62: Build a neural network global compensation model based on the dynamic compensation feature data to obtain a microphone global compensation model;

[0059] Step S63: Obtain user microphone interaction data and perform user audio preference analysis on the user microphone interaction data to obtain user audio preference data;

[0060] Step S64: Select a user personalized compensation strategy for the user audio preference data through the microphone global compensation model to obtain a user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

[0061] Through the dynamic compensation and personalized optimization of the microphone frequency response and environmental noise, the accuracy and quality of microphone audio acquisition and output are significantly improved. By obtaining the microphone output signal data and multi-dimensional environmental data, and performing frequency response curve conversion and feature integration, the frequency response characteristics of the microphone in a specific environment can be effectively depicted. Combining the environmental impact with the frequency response characteristics of the microphone itself provides a comprehensive feature data basis for subsequent compensation. Next, based on the screening of the main component features of the frequency band and the correlation analysis of the frequency response-environment parameters, it helps to extract the key features affecting the frequency response, and combines these features with environmental factors, thus laying a foundation for a more accurate compensation strategy. Through the monitoring and calibration of environmental noise, not only can the noise interference be effectively quantified, but also the noise impact can be adaptively adjusted under different environmental conditions, and then the compensation model can be optimized to ensure the efficiency and accuracy of the compensation effect. The optimization of the dynamic compensation strategy adopts Monte Carlo multi-dimensional parameter optimization to ensure the adaptability and stability of the compensation strategy under different environments and noise conditions. Through the adaptive dynamic compensation and model modeling based on real-time feedback data, the compensation strategy can be adjusted in a timely manner according to the actual feedback, so as to ensure that the microphone can provide ideal audio acquisition effects in various environments. Especially through the real-time audio compensation scoring mechanism, the effectiveness of the optimized compensation strategy can be dynamically evaluated, and the compensation strategy can be further adjusted according to the real-time feedback to ensure that each compensation can meet the user's audio preferences as much as possible. The core advantage of this process lies in that through real-time monitoring and adjustment, the microphone can adapt to complex and changing audio environments, thus significantly improving the accuracy and quality of audio processing. By analyzing the user preference data and combining the global compensation model, personalized compensation strategies can be provided according to the audio needs of different users. The application of this personalized compensation strategy greatly improves the user experience, enabling the microphone not only to compensate in a noisy environment, but also to adjust its frequency response characteristics according to different user preferences to provide the best audio output effect. Through this series of beneficial technical steps, the goal of providing accurate and high-quality audio acquisition and processing under different environments and user needs is finally achieved.

[0062] Optionally, this specification also provides a system for compensating the frequency response of a microphone, which is used to execute the method for compensating the frequency response of a microphone as described above. The system for compensating the frequency response of a microphone includes:

[0063] A frequency response coupling module, configured to obtain microphone output signal data and multi-dimensional environmental data, and perform frequency response curve conversion on the microphone output signal data to obtain a microphone frequency response curve; perform frequency response multi-level feature integration according to the microphone frequency response curve and the multi-dimensional environmental data to obtain coupled frequency response data;

[0064] The principal component feature screening module is used to screen the principal component features of the frequency bands based on the coupled frequency response data, so as to obtain the principal component feature data of the frequency bands of the frequency, and perform frequency response - environmental parameter correlation analysis according to the principal component feature data of the frequency bands of the frequency, so as to obtain the frequency feature matrix;

[0065] The environmental noise calibration module is used to obtain the environmental noise monitoring data, and map the environmental noise monitoring data to the frequency feature matrix, so as to obtain the noise signal mapping matrix; perform frequency band environmental noise calibration on the noise signal mapping matrix, so as to obtain the noise calibration matrix;

[0066] The compensation strategy generation module is used to perform adaptive dynamic compensation strategy analysis based on the frequency feature matrix and the noise calibration matrix, so as to obtain the initial adaptive dynamic compensation strategy; perform Monte Carlo multi - dimensional parameter optimization on the initial adaptive dynamic compensation strategy according to the coupled frequency response data, so as to obtain the optimized compensation strategy, and transmit the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy;

[0067] The compensation strategy feedback module is used to obtain the real - time microphone output signal data, and perform compensation effect evaluation on the real - time microphone output signal data and the optimized compensation strategy, so as to obtain the real - time audio compensation score data; perform dynamic compensation feedback on the optimized compensation strategy based on the real - time audio compensation score data, so as to obtain the dynamic compensation feedback data;

[0068] The personalized compensation strategy selection module is used to perform global compensation model modeling according to the dynamic compensation feedback data, so as to obtain the microphone global compensation model; perform user - personalized compensation strategy selection based on the microphone global compensation model, so as to obtain the user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

[0069] The system for compensating the frequency response of a microphone according to the present invention can implement any method for compensating the frequency response of a microphone according to the present invention. It is a medium for coordinating the operations and signal transmissions between various modules to complete the method for compensating the frequency response of a microphone. The internal modules of the system cooperate with each other, thereby improving the accuracy and quality of audio processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Other features, objects, and advantages of the present invention will become more apparent by reading the detailed description of the non - restrictive embodiments with reference to the following drawings:

[0071] Figure 1 It is a schematic flowchart of the steps of the method for compensating the frequency response of a microphone according to the present invention;

[0072] Figure 2It is a detailed step - by - step flowchart of step S1 in the present invention;

[0073] Figure 3 It is a detailed step - by - step flowchart of step S2 in the present invention;

[0074] The realization of the object, functional features and advantages of the present invention will be further described in conjunction with embodiments with reference to the accompanying drawings. Specific embodiments

[0075] The technical method of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0076] In addition, the accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0077] It should be understood that although the terms "first", "second", etc. may be used here to describe each unit, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit can be called the second unit, and similarly the second unit can be called the first unit. The term "and / or" used here includes any and all combinations of one or more of the listed related items.

[0078] To achieve the above object, please refer to Figures 1 to 3 , the present invention provides a method for compensating the frequency response of a microphone, and the method includes the following steps:

[0079] Step S1: Obtain the microphone output signal data and multi - dimensional environmental data, and perform a frequency response curve conversion on the microphone output signal data to obtain the microphone frequency response curve; perform a multi - level feature integration of the frequency response according to the microphone frequency response curve and the multi - dimensional environmental data to obtain the coupled frequency response data;

[0080] In this embodiment, an environmental audio signal is collected by a microphone device. The microphone is built-in with an analog-to-digital converter (ADC) to convert the audio signal into a digital signal, and a fast Fourier transform (FFT) is used to perform frequency-domain analysis on it to obtain a frequency response curve. This curve shows the response of the microphone to signals of different frequencies. On this basis, multi-dimensional data in the environment, including temperature, humidity, air pressure, background noise, etc., is further collected through sensors built in the microphone, and these data are obtained in real time through various sensors. Subsequently, the microphone frequency response curve and environmental data are fed into a deep neural network model for multi-level feature integration to obtain coupled frequency response data. This data combines the comprehensive effects of audio signals and environmental factors and is used for the design of subsequent compensation strategies.

[0081] Step S2: Screen the principal component features of the frequency bands based on the coupled frequency response data to obtain the principal component feature data of the frequency bands, and perform frequency response-environment parameter correlation analysis based on the principal component feature data of the frequency bands to obtain a frequency feature matrix;

[0082] In this embodiment, based on the coupled frequency response data, the principal component analysis (PCA) algorithm is used to perform dimensionality reduction processing on the data to extract the principal component features of the frequency bands. These features represent the most important change dimensions of the frequency response data. Then, using the principal component feature data of these frequency bands, the correlation between the frequency response and environmental parameters (such as temperature, humidity, noise, etc.) is analyzed through machine learning methods (such as random forest or support vector machine) to identify the laws of frequency response changes under different environmental conditions. In this way, the obtained frequency feature matrix can accurately reflect the influence of environmental factors on the frequency response of the audio signal, providing a reliable theoretical support for the next noise calibration and dynamic compensation.

[0083] Step S3: Obtain environmental noise monitoring data, map the environmental noise monitoring data to the frequency feature matrix to obtain a noise signal mapping matrix; perform frequency band environmental noise calibration on the noise signal mapping matrix to obtain a noise calibration matrix;

[0084] In this embodiment, environmental noise data is collected in real time through an environmental noise monitoring device (such as an environmental noise sensor) built in the microphone. The collected data is converted into a frequency-domain signal through FFT to obtain the noise intensity of each frequency band. Then, using the obtained frequency feature matrix, the environmental noise data is mapped into the frequency feature matrix to form a noise signal mapping matrix. In order to eliminate the interference of environmental noise on the microphone output signal, a frequency band-based noise calibration algorithm (such as a Kalman filter) is used to calibrate the noise signal mapping matrix. Through noise calibration, a noise calibration matrix is obtained, which can effectively reduce the influence of noise on the audio signal and lay a foundation for the optimization of subsequent compensation strategies.

