Micro-electro-mechanical loudspeaker system and control method

Abstract

The invention provides a micro-electro-mechanical loudspeaker system and a control method, and the method comprises the steps: determining the recognition sensitivity of each audio signal segment through micro-electro-mechanical audio information, and determining a frequency spectrum equalization relation according to all recognition sensitivities and the frequency band of an output signal of a frequency spectrum in each micro-electro-mechanical loudspeaker; performing masking sensing on each audio output signal in the micro-electro-mechanical audio information to obtain a masking interval of a frequency spectrum in the micro-electro-mechanical system, and determining a suppression gain of each micro-electro-mechanical loudspeaker when the target sound source is distorted according to an equilibrium relationship between the masking interval and each frequency spectrum; determining the audio distortion degree of the target sound source under the current sound wave phase through the phase shift delay information and all the spectrum equalization relations; and when the audio distortion degree is greater than a preset distortion threshold value, performing signal equalization distribution on each micro-electro-mechanical loudspeaker in the micro-electro-mechanical system according to each suppression gain. Based on the scheme, the distortion control capability of the micro-electro-mechanical loudspeaker system under different sound wave phases can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of micro-electromechanical speakers, and more specifically, to a micro-electromechanical speaker system and a control method. Background Art

[0002] A Micro-Electro-Mechanical Systems (MEMS) speaker is a speaker manufactured based on MEMS technology. MEMS speakers use micro-manufacturing technology to integrate mechanical structures and electronic components on a tiny chip. They usually have the advantages of small size, light weight, low power consumption, and fast response speed. They are widely used in mobile devices, headphones, hearing aids and other fields. The size of MEMS speakers is much smaller than traditional speakers, which can greatly save space and make the design more flexible, especially suitable for portable devices.

[0003] In a multi-speaker system, due to the different speaker installation positions and environmental factors, phase mismatch may cause sound interference and distortion, affecting the consistency of sound quality. The frequency response characteristics of different speakers may be inconsistent, resulting in insufficient spectrum balance, thereby making the sound quality performance uneven. The existing technology is difficult to adjust the speaker output in real time to adapt to environmental changes or dynamic changes in the sound source, making it difficult to maintain audio output consistency in different usage scenarios. In addition, the speaker adjustment process in the existing technology is usually complicated and requires manual settings, which makes it difficult to achieve fast and accurate optimization, limiting the performance of the multi-speaker system in actual applications. Automated adjustment technology is urgently needed to improve the overall consistency of the system. Therefore, how to achieve balanced signal distribution for each MEMS speaker in a MEMS system to ensure the consistency of audio output between each MEMS speaker is a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a micro-electromechanical speaker system and a control method, which can realize balanced signal distribution of each micro-electromechanical speaker in the micro-electromechanical system, thereby improving the distortion control capability of the micro-electromechanical speaker system under different sound wave phases and ensuring the consistency of audio output between each micro-electromechanical speaker.

[0005] In a first aspect, the present application provides a micro-electromechanical speaker control method, comprising: Using the audio sensor in the MEMS to monitor the audio input signal of each MEMS speaker when playing the target sound source, and then obtaining the MEMS audio information; Divide the target sound source into a plurality of audio signal segments by the power amplifier gradient of each MEMS speaker during frequency division output, and then determine the recognition sensitivity of each audio signal segment by the MEMS audio information, and determine the spectrum balance relationship of the signal distribution in each MEMS speaker according to all the recognition sensitivities and the output signal frequency band of the spectrum in each MEMS speaker; Based on the auditory masking effect of the neural network, each audio output signal in the micro-electromechanical system audio information is masked and sensed, thereby obtaining the masking interval of the spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and the equalization relationship of each spectrum; Acquire phase shift delay information of the sound wave phase between the micro-electromechanical speakers, and determine the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships; When the audio distortion is greater than a preset distortion threshold, signal equalization distribution is performed on each MEMS speaker in the MEMS system according to each suppression gain.

[0006] In some embodiments, dividing the target sound source into a plurality of audio signal segments by the power amplification gradient of each micro-electromechanical speaker during frequency division output specifically includes: Obtaining a received signal segment of a target sound source in each micro-electromechanical speaker; Determine the output stability of each MEMS speaker through each received signal segment and the power amplifier gradient of the corresponding MEMS speaker at the frequency division output; The received signal segments of the target sound source in each micro-electromechanical loudspeaker are re-divided according to all output stabilities to obtain a plurality of audio signal segments.

[0007] In some embodiments, determining the recognition sensitivity of each audio signal segment based on the micro-electromechanical audio information specifically includes: For each audio signal segment, obtaining all input signal segments of the audio signal segment in each audio input signal of the micro-electromechanical audio information; Determine the recognition delay of each audio recognition in the audio signal segment through each input signal segment; The recognition sensitivity of the audio signal segment is determined according to all the recognition delays, and then the recognition sensitivity of each audio signal segment is obtained.

[0008] In some embodiments, determining the spectrum balance relationship of signal distribution in each MEMS speaker according to all recognition sensitivities and the frequency bands of output signals of the spectrum in each MEMS speaker specifically includes: For each MEMS speaker, obtaining all audio signal segments included in the output signal frequency band of the frequency spectrum in the MEMS speaker; determining signal allocation information of the frequency spectrum in the MEMS speaker according to the recognition sensitivity of each audio signal segment; The spectrum balance relationship of the signal distribution in the micro-electro-mechanical speakers is determined by the signal distribution information, and then the spectrum balance relationship of the spectrum in each micro-electro-mechanical speaker is obtained.

