An intelligent noise reduction method and system for a MEMS speaker

Abstract

The present invention discloses an intelligent noise reduction method and system for a MEMS speaker. The method includes: collecting the original audio signal and environmental noise signal data emitted by the MEMS speaker; extracting the acoustic fingerprint of the original audio signal to dynamically distinguish the sound wave components in the audio signal; deconstructing the extracted acoustic fingerprint and generating a noise reduction control vector for the noise signal; performing noise reduction processing on the original audio signal according to the noise reduction control vector; calculating the sound wave kinetic energy of the environmental noise signal, and generating a corresponding reverse sound wave kinetic energy signal according to the sound wave kinetic energy; converting the reverse sound wave kinetic energy signal into an electrical signal and converting and outputting it through the MEMS speaker to achieve noise reduction of the environmental noise signal. By using the embodiments of the present invention, the characteristics of the noise signal can be accurately captured, and a more effective noise reduction signal can be generated to improve the real-time performance and adaptability of the intelligent noise reduction method.

Description

Technical Field

[0001] The present invention belongs to the technical field of loudspeakers, and particularly relates to an intelligent noise reduction method and system for MEMS loudspeakers. Background Art

[0002] In modern society, with the accelerating process of industrialization and urbanization, noise pollution has become a major problem affecting people's quality of life. Especially in dense urban environments, transportation hubs, industrial areas, and modern homes and workplaces, noise interference has become an inescapable phenomenon. Prolonged exposure to noise not only affects an individual's work efficiency and emotional state but may also cause permanent damage to hearing. Therefore, effective noise control technologies are of great significance for improving the comfort of living and working environments.

[0003] Traditional passive noise control methods, such as sound insulation materials and sound insulation walls, although they have played a certain role in reducing environmental noise, have disadvantages such as limited application scope and sound insulation effects being restricted by physical space. To solve these problems, people have gradually turned to the use of active noise control (ANC) technologies. ANC technologies can cancel out the original noise by emitting sound waves with opposite phases to the noise, thereby achieving a noise reduction effect.

[0004] MEMS (Micro-Electro-Mechanical System) loudspeakers have become an ideal choice in ANC systems due to their advantages such as small size, light weight, fast response, and low power consumption. MEMS loudspeakers can precisely control the emission of sound waves and are applicable to various electronic products such as smartphones, wearable devices, and smart home devices. However, improving the performance of MEMS loudspeakers in ANC systems remains a difficult point and development direction in technical research. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent noise reduction method and system for MEMS loudspeakers to solve the deficiencies in the prior art, capable of accurately capturing the characteristics of noise signals and generating more effective noise reduction signals to improve the real-time performance and adaptability of the intelligent noise reduction method.

[0006] An embodiment of the present application provides an intelligent noise reduction method for MEMS loudspeakers, the method comprising:

[0007] Collecting the original audio signal and environmental noise signal data emitted by the MEMS loudspeaker;

[0008] Performing acoustic fingerprint extraction on the original audio signal to dynamically distinguish the sound wave components in the audio signal, the sound wave components including the characteristic spectra of the desired signal and the original noise signal;

[0009] Deconstructing the extracted acoustic fingerprint and generating a noise reduction control vector for the original noise signal;

[0010] Denoise the original audio signal according to the denoising control vector;

[0011] Calculate the acoustic wave kinetic energy of the environmental noise signal, and generate a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy;

[0012] Convert the reverse acoustic wave kinetic energy signal into an electrical signal and output it through the MEMS speaker to achieve denoising of the environmental noise signal.

[0013] Optionally, extracting an acoustic fingerprint from the original audio signal to dynamically distinguish the acoustic wave components in the audio signal includes:

[0014] Convert the time-domain signal corresponding to the original audio signal into a frequency-domain signal;

[0015] Divide the frequency-domain signal into multiple frequency bands, where each frequency band represents a part of the frequency components in the signal;

[0016] Calculate the energy spectral density in each frequency band to quantify the energy of the signal in that frequency band;

[0017] Analyze the energy spectral density of each frequency band to extract the eigenvalue of the acoustic wave component, and the eigenvalue constitutes the acoustic fingerprint of the acoustic wave component.

[0018] Optionally, deconstructing the extracted acoustic fingerprint and generating a denoising control vector for the original noise signal includes:

[0019] Classify the set of eigenvalues of the acoustic wave components included in the acoustic fingerprint to obtain the desired signal feature and the original noise signal feature;

[0020] Combine the noise signal features to construct a one-dimensional or multi-dimensional noise signal model representing the power spectral density of the actual noise signal;

[0021] Create a denoising control vector using the noise signal model, and the denoising control vector is used to indicate the amount of energy that needs to be suppressed at each frequency.

[0022] Optionally, the denoising process of the original audio signal according to the denoising control vector includes:

[0023] Generate a reverse acoustic wave signal matching the power spectral density of the noise signal using the denoising control vector as the denoising signal;

[0024] Perform denoising processing on the original audio signal based on the denoising signal, where the denoising signal is opposite in phase to the noise signal in the original audio signal.