[0085] Step S4: Perform an adaptive dynamic compensation strategy analysis based on the frequency feature matrix and the noise calibration matrix to obtain an initial adaptive dynamic compensation strategy; perform Monte Carlo multi-dimensional parameter optimization on the initial adaptive dynamic compensation strategy according to the coupled frequency response data to obtain an optimized compensation strategy, and transmit the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy;

[0086] In this embodiment, based on the obtained frequency feature matrix and noise calibration matrix, an adaptive filtering algorithm is used for analysis to generate an initial adaptive dynamic compensation strategy. This strategy mainly automatically adjusts the microphone gain, frequency response, and noise suppression parameters according to environmental parameters and frequency characteristics to optimize the audio quality. To further improve the compensation effect, the Monte Carlo simulation method is used to perform multi-dimensional optimization on each parameter of the initial compensation strategy, simulate the compensation effect in different environments, and adjust the compensation strategy. After optimization, the obtained compensation strategy is sent to the microphone device controller by wireless transmission and executed in the controller. This optimized compensation strategy can adjust the working state of the microphone according to real-time environmental conditions to ensure that the audio signal is always clear and stable.

[0087] Step S5: Obtain the real-time microphone output signal data, and evaluate the compensation effect of the real-time microphone output signal data and the optimized compensation strategy to obtain real-time audio compensation score data; perform dynamic compensation feedback on the optimized compensation strategy based on the real-time audio compensation score data to obtain dynamic compensation feedback data;

[0088] In this embodiment, the obtained microphone output signal data is subjected to frequency domain analysis through a digital signal processing (DSP) module, combined with the optimized compensation strategy, and the compensation effect is evaluated. The compensation effect is evaluated by calculating the error between the optimized audio signal and the uncompensated audio signal to obtain real-time audio compensation score data. This score data is analyzed through a machine learning algorithm (such as a neural network) to provide real-time feedback on the effectiveness of the optimized compensation strategy. If the score data is lower than a preset threshold, the system will automatically adjust the compensation strategy to optimize the signal processing process. The dynamic compensation feedback data further adjusts the compensation strategy through an adaptive algorithm to ensure continuous improvement of the compensation effect of the real-time audio signal.

[0089] Step S6: Perform global compensation model modeling according to the dynamic compensation feedback data to obtain a microphone global compensation model; perform user personalized compensation strategy selection based on the microphone global compensation model to obtain a user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

[0090] In this embodiment, a global compensation model is constructed based on dynamic compensation feedback data. This model uses the regression analysis method, comprehensively considering frequency characteristics, environmental factors, noise conditions, and real-time compensation feedback data, to generate a global compensation model. Through this model, personalized compensation strategies can be automatically selected according to different user requirements. The personalized compensation strategy is adjusted according to the specific needs of the user (such as preferences for specific frequency band audio, different usage environments), such as enhancing the high-frequency part, reducing low-frequency noise, etc. Finally, the personalized compensation strategy is transmitted to the microphone device controller to ensure that the device executes the compensation strategy most suitable for the user, thereby providing the best audio experience.

[0091] Optionally, step S1 is specifically as follows:

[0092] Step S11: Obtain the microphone output signal data and multi-dimensional environmental data, and perform data preprocessing on the microphone output signal data and multi-dimensional environmental data respectively, so as to obtain the microphone output signal data to be analyzed and the multi-dimensional environmental data to be analyzed;

[0093] In this embodiment, the audio output signal data is obtained from the microphone device, and this data is collected and converted into a digital signal by an analog-to-digital converter (ADC). At the same time, environmental data is collected in real time through multiple environmental sensors (such as temperature and humidity sensors, barometric pressure sensors, noise sensors, etc.) built into the microphone device. These sensor data include multi-dimensional data such as environmental temperature, humidity, and noise level. Then, preprocessing is performed on these two types of data: the microphone output signal data undergoes denoising and filtering to remove high-frequency noise and unnecessary interference to ensure the validity of the signal; while the multi-dimensional environmental data fills in missing values through interpolation, and performs unit conversion and standardization to ensure that each environmental parameter has the same dimension and consistency. The preprocessed data is respectively marked as the microphone output signal data to be analyzed and the environmental data to be analyzed, and is ready to enter the subsequent analysis process.

[0094] Step S12: Perform Fourier transform on the microphone output signal data to be analyzed to obtain the microphone output signal spectrum, and perform frequency domain data standardization on the microphone output signal spectrum to obtain the microphone output signal standardized spectrum;

[0095] In this embodiment, the Fourier transform (FFT) is performed on the microphone output signal data to be analyzed, converting the signal from the time domain to the frequency domain. The Fourier transform can decompose the time-domain waveform of the audio signal into different frequency components, obtaining the spectrum of the microphone output signal. After obtaining the spectrum data, further normalization processing is performed on the spectrum data. Normalization includes normalizing the spectrum data to ensure that the signal intensity within the frequency range is on a unified scale, removing the amplitude differences brought by different devices or recording environments, thereby improving the stability and consistency of data processing. After normalization, the obtained spectrum data is the normalized spectrum of the microphone output signal for subsequent feature extraction and analysis.

[0096] Step S13: Integrate the frequency response characteristics of the normalized spectrum of the microphone output signal to obtain the microphone frequency response characteristic data, and perform a frequency response curve conversion based on the microphone frequency response characteristic data to obtain the microphone frequency response curve;

[0097] In this embodiment, based on the obtained normalized spectrum of the microphone output signal, it is analyzed by means of frequency response characteristic integration. By weighted synthesis of the signals in different frequency bands, the frequency response characteristic data representing the microphone is extracted. At this time, the responses in different frequency bands are integrated according to the design characteristics of the microphone (such as directivity, frequency response curve, etc.), so as to reflect the response characteristics of the microphone under different environmental conditions. Then, a frequency response curve conversion is performed based on the extracted frequency response characteristic data. This conversion process uses a mathematical model (such as linear regression or neural network) to generate the frequency response curve of the microphone according to the frequency response characteristic data. The frequency response curve describes the sensitivity of the microphone to different frequency signals, providing a basis for subsequent optimization and compensation strategies.

[0098] Step S14: Perform multi-dimensional environmental data fusion on the multi-dimensional environmental data to be analyzed to obtain the environmental fusion data to be analyzed;

[0099] In this embodiment, the multi-dimensional environmental data to be analyzed is subjected to fusion processing. First, the environmental data collected by multiple sensors (such as temperature, humidity, air pressure, noise, etc.) is subjected to dimensionality reduction and fusion through weighted average or principal component analysis (PCA). In this way, data from different sources and of different types can be integrated into a unified environmental fusion data. Specifically, if the environmental noise data has a high correlation, it can be weighted and fused with the temperature and humidity data to reduce the volatility brought by a single data source. At the same time, the environmental data fusion method also considers the accuracy and reliability of each sensor to ensure the representativeness and accuracy of the final fused data. The fused data is marked as the environmental fusion data to be analyzed, providing a basis for subsequent feature integration.

[0100] Step S15: Integrate the multi-level features of the frequency response based on the microphone frequency response curve and the environment fusion data to be analyzed, so as to obtain the coupled frequency response data.

[0101] In this embodiment, based on the microphone frequency response curve and the environment fusion data to be analyzed, the multi-level features of the frequency response are integrated. Through the multi-level feature fusion method, the frequency response characteristics of the microphone and the environmental data are combined to extract representative features. This process uses ensemble learning algorithms such as random forest or XGBoost to fuse the frequency response characteristics of the microphone and the environmental parameters to capture the complex relationship between the audio signal and the environmental factors. After integration, the obtained coupled frequency response data contains the comprehensive response characteristics of the microphone under different environmental conditions, providing data support for subsequent compensation and optimization processing.

