Confidential place sound interference device

Abstract

The invention relates to the technical field of information security, and discloses a secret-related place sound interference device, which comprises the following steps of: performing voice feature extraction on a target voice signal acquired from a secret-related place; designing an interference signal spectrum based on the extracted speech features; performing phase modulation on the interference signal frequency spectrum; and carrying out time-varying characteristic design on the interference signal subjected to phase modulation so as to generate an interference signal for interfering the target voice of the secret-related place. According to the method, the target voice features can be comprehensively and accurately obtained to provide a basis for interference signal design, the interference signals effectively destroy key elements of tone and phase of the target voice, and the interference signals can be dynamically adjusted along with changes of the target voice and environmental noise. And the adaptability of sound interference and the protection effect on sound information of a confidential place are enhanced. The problem that a traditional sound interference means is difficult to effectively deal with complex and changeable conditions is solved.

Description

Technical Field

[0001] The present invention relates to the field of information security technology, and in particular to a sound interference device for confidential places. Background Art

[0002] As information security in confidential places becomes increasingly important, the protection of sound information is of vital importance. Traditional sound interference methods have obvious limitations, as they are mostly based on designing interference signals with fixed parameters. On the one hand, when faced with changes in the target voice itself, such as changes in the speaker's speaking speed, pitch, and timbre, the interference strategy cannot be adjusted accordingly, making it difficult to ensure a stable and effective interference effect, resulting in a continuous reduction in the intelligibility of the target voice. On the other hand, traditional methods lack adaptability to the complex and changing environmental noise in confidential places, such as changes in intensity and frequency distribution. There may even be situations where the interference signal is masked or the interference effectiveness is reduced. Traditional sound interference methods are unable to cope with these complex and changing situations, resulting in the risk of sound information in confidential places being eavesdropped and leaked. Summary of the Invention

[0003] In order to make up for the above shortcomings, the present invention provides a sound interference device for confidential places, aiming to improve the problem that traditional sound interference means are difficult to effectively deal with complex and changing situations.

[0004] In a first aspect, the present invention provides the following technical solution: a method for sound interference in a confidential place, comprising the following steps: S1. Extracting speech features from target speech signals collected from confidential places; S2. Design interference signal spectrum based on the extracted speech features; S3, phase modulating the interference signal spectrum; S4. Design a time-varying characteristic for the phase-modulated interference signal to generate an interference signal for interfering with target speech in confidential places.

[0005] By adopting the above technical solution, the target speech features can be accurately extracted and the interference signal can be designed and modulated accordingly to reduce the intelligibility of the target speech. The parameters can also be adjusted in real time as the environment and speech changes, thereby enhancing adaptability, ensuring information confidentiality, making up for the shortcomings of traditional methods, and optimizing the interference effect.

[0006] Preferably, the step S1 specifically includes: performing time-frequency analysis on the target speech signal using short-time Fourier transform, wherein the short-time Fourier transform is used for the discrete target speech signal. ,in Represents discrete time sample points, using a Hamming window , as a window function, according to the formula The spectrum of the speech signal changing over time is calculated, where represents the frequency discrete sample points, Indicates the position of the time window; then extracts key speech features including spectral features, fundamental frequency and formant frequency from the spectrogram, wherein the spectral features are the Amplitude , the fundamental frequency is calculated by the speech signal The autocorrelation function ,in is the delay time sample point, and according to the formula It is calculated from the delay time corresponding to the peak of the autocorrelation function, where is the sampling frequency of the speech signal, and the formant frequency is determined by performing peak detection on the spectrum graph as the frequency point corresponding to the peak with a large amplitude and a prominent peak and marked as ,in are different resonance peak numbers.

[0007] Preferably, the S2 specifically includes: generating a narrowband interference signal based on the speech formant, for each formant frequency , with its center frequency , construct the narrowband interference signal component , whose expression is ,in Determined based on the energy of the target speech at the corresponding formant and the expected interference intensity, Initially set to a random value, Select according to the voice characteristics and the expected interference frequency range; then superimpose the narrowband interference signal components and calculate the interference frequency according to the formula Get the spectrum of the entire interference signal .