[0025] Optionally, calculating the acoustic wave kinetic energy of the ambient noise signal and generating a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy includes:

[0026] Converting the ambient noise signal from the time domain to the frequency domain to obtain the spectral distribution of the ambient noise signal;

[0027] Calculating the power spectral density of the ambient noise signal at each frequency point according to the spectral distribution;

[0028] Integrating the power spectral density over all or a selected frequency range to obtain the total acoustic wave kinetic energy of the ambient noise signal;

[0029] Identifying specific high-energy frequency components or characteristic frequency components in the power spectral density, and generating a sound pressure level diagram using the identified frequency components and the total acoustic wave kinetic energy data, where the sound pressure level diagram describes the sound pressure levels of the frequency components;

[0030] Synthesizing a filtered signal with the same amplitude but opposite phase corresponding to the sound pressure level of the specific frequency component according to the sound pressure level diagram as the reverse acoustic wave kinetic energy signal.

[0031] Another embodiment of the present application provides an intelligent noise reduction system for a MEMS speaker, and the system includes:

[0032] An acquisition module for acquiring the original audio signal and ambient noise signal data emitted by the MEMS speaker;

[0033] An extraction module for extracting the acoustic fingerprint of the original audio signal to dynamically distinguish the sound wave components in the audio signal, where the sound wave components include the characteristic spectra of the desired signal and the original noise signal;

[0034] A generation module for deconstructing the extracted acoustic fingerprint and generating a noise reduction control vector for the original noise signal;

[0035] A first noise reduction module for performing noise reduction processing on the original audio signal according to the noise reduction control vector;

[0036] A calculation module for calculating the acoustic wave kinetic energy of the ambient noise signal and generating a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy;

[0037] A second noise reduction module for converting the reverse acoustic wave kinetic energy signal into an electrical signal and converting and outputting it through the MEMS speaker to achieve noise reduction of the ambient noise signal.

[0038] Another embodiment of the present application provides a storage medium, in which a computer program is stored, and the computer program is set to execute the method described in any one of the above when running.

[0039] Another embodiment of the present application provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the method described in any one of the above.

[0040] Compared with the prior art, an intelligent noise reduction method for a MEMS speaker provided by the present invention includes collecting original audio signal and ambient noise signal data emitted by the MEMS speaker; extracting an acoustic fingerprint from the original audio signal to dynamically distinguish the sound wave components in the audio signal; deconstructing the extracted acoustic fingerprint and generating a noise reduction control vector for the noise signal; performing noise reduction processing on the original audio signal according to the noise reduction control vector; calculating the sound wave kinetic energy of the ambient noise signal, and generating a corresponding reverse sound wave kinetic energy signal according to the sound wave kinetic energy; converting the reverse sound wave kinetic energy signal into an electrical signal and converting and outputting the electrical signal through the MEMS speaker to achieve noise reduction of the ambient noise signal, so as to accurately capture the characteristics of the noise signal and generate a more effective noise reduction signal, thereby improving the real-time performance and adaptability of the intelligent noise reduction method. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a hardware structure block diagram of a computer terminal for an intelligent noise reduction method of a MEMS speaker provided by an embodiment of the present invention;

[0042] Figure 2 It is a schematic flowchart of an intelligent noise reduction method of a MEMS speaker provided by an embodiment of the present invention;

[0043] Figure 3 It is a schematic structural diagram of an intelligent noise reduction system of a MEMS speaker provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The embodiments described below by referring to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0045] An embodiment of the present invention first provides an intelligent noise reduction method for a MEMS speaker. This method can be applied to an electronic device, such as a computer terminal, specifically, such as an ordinary computer, etc.

[0046] The following takes running on a computer terminal as an example to describe it in detail. Figure 1 It is a hardware structure block diagram of a computer terminal for an intelligent noise reduction method of a MEMS speaker provided by an embodiment of the present invention. As Figure 1 shown, the computer terminal may include one or more ( Figure 1Only one processor 102 (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a field programmable gate array FPGA) and a memory 104 for storing data are shown. Optionally, the computer terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those of ordinary skill in the art can understand that Figure 1 The structure shown is only schematic and does not limit the structure of the computer terminal. For example, the computer terminal may further include more or fewer components than Figure 1 shown therein, or have a different configuration from Figure 1 that shown.