[0102] Optionally, step S15 is specifically as follows:

[0103] Step S151: Align the frequency features of the microphone frequency response curve and the environment fusion data to be analyzed, so as to obtain the frequency-aligned response curve and the frequency-aligned environment fusion data;

[0104] In this embodiment, the microphone frequency response curve is the frequency-domain signal obtained through Fourier transform and filtering. Relevant parameters such as temperature, humidity, noise level, etc. are extracted from the multi-dimensional environmental data collected by the environmental sensor and mapped to the frequency range corresponding to the microphone frequency response curve. Through the alignment method of the time axis and the frequency axis (such as interpolation or time-frequency transformation), the frequency features of these two types of data are aligned to ensure that each environmental data point corresponds to a specific frequency band. After alignment, the generated frequency-aligned response curve and frequency-aligned environment fusion data will provide a unified frequency reference for subsequent analysis.

[0105] Step S152: Perform a correlation analysis on the frequency-aligned response curve and the frequency-aligned environment fusion data to obtain the microphone frequency-environment correlation data, and perform feature selection on the frequency-aligned environment fusion data according to the microphone frequency-environment correlation data to obtain the frequency-related environmental data;

[0106] In this embodiment, by calculating the correlation between the frequency-aligned response curve and the frequency-aligned environmental fusion data, and using statistical methods such as Pearson correlation coefficient or mutual information, the impact of environmental factors (such as temperature and humidity changes, background noise, etc.) on the microphone frequency response is evaluated. For example, when the background noise is large, the high-frequency response of the microphone may be suppressed, while the low-frequency response changes less. According to the calculated correlation data, the environmental data features that have a greater impact on the microphone frequency response are selected. For example, only temperature and humidity are selected as the influencing factors, and the factors that contribute less to the frequency response are removed. The finally obtained frequency-related environmental data will contain environmental factors that are closely related to the frequency response.

[0107] Step S153: Perform multi-dimensional feature vector splicing on the frequency-aligned response curve and the frequency-related environmental data to obtain frequency response-environmental feature composite data;

[0108] In this embodiment, the frequency-aligned response curve and the frequency-related environmental data are subjected to feature vector splicing. The specific operation is to combine the response data (such as amplitude, phase, etc.) at each frequency point with the corresponding environmental data (such as temperature, humidity, noise level, etc.) to form a composite feature vector. For example, for a specific frequency band (such as 1 kHz to 3 kHz), the microphone response data in this frequency band is spliced with the temperature and humidity data in this frequency band period to form a new feature vector. This process can be achieved by point-by-point splicing or by weighted splicing of different frequency bands to form composite data containing environmental factors and frequency response information. These composite data will provide rich information for the next step of environmental impact analysis.

[0109] Step S154: Perform multi-frequency response hierarchical analysis on the frequency response-environmental feature composite data to obtain frequency response hierarchical environmental association data;

[0110] In this embodiment, the multi-frequency response hierarchical analysis method (such as hierarchical clustering analysis, principal component analysis, etc.) is used to process the frequency response-environmental feature composite data. The specific approach is to hierarchically divide the frequency response data and the environmental data by different frequency bands and analyze the response characteristics of each frequency band under different environmental conditions. For example, the response in the low-frequency band (20 Hz to 200 Hz) may have a greater relationship with humidity changes, while the middle-frequency band (200 Hz to 2 kHz) may be more affected by temperature changes. By performing hierarchical analysis on these frequency bands, it is possible to identify which frequency bands are more significantly affected under different environments, thereby obtaining frequency response hierarchical environmental association data. These data can help further understand the specific associations between the microphone responses of different frequency bands and environmental factors.

[0111] Step S155: Perform frequency response hierarchical relationship mapping and coupling on the frequency response-environment feature composite data based on the frequency response hierarchical environment correlation data, so as to obtain coupled frequency response data.

[0112] In this embodiment, based on the obtained frequency response hierarchical environment correlation data, frequency response hierarchical relationship mapping and coupling are performed. Specifically, by constructing a mapping model, the hierarchical correlation relationship between frequency response features and environmental factors is mapped into a unified coupling model. The mapping model adopts methods such as regression models, support vector machines (SVMs), or neural networks. Through this model, the environmental factors of each data point in the frequency response-environment feature composite data are mapped onto the corresponding frequency response data, thereby obtaining coupled frequency response data. At this time, the coupled frequency response data contains the influence of environmental factors on the frequency response of each frequency band, and can effectively reflect how environmental changes affect the frequency response performance of the microphone.

[0113] Optionally, step S2 is specifically as follows:

[0114] Step S21: Perform frequency band division on the coupled frequency response data, so as to obtain frequency band division coupled frequency response data;

[0115] In this embodiment, complete frequency response information is extracted from the coupled frequency response data, and then according to the required frequency band division criteria (such as low frequency band: 20 Hz - 200 Hz, middle frequency band: 200 Hz - 2 kHz, high frequency band: 2 kHz - 20 kHz), these data are divided into different frequency bands. The response data in each frequency band is processed separately to ensure that the characteristics of each frequency band can be analyzed independently. For example, the low frequency band may be mainly affected by temperature and humidity, while the high frequency band is more affected by air pressure and noise. Through this frequency band division, subsequent analysis can be carried out for each frequency band respectively, providing data support for frequency band principal component feature analysis.

[0116] Step S22: Perform frequency band principal component feature analysis on the frequency band division coupled frequency response data, so as to obtain frequency band principal component feature data;

[0117] In this embodiment, the principal component analysis (PCA) technique is applied to the coupled frequency response data of each frequency band to extract the principal component feature data. In this process, for the response data of the low frequency band, middle frequency band, and high frequency band, covariance matrix calculation and eigenvalue decomposition are respectively performed to obtain the principal component features of each frequency band. Through PCA analysis, the feature dimensions that have the greatest influence on the frequency response change can be identified. For example, the principal component of a certain frequency band may reveal that this frequency band is particularly sensitive to humidity changes, while another frequency band may be more affected by temperature changes. The principal component feature data of each frequency band will provide core information for the subsequent calculation of frequency response influencing factors.

[0118] Step S23: Calculate the frequency response impact factors of the frequency bands based on the frequency band principal component feature data, so as to obtain the frequency response impact factors of the frequency bands;

[0119] In this embodiment, based on the frequency band principal component feature data, regression analysis or multivariate statistical methods are used to calculate the response impact factors of each frequency band on environmental factors (such as temperature, humidity, air pressure, etc.). Taking temperature as an example, assuming that the principal component feature data of the low frequency band shows the trend of frequency response changing with temperature, a linear regression model can be used to quantify the impact of temperature change on the frequency response of the frequency band. Specifically, the environmental parameters can be set as independent variables, and the principal component features of the frequency band as dependent variables, and the influence degree of each environmental parameter on each frequency band can be calculated. For example, the calculated temperature impact factor of the low frequency band is 0.35, and the humidity impact factor is 0.22, which means that the temperature change has a greater impact on the frequency response of the low frequency band. Through this method, clear impact factor data can be provided for subsequent environmental impact quantification.

[0120] Step S24: Quantify the impact of environmental parameters according to the frequency response impact factors of the frequency bands, so as to obtain the environmental parameter impact degree data;

[0121] In this embodiment, the impact of each environmental parameter on the microphone frequency response is quantified by analyzing the calculated frequency response impact factors of the frequency bands. For example, through statistical methods (such as standard deviation analysis, regression model, etc.), the relationship between each frequency band frequency response impact factor and the environmental parameter is quantified, and the impact degree of each environmental parameter on the frequency response is obtained. For example, if the humidity change has a greater impact on the low frequency band impact factor, then the impact degree of humidity on the low frequency band may reach 60%, while the impact degree of temperature is 40%. These impact degree data help to understand the priority of environmental factors and how to optimize the frequency response by adjusting parameters.

[0122] Step S25: Perform frequency response - environmental parameter association on the environmental parameter impact degree data and the frequency band principal component feature data, so as to obtain a frequency response - environment association matrix, and construct a frequency feature matrix according to the frequency response - environment association matrix.

[0123] In this embodiment, based on the environmental parameter influence degree data and the frequency band principal component feature data, a frequency response - environment correlation matrix is constructed through correlation analysis or weighted calculation. Each row represents the response of a frequency band, and each column represents an environmental parameter. Through this matrix, it can be clearly seen how each frequency band is affected by different environmental factors. For example, the low - frequency band is very sensitive to humidity and temperature changes, while the high - frequency band is more affected by noise. A frequency feature matrix is constructed based on the frequency response - environment correlation matrix, which integrates the relationship between the frequency band and the environmental parameters, facilitating subsequent frequency optimization, compensation, and other processing. The correlation matrix is used to weight the response of each frequency band to obtain the frequency feature matrix. For example, for a certain frequency band, its frequency feature can be expressed as a weighted sum of factors such as environmental temperature and humidity. The specific implementation method can be completed through matrix multiplication.