[0008] Preferably, the S2 further includes: adjusting the interference signal spectrum based on information entropy, calculating the interference signal spectrum Information entropy , the calculation formula is ,in At the same time, the spectrum information entropy of the target speech is calculated in a similar way to the above-mentioned calculation of the interference signal information entropy. ; Set the information entropy of the interference signal Information entropy of target speech The ratio is , by adjusting the amplitude of each component of the interference signal spectrum Parameters, recalculate information entropy and adjust iteratively until it satisfies , to obtain the adjusted interference signal spectrum .

[0009] Preferably, the S3 specifically includes: Hilbert transform application, for discrete interference signals, analogy continuous signal , according to the Hilbert transform definition The discretization method is used to calculate the Hilbert transform ,in is a discrete time sample point, and then the analytical signal is constructed , and then through the formula Calculate the phase of the analytic signal , obtain the phase information corresponding to each discrete time point of the interference signal, and prepare for subsequent phase adjustment.

[0010] Preferably, the S3 further includes: adjusting the interference signal phase based on phase interference, The components of the target speech at each frequency Phase , according to the formula Adjust its phase ,in is a small phase adjustment amount, and its value range is determined according to the actual interference experiment; for the entire interference signal spectrum After the phase adjustment of each component is completed, the frequency domain interference signal after phase adjustment is converted back to the time domain interference signal through inverse Fourier transform. .

[0011] Preferably, the step S4 specifically includes: deploying an adaptive filtering algorithm to filter a mixed input signal containing target speech and ambient noise. and the expected output signal As input, set the coefficient vector of the adaptive filter ,in is the filter order, according to the input signal , according to the formula Calculate filter output ; Define the error signal , and based on the update formula of the least mean square adaptive filtering algorithm To update the filter coefficients, so as to dynamically adjust the filter coefficients according to the input signal characteristics and the expected convergence speed, where is the step size factor.

[0012] In a second aspect, the present invention provides the following technical solution: a sound interference system for a confidential place, used in the above-mentioned sound interference method for a confidential place, the system comprising: A speech feature extraction module is used to extract speech features from target speech signals collected from confidential places; an interference signal spectrum design module, connected to the speech feature extraction module, for designing an interference signal spectrum based on the extracted speech features; A phase modulation module, connected to the interference signal spectrum design module, for performing phase modulation on the interference signal spectrum; A time-varying characteristic design module, connected to the phase modulation module, for performing time-varying characteristic design on the phase-modulated interference signal; The sound-generating unit is connected to the time-varying characteristic design module and is used to transmit the interference signal finally generated to the confidential place space to interfere with the target voice.

[0013] In the third aspect, the invention provides the following technical solution: a sound interference device for confidential places, comprising a memory, a processor, and a computer program stored on the memory and runnable on the processor, wherein the processor implements the above-mentioned sound interference method for confidential places when executing the computer program.

[0014] In a fourth aspect, the present invention provides the following technical solution: a readable storage medium having a computer program stored thereon, and the computer program, when executed by a processor, implements the above-mentioned method for sound interference in confidential places.

[0015] The present invention has the following beneficial effects: 1. The present invention's method comprehensively and accurately captures target speech characteristics, providing a basis for jamming signal design. This allows the jamming signal to effectively destroy the key elements of the target speech, such as timbre and phase. Furthermore, the jamming signal can be dynamically adjusted as the target speech and ambient noise change, enhancing the adaptability of the sound jamming and its effectiveness in protecting sound information in confidential locations. This solves the problem that traditional sound jamming methods are unable to effectively respond to complex and changing situations.

