Square dance directional noise reduction method and system based on self-sound cancellation

Abstract

This invention discloses a method and system for directional noise reduction in square dancing based on self-voice cancellation, belonging to the field of sound recognition technology. The method includes deploying an environmental microphone behind a directional speaker, playing a test signal, measuring and storing the sound field transfer function behind the speaker; during normal operation, real-time acquisition of the total noise signal behind the speaker and the currently played audio signal; convolving the audio signal with the time-domain impulse response corresponding to the transfer function to obtain the instantaneous sound pressure contribution value generated by the speaker behind the speaker; subtracting this contribution value from the total noise to obtain the true environmental noise signal; calculating the equivalent continuous sound pressure level and comparing it with a threshold, and performing graded noise reduction operations according to the degree of exceedance; the system includes a directional speaker body, an environmental microphone, a signal processing unit, a subtraction module, a comparison control unit, and a graded noise reduction execution unit. This invention can accurately separate self-voice leakage from environmental background noise, avoiding false triggering and missed judgments, and achieving precise, automatic, and traceable noise control for square dancing.

Description

Technical Field

[0001] This invention relates to the field of sound recognition technology, and in particular to a method and system for directional noise reduction in square dancing based on self-voice cancellation. Background Technology

[0002] The directional speakers used in square dancing employ phased array beamforming technology to control the sound within a specific angular range (e.g., 30° to 120° in front), significantly reducing sound intensity from the rear and sides. However, in actual use, due to factors such as ground reflection, building diffraction, and installation errors, a certain amount of sound leakage still exists behind the directional speakers. Actual measurements show that the sound pressure level 2 meters behind the directional speakers can still reach approximately 80 dB, which may exceed the noise limits for residential areas at night in quiet environments as stipulated in the national "Environmental Noise Quality Standard" (GB 3096-2008) (generally 45 dB to 50 dB).

[0003] To monitor whether the noise level behind a directional speaker exceeds the limit, existing technologies mainly focus on hardware improvements. For example, optimizing beamforming algorithms to increase the sound pressure difference between the front and rear, or adding sound-absorbing materials behind the device to absorb leaked sound. However, these hardware improvements fail to address a fundamental problem: when monitoring noise behind a directional speaker, the total noise signal collected contains two components—the component of sound leaked from the speaker itself, and the actual ambient background noise (such as vehicle traffic, human voices, and wind noise). Because these two components highly overlap in both the frequency and time domains, existing technologies cannot accurately separate the speaker's own sound leakage component from the total noise signal. Therefore, it is impossible to determine whether the excessive noise level behind the speaker is caused by its own sound leakage or by ambient noise.

[0004] If the two components mentioned above cannot be distinguished, two types of misjudgments may occur: First, the sound leakage itself may not exceed the standard, but the total noise exceeds the standard after being combined with environmental noise. The system may trigger unnecessary noise reduction operations, affecting the dancer's experience. Second, the environmental noise may exceed the standard (such as the noise from surrounding roads and vehicles), but the sound leakage itself may be small. The system may fail to identify the real risk of disturbing the peace and continue to play normally, causing continuous interference to the surrounding residents.

[0005] Therefore, there is an urgent need for a directional noise reduction method for square dancing that can accurately separate and eliminate its own sound leakage components from the total noise signal behind it and respond only to real environmental noise. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for directional noise reduction in square dancing based on self-voice cancellation, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, this invention provides a method for directional noise reduction in square dancing based on self-sound cancellation, applicable to directional speakers used in square dancing, comprising the following steps: S1. After the installation of the directional sound system for square dancing is completed, an environmental microphone is deployed behind the directional sound system to play test signals. The response signals are collected through the environmental microphones. The rear sound field transfer function is extracted by the cross-correlation method or the spectrum division method, and the rear sound field transfer function and its corresponding time-domain impulse response are stored. S2. During normal operation, the total noise signal behind the ambient microphone is collected in real time, and the audio signal currently played by the square dance directional speaker is also acquired. S3. Convolve the currently playing audio signal with the time-domain impulse response corresponding to the rear sound field transfer function to calculate the instantaneous sound pressure contribution value generated by the sound itself in the rear. S4. Subtract the instantaneous sound pressure contribution value from the total noise signal at the rear to obtain the real environmental noise signal; S5. Calculate the equivalent continuous sound pressure level of the real environmental noise signal and compare it with the preset noise threshold. S6. When the equivalent continuous sound pressure level exceeds the noise threshold, graded noise reduction operation is performed according to the degree of exceedance. The graded noise reduction operation includes reducing the output power of the square dance directional speaker, switching to low frequency cut-off mode, or forced mute.

