Intelligent noise reduction processing method for wireless microphone chip integration

Abstract

The invention discloses an intelligent noise reduction processing method for wireless microphone chip integration, which relates to the technical field of intelligent noise reduction, is used for solving the problems that noise in a quiet recording environment is difficult to collect, a denoising algorithm is single and the like, and comprises the steps of performing matching identification on microphone models, confirming a connection range and a sound source collection range of the microphone, and performing noise reduction. And integrating the connection range of the microphone and the sound source acquisition range, and calculating a sound acquisition position by using a K-nearest neighbor algorithm. The method comprises the following steps: determining a real-time position of a user, collecting a sound source, obtaining microphone output voltage, calculating microphone sensitivity and frequency response, comparing recording adaptation sensitivity of a noise file by using a logistic regression algorithm according to the microphone sensitivity and the frequency response, and selecting the noise file for playing. And acquiring sound source data of the noise file through a microphone, calculating a noise file preceding SNR value at a sound collection position, comparing the noise file preceding SNR value with the SNR stored in the recording database, analyzing environmental noise, selecting different algorithms according to the environmental noise to carry out de-noising processing, and storing the data.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent noise reduction, and more specifically, to an intelligent noise reduction processing method integrated with a wireless microphone chip. Background Art

[0002] Intelligent noise reduction technology uses advanced algorithms and models to automatically identify and reduce unnecessary noise in audio or video signals. Applying intelligent noise reduction technology to chip-integrated wireless microphones can effectively reduce various types of noise in wireless microphones.

[0003] The prior art has the following deficiencies:

[0004] In the past, when performing noise reduction processing on wireless microphones, a set denoising algorithm was used to uniformly process the audio source signal. However, the ambient noise needed to be detected in the environment, and then processed according to the detection results during denoising. When the detection environment is quiet enough, the ambient noise is difficult to detect, and the denoising accuracy of the unified denoising algorithm is low. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent noise reduction processing method integrated with a wireless microphone chip, which calculates the sound collection position by analyzing the microphone performance, compares the file noise of the measured sound collection position with the file noise in the database, analyzes the degree of environmental noise, and then selects different algorithms for denoising to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent noise reduction processing method for integrating a wireless microphone chip comprises the following steps:

[0008] Step S1: Connect the microphone to the host computer, the host computer identifies the microphone model, confirms the connection range of the microphone and the sound source collection range, and uses the K-nearest neighbor algorithm to calculate the sound collection position based on the connection range of the microphone and the sound source collection range;

[0009] Step S2: Determine the user's real-time location, prompt the user to perform initial sound source acquisition, obtain the microphone output voltage to calculate the microphone sensitivity and frequency response; use a logistic regression algorithm based on the microphone sensitivity and frequency response to select a pre-recorded noise file in the recording database and provide it to the user;

[0010] Step S3: Play the noise file at the sound collection location, obtain the noise file at the sound collection location through the microphone to determine the collected signal power, and calculate the SNR value through the noise file signal power recorded in the recording database to determine the ambient noise;

[0011] Step S4: according to the degree of environmental noise, a time domain denoising algorithm or an adaptive filtering algorithm is selected to denoise the environmental noise of the file collected by the microphone.

[0012] In a preferred embodiment, in step S1, the wireless microphone is connected to the host computer, and the host computer matches the microphone model with various microphone models stored in the software database, queries the best sound collection position of the connected microphone and marks it. If the connected microphone does not record the best sound collection position, the connection range and sound source collection range of the connected microphone are obtained;

[0013] The software database stores microphone models and basic information about the corresponding models, including connection range, sound source collection range, and optimal sound collection position.

[0014] In a preferred embodiment, in step S1, the connection range is the farthest distance at which the microphone can be wirelessly connected to the host computer, the sound source collection range is the farthest distance at which the microphone can collect the sound of the sound source position, and the optimal sound collection position is the stable collection distance between the sound source and the microphone obtained through testing and stored in the software database together with the microphone model.