[0016] 2. In this invention, by designing the interference signal spectrum based on extracted speech features, a narrowband interference signal is first generated based on the speech formants, allowing the interference signal spectrum to be targeted at the key timbre area of ​​the target speech, destroying the timbre characteristics and reducing intelligibility. The spectrum is then adjusted based on information entropy, increasing the chaos of the mixed signal, making it more difficult for eavesdroppers to extract information and further improving the interference effect. This solves the problem that previous interference methods have difficulty effectively interfering with the timbre characteristics of the target speech and that the overall information of the mixed signal is easily analyzed and restored.

[0017] 3. In this invention, by phase-modulating the interference signal spectrum, the Hilbert transform is first used to obtain the phase information of the interference signal at each discrete time point, providing a basis for subsequent adjustments. Then, phase adjustment is performed based on phase interference to achieve optimal phase matching between the interference signal and the target voice, generating destructive interference to weaken the target voice energy and reduce intelligibility. The frequency domain signal is also converted into a time domain signal for practical deployment, effectively enhancing the sound interference effect in confidential places. This solves the problems of being unable to accurately obtain the phase information of the interference signal and the difficulty in effectively weakening the target voice through phase interference.

[0018] 4. This invention utilizes an adaptive filtering algorithm designed for time-varying characteristics of phase-modulated interference signals. This allows the interference signal to dynamically adjust its parameters based on changes in the target speech and ambient noise, continuously reducing the target speech's intelligibility. This enhances the protection of sound information in classified locations and improves the adaptability and robustness of the sound interference. This solves the problem that traditional fixed-parameter interference signals are difficult to adapt to the dynamic changes in classified locations and suffer from poor interference effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a method for sound interference in confidential places proposed by the present invention; Figure 2 This is a system architecture diagram of a sound interference system for confidential places proposed by the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] Example 1 Reference Figure 1 In a first embodiment of the present invention, the present invention provides a method for sound interference in a confidential place, comprising the following steps: S1. Extract speech features from target speech signals collected from confidential places.

[0022] S1 specifically includes: using short-time Fourier transform to perform time-frequency analysis on the target speech signal, short-time Fourier transform for discrete target speech signals ,in Represents discrete time sample points, using a Hamming window , as a window function, according to the formula The spectrum of the speech signal changing over time is calculated, where represents the frequency discrete sample points, Indicates the position of the time window; then extracts key speech features including spectral features, fundamental frequency and formant frequency from the spectrogram, where the spectral features are Amplitude , the fundamental frequency is calculated by the speech signal The autocorrelation function ,in is the delay time sample point, and according to the formula It is calculated from the delay time corresponding to the peak of the autocorrelation function, where is the sampling frequency of the speech signal, and the formant frequency is determined by peak detection on the spectrum graph as the frequency point corresponding to the peak with a large and prominent amplitude and marked as ,in are different resonance peak numbers.

[0023] Specifically, short-time Fourier transform is used to perform time-frequency analysis on discrete target speech signals. ,in represents discrete time sample points, , is the total length of the signal, using a Hamming window As a window function, its expression is The Hamming window is chosen because it can reduce spectrum leakage to a certain extent, so that the time-frequency analysis results can more accurately reflect the changes in the real frequency components of the speech signal over time.

[0024] According to the formula Perform short-time Fourier transform operation. represents the frequency discrete sample points, , represents the position of the time window, , is the total number of time windows. The principle of this formula is to convert the speech signal With the sliding Hamming window function Multiply, intercept the signal in each time window and perform discrete Fourier transform, that is, multiply by the complex exponential And sum them up to get different time window positions Next, different frequencies The spectrum information at each location is finally integrated to form a spectrum diagram of the speech signal changing over time. .

[0025] Spectral feature extraction, from the spectrum obtained by short-time Fourier transform In the This is the extracted spectrum feature. The amplitude represents the and frequency By analyzing the amplitude at different times and frequency positions in the entire spectrum, we can understand the change pattern of the strength of each frequency component of the speech signal over time.