[0009] In some embodiments, performing masking perception on each audio output signal in the micro-electromechanical system audio information based on the auditory masking effect of a neural network, and then obtaining the masking interval of the spectrum in the micro-electromechanical system specifically includes: Determine the interference information of the output frequency bands between the respective MEMS speakers through the respective audio output signals in the MEMS audio information; Based on the auditory masking effect of neural network, the spectrum masking evaluation of each interference information is carried out to obtain the critical frequency band of the target sound source in each MEMS speaker; Filtering out the spectrum masking range of each MEMS speaker in each audio output signal in the MEMS audio information according to each critical frequency band; The masking interval of the spectrum in the MEMS is determined by all the spectrum masking ranges.

[0010] In some embodiments, determining the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships specifically includes: Determine the phase shift interference amount of the current sound wave phase on the target sound source according to the phase shift delay information; Determine the frequency response distortion of the target sound source by the signal distribution under the current sound wave phase through all spectrum equalization relationships; The audio distortion degree of the target sound source at the current sound wave phase is determined based on the phase shift interference amount and the frequency response distortion.

[0011] In some embodiments, the signal distribution to each MEMS speaker according to each suppression gain is to use each suppression gain as a balance adjustment amount when performing signal distribution to the corresponding MEMS speaker, thereby completing the balance adjustment of the signal distribution.

[0012] In a second aspect, the present application provides a micro-electromechanical speaker system, comprising a control unit, wherein the control unit comprises: An acquisition module is used to use an audio sensor in a micro-electro-mechanical system to monitor the audio input signal of each micro-electro-mechanical speaker when playing a target sound source, thereby obtaining micro-electro-mechanical audio information; A processing module, used to divide the target sound source into a plurality of audio signal segments by the power amplifier gradient of each MEMS speaker during frequency division output, and then determine the recognition sensitivity of each audio signal segment by the MEMS audio information, and determine the spectrum balance relationship of the signal distribution in each MEMS speaker according to all the recognition sensitivities and the output signal frequency band of the spectrum in each MEMS speaker; The processing module is also used to perform masking perception on each audio output signal in the micro-electromechanical audio information based on the auditory masking effect of the neural network, thereby obtaining a masking interval of the spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and each spectrum balance relationship; The processing module is also used to obtain phase shift delay information of the sound wave phase between the micro-electromechanical speakers, and determine the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships; The execution module is used for performing signal balancing distribution to each MEMS speaker in the MEMS system according to each suppression gain when the audio distortion is greater than a preset distortion threshold.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the above-mentioned micro-electromechanical speaker control method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions or codes, and when the instructions or codes are run on a computer, the computer implements the above-mentioned micro-electromechanical speaker control method when executed.

[0015] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects: In a micro-electromechanical speaker system and control method provided by the present application, an audio sensor in the micro-electromechanical system is used to monitor the audio input signal of each micro-electromechanical speaker when playing a target sound source, thereby obtaining micro-electromechanical audio information; the target sound source is divided into multiple audio signal segments by the power amplifier gradient of each micro-electromechanical speaker during frequency division output, and the recognition sensitivity of each audio signal segment is determined by the micro-electromechanical audio information, and the spectrum balance relationship of the signal distribution in each micro-electromechanical speaker is determined according to all the recognition sensitivities and the output signal frequency band of the spectrum in each micro-electromechanical speaker; the auditory masking based on the neural network is used to generate a plurality of audio signal segments; the recognition sensitivity of each audio signal segment is determined according to the micro-electromechanical audio information; the recognition sensitivity of each audio signal segment is determined according to the micro-electromechanical audio information; the recognition sensitivity of each audio signal segment is determined according to all the recognition sensitivity and the output signal frequency band of the spectrum in each micro-electromechanical speaker; the recognition sensitivity of each audio signal segment is determined according to the micro-electromechanical audio information ... The masking effect is used to perform masking perception on each audio output signal in the micro-electromechanical audio information, thereby obtaining the masking interval of the frequency spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and each spectrum equalization relationship; obtaining the phase shift delay information of the sound wave phase between the micro-electromechanical speakers, and determining the audio distortion degree of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships; when the audio distortion degree is greater than a preset distortion threshold, performing signal equalization distribution on each micro-electromechanical speaker in the micro-electromechanical system according to each suppression gain.

[0016] It can be seen that in the present application, when the audio distortion is greater than the preset distortion threshold, the signal of each micro-electromechanical speaker in the micro-electromechanical system can be evenly distributed according to each suppression gain; firstly, by determining the recognition sensitivity of each audio signal segment, the signal segment that is easy to cause distortion can be processed preferentially, thereby improving the consistency of the audio signal between each micro-electromechanical speaker, and determining the spectrum balance relationship of the signal distribution in each micro-electromechanical speaker according to all the recognition sensitivities and the output signal bands of the spectrum in each micro-electromechanical speaker, thereby ensuring the coordinated work between the micro-electromechanical speakers, thereby improving the distortion control capability of the micro-electromechanical speaker system under different sound wave phases; secondly, by combining the masking interval and the spectrum balance relationship to determine the suppression gain, the output intensity of each micro-electromechanical speaker can be dynamically adjusted, so as to facilitate the subsequent adjustment when the target sound source is distorted. In real time, effective gain suppression measures are taken to reduce audio distortion caused by excessive or weak output; finally, the influence of the current sound wave phase on the audio signal is judged by monitoring the phase shift delay between the MEMS speakers. Since sound waves at different phases may cause different types of distortion, the audio distortion at the current sound wave phase can be accurately determined in combination with the phase shift delay. When the distortion exceeds the threshold, the signal of each MEMS speaker is balanced and distributed, which can more effectively control the quality of the audio output and ensure the signal consistency between different MEMS speakers. In summary, based on the above scheme, the signal balance distribution of each MEMS speaker in the MEMS system can be achieved, thereby improving the distortion control capability of the MEMS speaker system at different sound wave phases and ensuring the consistency of the audio output between each MEMS speaker. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0018] Figure 1 is an exemplary flow chart of a micro-electromechanical speaker control method according to some embodiments of the present application; Figure 2 is a schematic diagram of a process for determining a masking interval according to some embodiments of the present application; Figure 3 is a schematic diagram of a process for determining audio distortion according to some embodiments of the present application; Figure 4 is a schematic diagram of the structure of a control unit according to some embodiments of the present application; Figure 5 It is a structural schematic diagram of a computer device for implementing a micro-electromechanical speaker control method according to some embodiments of the present application. DETAILED DESCRIPTION