[0124] Optionally, step S3 is specifically as follows:

[0125] Step S31: Obtain environmental noise monitoring data and perform data pre - processing on the environmental noise monitoring data to obtain the environmental noise monitoring data to be analyzed;

[0126] In this embodiment, the noise signal in the environment is collected in real - time through the noise monitoring device built into the microphone and is usually output in the form of sound intensity, frequency, etc. The obtained data often contains various noise sources such as background noise, traffic noise, and industrial noise. In the pre - processing stage, first, these raw data need to be filtered to remove high - frequency or low - frequency interference noise. Common methods include low - pass filtering, high - pass filtering, band - pass filtering, etc. The specific filter selection depends on the frequency range of the noise to be analyzed. For example, in an industrial environment, the noise in the range of 50 - 5000 Hz may be of concern, so a band - pass filter needs to be used to filter out irrelevant frequency bands. The pre - processing may also include operations such as noise reduction, outlier removal, and missing data supplementation to ensure that the data to be analyzed has high quality.

[0127] Step S32: Perform Fourier transform on the environmental noise monitoring data to be analyzed to obtain the environmental noise signal spectrum, and perform statistics on the energy distribution of the environmental noise signal spectrum in frequency bands to obtain the energy distribution diagram of the environmental noise signal in frequency bands;

[0128] In this embodiment, the environmental noise signal is transformed into frequency-domain data through Fourier transform. Specifically, first, the processed time-domain noise signal is subjected to a fast Fourier transform (FFT) to convert it from the time domain to the frequency domain, so as to obtain the energy distribution of the noise signal at each frequency. This process can utilize existing mathematical tools, such as numpy.fft in Python or the fft function in MATLAB. The spectral data obtained after Fourier transform contains the energy of the environmental noise signal in different frequency bands. Then, the energy proportion of each frequency band is statistically calculated to form an energy distribution diagram of the noise signal frequency band. For example, the spectrogram can show the energy distribution in the low-frequency (20 Hz to 500 Hz), medium-frequency (500 Hz to 2000 Hz), and high-frequency (2000 Hz to 10000 Hz) regions. In order to have a clear understanding of the frequency distribution of the noise signal, the spectrogram can be divided into several frequency bands, and the energy value of each frequency band can be calculated. This graph can help analyze the contribution of specific frequency bands to the noise and which frequency bands have a greater impact on the environment.

[0129] Step S33: Integrate the frequency distribution of the noise signal according to the energy distribution diagram of the noise signal frequency band, so as to obtain a noise signal frequency distribution matrix;

[0130] In this embodiment, by analyzing the energy distribution diagram of the noise signal frequency band, the energy distribution of the frequency band is integrated into a comprehensive noise signal frequency distribution matrix. The specific method is to allocate the energy of each frequency band to the corresponding position in the matrix. For example, the rows of the matrix represent different time periods or noise monitoring positions, and the columns represent different frequency bands. Each element in the matrix represents the noise energy of a certain frequency band at a certain moment (or a certain measurement point). In this way, a multi-dimensional noise frequency distribution matrix can be formed, which can not only reflect the changes in the time series but also show the noise energy distribution under different positions or different environmental conditions.

[0131] Step S34: Perform time series mapping on the noise signal frequency distribution matrix and the frequency feature matrix to obtain an initial signal mapping matrix, and perform frequency-axis feature alignment on the initial signal mapping matrix to obtain a noise signal mapping matrix;

[0132] In this embodiment, a time-series mapping is performed on the noise signal frequency distribution matrix and the frequency feature matrix. Time-series mapping refers to aligning the time-series data of the noise frequency distribution matrix with the data of the frequency feature matrix according to time. The purpose of doing this is to associate the noise data from different time periods with the corresponding frequency response characteristics. For example, assume that the frequency feature matrix reflects the response of the microphone to environmental noise in different frequency bands, while the noise signal frequency distribution matrix contains the noise energy distribution data of each frequency band at a certain moment. During mapping, the noise signal spectrum at each moment is mapped to the corresponding frequency band data of the frequency feature matrix to form a preliminary signal mapping matrix. Next, frequency-axis feature alignment is performed to ensure the alignment of the frequency response data and the noise data. Even if the frequency range of the noise spectrum does not exactly coincide with the frequency bands of the microphone frequency response, they can be aligned through methods such as interpolation, weighting, or normalization to make them match on the frequency axis.

[0133] Step S35: Calibrate the noise signal mapping matrix for the environmental noise in each frequency band to obtain a noise calibration matrix.

[0134] In this embodiment, the noise signal mapping matrix will be calibrated so that the frequency response of the noise signal is consistent with the noise level in the real environment. The purpose of calibration is to adjust the values in the noise signal mapping matrix so that it reflects the actual noise situation in the environment, rather than just the original measurement data. The calibration method can include using a known standard noise source to adjust the values of the mapping matrix. For example, in the laboratory, a white noise source with known intensity and frequency can be used to calibrate the measured noise data. By comparing with the standard noise source, each element in the noise signal mapping matrix is corrected to make it closer to the actual noise level in the environment. The influence of factors such as the temperature and humidity of the environment can also be considered, and by establishing a compensation model, the noise signal is further corrected. The finally obtained noise calibration matrix can more accurately reflect the frequency characteristics of the environmental noise and provide a basis for subsequent noise suppression or optimization.

[0135] Optionally, step S35 is specifically:

[0136] Step S351: Divide the frequency data of the noise signal mapping matrix into frequency bands to obtain the noise signal band spectrum;

[0137] In this embodiment, the frequency data in the noise signal mapping matrix first needs to be divided into frequency bands to better analyze the noise characteristics of different frequency bands. According to the frequency range of the environmental noise signal and the preset frequency band division strategy, the entire frequency range is divided into multiple sub-frequency bands. For example, the range from 20 Hz to 1000 Hz may be divided into a low-frequency band (20 Hz - 250 Hz), a mid-low frequency band (250 Hz - 500 Hz), a mid-frequency band (500 Hz - 750 Hz), and a high-frequency band (750 Hz - 1000 Hz). Each frequency band contains different frequency information, enabling more refined analysis of the noise signals in different frequency bands. The frequency data of each frequency band is extracted from the noise signal mapping matrix, and the corresponding noise signal frequency band spectrum is generated. Each frequency band spectrum reflects the noise intensity distribution within that frequency band. For spectrum analysis, the fast Fourier transform (FFT) can be used to perform frequency domain conversion on the noise signal, and then the energy distribution within each frequency band is statistically analyzed to obtain the frequency band spectrum. For complex environments, the frequency band division needs to be dynamically adjusted to cope with the changes of different noise sources.

[0138] Step S352: Perform environmental condition - frequency band noise signal time series correlation based on the noise signal frequency distribution matrix and the noise signal frequency band spectrum, so as to obtain environmental condition - frequency band noise signal change data;

[0139] In this embodiment, the time series data in the noise signal frequency distribution matrix is combined with the frequency band spectrum, and the frequency characteristics of the noise signal are organized in chronological order. For example, assume that spectrum data is obtained during different time periods of noise monitoring, and this spectrum data needs to be matched and analyzed with the environmental conditions at that time (such as temperature, humidity, wind speed, etc.). By calculating the changes in the noise frequency bands at different time points and performing time series correlation in combination with environmental factors (such as temperature changes, wind speed, etc.), the change trend of the noise signal under different environmental conditions can be revealed. The time series correlation process can adopt the sliding window analysis method, select the frequency band data within a certain time period, and perform correlation analysis in combination with the environmental conditions at that time, so as to obtain environmental condition - frequency band noise signal change data. This data reflects the impact of environmental changes on the noise signals in different frequency bands and the change characteristics of the noise signals under different environmental conditions.