[0026] fundamental frequency Extract and calculate speech signals The autocorrelation function , is the delay time sample point. The autocorrelation function reflects the similarity between the signal and itself at different delay times. For periodic signals, speech signals have quasi-periodicity, and their fundamental frequency reflects this periodicity. The autocorrelation function will have a peak at the integer multiple delay of the period. Based on this characteristic, by finding the autocorrelation function The peak value, then according to the formula To calculate the fundamental frequency That is, using the sampling frequency Delay time corresponding to the peak of the autocorrelation function The fundamental frequency is determined by the ratio of the sampling frequency to the actual physical frequency. Because there is a fixed conversion relationship between the sampling frequency and the actual physical frequency, the fundamental frequency information of the speech signal can be accurately obtained in this way.

[0027] Formant frequency Extract, in the obtained spectrum The peak detection operation is performed on the spectrum, specifically to find the frequency points corresponding to the peaks with relatively large amplitude and more prominent peaks in the entire spectrum range, and mark these frequency points as the resonance peak frequencies. ,in Indicates different formant numbers. In the spectrum of a speech signal, formants are frequency regions where energy is relatively concentrated. They are determined by the resonance characteristics of the vocal tract. Different vowels and consonants have distinct differences in these formant frequencies, which are an important reflection of the timbre of the speech.

[0028] By extracting speech features from target speech signals collected from classified locations, the short-time Fourier transform is first used to obtain a time-varying spectrum. Key speech features, such as spectral features, fundamental frequency, and formant frequency, are then extracted from this spectrum. This allows for a clear understanding of the frequency components, energy distribution, pitch, and timbre-related characteristics of the target speech at different times, providing an accurate basis for subsequent jamming signal design. This allows for targeted placement of jamming signals in key frequency regions, disrupting speech intelligibility in multiple ways and effectively improving the accuracy and effectiveness of sound jamming. This solves the problem of difficulty in accurately designing jamming signals due to a lack of understanding of the target speech characteristics when jamming in classified locations.

[0029] S2. Design interference signal spectrum based on the extracted speech features.

[0030] S2 specifically includes: generating narrowband interference signals based on speech formants, and , with its center frequency , construct the narrowband interference signal component , whose expression is ,in Determined based on the energy of the target speech at the corresponding formant and the expected interference intensity, Initially set to a random value, Select according to the voice characteristics and the expected interference frequency range; then superimpose the narrowband interference signal components and calculate the interference frequency according to the formula Get the spectrum of the entire interference signal .

[0031] S2 also includes: adjusting the interference signal spectrum based on information entropy, calculating the interference signal spectrum Information entropy , the calculation formula is ,in At the same time, the spectrum information entropy of the target speech is calculated in a similar way to the above-mentioned calculation of the interference signal information entropy. ; Set the information entropy of the interference signal Information entropy of target speech The ratio is , by adjusting the amplitude of each component of the interference signal spectrum Parameters, recalculate information entropy and adjust iteratively until it satisfies , to obtain the adjusted interference signal spectrum .

[0032] Specifically, when generating a narrowband interference signal based on speech formants, the focus is on the formant frequency of the target speech extracted previously. For each determined formant frequency , taking it as the center frequency To construct the narrowband interference signal component The specific construction method is through the expression Reflect. The determination of needs to consider two factors comprehensively. On the one hand, the energy of the target speech at the corresponding resonance peak can be judged by checking the amplitude of the resonance peak frequency in the speech spectrum graph obtained previously. On the other hand, the desired interference intensity is the desired interference effect on the voice timbre characteristics reflected by the resonance peak. For example, if the target speech has a strong energy at a certain resonance peak, and a strong interference effect is expected, then the corresponding interference intensity should be set. The value of is set to be larger, so that the interference signal in this frequency region can have enough energy to affect the timbre performance of the target speech.