[0047] The memory 104 can be used to store software programs and modules of application software, such as program instructions / modules corresponding to the intelligent noise reduction method of the MEMS speaker in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the above method. The memory 104 may include a high-speed random access memory, and may further include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely provided with respect to the processor 102, and these remote memories may be connected to the computer terminal through a network. Examples of the above network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0048] The transmission device 106 is used to receive or send data via a network. Specific examples of the above network may include a wireless network provided by a communication provider of the computer terminal. In one instance, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices through a base station and thus communicate with the Internet. In one instance, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0049] See Figure 2 , embodiments of the present invention provide an intelligent noise reduction method for a MEMS speaker, which may include the following steps:

[0050] S201, collect the original audio signal and environmental noise signal data emitted by the MEMS speaker;

[0051] Specifically, a method of high-precision and synchronous acquisition can be adopted. This method can not only efficiently capture the original audio signal and ambient noise, but also ensure the time alignment of these two signals, providing accurate data for subsequent processing. The following are the specific steps of this method:

[0052] Step 1: Dual-channel synchronous acquisition setup Configure two high-sensitivity MEMS microphones near the MEMS speaker. One is used to capture the original audio signal emitted by the MEMS speaker, and the other is used to capture the ambient noise signal. These two microphones should be as close as possible to the MEMS speaker and ensure that their positions are fixed to reduce the impact of sound propagation time differences on signal synchronization.

[0053] Step 2: High-precision clock synchronization Adopt high-precision clock synchronization technology to ensure that the acquisition actions of the two microphones are strictly synchronized. This can be achieved by using the same clock source or clock synchronization signal to ensure that the data acquisition of the two microphones is completely consistent in time. This step is the key to ensuring the accuracy of subsequent signal processing.

[0054] Step 3: Signal preprocessing and enhancement The acquired signals are preprocessed, including filtering, enhancement, etc., to improve the signal quality. For the original audio signal, a low-pass filter can be used to remove the noise outside the speaker playback frequency range. For the ambient noise signal, a band-pass filter can be used to retain the ambient noise components similar to the frequency range of the original audio signal to more accurately analyze the impact of noise on the original signal.

[0055] Step 4: Digitization and storage Convert the preprocessed analog signal into a digital signal through a high-precision analog-to-digital converter (ADC) and store it in the data memory at a high sampling rate. This step not only ensures the high fidelity of the signal but also facilitates subsequent digital signal processing.

[0056] Step 5: Data calibration and correction To ensure the accuracy of the acquired data, the acquired data is calibrated and corrected. This includes adjusting the amplitude of the signal to ensure that the absolute intensity of the signal can reflect the actual sound pressure level and timestamping the acquired signal so that it can be accurately corresponded to a specific acquisition moment in subsequent processing.

[0057] Step 6: Synchronization verification and calibration Finally, perform synchronization verification on the acquired original audio signal and ambient noise signal, check the time alignment of the two signals, and perform calibration if necessary. This may involve adjusting the small time offset of the signal to ensure that the two signals are completely synchronized in time.

[0058] Through the above steps, the original audio signal emitted by the MEMS speaker and the environmental noise signal data can be collected efficiently and accurately. This method makes full use of the high sensitivity of modern MEMS microphones and the high precision of digital signal processing technology, providing a solid data foundation for the implementation of intelligent noise reduction methods.

[0059] S202, perform acoustic fingerprint extraction on the original audio signal to dynamically distinguish the sound wave components in the audio signal, where the sound wave components include the characteristic spectra of the desired signal and the original noise signal;

[0060] Specifically, the time-domain signal corresponding to the original audio signal can be converted into a frequency-domain signal; the frequency-domain signal is divided into multiple frequency bands, where each frequency band represents a part of the frequency components in the signal; calculate the energy spectral density in each frequency band to quantify the energy of the signal in that frequency band; analyze the energy spectral density of each frequency band to extract the characteristic values of the sound wave components, and the characteristic values constitute the acoustic fingerprint of the sound wave components. One implementation method may include:

[0061] Step 1: Conversion from time domain to frequency domain First, convert the original audio signal collected by the MEMS speaker from the time domain to the frequency domain. This can be achieved by applying the Fast Fourier Transform (FFT). The FFT converts the signal within a duration into a signal composed of frequencies, revealing the spectral characteristics of the signal.

[0062] Step 2: Frequency band segmentation of the frequency-domain signal Divide the frequency-domain signal into multiple frequency bands. The division can be based on prior knowledge of the required frequency resolution and signal characteristics, such as the basic frequency bands of sound, such as low-frequency, mid-frequency, and high-frequency bands. Each frequency band represents the sound wave information of different frequency components in the original audio signal.

[0063] Step 3: Calculate the energy spectral density In each divided frequency band, calculate the power spectral density (PSD) of the signal. PSD is a way to quantify the energy of the signal's spectral distribution, and it is directly related to the power of the signal in each frequency band.

[0064] Step 4: Feature extraction and acoustic fingerprint construction Use the calculated energy spectral density to extract key sound wave features in each frequency band, such as statistics such as peak values, average energy, skewness, and kurtosis of the energy distribution. Combine these characteristic values to construct the acoustic fingerprint of each frequency band. The acoustic fingerprints of several frequency bands together constitute the acoustic fingerprint of the entire original audio signal.

[0065] This acoustic fingerprint contains the characteristic spectra of the desired signal (such as human voice, music, etc.) and the original noise signal (such as traffic noise, wind noise, etc.) in the audio signal. The acoustic fingerprint captures the detailed characteristics of the audio signal and provides the necessary information for subsequent noise reduction processing.