[0140] Step S353: Calculate the environmental noise impact factor based on the environmental condition - frequency band noise signal change data, so as to obtain the environmental noise impact factor;

[0141] In this embodiment, an environmental noise impact factor is calculated based on the environmental condition - frequency band noise signal change data. This impact factor describes the degree of influence of environmental conditions on noise signals in different frequency bands. Regression analysis or correlation analysis can be performed on the relationship between environmental conditions (such as temperature, humidity, air pressure, etc.) and the change of noise signals. Through statistical analysis, the influence coefficient of environmental factor changes on the frequency band change of noise signals is calculated. For example, when the temperature rises, the low - frequency noise signal will increase due to the change of the air propagation speed, and when the humidity is high, the absorption of medium - and high - frequency noise is enhanced. In this case, methods such as multiple linear regression models or support vector machines (SVM) can be used to establish a quantitative relationship model between environmental conditions and the frequency bands of noise signals. Finally, these relationships are converted into an environmental noise impact factor, which represents the influence intensity of environmental factors on the noise in a specific frequency band.

[0142] Step S354: Calculate the frequency band calibration coefficient for the frequency band spectrum of the noise signal according to the environmental noise impact factor, so as to obtain the frequency band calibration coefficient;

[0143] In this embodiment, the frequency band spectrum of the noise signal is calibrated based on the environmental noise impact factor. According to the calculated environmental noise impact factor, the calibration factor for each frequency band is first determined. These factors indicate how the noise signals in each frequency band should be adjusted under different environmental conditions. For example, a certain low - frequency band is strongly affected by temperature changes and needs to be multiplied by a relatively large calibration factor, while another high - frequency band is less affected by humidity changes and the calibration factor is relatively small. The calculation method of the calibration coefficient can adopt the weighted average method, the least - squares method or regression analysis based on neural networks. The specific approach is: for each frequency band, the relationship between the environmental noise impact factor and the frequency band noise signal is used to estimate the calibration factor, and this process requires composite calculations of multiple environmental variables. For example, within the temperature change range, the calibration factor can represent the increment of the noise intensity in a certain frequency band when the temperature rises, and the change of humidity corresponds to another type of calibration factor.

[0144] Step S355: Calibrate the noise signal frequency band of the noise signal mapping matrix according to the frequency band calibration coefficient, so as to obtain the noise calibration matrix.

[0145] In this embodiment, the obtained frequency band calibration coefficients are used to calibrate the noise signal mapping matrix. Specifically, the noise signals of each frequency band are multiplied by the corresponding frequency band calibration coefficients to correct the amplitudes of the noise signals, so as to more accurately reflect the influence of environmental conditions on the noise signals. Assuming that the noise signal mapping matrix contains the noise data of each moment and frequency band, by applying the calibration coefficients to the data of each frequency band, the noise values of each time point and frequency band can be adjusted. For example, if the calibration coefficient of a certain frequency band is 1.2, indicating that the noise signal intensity of this frequency band should be increased by 20%, then the corresponding value of this frequency band in the noise signal mapping matrix is multiplied by 1.2. In this way, the influence of the noise signal under different environmental conditions is effectively compensated and corrected, and finally a noise calibration matrix is obtained. This calibration matrix can be used for subsequent noise optimization, compensation and other processes to ensure the accurate expression and adjustment of environmental noise.

[0146] Optionally, step S4 is specifically as follows:

[0147] Step S41: Compare the frequency feature matrix and the noise calibration matrix to obtain frequency response difference data, and integrate the noise influence factors according to the frequency response difference data to obtain the noise influence factors;

[0148] In this embodiment, the data in the frequency feature matrix and the noise calibration matrix are compared, especially the eigenvalues of each frequency band and the calibrated noise signal intensities. The purpose of the comparison is to identify the differences in frequency response, that is, to compare the changes of the noise signal in different frequency bands. Based on the frequency band spectrum, the difference data between the same frequency bands in the frequency feature matrix and the noise calibration matrix are calculated to form the frequency response difference data. For example, if a certain frequency band shows a strong response in the frequency feature matrix but a weak response in the noise calibration matrix, then the difference data of this frequency band reflects the influence of the environmental noise on this frequency band. After obtaining the frequency response difference data, the noise influence factors are integrated according to these data. The noise influence factor is a quantitative description of how environmental noise affects the frequency response of the microphone device. Usually, the difference data of multiple frequency bands need to be weighted averaged or processed by other statistical methods to generate a comprehensive noise influence factor. This factor can be used for the subsequent construction of the noise compensation model.

[0149] Step S42: Construct an initial noise compensation model based on the frequency response difference data and the noise influence factors;

[0150] In this embodiment, an initial noise compensation model is constructed based on the calculation results of the frequency response difference data and the noise impact factor. The goal of this step is to establish a mathematical model that can automatically adjust the response characteristics of the microphone according to environmental conditions and frequency band noise differences. Common models can adopt regression analysis, neural networks, or machine learning algorithms. For example, first, determine the key frequency bands in the frequency response difference data, and then weight and adjust these frequency bands according to the noise impact factor to construct a preliminary noise compensation strategy. Through experiments, it is found that the noise in the low-frequency band has a greater impact, while the high-frequency band is relatively small. A polynomial regression model or a deep learning-based model can be constructed to learn the non-linear relationships of these frequency bands, so as to effectively predict the impact of noise on the microphone response and perform compensation. The preliminary compensation model also needs to be verified and adjusted through experimental data.

[0151] Step S43: Through the initial noise compensation model, and in combination with the microphone frequency response characteristic data, the frequency band calibration coefficient, and the data fused with the environment to be analyzed, design a compensation strategy, so as to obtain an initial adaptive dynamic compensation strategy;

[0152] In this embodiment, by combining the initial noise compensation model with the microphone frequency response characteristic data, the frequency band calibration coefficient, and the data fused with the environment to be analyzed, a preliminary compensation strategy is designed. The microphone frequency response characteristic data provides the response ability of the microphone in different frequency bands, while the frequency band calibration coefficient helps to adjust the signal strength of each frequency band. The data fused with the environment (such as temperature, humidity, etc.) provides the conditions for dynamic adjustment of the compensation strategy. Combining these data, through dynamic calculation by the initial compensation model, an adaptive compensation strategy is obtained. For example, in a high-temperature or high-humidity environment, the high-frequency response of the microphone will be suppressed, so the compensation strategy will increase the gain in the high-frequency band. At low temperatures, the response in the low-frequency band is stronger, and the compensation strategy will reduce the low-frequency gain. Through this strategy, the microphone can optimize its frequency response under various environmental conditions.

[0153] Step S44: Set compensation parameters according to the coupled frequency response data, so as to obtain a set of compensation parameters, and perform Monte Carlo random sampling on the set of compensation parameters, so as to obtain a set of random compensation parameters;

[0154] In this embodiment, the compensation parameter set needs to be set according to the coupled frequency response data. The coupled frequency response data reflects the actual influence of environmental conditions on the microphone frequency response, including the interaction between device and environmental factors. By analyzing these data, the initial compensation parameters in the compensation parameter set are determined, such as gain, attenuation coefficient, etc. The Monte Carlo method is used to randomly sample the compensation parameter set. Monte Carlo sampling is a statistical method used to randomly draw a set of parameters within a given parameter space and evaluate their effects. For example, different combinations of gain values and attenuation coefficients are used to simulate the response of the microphone in different environments. Through multiple random samplings, a series of potential compensation parameter combinations are generated, and these combinations represent the possible optimal compensation strategies under different conditions.

[0155] Step S45: Perform signal compensation simulation on the random compensation parameter set according to the initial adaptive dynamic compensation strategy, so as to obtain signal compensation simulation data, and evaluate the compensation signal error of the signal compensation simulation data, so as to obtain compensation signal error data;

[0156] In this embodiment, according to the initial adaptive dynamic compensation strategy, signal compensation simulation is performed on the random compensation parameter set. By applying different combinations of compensation parameters, the output signals of the microphone in different environments are simulated, and the compensated signals are calculated. Specifically, for each set of compensation parameters, a noise signal is generated in the simulated environment and corrected by the compensation model. During the simulation process, there will be an error between the compensated signal and the real signal. Therefore, it is necessary to evaluate the compensation signal error of each simulation, and error analysis methods such as mean square error (MSE) or signal distortion degree evaluation are used. By comparing the difference between the compensated signal and the ideal signal, the compensation signal error data is obtained. These data will be used to optimize the compensation parameters later.

[0157] Step S46: Select a compensation parameter group from the random compensation parameter set based on the compensation signal error data, so as to obtain an optimal compensation parameter group;

[0158] In this embodiment, the best compensation parameter group is selected according to the compensation signal error data. By analyzing the compensation signal error data, the compensation parameter combination with the smallest error is identified, and the optimal parameter set is selected. Sorting and screening algorithms can be used to select the compensation parameter combination with the smallest error. For example, an error threshold can be set, and for the compensation parameter combination below this threshold, it is considered as the optimal compensation group. At the same time, optimization algorithms such as genetic algorithms and particle swarm optimization (PSO) are used to further screen the compensation parameter set to improve the compensation effect. Finally, the compensation parameter group most suitable for the current environmental conditions is selected.