[0019] The embodiment of the present application provides a micro-electromechanical speaker system and control method, the core of which is to determine the recognition sensitivity of each audio signal segment through micro-electromechanical audio information, determine the spectrum balance relationship according to all the recognition sensitivities and the output signal frequency bands of the spectrum in each micro-electromechanical speaker; perform masking perception on each audio output signal in the micro-electromechanical audio information, and then obtain the masking interval of the spectrum in the micro-electromechanical system, and determine the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and each spectrum balance relationship; determine the audio distortion of the target sound source at the current sound wave phase through phase shift delay information and all spectrum balance relationships; when the audio distortion is greater than the preset distortion threshold, perform signal balance distribution for each micro-electromechanical speaker in the micro-electromechanical system according to each suppression gain. Based on the above scheme, the distortion control capability of the micro-electromechanical speaker system at different sound wave phases can be improved.

[0020] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 , which is an exemplary flow chart of a micro-electromechanical speaker control method according to some embodiments of the present application, and the micro-electromechanical speaker control method mainly includes the following steps: In step 101, an audio sensor in a micro-electro-mechanical system is used to monitor an audio input signal of each micro-electro-mechanical speaker when playing a target sound source, thereby obtaining micro-electro-mechanical audio information.

[0021] It should be noted that, in the present application, microelectromechanical audio information refers to the data set of audio input signals of each microelectromechanical speaker, and the audio input signal refers to the spectrum change signal received by a microelectromechanical speaker when playing the target sound source; in the MEMS speaker, a high-precision MEMS microphone is deployed as an audio sensor. The MEMS microphone can capture sound wave signals in a wide frequency range, and the sensor should be located close to the output end of the speaker to ensure that it can accurately reflect the input signal of the speaker; in specific implementation, the audio sensor in the microelectromechanical system can be used to monitor the audio input signal of each microelectromechanical speaker when playing the target sound source at each preset acquisition interval (the default is 1s), and the collection of all audio input signals can be used as the microelectromechanical audio information.

[0022] In step 102, the target sound source is divided into a plurality of audio signal segments by the power amplifier gradient of each micro-electromechanical speaker at the time of frequency division output, and then the recognition sensitivity of each audio signal segment is determined by the micro-electromechanical audio information, and the spectrum balance relationship of the signal distribution in each micro-electromechanical speaker is determined according to all the recognition sensitivities and the output signal frequency band of the spectrum in each micro-electromechanical speaker.

[0023] It should be noted that, in the present application, the power amplifier gradient refers to the amplification factor of the microelectromechanical speaker at different frequencies. The more stable the power amplifier gradient is, the more stable the output of the microelectromechanical speaker in this frequency band is. In specific implementation, for each microelectromechanical speaker, the ratio of the input intensity of the microelectromechanical speaker under input signals of different frequencies in the historical playback records and the output intensity after passing through the power amplifier can be calculated as the power amplifier gradient value. The set of all power amplifier gradient values ​​can be used as the value range of the power amplifier gradient of the microelectromechanical speaker when the frequency is divided and output. The power amplifier gradient of each microelectromechanical speaker at the time of frequency division and output can be obtained in the above manner. Other methods can also be used to implement it in other embodiments, which are not limited here.

[0024] In some embodiments, dividing the target sound source into multiple audio signal segments by the power amplifier gradient of each MEMS speaker during frequency division output can be implemented by the following steps: Obtaining a received signal segment of a target sound source in each micro-electromechanical speaker; Determine the output stability of each MEMS speaker through each received signal segment and the power amplifier gradient of the corresponding MEMS speaker at the frequency division output; The received signal segments of the target sound source in each micro-electromechanical loudspeaker are re-divided according to all output stabilities to obtain a plurality of audio signal segments.

[0025] It should be noted that, in the present application, the audio signal segment refers to the range of the stable signal frequency band of each MEMS speaker. By redividing the signal segments, each MEMS speaker can maintain maximum stability and consistency when processing the assigned frequency band in a complex audio environment, which helps to reduce interference between MEMS speakers and optimize the overall audio output quality. The output stability is a measure of whether the output signal of the speaker at a specific frequency is stable. The higher the output stability, the smaller the volatility of the MEMS speaker output signal, and the more stable the playback sound quality of the target sound source.