[0066] In summary, the acoustic components in the audio signal can be dynamically distinguished and extracted, and the acoustic fingerprints representing these components can be constructed. This process can provide key information for subsequent noise control and denoising processing, thereby realizing the intelligent denoising function of the MEMS speaker.

[0067] S203, deconstruct the extracted acoustic fingerprint and generate a denoising control vector for the original noise signal;

[0068] Specifically, the set of eigenvalue of the acoustic components included in the acoustic fingerprint can be classified to obtain the desired signal characteristics and the original noise signal characteristics; the noise signal characteristics are combined to construct a one-dimensional or multi-dimensional noise signal model representing the power spectral density of the actual noise signal; a denoising control vector is created using the noise signal model, and the denoising control vector is used to indicate the amount of energy that needs to be suppressed at each frequency. One implementation may include:

[0069] Step 1: Acoustic fingerprint classification and eigenvalue set In the already obtained acoustic fingerprint, machine learning methods (such as support vector machine (SVM), neural network or decision tree) are used to classify the acoustic fingerprint. First, it is necessary to learn from a large number of labeled training data sets. In this data set, the acoustic fingerprint of each sample is associated with its corresponding signal type (desired signal or noise signal) label. The classification model can distinguish and output the desired signal characteristics and the original noise signal characteristics.

[0070] Step 2: Establish a noise signal model The original noise signal eigenvalue identified in the classification process is combined, and a one-dimensional or multi-dimensional noise signal model is constructed by statistical methods (such as principal component analysis (PCA) or clustering analysis). This model depicts the power spectral density distribution of the actual noise signal and can simulate the noise signal in different environments.

[0071] Step 3: Generate a denoising control vector Using the established noise signal model, a denoising control vector is created through an optimization algorithm (such as genetic algorithm, particle swarm optimization or Bayesian optimization). The denoising control vector specifically defines the energy level and phase that need to be suppressed in each frequency component of the audio signal. This step needs to consider the dynamic range of the signal and the sensitivity of the human ear to noise at different frequencies, so as to minimize the impact of the noise signal without affecting the desired signal.

[0072] Step 4: Optimize the control vector According to the changes in the environmental noise and the desired signal, the denoising control vector is updated in real time. This step uses adaptive filtering techniques (such as the least mean square error (LMS) algorithm) to ensure that as the environmental noise changes dynamically, the denoising control vector can adapt in real time to achieve the best denoising effect.

[0073] Step 5: Verification and feedback Finally, apply the denoising control vector to the actual denoising process and evaluate the denoising effect through the microphone array in the environment. If the denoising effect does not meet the expectation, use the feedback loop to adjust the parameters for generating the denoising control vector, and further optimize the model through iterative learning until the denoising performance requirements are met.

[0074] Through the above steps, not only can the acoustic fingerprint be deeply deconstructed to identify the key features of the noise signal, but also a denoising control vector can be intelligently generated and adjusted and optimized according to the actual effect, so as to provide an efficient and adaptive intelligent denoising solution for MEMS speakers.

[0075] S204, perform denoising processing on the original audio signal according to the denoising control vector;

[0076] Specifically, the denoising control vector can be used to generate a reverse acoustic wave signal that matches the power spectral density of the noise signal as the denoising signal; perform denoising processing on the original audio signal based on the denoising signal, where the denoising signal is opposite in phase to the noise signal in the original audio signal. One implementation may include:

[0077] Step 1: Denoising signal synthesis Synthesize the denoising signal according to the already generated denoising control vector. This includes generating a series of reverse acoustic wave signals that match the power spectral density of the original noise signal. Specifically, for each frequency band, use the energy magnitude and phase indicated by the control vector to generate a series of sine wave signals with opposite phases, so as to cancel the energy of the corresponding frequency band in the original noise.

[0078] Step 2: Time alignment of the denoising signal In order to effectively cancel the interference between the denoising signal and the noise part in the original audio signal, time alignment is required. Use digital signal processing techniques, such as delay filters, to adjust the time delay of the denoising signal to ensure its synchronization with the original noise signal in the time domain.

[0079] Step 3: Adaptive denoising processing Add the synthesized denoising signal to the original audio signal. Since the denoising signal is in antiphase with the noise component in the original audio signal, the addition operation in this step actually cancels the noise part in the corresponding frequency band. The amplitude of the denoising signal should be adjusted according to the control vector to achieve an appropriate noise cancellation level.

[0080] Step 4: Real-time feedback adjustment Monitor the denoised audio signal in real time through the microphone array, and compare its analysis result with the original audio signal. If the expected noise cancellation effect is not achieved, adjust the denoising control vector and resynthesize the denoising signal, which may involve changing the denoising energy level of a specific frequency band or adjusting the phase of the denoising signal to achieve more refined noise cancellation.

[0081] Step 5: Output the processed signal. The adjusted denoised signal is output through the MEMS speaker to effectively reduce the noise level in the environment. Through the precise control of the MEMS speaker, the denoised signal can be accurately played to ensure the maximization of the noise cancellation effect.