[0008] Preferably, the specific steps for S1 to extract the rear sound field transfer function using the cross-correlation method include: Calculate the cross-correlation function between the test signal and the response signal acquired by the environmental microphone. When the test signal is white noise or a maximum-length sequence, the cross-correlation function is approximated as the impulse response of the rear sound field. The calculation formula is as follows: ; in, For test signals, The response signal collected by the environmental microphone. For time variables, For time delay variables, for and The cross-correlation function; Performing a Fourier transform on the impulse response yields the rear acoustic field transfer function, calculated as follows: ; in, For the rear sound field transfer function, For frequency, It is the imaginary unit.

[0009] Preferably, the specific steps for S1 to extract the rear acoustic field transfer function using the spectrum division method include: Perform Fourier transforms on the test signal and the response signal respectively to obtain the input spectrum. and output spectrum ; The rear sound field transfer function can be directly calculated using the ratio of the output spectrum to the input spectrum. The calculation formula is as follows: ; Introducing regularization parameters Suppress noise; ; in, for The complex conjugate, Power spectrum, The regularization parameter is greater than 0.

[0010] Preferably, the formula for calculating the instantaneous sound pressure contribution value obtained through convolution operation in S3 is as follows: ; in, For discrete-time indexing, The contribution value of instantaneous sound pressure. This is a finite impulse response sequence of the rear sound field. For tap index, Let be the length of the finite impulse response sequence. For the currently playing audio signal at time The sampled values.

[0011] Preferably, the formula for calculating the equivalent continuous sound pressure level in S5 is as follows: ; in, Equivalent continuous sound pressure level, in decibels; It is a discrete sequence of real environmental noise signals; For reference sound pressure level; The calculation involves counting the number of sampling points within the window, and a sliding time window is used with a length of 1 second, updating the calculation results every 0.2 seconds.

[0012] Preferably, S6 performs graded noise reduction operations based on the degree of noise exceeding the limit, specifically including: Determine the noise reduction amount using the following formula: ; in, To exceed the limit, For equivalent continuous sound pressure level, The preset noise threshold, The maximum allowable attenuation, and Preset parameters and ; The output sound pressure level after noise reduction is , This represents the current output sound pressure level. Furthermore, when the noise level exceeds 3 dB, only the output power is reduced; when the noise level exceeds 3 dB but does not exceed 6 dB, the system switches to low-frequency cut-off mode while reducing noise, cutting off the frequency band below 100 Hz; when the noise level exceeds 6 dB, the system forces silence for 30 seconds and sends an alarm message to the management terminal via the wireless network.

[0013] Preferably, it also includes hysteresis comparison and anti-shake processing: noise threshold Configured as a comparator with hysteresis, the threshold for triggering noise reduction is [value missing]. The threshold for stopping noise reduction is ,in ; Furthermore, noise reduction is triggered only when the equivalent continuous sound pressure level exceeds the activation threshold three times consecutively; the original output power is restored only when the equivalent continuous sound pressure level falls below the deactivation threshold three times consecutively.

[0014] A directional noise reduction system for square dancing based on self-voice cancellation includes: The main body of the directional speaker for square dancing; An environmental microphone is fixedly installed behind the directional speakers used for square dancing to collect the total noise signal from behind. The signal processing unit is connected to the audio input terminals of the environmental microphone and the directional speaker for square dancing, respectively. It is used to store the pre-measured rear sound field transfer function and its corresponding time-domain impulse response, acquire the currently playing audio signal, and convolve the audio signal with the time-domain impulse response to calculate the instantaneous sound pressure contribution value generated by its own sound in the rear. The subtraction module is used to subtract the instantaneous sound pressure contribution value from the total noise signal and output the real environmental noise signal. The comparison control unit is used to calculate the equivalent continuous sound pressure level of the real environmental noise signal and compare it with a preset noise threshold, and output a noise reduction control signal based on the comparison result. The graded noise reduction execution unit is used to reduce output power, switch low-frequency cut-off mode, or force mute according to the noise reduction control signal.