[0159] Step S47: Optimize the multi-dimensional parameters of the initial adaptive dynamic compensation strategy according to the optimal compensation parameter group, so as to obtain an optimized compensation strategy, and transmit the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy.

[0160] In this embodiment, based on the optimal compensation parameter group, multi-dimensional parameter optimization is performed on the initial adaptive dynamic compensation strategy. By further optimizing the compensation parameters, such as adjusting the gain, frequency range, etc., a more accurate compensation strategy is finally obtained. This optimized compensation strategy should adapt to various environmental changes and be able to be adjusted in real time. The optimized compensation strategy is finally transmitted to the microphone device controller for actual execution. Specifically, the compensation strategy can be transmitted to the controller through wireless communication or wired connection, and the compensation strategy is applied in real time during the operation of the device to ensure that the microphone always maintains the best audio capture effect in different environments.

[0161] Optionally, step S5 is specifically as follows:

[0162] Step S51: Obtain the real-time microphone output signal data, and perform data preprocessing on the real-time microphone output signal data to obtain the real-time microphone output signal data to be analyzed;

[0163] In this embodiment, the microphone device continuously captures the sound in the environment and converts it into an electrical signal. These signals need to be sampled, usually at a certain sampling rate (such as 44.1 kHz or 48 kHz). During the preprocessing process, operations such as denoising, removing DC offset, and gain adjustment are required for the audio signal. Specifically, a low-pass filter can be applied to remove high-frequency noise, mean filtering can be used to remove large mutation signals, and even time-domain or frequency-domain denoising algorithms (such as Wiener filtering) can be applied to improve the signal quality and remove unnecessary background noise. The preprocessed signal can be used for subsequent frequency analysis and compensation.

[0164] Step S52: Extract the frequency characteristics of the audio signal from the real-time microphone output signal data to be analyzed to obtain the real-time audio signal spectrum;

[0165] In this embodiment, frequency characteristic extraction is performed on the processed microphone output signal. The audio signal in the time domain is converted to the frequency domain through the fast Fourier transform (FFT). Specifically, during implementation, the preprocessed signal is first divided into multiple short-time frames, each frame of the signal is usually dozens of milliseconds to hundreds of milliseconds long, and a window function (such as a Hamming window) is applied to reduce spectral leakage. Then, the FFT is applied to each frame to extract the frequency components of each frequency band and obtain the spectrum data. For example, the FFT result of a certain frame may show the amplitudes of frequency components such as 1 kHz, 2 kHz, and 4 kHz. Through these spectrum data, the frequency distribution of the current microphone signal can be evaluated, providing a basis for subsequent compensation and optimization.

[0166] Step S53: Perform compensation strategy backtracking on the real-time audio signal spectrum according to the optimized compensation strategy, so as to obtain the real-time uncompensated audio signal spectrum;

[0167] In this embodiment, before applying the optimized compensation strategy retrospectively, the original spectrum of the real-time audio signal is evaluated. The optimized compensation strategy usually compensates by means of gain adjustment, attenuation, frequency correction, etc. for specific frequency bands. To evaluate the effect of the compensation, it is necessary to backtrack to the uncompensated original audio signal spectrum. By comparing the real-time audio signal spectrum, the state of the audio signal before compensation is identified. For example, if the goal of the optimized compensation strategy is to improve the low-frequency response, the backtracking process will show that the gain in the low-frequency band before compensation is insufficient. This backtracking step provides a comparison basis for subsequent effect evaluation.

[0168] Step S54: Evaluate the compensation effect based on the real-time audio signal spectrum and the real-time uncompensated audio signal spectrum, so as to obtain audio compensation effect evaluation data, and weight the evaluation results of the audio compensation effect evaluation data, so as to obtain real-time audio compensation score data;

[0169] In this embodiment, the effect of the compensation strategy is evaluated by comparing the differences between the real-time audio signal spectrum and the uncompensated spectrum. Specifically, it is quantified using indicators such as frequency response error, signal enhancement degree, signal distortion degree, etc. For example, if the compensation strategy is effective, the gain in the low-frequency and mid-frequency bands should increase, and the spectrum should be closer to the ideal form. The evaluation data can be obtained by calculating the mean square error (MSE) or signal enhancement ratio between the real-time spectrum and the uncompensated spectrum. In addition, according to the importance or target of different frequency bands, the evaluation results of different frequency bands can be weighted. For example, for voice signals, some key frequency bands (such as the mid-frequency band) within the frequency range may be given higher weights, and finally the comprehensive real-time audio compensation score data is obtained.

[0170] Step S55: Classify the real-time audio compensation score data according to a preset score threshold. If the real-time audio compensation score data is greater than or equal to the score threshold, send an instruction to the microphone device controller to continue executing the optimized compensation strategy; if the real-time audio compensation score data is less than the score threshold, mark the corresponding optimized compensation strategy as an abnormal compensation strategy, and send an instruction to the microphone device controller to stop executing the optimized compensation strategy;

[0171] In this embodiment, by comparing the real-time audio compensation score data with a preset score threshold, it is determined whether to continue executing the current optimization compensation strategy. If the real-time compensation score data is greater than or equal to the score threshold, indicating that the current compensation effect has reached the expected level, the compensation strategy is continued or optimized. For example, if the score threshold is set to 80 points and the real-time audio compensation score data is 85 points, it shows that the compensation strategy is effective, and the controller can continue to execute the optimization compensation strategy. On the contrary, if the compensation score data is lower than the score threshold, it indicates that the compensation strategy has not achieved the expected effect, which may lead to a decline in audio quality. The strategy will be marked as abnormal, and an instruction to stop executing the strategy will be sent to the microphone controller. In this way, it is ensured that only effective compensation strategies are applied, avoiding the influence of invalid or miscompensated strategies on the microphone performance.

[0172] Step S56: Adjust the abnormal compensation strategy in real time according to the real-time audio compensation score data to obtain dynamic compensation feedback data.

[0173] In this embodiment, if an abnormal compensation strategy is detected, by analyzing the real-time audio compensation score data, the strategy is adjusted to improve the compensation effect. For example, if the score data is too low, it may be that the compensation gain of a certain frequency band is too strong or too weak, resulting in audio distortion or noise enhancement. At this time, the strategy can be corrected by adjusting the frequency band gain, attenuation, or other compensation parameters. During the dynamic adjustment process, it is necessary to analyze the frequency characteristics of the microphone response and the current environmental factors (such as temperature and humidity changes) in real time, and adjust the compensation parameters according to the feedback. The adjusted strategy includes strengthening the compensation of certain frequency bands, reducing the gain of other frequency bands, or adjusting the bandwidth of the filter, etc. These adjustments will be fed back to the compensation strategy in real time, and through continuous optimization, ensure that the quality of the audio signal is continuously improved.

[0174] Optionally, step S6 is specifically as follows:

[0175] Step S61: Perform principal component feature selection on the dynamic compensation feedback data to obtain dynamic compensation feature data;

[0176] In this embodiment, the dynamic compensation feedback data is subjected to feature selection through principal component analysis (PCA). Principal component analysis is a commonly used dimensionality reduction technique, aiming to extract the most representative features from high-dimensional data. In actual operation, the dynamic compensation feedback data contains information such as compensation errors in multiple frequency bands and noise signal changes, and the dimension of these data is relatively high. Through PCA, the main change directions in the data can be identified, and these principal components are selected as new feature representations. For example, assume that the dynamic compensation feedback data contains the signal characteristics of 10 frequency bands. Through PCA analysis, 3 main components may be extracted, representing the core information of signal changes. These principal components will become the basis for subsequent modeling and analysis, helping to reduce the computational complexity and improve the model efficiency.

[0177] Step S62: Build a neural network global compensation model based on the dynamic compensation feature data, thereby obtaining a microphone global compensation model;

[0178] In this embodiment, the neural network model is trained using the dynamic compensation feature data to establish a global compensation model. First, an appropriate neural network architecture is selected, such as a multi-layer perceptron (MLP) or a convolutional neural network (CNN). The input of the network is the extracted dynamic compensation feature data, and the output is the corresponding compensation parameter or frequency band correction value. The training dataset includes different audio environments and compensation feedback data. After multiple trainings, the neural network learns how to output effective compensation parameters based on the input dynamic compensation feature data. For example, assume that through training, the network can learn how to adjust the gain or attenuation of the low frequency band under specific environmental noises, thereby optimizing the sound quality of the microphone. After the model training is completed, a global compensation model for compensating the microphone response is obtained.