[0082] Through the above steps, combined with the denoising control vector, the original audio signal can be denoised in real time and effectively, adapt to different noise environments, achieve high-quality sound output, and protect the integrity of the desired signal.

[0083] S205, calculate the acoustic wave kinetic energy of the environmental noise signal, and generate a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy.

[0084] Specifically, the environmental noise signal can be transformed from the time domain to the frequency domain to obtain the spectral distribution of the environmental noise signal; according to the spectral distribution, calculate the power spectral density of the environmental noise signal at each frequency point; integrate the power spectral density within all or selected frequency ranges to obtain the total acoustic wave kinetic energy of the environmental noise signal; identify specific high-energy frequency components or characteristic frequency components in the power spectral density, and use the identified frequency components and total acoustic wave kinetic energy data to generate a sound pressure level map, where the sound pressure level map describes the sound pressure levels of each frequency component; according to the sound pressure level map, synthesize a filtering signal with the same amplitude but opposite phase corresponding to the sound pressure level of the specific frequency component as the reverse acoustic wave kinetic energy signal. One implementation method may include:

[0085] Step 1: Frequency domain conversion of the environmental noise signal. First, use the Fourier transform (such as the fast Fourier transform FFT) to convert the environmental noise signal from the time domain to the frequency domain, so as to obtain the spectral distribution of the environmental noise signal. This can convert the continuous or discrete environmental noise time domain signal into a set of frequency domain signals that describe its frequency components and corresponding energies.

[0086] Step 2: Calculation of the power spectral density. Then, use the obtained spectral distribution to calculate the power spectral density of the environmental noise signal at each frequency point. The calculation of the power spectral density reveals the contribution of each frequency component to the total energy of the signal. Here, Parseval's theorem can be applied to extract energy information from the frequency domain signal. Specifically, analyze the amplitudes at each frequency point and square them to obtain the power spectral density.

[0087] Step 3: Acoustic Wave Kinetic Energy Integration Subsequently, select all or a specific frequency range and integrate the power spectral density to calculate the total acoustic wave kinetic energy of the ambient noise signal. At this time, the frequency range of integration can be restricted to focus on those frequency components that are sensitive to the human ear or significantly interfere with the sound quality. For example, based on the calculated power spectral density, frequency points with high energy within the selected frequency range can be identified by setting a threshold. These frequency points are the key targets for noise reduction processing because they represent the most significant interfering components in the ambient noise. Integrate the identified high-energy frequency points to calculate the total acoustic wave kinetic energy of these points. This step is achieved by accumulating the power spectral density within the selected frequency range.

[0088] The total acoustic wave kinetic energy data can provide a quantitative baseline for the overall level of ambient noise. This baseline is crucial for evaluating the effectiveness of noise reduction or control because it can serve as a reference point for the success of subsequent intervention measures.

[0089] Step 4: Synthesize the Reverse Acoustic Wave Kinetic Energy Signal Based on the obtained total acoustic wave kinetic energy and the identified frequency components, synthesize a sound pressure level map that describes the sound pressure levels of each frequency component. Use the power spectral density values in the identified frequency components and their corresponding total acoustic wave kinetic energies to convert the energy of each frequency component into sound pressure level (dB). The formula for sound pressure level is:

[0090] L = 10*log_{10}(frac{P}{P_0})

[0091] where L is the sound pressure level (unit: decibel), P is the sound power corresponding to the power spectral density, and P_0 is the reference sound power. Generate a sound pressure level map based on the sound pressure level of each frequency component. This chart describes the sound pressure levels of each frequency component in the entire collected ambient noise and provides the necessary data basis for the next step of synthesizing the reverse acoustic wave kinetic energy signal. Use this sound pressure level map to generate the corresponding reverse acoustic wave kinetic energy signal, that is, for each significant frequency component in the spectrum, synthesize a filtered signal with the same sound pressure level but opposite phase. Specifically, this step can be achieved by adding a phase shift of π (180 degrees) to the synthesized signal. The resulting reverse acoustic wave will produce a destructive interference effect when it meets the original noise signal.

[0092] Step 5: Optimization of the Reverse Acoustic Wave Kinetic Energy Signal Considering environmental changes and the characteristics of MEMS speakers, the reverse acoustic wave kinetic energy signal may need to be optimized to adapt to the actual application conditions. Optimization can be based on real-time monitoring data or a preset environmental model to adjust the frequency components, amplitude, and phase of the filtered signal to ensure the maximization of the noise reduction effect.

[0093] Step 6: Output the denoised signal. Convert the optimized reverse acoustic wave kinetic energy signal into an electrical signal and output it through the MEMS speaker. Due to the physical properties of sound waves, when the reverse acoustic wave kinetic energy signal is mixed with ambient noise, it will effectively reduce or even eliminate the noise, thus achieving the purpose of noise reduction.