[0026] In specific implementation, first, the receiving signal segment of the target sound source in each micro-electromechanical speaker can be obtained from the instruction manual of each micro-electromechanical speaker; then, for each micro-electromechanical speaker, all input signals with frequencies within the receiving signal segment of the micro-electromechanical speaker are obtained as power amplifier input signals, and a set of all values ​​of each power amplifier input signal is obtained in the power amplifier gradient of the micro-electromechanical speaker when the frequency is divided and output as a gradient set of the corresponding power amplifier input signal, the standard deviation of all values ​​in each gradient set can be used as the output stability value of the corresponding power amplifier input signal, and the mean of all output stability values ​​can be used as the output stability of the micro-electromechanical speaker. The output stability of the micro-electromechanical speaker can be obtained in the above manner; finally, a linear allocation model is initialized to convert each The output stability of each MEMS speaker is used as the basic parameter of the linear allocation model, the preset difference threshold is used as the allocation evaluation standard of the linear allocation model, the received signal segment of the target sound source is used as the allocation target of the linear allocation model, and the received signal segment of the target sound source is re-divided using the linear allocation model to obtain all the audio signal segments. In other embodiments, other models of the received signal segments of the target sound source can also be used to re-divide. It is only necessary to ensure that the difference in the output stability of each audio signal segment is within the preset difference threshold, which is not limited here. The difference threshold represents the upper limit of the difference in output stability in each audio signal segment, which can be preset through historical experience and the requirements for sound quality in the usage scenario of the MEMS speaker.

[0027] In some embodiments, determining the recognition sensitivity of each audio signal segment from the micro-electromechanical audio information may be implemented by the following steps: For each audio signal segment, obtaining all input signal segments of the audio signal segment in each audio input signal of the micro-electromechanical audio information; Determine the recognition delay of each audio recognition in the audio signal segment through each input signal segment; The recognition sensitivity of the audio signal segment is determined according to all the recognition delays, and then the recognition sensitivity of each audio signal segment is obtained.

[0028] It should be noted that, in the present application, recognition sensitivity is an indicator of the response speed of a microelectromechanical speaker when processing an audio signal in a specific frequency band. The higher the recognition sensitivity, the faster the response speed of the microelectromechanical speaker when processing audio in the frequency band. In specific implementation, first, for each audio signal segment, all input signal segments whose frequencies are within the audio signal segment are screened out from each audio input signal of the microelectromechanical audio information. Then, for each audio recognition, the input signal segment under the audio recognition is obtained. The time difference from input to output of the input signal segment can be obtained in the sensor of the microelectromechanical system as the recognition delay during audio recognition. The recognition delay for each audio recognition in the audio signal segment can be obtained in the above manner. The recognition delay refers to the time difference in the process from input to output of the audio signal. Finally, the inverse of each recognition delay is used as the recognition sensitivity value for each audio recognition in the audio signal segment. The average of all recognition sensitivity values ​​can be used as the recognition sensitivity of the audio signal segment. The recognition sensitivity of each audio signal segment can be obtained in the above manner.

[0029] In some embodiments, determining the spectrum balance relationship of signal distribution in each MEMS speaker according to all recognition sensitivities and the frequency bands of output signals of the spectrum in each MEMS speaker can be implemented by the following steps: For each MEMS speaker, obtaining all audio signal segments included in the output signal frequency band of the frequency spectrum in the MEMS speaker; determining signal allocation information of the frequency spectrum in the MEMS speaker according to the recognition sensitivity of each audio signal segment; The spectrum balance relationship of the signal distribution in the micro-electro-mechanical speakers is determined by the signal distribution information, and then the spectrum balance relationship of the spectrum in each micro-electro-mechanical speaker is obtained.

[0030] It should be noted that in the present application, the spectrum balance relationship refers to the load distribution relationship of each MEMS speaker when processing signals in different frequency bands. It is calculated by integrating the signal distribution information of all speakers to ensure the spectrum load balance of the audio system; the signal distribution information refers to the distribution ratio information of each audio signal segment in the MEMS speaker. Adjusting the distribution based on the recognition sensitivity can match the processing capabilities of each MEMS speaker for different signal segments.

[0031] In the specific implementation, first, for each micro-electromechanical speaker, a micro-electromechanical audio signal of the micro-electromechanical speaker is obtained in the micro-electromechanical audio information, and all audio signal segments contained in the output signal frequency band can be screened out from the micro-electromechanical audio signal; then, the recognition sensitivity of each audio signal segment is obtained, and the ratio of each recognition sensitivity to the sum of all recognition sensitivities can be used as the signal distribution ratio of each audio signal segment in the micro-electromechanical speaker to the frequency spectrum, and the set of all signal distribution ratios can be used as the signal distribution information; finally, a balance relationship model based on deep learning is initialized, and the signal distribution information of the frequency spectrum in the micro-electromechanical speaker is used as an evaluation parameter of the balance relationship model, and the balance relationship model is used to evaluate the balance of the signal distribution in the micro-electromechanical speaker, and the evaluation result can be used as the spectrum balance relationship of the signal distribution in the micro-electromechanical speaker. The spectrum balance relationship of the frequency spectrum in each micro-electromechanical speaker can be obtained in the above manner.

[0032] In step 103, masking perception is performed on each audio output signal in the micro-electromechanical audio information based on the auditory masking effect of the neural network, thereby obtaining a masking interval of the frequency spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted based on the masking interval and each spectrum equalization relationship.

[0033] In some embodiments, masking perception is performed on each audio output signal in the micro-electromechanical system audio information based on the auditory masking effect of the neural network, thereby obtaining the masking interval of the spectrum in the micro-electromechanical system, referring to Figure 2 As described above, the figure is a schematic diagram of a process for determining a masking interval in some embodiments of the present application. In this embodiment, determining the masking interval can be implemented by the following steps: In step 1031, interference information of output frequency bands between each MEMS speaker is determined through each audio output signal in the MEMS audio information; In step 1032, spectrum masking evaluation is performed on each interference information based on the auditory masking effect of the neural network to obtain the critical frequency band of the target sound source in each MEMS speaker; In step 1033, the spectrum masking range of each MEMS speaker is screened out from each audio output signal in the MEMS audio information according to each critical frequency band; In step 1034 , the masking interval of the spectrum in the micro-electromechanical system is determined through all the spectrum masking ranges.