[0015] In a preferred embodiment, in step S1, the microphone models storing the best sound collection positions in the software database are screened and marked, the connection range and the sound source collection range of the marked microphone are respectively merged into a connection range data set and a collection range data set, and the sound collection position is calculated using the K-nearest neighbor algorithm. The specific steps are as follows:

[0016] Define input and output: Set the connection range of the connected microphone and the sound source collection range as input features, and set the sound collection position of the connected microphone as output features;

[0017] Select parameter K: Set parameter K by taking the average value based on the connection range of the connected microphone and the sound source collection range and multiplying it by the preset reference ratio;

[0018] Calculate the distance: select the connection range and sound source collection range corresponding to the same model of marked microphones in the connection range data set and the collection range data set, and take the average value as the nearest neighbor distance;

[0019] Find K nearest neighbors: take the average of the two input features as the feature average, select the K labeled microphone models whose nearest neighbor distances are closest to the feature average, obtain the best sound collection position of the K nearest neighbor distances and select the average value as the candidate position corresponding to the K value;

[0020] Prediction output: Set multiple different K values ​​through a preset number of repetitions and obtain multiple alternative adoption positions, calculate their average value and output it as the sound collection position of the connected microphone.

[0021] In a preferred embodiment, in step S2, an application is made to the user end for mobile device connection, and the user's real-time position is obtained to determine whether the user is in the sound collection position. When the user adjusts the position to the sound collection position, the host computer sends a text prompt to the mobile device, and the microphone collects sound source data through the user reading the text prompt. The host computer calculates the microphone sensitivity and frequency response through the output voltage of the microphone.

[0022] In a preferred embodiment, in step S2, when calculating the microphone sensitivity, the host computer calculates the microphone sensitivity using a sensitivity formula: , where dB is the microphone sensitivity, is the microphone output voltage, For reference sound pressure, the reference sound pressure is set to a commonly used standard sound pressure.

[0023] Frequency response is the change in microphone sensitivity at different frequencies. Multiple different sound frequencies are obtained through text prompts, and the microphone sensitivities at different sound frequencies are calculated respectively. The obtained microphone sensitivities are subtracted from each other, and the calculation result obtained by the subtraction is used as the frequency response.

[0024] The calculated microphone sensitivity and frequency response are used to construct a logistic regression equation using the logistic regression algorithm to calculate the logistic regression coefficient: , where L is the logistic regression coefficient, e is the natural base, and D is the microphone sensitivity coefficient, which is obtained by summing the microphone sensitivity and frequency response.

[0025] In a preferred embodiment, in step S3, the SNR value is the signal-to-noise ratio transmitted from the microphone to the host computer. The larger the SNR value, the lower the ambient noise intensity. The noise file at the sound collection position obtained by the microphone determines that the collected signal power includes the power of the audio source and the power of the background noise. The noise file signal power recorded in the recording database includes the power of the same audio source and the power of the background noise.

[0026] The SNR value is calculated by subtracting the signal power obtained by the connected microphone from the signal power recorded and saved in advance.

[0027] In a preferred embodiment, in step S4, when the calculated SNR value exceeds the preset SNR threshold, it is judged that the environmental noise level is low, and the environmental noise is denoised using a time domain denoising algorithm; when the calculated SNR value is lower than the preset SNR threshold, it is judged that the environmental noise level is high, and the environmental noise is denoised using an adaptive filtering algorithm.

[0028] In a preferred embodiment, in step S4, the specific steps of denoising the ambient noise using a time domain denoising algorithm are as follows:

[0029] The signal is represented by the acquisition time as the horizontal axis as the analysis window, and the environmental noise is processed by the moving average filter formula: ,in is the value of the audio source signal after denoising the nth segment, N is the analysis window size, is the frequency band of the original audio source signal that has not been denoised. The value is exchanged with the center value of the analysis window during denoising to perform denoising;

[0030] The specific steps for denoising environmental noise using the adaptive filtering algorithm are as follows:

[0031] The normalized least mean square formula is selected to process the original audio source signal, and its calculation formula is: , the environmental noise is denoised by processing each time point of the acquisition time, where w(n) is the filter coefficient, is the step size factor, It is the tolerance error. It is the value of the original audio sound source signal. The filter coefficient of the next time point is adjusted by each original audio sound source signal value, thereby adjusting the original audio sound source signal value at each time point for denoising.