[0094] Through the above steps, not only the acoustic wave kinetic energy of the ambient noise is accurately calculated, but also an effective anti-phase noise reduction signal can be created based on the measured energy data, and intelligent noise reduction of the ambient noise can be achieved through the MEMS speaker.

[0095] S206, convert the reverse acoustic wave kinetic energy signal into an electrical signal and convert and output it through the MEMS speaker to achieve noise reduction of the ambient noise signal.

[0096] Specifically, one implementation method may include:

[0097] Step 1: Synthesis of the reverse acoustic wave kinetic energy signal. Based on the sound pressure level map generated in the previous step, first, a filtered signal with a reverse phase needs to be synthesized for each specific frequency component. The amplitude of these filtered signals should be equal to the sound pressure level of the corresponding frequency component in the original noise signal, but with the opposite phase. To achieve this, digital signal processing software (such as the signal processing toolbox in MATLAB or Python) can be used for precise filtered signal design. During the design process, the key is to adjust the phase of each filtered signal to be exactly opposite to the phase of the corresponding frequency component in the original noise signal.

[0098] Step 2: Synthesis of the digital filtered signal. Synthesize all the designed reverse filtered signals in the digital domain to form a complete reverse acoustic wave kinetic energy signal. This step ensures that the signals of all frequency components can be accurately superimposed in the time domain to form a continuous digital signal. This continuous signal will be used to interfere with the noise signal in the environment, thus achieving the noise reduction effect.

[0099] Step 3: Digital-to-analog conversion. Convert the synthesized digital reverse acoustic wave kinetic energy signal into an analog signal through a digital-to-analog converter (DAC). During this process, it is necessary to ensure that the sampling rate and conversion rate of the signal are high enough to ensure that the converted analog signal can faithfully reflect all the details of the digital signal and avoid introducing additional distortion or noise.

[0100] Step 4: Signal amplification and filtering. Before sending the signal to the MEMS speaker, amplify the analog signal through an amplifier to ensure that the signal has enough power to drive the speaker. At the same time, an analog filter can be used to remove the high-frequency noise that may be introduced during the DAC conversion process to ensure the purity and effectiveness of the output signal.

[0101] Step 5: Output of the MEMS Speaker Finally, the processed analog signal is sent to the MEMS speaker. The MEMS speaker converts the electrical signal into an acoustic signal, outputs the corresponding reverse sound wave, and meets the original noise signal in the environment. Since the reverse sound wave signal is opposite in phase to the original noise signal, the two will cancel each other out when they meet in space, thus achieving the noise reduction effect. Among them, it should be noted that:

[0102] - Signal Synchronization: When synthesizing the reverse acoustic wave kinetic energy signal, ensuring synchronization with the environmental noise signal is crucial. Any time delay may reduce the noise reduction effect.

[0103] - Signal Accuracy: The accuracy in the digital-to-analog conversion process is of vital importance. It is necessary to ensure that the converted analog signal can accurately reflect the characteristics of the digital signal.

[0104] - Speaker Performance: The performance limitations of the MEMS speaker (such as the frequency response range and the maximum output power) will directly affect the noise reduction effect. Therefore, these factors must be considered when designing the noise reduction signal.

[0105] Through the above steps, the process of converting the reverse acoustic wave kinetic energy signal into an electrical signal and outputting it through the MEMS speaker can be realized, thereby effectively achieving the noise reduction of the environmental noise signal. This method combines modern digital signal processing technology and traditional acoustic principles, providing an efficient and accurate intelligent noise reduction solution.

[0106] It can be seen that by collecting the original audio signal emitted by the MEMS speaker and the environmental noise signal data; extracting the acoustic fingerprint of the original audio signal to dynamically distinguish the acoustic wave components in the audio signal; deconstructing the extracted acoustic fingerprint and generating the noise reduction control vector of the noise signal; performing noise reduction processing on the original audio signal according to the noise reduction control vector; calculating the acoustic wave kinetic energy of the environmental noise signal and generating the corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy; converting the reverse acoustic wave kinetic energy signal into an electrical signal and converting and outputting it through the MEMS speaker to achieve the noise reduction of the environmental noise signal, so as to accurately capture the characteristics of the noise signal and generate a more effective noise reduction signal to improve the real-time performance and adaptability of the intelligent noise reduction method.

[0107] Another embodiment of the present invention provides an intelligent noise reduction system for a MEMS speaker. Refer to Figure 3 , the system may include:

[0108] A collection module 301, configured to collect the original audio signal emitted by the MEMS speaker and the environmental noise signal data;

[0109] An extraction module 302, configured to extract an acoustic fingerprint from the original audio signal to dynamically distinguish the sound wave components in the audio signal;

[0110] A generation module 303, configured to deconstruct the extracted acoustic fingerprint and generate a denoising control vector for the noise signal;

[0111] A first denoising module 304, configured to perform denoising processing on the original audio signal according to the denoising control vector;

[0112] A calculation module 305, configured to calculate the sound wave kinetic energy of the environmental noise signal and generate a corresponding reverse sound wave kinetic energy signal according to the sound wave kinetic energy;

[0113] A second denoising module 306, configured to convert the reverse sound wave kinetic energy signal into an electrical signal and output it through the MEMS speaker to achieve denoising of the environmental noise signal.