[0034] It should be noted that, in the present application, the masking interval refers to the signal interval in which the target sound source is masked; the interference information refers to the mutual interference between the audio signals generated by each MEMS speaker due to the overlapping frequency bands; the critical frequency band refers to the frequency range in which a specific frequency band is masked by other signals under the auditory masking effect; and the spectrum masking range refers to the audio signal range masked in each MEMS speaker.

[0035] In specific implementation, first, all frequency band overlaps of output frequency bands between various micro-electromechanical speakers are screened out from various audio output signals in the micro-electromechanical audio information, and a set of amplitude difference values ​​of the overlapping parts of various frequency bands can be used as interference information; secondly, for each micro-electromechanical speaker, the corresponding part of the audio signal segment in the micro-electromechanical speaker is screened out from all interference information, and each interference information is analyzed using a pre-trained neural network model to simulate the auditory masking effect of the human ear, and the spectrum information and time series information can be captured by the long short-term memory network in the neural network model, and the output result of the neural network model can be used as the critical frequency band of the micro-electromechanical speaker, and the critical frequency band of the micro-electromechanical speaker can be obtained by the above method; then, for each micro-electromechanical speaker, the signal interval in the audio output signal of the micro-electromechanical speaker in the micro-electromechanical audio information that is lower than the critical frequency band of the micro-electromechanical speaker can be used as the spectrum masking range of the micro-electromechanical speaker, and the spectrum masking range of each micro-electromechanical speaker can be obtained by the above method; finally, the set of all spectrum masking ranges can be used as the masking interval of the spectrum in the micro-electromechanical system.

[0036] It should be noted that the simulation of the auditory masking effect through a neural network can more accurately identify the critical frequency bands that are easily masked in the spectrum, thereby reducing unnecessary audio processing overhead, and then clearly identify the frequency band areas that require special processing during output, in order to avoid the degradation of audio quality due to the masking effect, thereby improving the overall audio performance.

[0037] In some embodiments, determining the suppression gain of each MEMS speaker when the target sound source is distorted according to the masking interval and each spectrum equalization relationship can be implemented by the following steps: For each MEMS speaker, determining the masking spectrum information and the masked signal interval of the MEMS speaker according to the spectrum of the MEMS speaker in the masking interval; Determine a suppression curve of the MEMS speaker when the target sound source is distorted according to the spectrum balance relationship of the MEMS speaker and the masking spectrum information; Determining a gain curve of the MEMS speaker when the target sound source is distorted according to the spectrum balance relationship of the MEMS speaker and the masked signal interval; The suppression gain of the micro-electro-mechanical loudspeaker when the target sound source is distorted is determined according to the suppression curve and the gain curve, and then the suppression gain of each micro-electro-mechanical loudspeaker when the target sound source is distorted is obtained.

[0038] It should be noted that, in the present application, suppression gain refers to the loudspeaker output gain after comprehensive suppression and gain adjustment, which is used to optimize the audio output under distortion; masked spectrum information refers to the part of the output spectrum of each micro-electromechanical loudspeaker within the masking interval that masks other spectrums; the masked signal interval refers to the spectrum signal range actually masked in the output spectrum of each micro-electromechanical loudspeaker within the masking interval; the suppression curve refers to the curve of each micro-electromechanical loudspeaker adjusting the output spectrum within the masking interval to reduce distortion; the gain curve refers to the spectrum gain adjusted by each micro-electromechanical loudspeaker to enhance the output intensity of the non-masked signal under distortion.

[0039] In a specific implementation, firstly, for each MEMS speaker, the spectrum masking range of the MEMS speaker in the masking interval can be used as the signal interval of the masked signal interval, the signal difference between the audio output signals of the MEMS speaker and the remaining MEMS speakers in the signal interval of the masked signal interval can be calculated, and the signal intervals of each audio output signal whose absolute value is negative and greater than a preset masking threshold value can be screened out from the signal difference, and all signal intervals can be used as masked spectrum information, and the masking threshold value represents the lower limit value of the audio difference that will cause the auditory masking effect, which can be preset through historical experience; secondly, a Kalman filter based on state transition is initialized, the spectrum balance relationship of the MEMS speaker is used as the input control of the Kalman filter, and the masked spectrum information is used as the state transition information of the Kalman filter, and the Kalman filter can be used to estimate the signal change of the MEMS speaker when the target sound source is distorted, and the signal estimation results after the Kalman filter are connected in the predicted time order as the state transition information of the MEMS speaker. The suppression curve when the target sound source is distorted. In other embodiments, other models or algorithms may be used to determine the suppression curve of the micro-electromechanical speaker when the target sound source is distorted, which is not limited here. Then, a gain control model based on statistical characteristics (for example, a linear prediction coding model) is initialized, and the spectrum equalization relationship is used as the relationship between different frequency signals in the spectrum of the gain control model. The masked signal interval is used as the gain range of the gain control model. The gain control model is used to predict the gain of the spectrum characteristics in the masked signal interval. The gain prediction results can be connected in the order of prediction time as the gain curve of the micro-electromechanical speaker when the target sound source is distorted. In other embodiments, other models or algorithms may be used to determine the gain curve of the micro-electromechanical speaker when the target sound source is distorted, which is not limited here. Finally, the suppression curve and the gain curve can be combined as the suppression gain of the micro-electromechanical speaker when the target sound source is distorted. The suppression gain of each micro-electromechanical speaker when the target sound source is distorted can be obtained in the above manner.