[0033] It is initially set to a random value because it will be further precisely adjusted based on phase interference and other related content in acoustic principles. Therefore, a random value is first assigned to start the entire interference signal construction process. Subsequently, other operations are combined to optimize its value so that it can better interact with the target speech and achieve the ideal interference effect.

[0034] The selection of the frequency band should be determined based on the characteristics of the speech itself and the expected interference frequency range. From the perspective of speech characteristics, the half-power bandwidth of the target speech at the formant will be referred to. The half-power bandwidth can reflect the approximate range of the energy distribution of the formant frequency. Based on it, the appropriate The value should be selected to ensure that the constructed narrowband interference signal can cover the key frequency area closely related to the target speech formant. At the same time, it is also necessary to combine the actual desired interference frequency range, that is, to adjust the frequency range according to the specific frequency range to be interfered. Make appropriate adjustments so that it can focus on the key frequency area of ​​the target speech formant and meet the frequency range requirements of the overall interference strategy.

[0035] After constructing each narrowband interference signal component Afterwards, according to the formula Superimpose these components to obtain the spectrum of the entire interference signal The reason for this is that by adding multiple narrowband interference signal components that focus on different formant frequencies, the resulting interference signal spectrum can fully cover the key formant frequency regions of the target speech, thereby interfering with the target speech at multiple key frequency points related to timbre.

[0036] Interference signal spectrum adjustment based on information entropy, to calculate the interference signal spectrum Information entropy , the calculation formula is , here The principle is to first analyze the interference signal spectrum The amplitude of is normalized to obtain , which represents the interference signal at each frequency Then, based on the concept of information entropy in information theory, the above formula is used to calculate the information entropy of the interference signal, and then the degree of information uncertainty contained in the interference signal spectrum is quantified.

[0037] The spectral information entropy of the target speech is calculated using a method similar to that used to calculate the information entropy of the interference signal. That is, the frequency spectrum of the target speech is also processed accordingly to obtain its probability distribution at different frequencies, and then the corresponding information entropy is calculated to measure the degree of uncertainty of the information carried by the target speech itself.

[0038] Set the information entropy of the interference signal Information entropy of target speech The ratio is , here It is a fixed parameter determined by many factors such as the design requirements of the entire interference scheme and the expectation of the final interference effect. Parameters such as , recalculate the information entropy and adjust iteratively until it meets This condition. The specific adjustment process is that each time the amplitude After the parameters are adjusted, the information entropy of the interference signal must be calculated again according to the method described above. , and then compare it with the target speech information entropy Compare and determine whether it meets the set ratio relationship. If not, continue to adjust the parameters, calculate and compare again, and repeat this cycle until the requirements are met, and finally obtain the adjusted interference signal spectrum.

[0039] By designing the jamming signal spectrum based on extracted speech features, the method first generates a narrowband jamming signal based on the speech formants, allowing the jamming signal spectrum to be targeted at the key timbre areas of the target speech, disrupting the timbre characteristics and reducing intelligibility. The spectrum is then adjusted based on information entropy, increasing the chaos of the mixed signal, making it more difficult for eavesdroppers to extract information and further improving the jamming effect. This solves the problem that previous jamming methods had difficulty effectively jamming the timbre characteristics of the target speech, and that the overall information of the mixed signal was easily analyzed and restored.

[0040] S3. Phase modulate the interference signal spectrum.

[0041] S3 specifically includes: Hilbert transform application, analogy of continuous signal to discrete interference signal , according to the Hilbert transform definition The discretization method is used to calculate the Hilbert transform ,in is a discrete time sample point, and then the analytical signal is constructed , and then through the formula Calculate the phase of the analytic signal , obtain the phase information corresponding to each discrete time point of the interference signal, and prepare for subsequent phase adjustment.

[0042] S3 also includes: interference signal phase adjustment based on phase interference, for interference signal spectrum The components of the target speech at each frequency Phase , according to the formula Adjust its phase ,in is a small phase adjustment amount, and its value range is determined according to the actual interference experiment; for the entire interference signal spectrum After the phase adjustment of each component is completed, the frequency domain interference signal after phase adjustment is converted back to the time domain interference signal through inverse Fourier transform. .