[0114] It can be seen that by collecting the original audio signal and environmental noise signal data emitted by the MEMS speaker; extracting an acoustic fingerprint from the original audio signal to dynamically distinguish the sound wave components in the audio signal; deconstructing the extracted acoustic fingerprint and generating a denoising control vector for the noise signal; performing denoising processing on the original audio signal according to the denoising control vector; calculating the sound wave kinetic energy of the environmental noise signal and generating a corresponding reverse sound wave kinetic energy signal according to the sound wave kinetic energy; converting the reverse sound wave kinetic energy signal into an electrical signal and outputting it through the MEMS speaker to achieve denoising of the environmental noise signal, it is possible to accurately capture the characteristics of the noise signal and generate a more effective denoising signal, so as to improve the real-time performance and adaptability of the intelligent denoising method.

[0115] An embodiment of the present invention further provides a storage medium, in which a computer program is stored, and the computer program is configured to execute the steps in any one of the above method embodiments when running.

[0116] Specifically, in this embodiment, the above storage medium may be configured to store a computer program for executing the following steps:

[0117] S201, collect the original audio signal and environmental noise signal data emitted by the MEMS speaker;

[0118] S202, extract an acoustic fingerprint from the original audio signal to dynamically distinguish the sound wave components in the audio signal;

[0119] S203, deconstruct the extracted acoustic fingerprint and generate a denoising control vector for the noise signal;

[0120] S204, perform noise reduction processing on the original audio signal according to the noise reduction control vector;

[0121] S205, calculate the acoustic wave kinetic energy of the environmental noise signal, and generate a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy;

[0122] S206, convert the reverse acoustic wave kinetic energy signal into an electrical signal and convert and output it through the MEMS speaker to achieve noise reduction of the environmental noise signal.

[0123] It can be seen that by collecting the original audio signal and environmental noise signal data emitted by the MEMS speaker; performing acoustic fingerprint extraction on the original audio signal to dynamically distinguish the acoustic wave components in the audio signal; deconstructing the extracted acoustic fingerprints and generating a noise reduction control vector for the noise signal; performing noise reduction processing on the original audio signal according to the noise reduction control vector; calculating the acoustic wave kinetic energy of the environmental noise signal and generating a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy; converting the reverse acoustic wave kinetic energy signal into an electrical signal and converting and outputting it through the MEMS speaker to achieve noise reduction of the environmental noise signal, so as to accurately capture the characteristics of the noise signal and generate a more effective noise reduction signal to improve the real-time performance and adaptability of the intelligent noise reduction method.

[0124] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0125] Specifically, the above electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the above processor, and the input / output device is connected to the above processor.

[0126] Specifically, in this embodiment, the above processor may be configured to execute the following steps through a computer program:

[0127] S201, collect the original audio signal and environmental noise signal data emitted by the MEMS speaker;

[0128] S202, perform acoustic fingerprint extraction on the original audio signal to dynamically distinguish the acoustic wave components in the audio signal;

[0129] S203, deconstruct the extracted acoustic fingerprints and generate a noise reduction control vector for the noise signal;

[0130] S204, perform noise reduction processing on the original audio signal according to the noise reduction control vector;

[0131] S205, calculate the acoustic wave kinetic energy of the environmental noise signal, and generate a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy;

[0132] S206, convert the reverse acoustic wave kinetic energy signal into an electrical signal and convert and output it through the MEMS speaker to achieve noise reduction of the environmental noise signal.

[0133] Specifically, for the specific examples in this embodiment, reference may be made to the examples described in the above embodiments and optional implementation manners, and details are not described herein again.

[0134] It can be seen that by collecting the original audio signal and environmental noise signal data emitted by the MEMS speaker; extracting the acoustic fingerprint of the original audio signal to dynamically distinguish the acoustic wave components in the audio signal; deconstructing the extracted acoustic fingerprint and generating a noise reduction control vector for the noise signal; performing noise reduction processing on the original audio signal according to the noise reduction control vector; calculating the acoustic wave kinetic energy of the environmental noise signal, and generating a corresponding reverse acoustic wave kinetic energy signal according to the acoustic wave kinetic energy; converting the reverse acoustic wave kinetic energy signal into an electrical signal and converting and outputting it through the MEMS speaker to achieve noise reduction of the environmental noise signal, so as to accurately capture the characteristics of the noise signal and generate a more effective noise reduction signal, thereby improving the real-time performance and adaptability of the intelligent noise reduction method.

[0135] The structure, features and effects of the present invention have been described in detail based on the embodiments shown in the drawings. The above are only the preferred embodiments of the present invention, but the present invention is not limited to the scope defined by the drawings. Any changes made according to the concept of the present invention, or equivalent embodiments modified into equivalent changes, still fall within the spirit covered by the specification and the drawings, and should be within the protection scope of the present invention.