[0179] Step S63: Obtain the user microphone interaction data, and perform user audio preference analysis on the user microphone interaction data, thereby obtaining user audio preference data;

[0180] In this embodiment, the audio preferences of the user are obtained and understood by analyzing the interaction between the user and the microphone. The user interaction data can be obtained through the microphone controller, which can include operation information such as volume adjustment, frequency response adjustment, and gain setting. For example, the user may be more inclined to increase the gain in the low frequency band or improve the speech clarity in a specific noise environment. After collecting this data, by analyzing the preference patterns of the user in different environments, the specific requirements of the user for audio characteristics are obtained. For example, if the user frequently adjusts the volume control to a specific value or modifies the high frequency band response multiple times, the analysis tool can identify that the user prefers clear and high frequency sounds. These data will be used as the input for the design of the personalized compensation strategy.

[0181] Step S64: Select a user personalized compensation strategy for the user audio preference data through the microphone global compensation model, thereby obtaining a user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

[0182] In this embodiment, based on the obtained user audio preference data, a personalized compensation strategy will be generated in combination with the global compensation model. The global compensation model has been trained according to different dynamic compensation characteristics and contains general compensation rules. On this basis, the user's audio preference data is used to perform personalized adjustment on the compensation model. For example, assuming that the user prefers a strong low-frequency response and clear high frequencies, the model can moderately increase the gain of the low-frequency band according to these preferences and optimize the clarity of the high frequencies. By adjusting the parameters of the global compensation model, a compensation strategy that meets the user's personalized needs is obtained. Finally, this compensation strategy will be transmitted to the microphone device controller, and the controller adjusts the frequency response of the microphone according to the strategy to ensure that the user's audio experience reaches the best state. For example, if the user prefers speech clarity, the frequency response of the microphone in the range of 1 kHz to 5 kHz can be enhanced to suppress unnecessary background noise, thereby improving the intelligibility and clarity of the speech.

[0183] Optionally, this specification also provides a system for compensating the frequency response of a microphone, which is used to execute the method for compensating the frequency response of a microphone as described above. The system for compensating the frequency response of a microphone includes:

[0184] A frequency response coupling module, configured to obtain microphone output signal data and multi-dimensional environment data, perform a frequency response curve conversion on the microphone output signal data to obtain a microphone frequency response curve; perform frequency response multi-level feature integration based on the microphone frequency response curve and the multi-dimensional environment data to obtain coupled frequency response data;

[0185] A principal component feature screening module, configured to perform band frequency principal component feature screening based on the coupled frequency response data to obtain band frequency principal component feature data, and perform frequency response-environment parameter correlation analysis based on the band frequency principal component feature data to obtain a frequency feature matrix;

[0186] An environmental noise calibration module, configured to obtain environmental noise monitoring data, map the environmental noise monitoring data to the frequency feature matrix to obtain a noise signal mapping matrix; perform band environmental noise calibration on the noise signal mapping matrix to obtain a noise calibration matrix;

[0187] A compensation strategy generation module, configured to perform adaptive dynamic compensation strategy analysis based on the frequency feature matrix and the noise calibration matrix to obtain an initial adaptive dynamic compensation strategy; perform Monte Carlo multi-dimensional parameter optimization on the initial adaptive dynamic compensation strategy according to the coupled frequency response data to obtain an optimized compensation strategy, and transmit the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy;

[0188] A compensation strategy feedback module, which is used to obtain real-time microphone output signal data, evaluate the compensation effect of the real-time microphone output signal data and the optimized compensation strategy, so as to obtain real-time audio compensation score data; and perform dynamic compensation feedback on the optimized compensation strategy based on the real-time audio compensation score data, so as to obtain dynamic compensation feedback data.

[0189] A personalized compensation strategy selection module, which is used to build a global compensation model according to the dynamic compensation feedback data, so as to obtain a microphone global compensation model; select a user personalized compensation strategy based on the microphone global compensation model, so as to obtain a user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

[0190] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be included in the present invention.

[0191] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for compensating the frequency response of a microphone, characterized in that The following steps are involved: Step S1: Acquire microphone output signal data and multi-dimensional environmental data, and perform frequency response curve conversion on the microphone output signal data to obtain a microphone frequency response curve; perform multi-level frequency response feature integration based on the microphone frequency response curve and the multi-dimensional environmental data to obtain coupled frequency response data; Step S2: performing frequency band frequency principal component feature screening based on the coupled frequency response data, thereby obtaining frequency band frequency principal component feature data, and performing frequency response-environmental parameter correlation analysis based on the frequency band frequency principal component feature data, thereby obtaining a frequency feature matrix; Step S3: Acquire environmental noise monitoring data, and map the environmental noise monitoring data to a frequency characteristic matrix, thereby obtaining a noise signal mapping matrix; Perform frequency band environmental noise calibration on the noise signal mapping matrix to obtain a noise calibration matrix; Step S4: performing adaptive dynamic compensation strategy analysis based on the frequency characteristic matrix and the noise calibration matrix, thereby obtaining an initial adaptive dynamic compensation strategy; Performing Monte Carlo multi-dimensional parameter optimization on the initial adaptive dynamic compensation strategy according to the coupling frequency response data to obtain an optimized compensation strategy, and transmitting the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy; Step S5: acquiring real-time microphone output signal data, and evaluating the compensation effect of the real-time microphone output signal data and the optimized compensation strategy, thereby obtaining real-time audio compensation score data; performing dynamic compensation feedback on the optimized compensation strategy based on the real-time audio compensation score data, thereby obtaining dynamic compensation feedback data; Step S6: building a global compensation model according to the dynamic compensation feedback data, thereby obtaining a microphone global compensation model; A user personalized compensation strategy is selected based on the microphone global compensation model to obtain the user microphone personalized compensation strategy, and the strategy is transmitted to the microphone device controller to execute the user microphone personalized compensation strategy.

2. The method for compensating the frequency response of a microphone according to claim 1, characterized in that: Step S1 is specifically as follows: Step S11: Acquire microphone output signal data and multi-dimensional environmental data, and perform data preprocessing on the microphone output signal data and the multi-dimensional environmental data respectively, so as to obtain microphone output signal data to be analyzed and multi-dimensional environmental data to be analyzed; Step S12: performing Fourier transform on the microphone output signal data to be analyzed to obtain the microphone output signal spectrum, and performing frequency domain data standardization on the microphone output signal spectrum to obtain the microphone output signal standardized spectrum; Step S13: integrating the frequency response characteristics of the standardized frequency spectrum of the microphone output signal to obtain microphone frequency response characteristic data, and performing frequency response curve conversion based on the microphone frequency response characteristic data to obtain a microphone frequency response curve; Step S14: performing multi-dimensional environmental data fusion on the multi-dimensional environmental data to be analyzed, thereby obtaining environmental fusion data to be analyzed; Step S15: performing multi-level feature integration of frequency response according to the microphone frequency response curve and the environmental fusion data to be analyzed, thereby obtaining coupled frequency response data.

3. The method for compensating the frequency response of a microphone according to claim 2, characterized in that: Step S15 is specifically as follows: Step S151: performing frequency feature alignment on the microphone frequency response curve and the environment fusion data to be analyzed, so as to obtain a frequency-aligned response curve and frequency-aligned environment fusion data; Step S152: performing correlation analysis on the frequency alignment response curve and the frequency alignment environment fusion data to obtain microphone frequency-environment correlation data, and performing feature selection on the frequency alignment environment fusion data according to the microphone frequency-environment correlation data to obtain frequency-correlated environment data; Step S153: performing multi-dimensional feature vector splicing on the frequency alignment response curve and the frequency-related environmental data, so as to obtain frequency response-environmental feature composite data; Step S154: performing environmental impact multi-frequency response hierarchical analysis based on the frequency response-environmental feature composite data, thereby obtaining frequency response hierarchical environmental correlation data; Step S155: performing frequency response hierarchical relationship mapping coupling on the frequency response-environmental feature composite data based on the frequency response hierarchical environment association data, thereby obtaining coupled frequency response data.