[0015] Preferably, there are at least two environmental microphones, which are installed at different horizontal positions and different heights at the rear; each environmental microphone channel has an independent transfer function and subtraction module, and the final real environmental noise signal is the maximum value or weighted average of the results of each channel.

[0016] Preferably, the graded noise reduction execution unit is linked to the locking mechanism of the controller of the square dance directional speaker: when forced mute is triggered, the system automatically locks the controller to prohibit unauthorized volume increase operations; unlocking permission requires authorization signals from the square dance leader, security guard and management office at the same time before the mute state can be lifted; it also includes a wireless communication module, which is used to encrypt and upload the trigger time, exceedance level, noise reduction level executed, recovery time of each noise reduction operation and the noise sample collected by the environmental microphone to the cloud monitoring platform.

[0017] Therefore, the present invention employs the above-mentioned method and system for directional noise reduction in square dancing based on self-voice cancellation, which has the following beneficial effects: (1) By pre-measuring the rear sound field transfer function, the present invention subtracts the contribution value of its own sound leakage from the total rear noise signal, which can accurately distinguish the self-sound component and the environmental background noise component in the total rear noise, avoid false triggering of noise reduction due to the superposition of self-sound and environmental noise, and also avoid missed judgment due to the environmental noise exceeding the standard while the self-sound is small, thus improving the accuracy of noise monitoring.

[0018] (2) The present invention adopts a nonlinear graded noise reduction strategy, which automatically selects power reduction, low frequency cut-off or forced mute according to the degree of excess, and is combined with hysteresis comparison and anti-shake processing. While ensuring the dancer's listening experience, it effectively suppresses the excessive noise behind, so that the noise behind and around the directional sound source meets the national noise control standards for acoustic functional areas.

[0019] (3) This invention links noise reduction execution with the playback controller locking mechanism. After forced mute, it can only be unlocked with authorization from three parties, avoiding the possibility of unauthorized volume restoration. At the same time, the noise reduction event is encrypted and uploaded to the cloud, providing a traceable evidence chain for law enforcement supervision.

[0020] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0021] Figure 1 This is a flowchart of a square dance directional noise reduction method based on self-voice cancellation according to the present invention; Figure 2 This is a block diagram of a square dance directional noise reduction system based on self-voice cancellation according to the present invention. Detailed Implementation

[0022] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely illustrates selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0023] Example like Figure 1 As shown, this invention provides a method for directional noise reduction in square dancing based on self-voice cancellation, including the following steps: S1. After the installation of the directional speakers for square dancing is completed, an environmental microphone is deployed behind the speakers to play a test signal. The response signal is collected through the environmental microphone, and the rear sound field transfer function is extracted using the cross-correlation method or the spectrum division method. The rear sound field transfer function and its corresponding time-domain impulse response are stored. The purpose of this step is to pre-measure the propagation characteristics of the sound field behind the directional speakers, providing an accurate reference model for subsequent elimination of sound leakage. The test signal usually uses a swept frequency signal, white noise, or a maximum length sequence, and the playback duration is generally 5 to 10 seconds to ensure the stability of the measurement results. The extracted rear sound field transfer function reflects the acoustic path from the speakers to the rear microphone position, including the combined effects of direct sound, reflected sound, and diffracted sound.

[0024] The specific steps for extracting the rear acoustic field transfer function using the cross-correlation method include: Calculate the cross-correlation function between the test signal and the response signal acquired by the environmental microphone. When the test signal is white noise or a maximum-length sequence, the cross-correlation function is approximated as the impulse response of the rear sound field. The calculation formula is as follows: ; in, For test signals, The response signal collected by the environmental microphone. For time variables, For time delay variables, for and The cross-correlation function; Performing a Fourier transform on the impulse response yields the subsequent sound field transfer function. The Fourier transform converts the time-domain impulse response to the frequency domain, facilitating subsequent matching with the audio signal's spectrum or use for time-domain convolution. The calculation formula is as follows: ; in, For the rear sound field transfer function, For frequency, It is the imaginary unit.

[0025] The advantages of the cross-correlation method are its simple calculation and strong anti-interference ability and high signal-to-noise ratio for white noise or MLS signals, making it particularly suitable for installation scenarios with low environmental noise.