[0032] The technical effects and advantages of the intelligent noise reduction processing method integrated with a wireless microphone chip of the present invention are as follows:

[0033] The present invention connects a microphone and a host computer, matches and identifies the microphone model, confirms the connection range and sound source collection range of the microphone, and calculates the sound collection position based on the connection range and sound source collection range of the microphone. Determining the sound collection position according to the microphone performance improves the accuracy of the collected data, determines the real-time position of the user, collects the sound source, obtains the microphone output voltage to calculate the microphone sensitivity and frequency response, compares the recording adaptation sensitivity of the noise file according to the microphone sensitivity and frequency response, and selects the noise file for playback, obtains the noise file sound source data through the microphone, calculates the advance SNR value of the noise file at the sound collection position, and compares it with the SNR stored in the recording database, analyzes the environmental noise, and compares and calculates the SNR value, which is more suitable for a quiet environment than the traditional collection of environmental noise and calculation of the SNR value, selects different algorithms for denoising according to the environmental noise, and saves the data, so as to achieve high-quality denoising of the environmental noise in a specified environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A schematic diagram of an intelligent noise reduction processing method integrated with a wireless microphone chip according to the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] The present invention connects a microphone to a host computer, matches and identifies the microphone model, confirms the connection range and sound source collection range of the microphone, and calculates the sound collection position based on the connection range and sound source collection range of the microphone. The real-time position of the user is determined, sound source collection is performed, the microphone output voltage is obtained to calculate the microphone sensitivity and frequency response, the recording adaptation sensitivity of the noise file is compared with the microphone sensitivity and frequency response to select the noise file for playback, the noise file sound source data is obtained through the microphone, and the prior SNR value of the noise file at the sound collection position is calculated and compared with the SNR stored in the recording database, the environmental noise is analyzed, and different algorithms are selected according to the environmental noise for denoising and data is saved, so as to achieve high-quality denoising of the environmental noise in the specified environment.

[0037] Embodiment, a wireless microphone chip integrated intelligent noise reduction processing method, such as Figure 1 As shown, the following steps are included:

[0038] Step S1: Connect the microphone to the host computer, the host computer identifies the microphone model, confirms the connection range of the microphone and the sound source collection range, and uses the K-nearest neighbor algorithm to calculate the sound collection position based on the connection range of the microphone and the sound source collection range;

[0039] Step S2: Determine the user's real-time location, prompt the user to perform initial sound source acquisition, obtain the microphone output voltage to calculate the microphone sensitivity and frequency response; use a logistic regression algorithm based on the microphone sensitivity and frequency response to select a pre-recorded noise file in the recording database and provide it to the user;

[0040] Step S3: Play the noise file at the sound collection location, obtain the noise file at the sound collection location through the microphone to determine the collected signal power, and calculate the SNR value through the noise file signal power recorded in the recording database to determine the ambient noise;

[0041] Step S4: according to the degree of environmental noise, a time domain denoising algorithm or an adaptive filtering algorithm is selected to denoise the environmental noise of the file collected by the microphone.

[0042] The specific steps are as follows:

[0043] In step S1, the wireless microphone is connected to the host computer. The host computer matches the microphone model with various microphone models stored in the software database, queries the best sound collection position of the connected microphone and marks it. If the connected microphone does not record the best sound collection position, the connection range and sound source collection range of the connected microphone are obtained. The connection range is the farthest distance within which the microphone can be wirelessly connected to the host computer, and the sound source collection range is the farthest distance within which the microphone can collect the sound of the sound source. The best sound collection position is the distance between the sound source and the microphone obtained through testing and stored in the software database together with the microphone model.

[0044] If the various microphone models in the software database fail to match the connected microphone model, a microphone data collection pop-up window will pop up to the user, allowing the user to manually enter the connection range and sound source collection range of the connected microphone, and the connected microphone model, connection range and sound source collection range will be recorded in the software database.

[0045] The microphone models with the best sound collection positions stored in the software database are screened and marked. The connection range and sound source collection range of the marked microphones are merged into the connection range data set and the collection range data set respectively. The K-nearest neighbor algorithm is used to infer the sound collection position. The specific steps are as follows:

[0046] Define input and output: set the connection range of the connected microphone and the sound source collection range as input features, and set the sound collection position of the connected microphone as output features; it should be explained that both the input features and the output features are de-unitized data.

[0047] Select parameter K: Set a parameter K, which can be obtained by taking the average value of the connection range of the connected microphone and the sound source collection range and multiplying it by the preset reference ratio;

[0048] Calculate the distance: select the connection range and sound source collection range corresponding to the same model of marked microphones in the connection range data set and the collection range data set, and take the average value as the nearest neighbor distance;

[0049] Find K nearest neighbors: take the average of the two input features as the feature average, select the K labeled microphone models whose nearest neighbor distances are closest to the feature average, obtain the best sound collection position of the K nearest neighbor distances and select the average value as the candidate position corresponding to the K value;

[0050] Prediction output: Set multiple different K values ​​through a preset number of repetitions and obtain multiple alternative adoption positions, calculate their average value and output it as the sound collection position of the connected microphone.