[0043] Specifically, in the application of Hilbert transform, for discrete interference signals, the Hilbert transform definition should be used. The discretization method is used to calculate its Hilbert transform , here Represents discrete time sample points. When performing discretization processing, the continuous integral operation is converted into a summation form suitable for discrete signal processing based on the relevant methods in numerical calculation, so as to process discrete interference signals. In specific operations, it is necessary to refer to the sampling interval of the discrete signal and the distribution law of the discrete time sample points, use the appropriate discretization algorithm, and divide the integral interval into multiple discrete sub-intervals according to certain rules. Then, make an approximate calculation of the function value in each sub-interval, and then calculate the Hilbert transform at each discrete time sample point according to the corresponding weighted summation method. The value at .

[0044] Get the Hilbert transform Then, construct the analytical signal , which is constructed by This is the application of the theory of complex variable function to transform the original real discrete interference signal and the calculated Hilbert transform Combined into complex form analytical signals ,In this way, the characteristics of the interference signal can be more comprehensively displayed, especially in terms of phase information.

[0045] Finally, through the formula To calculate the phase of the analytical signal , and then obtain the phase information corresponding to the interference signal at each discrete time point. This formula is based on the relationship between the inverse tangent function in the trigonometric function, using the real part of the analytical signal and the imaginary part The phase angle is determined by the ratio of , so as to extract the phase information contained in the analytical signal, paving the way for the subsequent phase adjustment of the interference signal based on phase interference.

[0046] Interference signal phase adjustment based on phase interference. In the interference signal phase adjustment stage based on phase interference, the interference signal spectrum is adjusted. The components of the target speech should be based on the frequency Phase To make phase adjustment. The specific adjustment method is according to the formula To adjust the phase of each component Here, It is a small phase adjustment amount, and its value range must be determined based on the actual interference experiment. When doing the actual experiment, it is necessary to gradually change the phase in the simulated environment of the confidential place or the actual application scenario. Then observe the corresponding interference effect, such as how much the intelligibility of the target speech is reduced after interference, how the timbre changes, etc., in this way to determine the effect of different frequencies Under this condition, the interference effect can be best achieved. Value range.

[0047] The reason for adjusting the phase according to this formula is based on the phase interference principle in acoustics. When the interference signal and the target voice are superimposed in space, the phase of the interference signal is adjusted so that it meets a specific relationship with the phase of the target voice (that is, the phase of the interference signal is increased by 1 / 2 of the target voice phase). And the appropriate small adjustment , which can cause the two to interfere destructively at the corresponding frequencies. Destructive interference means that the amplitude of the composite wave becomes smaller after the two waves are superimposed. For the sound signal, this means that the energy is weakened, thereby achieving the purpose of destroying the intelligibility of the target speech.

[0048] Over the entire interference signal spectrum After all components of the signal have completed phase adjustment according to the above method, the frequency domain interference signal after phase adjustment should be converted into the time domain interference signal through inverse Fourier transform. The principle of inverse Fourier transform is the inverse operation of Fourier transform. It can restore the spectral information of the interference signal after phase adjustment in the frequency domain into a signal form in the time domain that can be directly applied to actual confidential places for interference, ensuring that the interference signal can play a role in the confidential place according to the expected time characteristics.

[0049] By phase-modulating the interference signal spectrum, the Hilbert transform is first applied to obtain the phase information of the interference signal at each discrete time point, providing a basis for subsequent adjustments. Phase adjustment is then performed based on phase interference, achieving an optimal phase match between the interference signal and the target voice. This generates destructive interference, weakening the target voice energy and reducing intelligibility. The frequency domain signal is also converted into a time domain signal for practical deployment, effectively enhancing the sound interference effect in confidential locations. This solves the problems of being unable to accurately obtain the phase information of the interference signal and the difficulty of effectively weakening the target voice through phase interference.