[0026] The specific steps for extracting the rear acoustic field transfer function using the spectrum division method include: Perform Fourier transforms on the test signal and the response signal respectively to obtain the input spectrum. and output spectrum ; The rear sound field transfer function can be directly calculated using the ratio of the output spectrum to the input spectrum. The calculation formula is as follows: ; Introducing regularization parameters Suppress noise; ; in, for The complex conjugate, Power spectrum, The regularization parameter is greater than 0. The purpose of the regularization parameter is to avoid division by zero error when the power spectrum is close to zero, and at the same time suppress the influence of high-frequency noise on the transfer function estimation. It is usually taken as 0.01 to 0.001 times the maximum power spectrum value.

[0027] The spectrum division method is suitable for sweep frequency test signals, and can directly obtain the transfer function of the entire frequency band, with a relatively fast measurement speed.

[0028] S2. During normal operation, the total noise signal behind the ambient microphone is collected in real time, and the audio signal currently played by the directional speaker for square dancing is also acquired. The total noise signal collected here contains both self-sound leakage and ambient background noise. The currently played audio signal is acquired as a reference signal for subsequent estimation of the self-sound leakage. Real-time acquisition usually uses a sampling rate of 48kHz or 44.1kHz to ensure the integrity of the audio signal and the real-time performance of the algorithm.

[0029] S3. Convolve the currently played audio signal with the time-domain impulse response corresponding to the rear sound field transfer function to calculate the instantaneous sound pressure contribution of the sound itself at the rear. The convolution operation simulates the physical process of sound propagating from the speaker to the rear microphone. Since the propagation and reflection of sound waves in the air can be approximated as a linear time-invariant system, the instantaneous sound pressure value generated by the currently played audio at the rear microphone can be accurately calculated through convolution. This step quantifies the sound leakage of the sound itself from the total noise, providing a basis for subsequent subtraction operations.

[0030] The formula for calculating the instantaneous sound pressure contribution value through convolution operation is as follows: ; in, For discrete-time indexing, The contribution value of instantaneous sound pressure. This is a finite impulse response sequence of the rear sound field. For tap index, Let be the length of the finite impulse response sequence. For the currently playing audio signal at time The sampled values; The value of determines the effective length of the time-domain impulse response; in this embodiment, is taken as . =1024, corresponding to an impulse response length of approximately 21 milliseconds, sufficient to cover the main reflection paths of the rear sound field. Tap Index From 0 to -1 represents the contribution weight under different delays.

[0031] S4. Subtract the instantaneous sound pressure contribution value from the total noise signal at the rear to obtain the real environmental noise signal. This subtraction operation is performed point by point in the time domain. Since the instantaneous sound pressure contribution value is the optimal estimate of its own sound leakage, the remaining signal after subtraction is the real environmental background noise, such as vehicle driving sound, human voice, wind sound, etc. This achieves effective separation of one's own sound from environmental noise and solves the problem that existing technologies cannot distinguish between the two.

[0032] S5. Calculate the equivalent continuous sound pressure level of the real environmental noise signal and compare it with the preset noise threshold. The equivalent continuous sound pressure level is an average measure of noise energy over a period of time, which meets the requirements for noise monitoring in the national sound environment quality standards. By comparison, it can be determined whether the current real environmental noise has exceeded the legal limit, thereby deciding whether noise reduction operation needs to be performed.

[0033] The formula for calculating the equivalent continuous sound pressure level is as follows: ; in, Equivalent continuous sound pressure level, in decibels; It is a discrete sequence of real environmental noise signals; For reference sound pressure level; The number of sampling points within the calculation window is calculated. A sliding time window is used, with a length of 1 second and the calculation results are updated every 0.2 seconds. The sliding time window enables continuous monitoring of noise levels, while the 1-second window length ensures statistical stability and the 0.2-second update frequency makes the system response speed moderate, neither too sensitive nor too sluggish.

[0034] Lag comparison and anti-shake processing: noise threshold Configured as a comparator with hysteresis, the threshold for triggering noise reduction is [value missing]. The threshold for stopping noise reduction is ,in The purpose of hysteresis comparison is to prevent frequent triggering and recovery of noise reduction when noise fluctuates around the threshold. For example, if the activation threshold is 2dB higher than the deactivation threshold, noise reduction will only be activated when the noise level significantly exceeds the standard, and will only recover when the noise level drops significantly below the standard, thus forming a stable control range.