Claims

1. An intelligent denoising method for a MEMS speaker, characterized in that: The method comprises: Collect the original audio signal and environmental noise signal data emitted by the MEMS speaker; The acoustic fingerprint of the original audio signal is extracted to dynamically distinguish the sound wave components in the audio signal, wherein the sound wave components include characteristic spectra of the desired signal and the original noise signal; wherein the time domain signal corresponding to the original audio signal is converted into a frequency domain signal; the frequency domain signal is divided into a plurality of frequency bands, wherein each frequency band represents a portion of the frequency components in the signal; the energy spectral density is calculated in each frequency band to quantify the energy of the signal in the frequency band; the energy spectral density of each frequency band is analyzed to extract the characteristic values ​​of the sound wave components, wherein the characteristic values ​​constitute the acoustic fingerprint of the sound wave components; Deconstructing the extracted acoustic fingerprint and generating a denoising control vector of the original noise signal; Performing denoising processing on the original audio signal according to the denoising control vector; Calculating the acoustic wave energy of the environmental noise signal, and generating a corresponding reverse acoustic wave energy signal according to the acoustic wave energy; The reverse acoustic wave energy signal is converted into an electrical signal and output through the MEMS speaker to achieve denoising of the environmental noise signal.

2. The method according to claim 1, characterized in that The step of deconstructing the extracted acoustic fingerprint and generating a denoising control vector of the original noise signal comprises: Classifying the feature value set of the sound wave components contained in the acoustic fingerprint to obtain expected signal features and original noise signal features; Combining the original noise signal features to construct a one-dimensional or multi-dimensional noise signal model representing the power spectral density of the actual noise signal; A denoising control vector is created using the noise signal model, the denoising control vector being used to indicate the amount of energy that needs to be suppressed at each frequency.

3. The method according to claim 2, characterized in that The performing denoising processing on the original audio signal according to the denoising control vector comprises: Using the denoising control vector to generate an inverse acoustic wave signal that matches the power spectrum density of the noise signal as a denoising signal; The original audio signal is subjected to denoising processing based on the denoising signal, wherein the denoising signal is opposite in phase to an original noise signal in the original audio signal.

4. The method according to claim 3, characterized in that The calculating the acoustic wave energy of the environmental noise signal and generating a corresponding reverse acoustic wave energy signal according to the acoustic wave energy comprises: Converting the ambient noise signal from the time domain to the frequency domain to obtain a frequency spectrum distribution of the ambient noise signal; According to the frequency spectrum distribution, the power spectrum density of the environmental noise signal at each frequency point is calculated; Integrate the power spectrum density in all or selected frequency ranges to obtain the total acoustic wave energy of the ambient noise signal; Identify high energy frequency components or characteristic frequency components in the power spectrum density, and generate a sound pressure level diagram using the identified frequency components and total sound wave energy data, wherein the sound pressure level diagram describes the sound pressure level of each frequency component; According to the sound pressure level graph, a filtered signal corresponding to the sound pressure level of the high-energy frequency component or the characteristic frequency component and having the same amplitude but opposite phase is synthesized as a reverse acoustic wave energy signal.

5. An intelligent denoising system for a MEMS speaker, characterized in that: The system comprises: An acquisition module is used to collect the original audio signal and environmental noise signal data emitted by the MEMS speaker; An extraction module is used to extract acoustic fingerprints from the original audio signal to dynamically distinguish the sound wave components in the audio signal, wherein the sound wave components include characteristic spectra of the desired signal and the original noise signal; wherein the time domain signal corresponding to the original audio signal is converted into a frequency domain signal; the frequency domain signal is divided into a plurality of frequency bands, wherein each frequency band represents a portion of the frequency components in the signal; the energy spectral density is calculated in each frequency band to quantify the energy of the signal in the frequency band; the energy spectral density of each frequency band is analyzed to extract characteristic values ​​of the sound wave components, wherein the characteristic values ​​constitute the acoustic fingerprints of the sound wave components; A generating module, used for deconstructing the extracted acoustic fingerprint and generating a denoising control vector of the original noise signal; A first denoising module, configured to perform denoising processing on the original audio signal according to the denoising control vector; A calculation module, used for calculating the acoustic wave energy of the environmental noise signal, and generating a corresponding reverse acoustic wave energy signal according to the acoustic wave energy; The second denoising module is used to convert the reverse acoustic wave energy signal into an electrical signal and output the converted signal through the MEMS speaker to achieve denoising of the environmental noise signal.

6. The system according to claim 5, characterized in that The generating module is specifically used for: Classifying the feature value set of the sound wave components contained in the acoustic fingerprint to obtain expected signal features and original noise signal features; Combining the original noise signal features to construct a one-dimensional or multi-dimensional noise signal model representing the power spectral density of the actual noise signal; A denoising control vector is created using the noise signal model, the denoising control vector being used to indicate the amount of energy that needs to be suppressed at each frequency.

7. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 4 when executed.

8. An electronic device, comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 4.

Images