[0050] S4. Design a time-varying characteristic for the phase-modulated interference signal to generate an interference signal for interfering with target speech in confidential places.

[0051] S4 specifically includes: Adaptive filtering algorithm deployment to contain mixed input signals of target speech and ambient noise and the expected output signal As input, set the coefficient vector of the adaptive filter ,in is the filter order, according to the input signal , according to the formula Calculate filter output ; Define the error signal , and based on the update formula of the least mean square adaptive filtering algorithm To update the filter coefficients, so as to dynamically adjust the filter coefficients according to the input signal characteristics and the expected convergence speed, where is the step size factor.

[0052] Specifically, the input signal and coefficient vector setting, in the deployment of the adaptive filtering algorithm, the input signal must be determined at the beginning. The input signal here consists of two parts, one is a mixed input signal containing the target speech and environmental noise, etc. This signal is collected in real time by sound sensors placed in confidential places. The other part is the expected output signal It refers to the expectation that after effective interference, the signal will present high confusion and low intelligibility, that is, the signal we expect to appear after the interference is completed.

[0053] To set the coefficient vector of the adaptive filter , here Represents the filter order. Filter order The determination of the filter order must take into account multiple factors, such as the complexity of the input signal. If the input signal contains many frequency components and the changes are quite complex, then in order to more finely process the signal and achieve a better interference effect, it is usually necessary to select a higher filter order. However, the computing resources must also be considered. If the order is set too high, the amount of calculation will become very large, which may affect real-time performance and other issues. Therefore, a good balance must be made between these factors.

[0054] Filter output calculation, based on the input signal , according to the formula To calculate the filter output The principle here is based on the vector multiplication operation in linear algebra, which is to multiply the coefficient vector and the input signal vector Do the inner product operation, so you can get a scalar value This value is the filter's output after processing the input signal based on the current coefficient vector. Through this calculation method, the filter can output the corresponding processing results in real time based on the input signal at different times and the current coefficient settings. This lays the foundation for subsequent judgment of interference effects and coefficient adjustments.

[0055] Error signal definition and coefficient update, to define the error signal , its calculation formula is This error signal reflects the difference between the desired output signal and the actual filter output signal. This difference is particularly important for measuring the current interference effect and guiding coefficient adjustment. Why do I say this? Because if the error signal is relatively large, it means that the current filter output result is quite different from the desired signal state after interference, which means that the filter coefficients must be adjusted to improve the interference effect.

[0056] Next, we need to follow the update formula of the least mean square adaptive filtering algorithm To update the filter coefficients, here The step size factor ranges from 0 to 1. The specific value of this step size factor is determined through experimental debugging and in combination with actual application scenarios. It controls the update speed of the filter coefficients. A small value will slow the coefficient update and make the filter less responsive to changes in the input signal, but this may result in a more stable coefficient adjustment process. A large value will increase the coefficient update speed, allowing the filter to more quickly keep up with changes in the input signal. However, this may cause the coefficient adjustment process to become unstable or even diverge.

[0057] During the update process, according to the current error signal And the input signal vector , according to the update formula mentioned above, the coefficient vector By updating the various elements in the filter, the filter coefficients can be dynamically adjusted based on the characteristics of the input signal and the desired convergence rate. Through continuous iterative updates, the filter gradually adapts to changes in the input signal, and the output becomes closer to the desired output signal. This allows for the design of the time-varying characteristics of the interference signal, allowing the interference signal to dynamically adjust its parameters based on real-time changes in the target speech and ambient noise.

[0058] By deploying an adaptive filtering algorithm based on the time-varying characteristics of phase-modulated interference signals, the interference signal can dynamically adjust its parameters based on changes in the target voice and ambient noise, continuously reducing the target voice's intelligibility, enhancing the protection of sound information in classified locations, and improving the adaptability and robustness of sound interference. This solves the problem that traditional fixed-parameter interference signals are difficult to adapt to the dynamic changes in classified locations and have poor interference effectiveness.