[0051] Mark the sound collection position of the connected microphone and enter it into the software database together with the model of the connected microphone, the connection range and the sound source collection range. The sound collection position of the connected microphone is preferentially calculated through the K-nearest neighbor algorithm, so that more detailed sound source data can be obtained to facilitate subsequent testing and analysis of environmental noise.

[0052] It should be noted that the software database involved in this example is a database that stores microphone models and basic information. It is mainly used to match and identify the acquired connected microphone models with the microphone models in the database. The K-nearest neighbor algorithm is a supervised learning algorithm that can be used for classification and regression tasks. The L-nearest neighbor algorithm can quickly calculate the output based on the data provided by the connected microphone as input features and the comparison features in the software database. It is fast and convenient, and greatly saves the time cost of calculating the sound collection position. The value of parameter K and the number of repetitions of setting the K value in the predicted output can be set according to the user's required accuracy requirements, and the setting method is not unique.

[0053] In step S2, after calculating the sound collection position of the connected microphone, an application for mobile device connection is made to the user end. The user is required to connect the mobile device he is using to the host computer to obtain the user's real-time position, that is, the user's positioning, which is used to determine whether the user is in the sound collection position.

[0054] After the user connects the mobile device to the host computer, a prompt window for the initial sound source collection is sent to the mobile device and the distance between the user and the host computer is detected. The prompt window will display the real-time distance between the user and the host computer and prompt the user to adjust the distance. When the user adjusts the distance so that the distance between the user and the host computer is the sound collection position, the host computer will send a text prompt to the mobile device, requiring the user to read the content in the text prompt at the sound collection position, collect it through the microphone and transmit it to the host computer as sound source data. The host computer calculates the microphone sensitivity and frequency response through the output voltage of the microphone.

[0055] The host computer calculates the microphone sensitivity through the sensitivity formula: The sensitivity is calculated using the microphone output voltage and reference sound pressure. The formula is: , where dB is the microphone sensitivity, is the microphone output voltage, For reference sound pressure, the reference sound pressure can be set to a commonly used standard sound pressure.

[0056] The frequency response is the change in microphone sensitivity at different frequencies. When the user reads out the content in the text prompt, multiple different sound frequencies are obtained, and the microphone sensitivities at different sound frequencies are calculated respectively. The obtained microphone sensitivities are then subtracted pairwise, and the calculation result obtained by the difference is used as the frequency response.

[0057] The calculated microphone sensitivity and frequency response are used to construct a logistic regression equation using a logistic regression algorithm to calculate the logistic regression coefficient. The calculation formula for the logistic regression method is: , where L is the logistic regression coefficient, e is the natural base, and D is the microphone sensitivity coefficient, which is obtained by summing the microphone sensitivity and frequency response. The larger the microphone sensitivity or frequency response, the larger the microphone sensitivity coefficient, the larger the logistic regression coefficient, and the more sensitive the connected microphone.

[0058] The noise file is a recording recorded in advance. The sensitivity coefficient and logistic regression coefficient of each noise file are calculated through the sensitivity and frequency response supported by each noise file and stored in the recording database. The microphone sensitivity coefficient and logistic regression coefficient of the connected microphone are calculated by connecting the sound source data obtained from the microphone and compared with the logistic regression coefficient in the recording database. When the logistic regression coefficient of the noise file in the recording database is higher, it means that the corresponding noise file has higher sensitivity requirements for the microphone. The noise file with a logistic regression coefficient smaller than the maximum value of the logistic regression coefficient calculated by the connected microphone is selected in the recording database, and the corresponding noise file is provided to the user's mobile device.

[0059] It should be noted that the recording database is a collection of various audio recordings for audio processing, speech recognition and analysis. The recorded noise files can be stored in the recording database.

[0060] In step S3, the signal power of the noise file at the sound sampling position is obtained by connecting the microphone, and compared with the signal power of the corresponding noise file in the recording database, and the SNR value is calculated to determine the ambient noise of the noise reduction environment. The SNR value is the signal-to-noise ratio transmitted to the host computer by the connected microphone, which is used to reflect the size of the signal strength and the noise strength. The larger the SNR value, the lower the ambient noise intensity.

[0061] The normal signal-to-noise ratio calculation process requires recording the power of the ambient sound source without human intervention as the power of the background noise, and then obtaining the power of the audio sound source with human intervention as the power of the target signal. However, when the environment is too quiet, the power detection accuracy of the ambient sound source is low, resulting in analysis errors of the ambient noise.