[0040] It should be noted that in the present application, the Kalman filter based on state transfer combines the powerful classification capability of the support vector machine and the real-time estimation capability of the Kalman filter, and can effectively utilize historical data and real-time data to generate a more accurate suppression curve. The classification and prediction functions of the support vector machine help provide a preliminary suppression curve, while the Kalman filter further improves the accuracy of the suppression curve through dynamic optimization, thereby further reducing the error of the final generated suppression curve; the gain control model based on statistical characteristics generates a gain curve that can more accurately adjust the gain values ​​of each signal frequency band in the masked signal interval, which helps to minimize audio distortion and improve the clarity and detail of the audio output.

[0041] In step 104, the phase shift delay information of the sound wave phase between the MEMS speakers is obtained, and the audio distortion of the target sound source at the current sound wave phase is determined through the phase shift delay information and all spectrum equalization relationships.

[0042] It should be noted that, in the present application, phase shift delay information refers to the time delay caused by the phase difference between micro-electromechanical speakers. The phase difference refers to the phase angle difference between two signals at the same frequency, which is often used to describe the relative delay between speakers; the sound wave phase is a parameter that describes the fluctuation state of the sound wave at a specific position; in specific implementation, all micro-electromechanical speakers are combined in pairs to obtain all speaker groups, and for each speaker group, a pulse signal is sent to the two micro-electromechanical speakers in the speaker group at the current sound wave phase, and the reception time of the pulse signal at the two micro-electromechanical speakers in the speaker group is measured. The absolute value of the difference between the reception times can be used as the phase shift delay, and the average of all phase shift delays obtained from a large number of pulse experiments can be used as the phase shift delay value of the sound wave phase between the two micro-electromechanical speakers in the speaker group. The phase shift delay value between every two micro-electromechanical speakers can be obtained in the above manner, and the set of all phase shift delay values ​​can be used as the phase shift delay information.

[0043] In some embodiments, the audio distortion of the target sound source at the current sound wave phase is determined by using the phase shift delay information and all spectrum equalization relationships, referring to Figure 3 The figure is a schematic diagram of a process for determining audio distortion in some embodiments of the present application. In this embodiment, the audio distortion can be determined by the following steps: In step 1041, the phase shift interference amount of the current sound wave phase on the target sound source is determined according to the phase shift delay information; In step 1042, the frequency response distortion of the signal distribution to the target sound source under the current sound wave phase is determined through all spectrum equalization relationships; In step 1043, the audio distortion degree of the target sound source at the current sound wave phase is determined based on the phase shift interference amount and the frequency response distortion.

[0044] It should be noted that in the present application, the audio distortion degree indicates the degree of distortion of the audio signal in the micro-electromechanical system under the current sound wave phase. By comprehensively considering the phase shift interference amount and the frequency response distortion, the audio distortion degree of the target sound source can be comprehensively evaluated, thereby facilitating the subsequent judgment of whether the target sound source is distorted; the phase shift interference amount refers to the phase change caused by the phase shift delay; the frequency response distortion refers to the deviation between the actual output frequency of the audio in the micro-electromechanical system and the ideal response frequency.

[0045] In specific implementation, first, for each phase shift delay value in the phase shift delay information, the frequency difference of each micro-electromechanical speaker in the speaker group corresponding to the phase shift delay value is obtained, and the product of the phase shift delay value, the frequency difference and 2π can be used as the phase shift interference value corresponding to the phase shift delay value. The phase shift interference value corresponding to each phase shift delay value in the phase shift delay information can be obtained in the above manner, and the average of all phase shift interference values ​​can be used as the phase shift interference amount of the current sound wave phase to the target sound source; then, a frequency response model based on a neural network can be used to evaluate the frequency response of all spectrum equalization relationships, and the result of the frequency response evaluation is used as the frequency response prediction value in the spectrum, and the ratio of the difference between the frequency response prediction value and the ideal frequency response value to the ideal frequency response value can be used as the frequency response distortion of the signal distribution to the target sound source under the current sound wave phase; finally, the sum of the phase shift interference amount and the frequency response distortion can be used as the audio distortion degree of the target sound source under the current sound wave phase.

[0046] It should be noted that in this application, the frequency response prediction value is an indicator used to measure the audio frequency processing capability of the micro-electromechanical system, and the ideal frequency response value refers to the ideal quantitative value of the audio frequency processing capability of the micro-electromechanical system, which is determined by the frequency response expected by the micro-electromechanical system design.

[0047] In step 105, when the audio distortion is greater than a preset distortion threshold, signal equalization distribution is performed on each MEMS speaker in the MEMS system according to each suppression gain.

[0048] In some embodiments, the signal distribution to each MEMS speaker according to each suppression gain is to use each suppression gain as a balance adjustment amount when performing signal distribution to the corresponding MEMS speaker, thereby completing the balance adjustment of the signal distribution.