[0035] Furthermore, noise reduction is triggered only when the equivalent continuous sound pressure level exceeds the activation threshold three times consecutively; the original output power is restored only when the equivalent continuous sound pressure level falls below the deactivation threshold three times consecutively. This three-times confirmation is an anti-jitter measure that effectively filters out false triggers caused by occasional noise spikes. The duration of each of the three confirmations is approximately 0.6 seconds, which effectively filters out transient interference without causing excessive response delay.

[0036] S6. When the equivalent continuous sound pressure level exceeds the noise threshold, graded noise reduction operations are performed according to the degree of exceedance. Graded noise reduction operations include reducing the output power of the square dance directional speakers, switching to low-frequency cutoff mode, or forced mute. The purpose of graded noise reduction is to minimize the impact on the dancers' experience when the disturbance is minor, and to take strong measures when the disturbance is severe, taking into account both user experience and noise control requirements.

[0037] The specific steps for performing graded noise reduction based on the degree of noise exceedance include: Determine the noise reduction amount using the following formula: ; in, To exceed the limit, For equivalent continuous sound pressure level, The preset noise threshold, The maximum allowable attenuation, and Preset parameters and ; The output sound pressure level after noise reduction is , This represents the current output sound pressure level. Furthermore, when the noise level exceeds 3 dB, only the output power is reduced; when the noise level exceeds 3 dB but does not exceed 6 dB, the system switches to low-frequency cut-off mode while reducing noise, cutting off the frequency band below 100 Hz; when the noise level exceeds 6 dB, the system forces silence for 30 seconds and sends an alarm message to the management terminal via the wireless network.

[0038] The low-frequency cutoff mode attenuates the high-penetration low-frequency rhythms in square dance music, significantly reducing interference to residents, while preserving the mid-to-high frequency components so dancers can still hear the rhythm. A 30-second forced silence is a last resort, simultaneously notifying the administrator for manual intervention.

[0039] like Figure 2 As shown, a directional noise reduction system for square dancing based on self-voice cancellation includes: The directional speaker for square dancing adopts phased array beamforming technology, with a sound emission angle of 30° to 120° in front, a maximum sound pressure level of over 110dB, and a sound pressure difference of no less than 25dB between the front and rear, which forms the basis of the directional sound field.

[0040] An environmental microphone is fixedly installed behind the directional speakers used for square dancing to collect the total noise signal from the rear. At least two environmental microphones are used, installed at different horizontal and height positions behind the sound system. Each microphone channel has an independent transfer function and subtraction module; the final real environmental noise signal is the maximum value or weighted average of the results from each channel. Using multiple microphones overcomes the standing wave nodes or local obstruction problems that may be encountered in single-point measurements, improving the robustness of the monitoring. Different heights and horizontal positions can cover the spatial variations of the rear sound field; taking the maximum value ensures that the most unfavorable position does not exceed the limit, while the weighted average represents the average noise level of the entire rear area.

[0041] The signal processing unit is connected to the audio input terminals of the environmental microphone and the directional speaker for square dancing, respectively. It stores the pre-measured rear sound field transfer function and its corresponding time-domain impulse response, acquires the currently playing audio signal, and convolves the audio signal with the time-domain impulse response to calculate the instantaneous sound pressure contribution value generated by its own sound in the rear. The signal processing unit typically uses a DSP chip or embedded processor, possessing real-time digital signal processing capabilities. The convolution operation is implemented in the time domain using an FIR filter, resulting in high computational efficiency and low latency.

[0042] The subtraction module is used to subtract the instantaneous sound pressure contribution value from the total noise signal and output the real environmental noise signal. The subtraction module can be an arithmetic operation unit inside the DSP, which performs point-by-point subtraction operation through software instructions and outputs the result for use by the subsequent comparison and control unit.

[0043] The comparison control unit is used to calculate the equivalent continuous sound pressure level of the real environmental noise signal and compare it with a preset noise threshold, and output a noise reduction control signal based on the comparison result. The comparison control unit contains a sliding window integrator, a logarithmic calculator and a hysteresis comparator, which can calculate the equivalent continuous sound pressure level in real time and perform hysteresis comparison and anti-jitter logic, and output a trigger signal to the graded noise reduction execution unit.