4. The method for compensating the frequency response of a microphone according to claim 1, characterized in that: Step S2 is specifically as follows: Step S21: dividing the coupling frequency response data into frequency bands, thereby obtaining frequency band-divided coupling frequency response data; Step S22: performing frequency band principal component characteristic analysis on the frequency band division coupled frequency response data, thereby obtaining frequency band principal component characteristic data; Step S23: calculating the frequency band frequency response influence factor based on the frequency band principal component characteristic data, thereby obtaining the frequency band frequency response influence factor; Step S24: quantifying the environmental parameter impact according to the frequency band frequency response impact factor, thereby obtaining environmental parameter impact degree data; Step S25: Perform frequency response-environment parameter association on the environmental parameter influence data and the frequency band principal component characteristic data, so as to obtain a frequency response-environment association matrix, and construct a frequency characteristic matrix according to the frequency response-environment association matrix.

5. The method for compensating the frequency response of a microphone according to claim 1, characterized in that: Step S3 is specifically as follows: Step S31: Acquire environmental noise monitoring data, and perform data preprocessing on the environmental noise monitoring data, so as to obtain environmental noise monitoring data to be analyzed; Step S32: performing Fourier transform of the environmental noise monitoring data to be analyzed, thereby obtaining an environmental noise signal spectrum, and performing frequency segment energy distribution statistics of the environmental noise signal spectrum, thereby obtaining a frequency segment energy distribution diagram of the noise signal; Step S33: integrating the frequency distribution of the noise signal according to the frequency segment energy distribution diagram of the noise signal, thereby obtaining a frequency distribution matrix of the noise signal; Step S34: performing time series mapping on the noise signal frequency distribution matrix and the frequency feature matrix to obtain an initial signal mapping matrix, and aligning the frequency axis features of the initial signal mapping matrix to obtain a noise signal mapping matrix; Step S35: performing frequency band environmental noise calibration on the noise signal mapping matrix, thereby obtaining a noise calibration matrix.

6. The method for compensating the frequency response of a microphone according to claim 5, characterized in that: Step S35 is specifically as follows: Step S351: dividing the noise signal mapping matrix into frequency bands, thereby obtaining a frequency band spectrum of the noise signal; Step S352: performing time series correlation of environmental condition-frequency band noise signal according to the noise signal frequency distribution matrix and the frequency band spectrum of the noise signal, thereby obtaining environmental condition-frequency band noise signal change data; Step S353: calculating the environmental noise impact factor based on the environmental condition-frequency band noise signal change data, thereby obtaining the environmental noise impact factor; Step S354: calculating the frequency band calibration coefficient of the frequency band spectrum of the noise signal according to the environmental noise impact factor, thereby obtaining the frequency band calibration coefficient; Step S355: performing noise signal frequency band calibration on the noise signal mapping matrix according to the frequency band calibration coefficient, thereby obtaining a noise calibration matrix.

7. The method for compensating the frequency response of a microphone according to claim 1, characterized in that: Step S4 is specifically as follows: Step S41: performing feature comparison on the frequency feature matrix and the noise calibration matrix to obtain frequency response difference data, and integrating the noise impact factor according to the frequency response difference data to obtain the noise impact factor; Step S42: constructing an initial noise compensation model based on the frequency response difference data and the noise impact factor; Step S43: designing a compensation strategy through an initial noise compensation model in combination with microphone frequency response characteristic data, frequency band calibration coefficients, and environmental fusion data to be analyzed, thereby obtaining an initial adaptive dynamic compensation strategy; Step S44: setting compensation parameters according to the coupling frequency response data to obtain a compensation parameter set, and performing Monte Carlo random sampling on the compensation parameter set to obtain a random compensation parameter set; Step S45: performing signal compensation simulation on the random compensation parameter set according to the initial adaptive dynamic compensation strategy to obtain signal compensation simulation data, and performing compensation signal error evaluation on the signal compensation simulation data to obtain compensation signal error data; Step S46: selecting a compensation parameter group for the random compensation parameter set based on the compensation signal error data, thereby obtaining an optimal compensation parameter group; Step S47: Optimizing the compensation strategy multi-dimensional parameters of the initial adaptive dynamic compensation strategy according to the optimal compensation parameter group, thereby obtaining an optimized compensation strategy, and transmitting the optimized compensation strategy to the microphone device controller to execute the optimized compensation strategy.

8. The method for compensating the frequency response of a microphone according to claim 1, characterized in that: Step S5 is specifically as follows: Step S51: acquiring real-time microphone output signal data, and performing data preprocessing on the real-time microphone output signal data, thereby obtaining real-time microphone output signal data to be analyzed; Step S52: extracting audio signal frequency features from the real-time microphone output signal data to be analyzed, thereby obtaining a real-time audio signal spectrum; Step S53: performing compensation strategy backtracking on the real-time audio signal spectrum according to the optimized compensation strategy, thereby obtaining the real-time uncompensated audio signal spectrum; Step S54: performing compensation effect evaluation based on the real-time audio signal spectrum and the real-time uncompensated audio signal spectrum to obtain audio compensation effect evaluation data, and weighting the evaluation results of the audio compensation effect evaluation data to obtain real-time audio compensation score data; Step S55: classifying the real-time audio compensation score data according to a preset score threshold, and if the real-time audio compensation score data is greater than or equal to the score threshold, sending an instruction to the microphone device controller to continue to execute the optimization compensation strategy; if the real-time audio compensation score data is less than the score threshold, marking the corresponding optimization compensation strategy as an abnormal compensation strategy, and sending an instruction to the microphone device controller to stop executing the optimization compensation strategy; Step S56: adjusting the abnormal compensation strategy in real time according to the real-time audio compensation score data, thereby obtaining dynamic compensation feedback data.

9. The method for compensating the frequency response of a microphone according to claim 1, characterized in that: Step S6 is specifically as follows: Step S61: performing principal component feature selection on the dynamic compensation feedback data to obtain dynamic compensation feature data; Step S62: Building a neural network global compensation model according to the dynamic compensation feature data, thereby obtaining a microphone global compensation model; Step S63: acquiring user microphone interaction data, and performing user audio preference analysis on the user microphone interaction data, thereby obtaining user audio preference data; Step S64: selecting a user personalized compensation strategy for the user audio preference data through the microphone global compensation model, thereby obtaining a user microphone personalized compensation strategy, and transmitting it to the microphone device controller to execute the user microphone personalized compensation strategy.

10. A system for compensating the frequency response of a microphone, characterized in that For executing the method for compensating the frequency response of a microphone as claimed in claim 1, the system for compensating the frequency response of a microphone comprises: A frequency response coupling module is used to obtain microphone output signal data and multi-dimensional environmental data, and perform frequency response curve conversion on the microphone output signal data to obtain a microphone frequency response curve; perform multi-level feature integration of frequency response according to the microphone frequency response curve and multi-dimensional environmental data to obtain coupled frequency response data; A principal component feature screening module is used to screen the frequency band frequency principal component features based on the coupled frequency response data, thereby obtaining the frequency band frequency principal component feature data, and to perform frequency response-environmental parameter correlation analysis based on the frequency band frequency principal component feature data, thereby obtaining a frequency feature matrix; The environmental noise calibration module is used to obtain environmental noise monitoring data and map the environmental noise monitoring data to a frequency characteristic matrix to obtain a noise signal mapping matrix; perform frequency band environmental noise calibration on the noise signal mapping matrix to obtain a noise calibration matrix; A compensation strategy generation module is used to perform adaptive dynamic compensation strategy analysis based on a frequency characteristic matrix and a noise calibration matrix to obtain an initial adaptive dynamic compensation strategy; perform Monte Carlo multi-dimensional parameter optimization on the initial adaptive dynamic compensation strategy according to coupled frequency response data to obtain an optimized compensation strategy, and transmit the optimized compensation strategy to a microphone device controller to execute the optimized compensation strategy; A compensation strategy feedback module is used to obtain real-time microphone output signal data, and evaluate the compensation effect of the real-time microphone output signal data and the optimized compensation strategy, so as to obtain real-time audio compensation score data; based on the real-time audio compensation score data, dynamic compensation feedback is performed on the optimized compensation strategy, so as to obtain dynamic compensation feedback data; The personalized compensation strategy selection module is used to model a global compensation model based on dynamic compensation feedback data to obtain a microphone global compensation model; select a user personalized compensation strategy based on the microphone global compensation model to obtain a user microphone personalized compensation strategy, and transmit it to the microphone device controller to execute the user microphone personalized compensation strategy.

Images