[0049] In the present application, when the audio distortion is greater than a preset distortion threshold, the signal of each micro-electromechanical speaker in the micro-electromechanical system can be evenly distributed according to each suppression gain; firstly, by determining the recognition sensitivity of each audio signal segment, the signal segment that is prone to distortion can be processed preferentially, thereby improving the consistency of the audio signal between each micro-electromechanical speaker, and determining the spectrum balance relationship of the signal distribution in each micro-electromechanical speaker according to all the recognition sensitivities and the output signal bands of the spectrum in each micro-electromechanical speaker, thereby ensuring the coordinated work between the micro-electromechanical speakers, thereby improving the distortion control capability of the micro-electromechanical speaker system under different sound wave phases; secondly, by combining the masking interval and the spectrum balance relationship to determine the suppression gain, the output intensity of each micro-electromechanical speaker can be dynamically adjusted, so as to facilitate the subsequent use of the target sound source when distortion occurs. Take effective gain suppression measures to reduce audio distortion caused by excessive or weak output; finally, determine the influence of the current sound wave phase on the audio signal by monitoring the phase shift delay between the MEMS speakers. Since sound waves at different phases may cause different types of distortion, the audio distortion at the current sound wave phase can be accurately determined in combination with the phase shift delay. When the distortion exceeds the threshold, the signal of each MEMS speaker is evenly distributed, which can more effectively control the quality of the audio output and ensure the signal consistency between different MEMS speakers. In summary, based on the above scheme, the signal balance distribution of each MEMS speaker in the MEMS system can be achieved, thereby improving the distortion control capability of the MEMS speaker system at different sound wave phases and ensuring the consistency of the audio output between each MEMS speaker.

[0050] In addition, in another aspect of the present application, in some embodiments, the present application provides a micro-electromechanical speaker system, the micro-electromechanical speaker system also includes a control unit, referring to Figure 4 , which is a schematic diagram of the structure of a control unit according to some embodiments of the present application, the control unit includes: an acquisition module 201, a processing module 202 and an execution module 203, which are described as follows: Acquisition module 201, in the present application, acquisition module 201 is mainly used to use the audio sensor in the micro-electro-mechanical system to monitor the audio input signal of each micro-electro-mechanical speaker when playing the target sound source, and then obtain the micro-electro-mechanical audio information; The processing module 202 in the present application is used to divide the target sound source into multiple audio signal segments by the power amplifier gradient of each micro-electromechanical speaker during frequency division output, and then determine the recognition sensitivity of each audio signal segment by the micro-electromechanical audio information, and determine the spectrum balance relationship of the signal distribution in each micro-electromechanical speaker according to all the recognition sensitivities and the output signal frequency band of the spectrum in each micro-electromechanical speaker; It should be noted that the processing module 202 is also used to perform masking perception on each audio output signal in the micro-electromechanical audio information based on the auditory masking effect of the neural network, thereby obtaining the masking interval of the spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and the equalization relationship of each spectrum; In addition, the processing module 202 is also used to obtain the phase shift delay information of the sound wave phase between the micro-electromechanical speakers, and determine the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships; The execution module 203 in the present application is mainly used to perform signal balancing distribution on each MEMS speaker in the MEMS system according to each suppression gain when the audio distortion is greater than a preset distortion threshold.

[0051] The above describes in detail the examples of the micro-electromechanical speaker system and control method provided by the embodiments of the present application. It can be understood that in order to realize the above functions, the corresponding device includes hardware structures and / or software modules corresponding to the execution of each function. It should be easily appreciated by those skilled in the art that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. 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 the present application.

[0052] In some embodiments, the present application also provides a computer device, which includes a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the above-mentioned micro-electromechanical speaker control method.

[0053] In some embodiments, reference Figure 5 , the dotted line in the figure indicates that the unit or the module is optional, and the figure is a schematic diagram of the structure of a computer device for implementing a micro-electromechanical speaker control method according to an embodiment of the present application. The micro-electromechanical speaker control method described in the above embodiment can be Figure 5 The computer device shown in the figure is implemented, and the computer device includes at least one processor 301, a memory 302 and at least one communication unit 305. The computer device can be a terminal device, a server or a chip.

[0054] The processor 301 may be a general-purpose processor or a special-purpose processor. For example, the processor 301 may be a central processing unit (CPU), which may be used to control the computer device, execute software programs, and process data of the software programs. The computer device may also include a communication unit 305 to implement signal input (reception) and output (transmission).

[0055] For example, the computer device may be a chip, the communication unit 305 may be an input and / or output circuit of the chip, or the communication unit 305 may be a communication interface of the chip, and the chip may be a component of a terminal device, a network device, or other devices.

[0056] For another example, the computer device may be a terminal device or a server, and the communication unit 305 may be a transceiver of the terminal device or the server, or the communication unit 305 may be a transceiver circuit of the terminal device or the server.

[0057] The computer device may include one or more memories 302, on which a program 304 is stored. The program 304 can be executed by the processor 301 to generate instructions 303, so that the processor 301 performs the method described in the above method embodiment according to the instructions 303. Optionally, data (such as a target audit model) can also be stored in the memory 302. Optionally, the processor 301 can also read the data stored in the memory 302, and the data can be stored at the same storage address as the program 304, or the data can be stored at a different storage address from the program 304.

[0058] The processor 301 and the memory 302 may be provided separately or integrated together, for example, integrated on a system on chip (SOC) of the terminal device.

[0059] It should be understood that each step of the above method embodiment can be completed by a hardware-based logic circuit or software-based instructions in the processor 301. The processor 301 can be a CPU, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, such as discrete gates, transistor logic devices, or discrete hardware components.

[0060] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0061] For example, in some embodiments, the present application also provides a computer-readable storage medium, in which instructions or codes are stored. When the instructions or codes are run on a computer, the computer implements the above-mentioned micro-electromechanical speaker control method when executed.