[0044] The graded noise reduction execution unit is used to reduce output power, switch low-frequency cut-off mode, or force mute according to the noise reduction control signal. The graded noise reduction execution unit achieves graded noise reduction by controlling the power amplifier gain, digital equalizer, and mute switch. Its control command response time should be less than 100 milliseconds to ensure user experience and timely noise control.

[0045] The graded noise reduction execution unit is linked to the locking mechanism of the speaker controller for square dancing: when forced mute is triggered, the system automatically locks the controller, prohibiting unauthorized volume increases; unlocking requires authorization signals from the square dance leader, security guard, and management office simultaneously. This linkage mechanism combines automated noise control with manual management, preventing dancers from circumventing noise reduction by privately increasing the volume. The tripartite authorization ensures that no single group can unilaterally unmute the system, enhancing the fairness and effectiveness of management.

[0046] The wireless communication module is used to encrypt and upload the trigger time, exceedance level, noise reduction level, recovery time, and noise samples collected by the environmental microphone for each noise reduction operation to the cloud monitoring platform. The wireless communication module typically uses 4G, 5G, or WiFi and supports encrypted data transmission. The uploaded data can be used as law enforcement evidence for urban noise management, and can also be used to analyze the patterns of square dancing activities and optimize the site layout.

[0047] Therefore, the present invention adopts the above-mentioned method and system for directional noise reduction of square dancing based on self-sound elimination, which can accurately separate and eliminate self-sound leakage from the total noise signal behind, and only respond to the noise exceeding the standard of the real environment, avoiding false triggering and missed judgment, and realizing accurate, automatic and traceable square dance noise control.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for directional noise reduction in square dancing based on self-sound cancellation, applied to directional speakers for square dancing, characterized in that, Includes the following steps: S1. After the installation of the directional sound system for square dancing is completed, an environmental microphone is deployed behind the directional sound system to play test signals. The response signals are collected through the environmental microphones. The rear sound field transfer function is extracted by the cross-correlation method or the spectrum division method, and the rear sound field transfer function and its corresponding time-domain impulse response are stored. S2. During normal operation, the total noise signal behind the ambient microphone is collected in real time, and the audio signal currently played by the square dance directional speaker is also acquired. S3. Convolve the currently playing audio signal with the time-domain impulse response corresponding to the rear sound field transfer function to calculate the instantaneous sound pressure contribution value generated by the sound itself in the rear. S4. Subtract the instantaneous sound pressure contribution value from the total noise signal at the rear to obtain the real environmental noise signal; S5. Calculate the equivalent continuous sound pressure level of the real environmental noise signal and compare it with the preset noise threshold. S6. When the equivalent continuous sound pressure level exceeds the noise threshold, graded noise reduction operation is performed according to the degree of exceedance. The graded noise reduction operation includes reducing the output power of the square dance directional speaker, switching to low frequency cut-off mode, or forced mute.

2. The method for directional noise reduction in square dancing based on self-voice cancellation according to claim 1, characterized in that, The specific steps for S1 to extract the rear sound field transfer function using the cross-correlation method include: Calculate the cross-correlation function between the test signal and the response signal acquired by the environmental microphone. When the test signal is white noise or a maximum-length sequence, the cross-correlation function is approximated as the impulse response of the rear sound field. The calculation formula is as follows: ； in, For test signals, The response signal collected by the environmental microphone. For time variables, For time delay variables, for and The cross-correlation function; Performing a Fourier transform on the impulse response yields the rear acoustic field transfer function, calculated as follows: ； in, For the rear sound field transfer function, For frequency, It is the imaginary unit.

3. The method for directional noise reduction in square dancing based on self-voice cancellation according to claim 1, characterized in that, The specific steps for S1 to extract the rear acoustic field transfer function using the spectrum division method include: Perform Fourier transforms on the test signal and the response signal respectively to obtain the input spectrum. and output spectrum ; The rear sound field transfer function can be directly calculated using the ratio of the output spectrum to the input spectrum. The calculation formula is as follows: ； Introducing regularization parameters Suppress noise; ； in, for The complex conjugate, Power spectrum, The regularization parameter is greater than 0.