[0062] Calculating the SNR value by subtracting the signal power obtained by the connected microphone from the signal power recorded and saved in advance can more accurately reflect the difference in environmental noise after switching.

[0063] The signal power recorded and saved in advance includes the sum of the power of the audio source and the power of the background noise as the overall target signal power. The signal power obtained by connecting the microphone also includes the power of the same audio source and the power of the background noise.

[0064] It should be noted that when recording a noise file, the recording environment must be quiet and the ambient noise must be low. The noise file not only contains the pre-recorded recording, but also includes the signal power of the noise file during recording.

[0065] In step S4, the calculated SNR value is compared with the preset SNR threshold to determine the degree of environmental noise. When the calculated SNR value exceeds the preset SNR threshold, it is judged that the degree of environmental noise is low, and the time domain denoising algorithm is used to denoise the environmental noise; when the calculated SNR value is lower than the preset SNR threshold, it is judged that the degree of environmental noise is high, and the adaptive filtering algorithm is used to denoise the environmental noise.

[0066] The time domain denoising algorithm performs denoising on a small range separately and then merges it to denoise the ambient noise. It is simple and quick and suitable for environments with low noise levels. The adaptive filtering algorithm performs denoising on each time point and merges it to denoise the ambient noise. It has higher denoising accuracy and is suitable for environments with high noise levels.

[0067] The specific steps for denoising environmental noise using the time domain denoising algorithm are as follows:

[0068] The signal is represented by the acquisition time as the horizontal axis. A short period of time is selected as the analysis window, and the environmental noise is processed by the moving average filter formula: ,in is the value of the audio source signal after denoising the nth segment, N is the analysis window size, is the frequency band of the original audio source signal that has not been denoised. The value is exchanged with the center value of the analysis window during denoising to smooth the audio source signal, and the denoising of the ambient noise is completed by merging the frequency bands of the denoised audio source signal.

[0069] The specific steps for denoising environmental noise using the adaptive filtering algorithm are as follows:

[0070] The normalized least mean square formula is selected to process the original audio source signal, and its calculation formula is: , the environmental noise is denoised by processing each time point of the acquisition time, where w(n) is the filter coefficient, is the step size factor, It is the tolerance error. It is the value of the original audio source signal. The filter coefficient of the next time point is adjusted by each original audio source signal value, so that convergence processing is performed on each time point. When the audio source is traversed, the environmental noise denoising is completed.

[0071] It should be noted that the SNR threshold can be set by averaging the SNR values ​​of all noise files in the recording database. The average filtering formula is only one of the time domain denoising algorithms, and the normalized least mean square formula is one of the adaptive filtering algorithms. This example is only used as an example. The filter coefficients, step size factors, and tolerance errors involved in the normalized least mean square formula are set by professionals in this field according to actual conditions and will not be analyzed here.

[0072] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0073] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0074] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0075] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0076] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent noise reduction processing method for wireless microphone chip integration, characterized in that: The following steps are included: Step S1: Connect the microphone to the host computer, the host computer identifies the microphone model, confirms the connection range of the microphone and the sound source collection range, and uses the K-nearest neighbor algorithm to calculate the sound collection position based on the connection range of the microphone and the sound source collection range; Step S2: Determine the user's real-time location, prompt the user to perform initial sound source acquisition, obtain the microphone output voltage to calculate the microphone sensitivity and frequency response; use a logistic regression algorithm based on the microphone sensitivity and frequency response to select a pre-recorded noise file in the recording database and provide it to the user; Step S3: Play the noise file at the sound collection location, obtain the noise file at the sound collection location through the microphone to determine the collected signal power, and calculate the SNR value through the noise file signal power recorded in the recording database to determine the ambient noise; Step S4: Select a time domain denoising algorithm or an adaptive filtering algorithm according to the degree of environmental noise to denoise the environmental noise of the file collected by the microphone.

2. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 1, characterized in that: In step S1, the wireless microphone is connected to the host computer, and the host computer matches the microphone model with various microphone models stored in the software database, queries the best sound collection position of the connected microphone and marks it. If the connected microphone does not record the best sound collection position, the connection range and sound source collection range of the connected microphone are obtained; The software database stores microphone models and basic information about the corresponding models, including connection range, sound source collection range, and optimal sound collection position.

3. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 2, characterized in that: In step S1, the connection range is the farthest distance at which the microphone can be wirelessly connected to the host computer, the sound source collection range is the farthest distance at which the microphone can collect the sound of the sound source, and the optimal sound collection position is the stable collection distance between the sound source and the microphone obtained through testing and stored in the software database along with the microphone model.

4. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 3 is characterized in that: In step S1, the microphone models with the best sound collection positions stored in the software database are screened and marked, the connection range and sound source collection range of the marked microphones are merged into a connection range data set and a collection range data set respectively, and the sound collection position is calculated using the K-nearest neighbor algorithm. The specific steps are as follows: Define input and output: Set the connection range of the connected microphone and the sound source collection range as input features, and set the sound collection position of the connected microphone as output features; Select parameter K: Set parameter K by taking the average value based on the connection range of the connected microphone and the sound source collection range and multiplying it by the preset reference ratio; Calculate the distance: select the connection range and sound source collection range corresponding to the same model of marked microphones in the connection range data set and the collection range data set, and take the average value as the nearest neighbor distance; Find K nearest neighbors: take the average of the two input features as the feature average, select the K labeled microphone models whose nearest neighbor distances are closest to the feature average, obtain the best sound collection position of the K nearest neighbor distances and select the average value as the candidate position corresponding to the K value; Prediction output: Set multiple different K values ​​through a preset number of repetitions and obtain multiple alternative adoption positions, calculate their average value and output it as the sound collection position of the connected microphone.

5. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 1, characterized in that: In step S2, an application is made to the user end for mobile device connection, and the user's real-time position is obtained to determine whether the user is in the sound collection position. When the user adjusts the position to the sound collection position, the host computer sends a text prompt to the mobile device, and the microphone collects sound source data through the user reading the text prompt. The host computer calculates the microphone sensitivity and frequency response through the output voltage of the microphone.

6. The intelligent noise reduction processing method for wireless microphone chip integration according to any one of claim 5, characterized in that: In step S2, when calculating the microphone sensitivity, the host computer calculates the microphone sensitivity using the sensitivity formula: , where dB is the microphone sensitivity, is the microphone output voltage, is the reference sound pressure, which is set to a commonly used standard sound pressure; Frequency response is the change in microphone sensitivity at different frequencies. Multiple different sound frequencies are obtained through text prompts, and the microphone sensitivities at different sound frequencies are calculated respectively. The obtained microphone sensitivities are subtracted from each other, and the calculation result obtained by the subtraction is used as the frequency response. The calculated microphone sensitivity and frequency response are used to construct a logistic regression equation using the logistic regression algorithm to calculate the logistic regression coefficient: , where L is the logistic regression coefficient, e is the natural base, and D is the microphone sensitivity coefficient, which is obtained by summing the microphone sensitivity and frequency response.

7. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 1, characterized in that: In step S3, the SNR value is the signal-to-noise ratio transmitted from the microphone to the host computer. The larger the SNR value, the lower the ambient noise intensity. The noise file of the sound collection position obtained by the microphone determines that the collected signal power includes the power of the audio source and the power of the background noise. The signal power of the noise file recorded in the recording database includes the power of the same audio source and the power of the background noise. The SNR value is calculated by subtracting the signal power obtained by the connected microphone from the signal power recorded and saved in advance.

8. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 1, characterized in that: In step S4, when the calculated SNR value exceeds the preset SNR threshold, it is judged that the environmental noise level is low, and the environmental noise is denoised using a time domain denoising algorithm; when the calculated SNR value is lower than the preset SNR threshold, it is judged that the environmental noise level is high, and the environmental noise is denoised using an adaptive filtering algorithm.

9. The intelligent noise reduction processing method for wireless microphone chip integration according to claim 8, characterized in that: In step S4, the environmental noise is denoised using a time domain denoising algorithm. The specific steps are as follows: The signal is represented by the acquisition time as the horizontal axis as the analysis window, and the environmental noise is processed by the moving average filter formula: ,in is the value of the audio source signal after denoising the nth segment, N is the analysis window size, is the frequency band of the original audio source signal that has not been denoised. The value is exchanged with the center value of the analysis window during denoising to perform denoising; The specific steps for denoising environmental noise using the adaptive filtering algorithm are as follows: The normalized least mean square formula is selected to process the original audio source signal, and its calculation formula is: , the environmental noise is denoised by processing each time point of the acquisition time, where w(n) is the filter coefficient, is the step size factor, It is the tolerance error. It is the value of the original audio sound source signal. The filter coefficient of the next time point is adjusted by each original audio sound source signal value, thereby adjusting the original audio sound source signal value at each time point for denoising.

Images