[0062] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0063] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for controlling a micro-electromechanical speaker, characterized in that: The steps include: Using the audio sensor in the MEMS to monitor the audio input signal of each MEMS speaker when playing the target sound source, and then obtaining the MEMS audio information; Divide the target sound source into a plurality of audio signal segments by the power amplifier gradient of each MEMS speaker during frequency division output, and then determine the recognition sensitivity of each audio signal segment by the MEMS audio information, and determine the spectrum balance relationship of the signal distribution in each MEMS speaker according to all the recognition sensitivities and the output signal frequency band of the spectrum in each MEMS speaker; Based on the auditory masking effect of the neural network, each audio output signal in the micro-electromechanical system audio information is masked and sensed, thereby obtaining the masking interval of the spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and the equalization relationship of each spectrum; Acquire phase shift delay information of the sound wave phase between the micro-electromechanical speakers, and determine the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships; When the audio distortion is greater than a preset distortion threshold, signal equalization distribution is performed on each MEMS speaker in the MEMS system according to each suppression gain.

2. The method according to claim 1, characterized in that The target sound source is divided into multiple audio signal segments by the power amplifier gradient of each MEMS speaker at the frequency division output, specifically including: Obtaining a received signal segment of a target sound source in each micro-electromechanical speaker; Determine the output stability of each MEMS speaker through each received signal segment and the power amplifier gradient of the corresponding MEMS speaker at the frequency division output; The received signal segments of the target sound source in each micro-electromechanical loudspeaker are re-divided according to all output stabilities to obtain a plurality of audio signal segments.

3. The method according to claim 1, characterized in that Determining the recognition sensitivity of each audio signal segment by the micro-electromechanical audio information specifically includes: For each audio signal segment, obtaining all input signal segments of the audio signal segment in each audio input signal of the micro-electromechanical audio information; Determine the recognition delay of each audio recognition in the audio signal segment through each input signal segment; The recognition sensitivity of the audio signal segment is determined according to all the recognition delays, and then the recognition sensitivity of each audio signal segment is obtained.

4. The method according to claim 1, characterized in that Determining the spectrum balance relationship of signal distribution in each MEMS speaker according to all recognition sensitivities and the output signal frequency bands of the spectrum in each MEMS speaker specifically includes: For each MEMS speaker, obtaining all audio signal segments included in the output signal frequency band of the frequency spectrum in the MEMS speaker; determining signal allocation information of the frequency spectrum in the MEMS speaker according to the recognition sensitivity of each audio signal segment; The spectrum balance relationship of the signal distribution in the micro-electro-mechanical speakers is determined by the signal distribution information, and then the spectrum balance relationship of the spectrum in each micro-electro-mechanical speaker is obtained.

5. The method according to claim 1, characterized in that The masking perception of each audio output signal in the MEMS audio information is performed based on the auditory masking effect of the neural network, and the masking interval of the spectrum in the MEMS system is obtained, which specifically includes: Determine the interference information of the output frequency bands between the respective MEMS speakers through the respective audio output signals in the MEMS audio information; Based on the auditory masking effect of neural network, the spectrum masking evaluation of each interference information is carried out to obtain the critical frequency band of the target sound source in each MEMS speaker; Filtering out the spectrum masking range of each MEMS speaker in each audio output signal in the MEMS audio information according to each critical frequency band; The masking interval of the spectrum in the MEMS is determined by all the spectrum masking ranges.

6. The method according to claim 1, characterized in that Determining the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships specifically includes: Determine the phase shift interference amount of the current sound wave phase on the target sound source according to the phase shift delay information; Determine the frequency response distortion of the target sound source by the signal distribution under the current sound wave phase through all spectrum equalization relationships; The audio distortion degree of the target sound source at the current sound wave phase is determined based on the phase shift interference amount and the frequency response distortion.

7. The method according to claim 1, characterized in that The signal distribution to each MEMS speaker according to each suppression gain is to use each suppression gain as a balance adjustment amount when performing signal distribution to the corresponding MEMS speaker, thereby completing the balance adjustment of the signal distribution.

8. A micro-electromechanical speaker system, comprising a control unit, characterized in that: The control unit comprises: An acquisition module is used to use an audio sensor in a micro-electro-mechanical system to monitor the audio input signal of each micro-electro-mechanical speaker when playing a target sound source, thereby obtaining micro-electro-mechanical audio information; A processing module, used to divide the target sound source into a plurality of audio signal segments by the power amplifier gradient of each MEMS speaker during frequency division output, and then determine the recognition sensitivity of each audio signal segment by the MEMS audio information, and determine the spectrum balance relationship of the signal distribution in each MEMS speaker according to all the recognition sensitivities and the output signal frequency band of the spectrum in each MEMS speaker; The processing module is also used to perform masking perception on each audio output signal in the micro-electromechanical audio information based on the auditory masking effect of the neural network, thereby obtaining a masking interval of the spectrum in the micro-electromechanical system, and determining the suppression gain of each micro-electromechanical speaker when the target sound source is distorted according to the masking interval and each spectrum balance relationship; The processing module is also used to obtain phase shift delay information of the sound wave phase between the micro-electromechanical speakers, and determine the audio distortion of the target sound source at the current sound wave phase through the phase shift delay information and all spectrum equalization relationships; The execution module is used for performing signal balancing distribution to each MEMS speaker in the MEMS system according to each suppression gain when the audio distortion is greater than a preset distortion threshold.

9. A computer device, characterized in that: The computer device comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the micro-electromechanical speaker control method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions or codes, and when the instructions or codes are executed on a computer, the computer implements the micro-electromechanical speaker control method according to any one of claims 1 to 7.

Images