4. The method for directional noise reduction in square dancing based on self-voice cancellation according to claim 1, characterized in that, The formula for calculating the instantaneous sound pressure contribution value obtained through convolution operation in S3 is as follows: ； in, For discrete-time indexing, The contribution value of instantaneous sound pressure. This is a finite impulse response sequence of the rear sound field. For tap index, Let be the length of the finite impulse response sequence. For the currently playing audio signal at time The sampled values.

5. A method for directional noise reduction in square dancing based on self-voice cancellation as described in claim 1, characterized in that, The formula for calculating the equivalent continuous sound pressure level in S5 is as follows: ； in, Equivalent continuous sound pressure level, in decibels; It is a discrete sequence of real environmental noise signals; For reference sound pressure level; The calculation involves counting the number of sampling points within the window, and a sliding time window is used with a length of 1 second, updating the calculation results every 0.2 seconds.

6. A method for directional noise reduction in square dancing based on self-voice cancellation as described in claim 1, characterized in that, S6 performs graded noise reduction operations based on the degree of noise exceeding the limit, specifically including: Determine the noise reduction amount using the following formula: ； in, To exceed the limit, For equivalent continuous sound pressure level, The preset noise threshold, The maximum allowable attenuation, and Preset parameters and ; The output sound pressure level after noise reduction is , This represents the current output sound pressure level. Furthermore, when the noise level exceeds 3 dB, only the output power is reduced; when the noise level exceeds 3 dB but does not exceed 6 dB, the system switches to low-frequency cut-off mode while reducing noise, cutting off the frequency band below 100 Hz; when the noise level exceeds 6 dB, the system forces silence for 30 seconds and sends an alarm message to the management terminal via the wireless network.

7. A method for directional noise reduction in square dancing based on self-voice cancellation according to claim 1, characterized in that, It also includes hysteresis comparison and anti-shake processing: noise threshold Configured as a comparator with hysteresis, the threshold for triggering noise reduction is [value missing]. The threshold for stopping noise reduction is ,in ; Furthermore, noise reduction is triggered only when the equivalent continuous sound pressure level exceeds the activation threshold three times consecutively; the original output power is restored only when the equivalent continuous sound pressure level falls below the deactivation threshold three times consecutively.

8. A square dance directional noise reduction system based on self-voice cancellation, used in the square dance directional noise reduction method based on self-voice cancellation as described in any one of claims 1-7, characterized in that, include: The main body of the directional speaker for square dancing; An environmental microphone is fixedly installed behind the directional speakers used for square dancing to collect the total noise signal from behind. The signal processing unit is connected to the audio input terminals of the environmental microphone and the directional speaker for square dancing, respectively. It is used to store the pre-measured rear sound field transfer function and its corresponding time-domain impulse response, acquire the currently playing audio signal, and convolve the audio signal with the time-domain impulse response to calculate the instantaneous sound pressure contribution value generated by its own sound in the rear. The subtraction module is used to subtract the instantaneous sound pressure contribution value from the total noise signal and output the real environmental noise signal. The comparison control unit is used to calculate the equivalent continuous sound pressure level of the real environmental noise signal and compare it with a preset noise threshold, and output a noise reduction control signal based on the comparison result. The graded noise reduction execution unit is used to reduce output power, switch low-frequency cut-off mode, or force mute according to the noise reduction control signal.

9. A square dance directional noise reduction system based on self-voice cancellation according to claim 8, characterized in that: There are at least two environmental microphones, which are installed at different horizontal positions and different heights at the rear. Each environmental microphone channel has an independent transfer function and subtraction module. The final real environmental noise signal is the maximum value or weighted average of the results of each channel.

10. A square dance directional noise reduction system based on self-voice cancellation according to claim 8, characterized in that: The graded noise reduction execution unit is linked to the locking mechanism of the controller of the square dance directional speaker: when forced mute is triggered, the system automatically locks the controller to prevent unauthorized volume increase operations; unlocking permission requires authorization signals from the square dance leader, security guard and management office at the same time before the mute state can be lifted; it also includes a wireless communication module, which is used to encrypt and upload the trigger time, degree of exceedance, noise reduction level executed, recovery time of each noise reduction operation and the noise sample collected by the environmental microphone to the cloud monitoring platform.

Images