[0059] In the speech feature extraction stage, short-time Fourier transform is used to analyze the target speech signal collected from confidential places, with the Hamming window as the window function. The discrete speech signal is subjected to time-frequency analysis through a specific formula, and then key information such as spectral characteristics, fundamental frequency, and resonance peak frequency are extracted, laying the foundation for subsequent steps.

[0060] Then, in the interference signal spectrum design stage, on the one hand, narrowband interference signal components are constructed based on the speech formants, and the component parameters are determined according to the target speech formant energy and the expected interference intensity, and then superimposed. On the other hand, the spectrum is optimized by calculating the information entropy of the interference signal and the target speech and adjusting the amplitude of each spectrum component according to the set ratio.

[0061] Phase modulation is then performed. First, the Hilbert transform is used to discretize the interference signal to obtain its phase information. Then, based on the phase interference principle, the phase of the interference signal is adjusted with reference to the target speech phase, and then converted into a time domain signal through an inverse Fourier transform.

[0062] Finally, in the time-varying characteristic design, the mixed input signal containing the target speech and environmental noise and the expected output signal are used as input, and the adaptive filter coefficient vector is set. According to the least mean square algorithm and relying on error signal feedback, the coefficient is dynamically updated to realize the function of adjusting parameters in real time as the interference signal changes with the external environment.

[0063] This method can comprehensively and accurately capture the characteristics of the target speech, providing a precise basis for jamming signal design. This allows the jamming signal to effectively destroy key elements of the target speech, such as timbre and phase, reducing intelligibility. It also dynamically adjusts the jamming signal to changes in the target speech and ambient noise, enhancing the adaptability and robustness of the sound jamming and its effectiveness in protecting sound information in confidential locations. This addresses the problem that traditional sound jamming methods struggle to effectively respond to complex and changing situations.

[0064] Example 2: Reference Figure 2 In a second embodiment of the present invention, the present invention provides a confidential place sound interference system, used in a confidential place sound interference method, the system comprising: A speech feature extraction module is used to extract speech features from target speech signals collected from confidential places; An interference signal spectrum design module, connected to the speech feature extraction module, for designing an interference signal spectrum based on the extracted speech features; A phase modulation module is connected to the interference signal spectrum design module and is used to perform phase modulation on the interference signal spectrum; A time-varying characteristic design module is connected to the phase modulation module and is used to design the time-varying characteristics of the phase-modulated interference signal; The sound-generating unit is connected to the time-varying characteristic design module and is used to transmit the finally generated interference signal to the confidential place space to interfere with the target voice.

[0065] Example 3 The third embodiment of the present invention is based on the same inventive concept. The present invention proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements a sound interference method for confidential places in the above embodiment.

[0066] Example 4 The fourth embodiment of the present invention is based on the same inventive concept. The present invention proposes a sound interference device for confidential places, including: a processor and a memory; the processor and the memory communicate with each other; the memory is used to store instructions; the processor is used to execute the instructions in the memory to execute a sound interference method for confidential places of the above embodiment.

[0067] It should be understood that various components of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof may be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0068] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A sound interference method for confidential places, characterized in that: The following steps are involved: S1. Extracting speech features from target speech signals collected from confidential places; S2. Design interference signal spectrum based on the extracted speech features; S3, phase modulating the interference signal spectrum; S4. Design a time-varying characteristic for the phase-modulated interference signal to generate an interference signal for interfering with target speech in confidential places.

2. A sound interference method for confidential places according to claim 1, characterized in that: The S1 specifically includes: using short-time Fourier transform to perform time-frequency analysis on the target speech signal, and the short-time Fourier transform is used for the discrete target speech signal. ,in Represents discrete time sample points, using a Hamming window , as a window function, according to the formula The spectrum of the speech signal changing over time is calculated, where represents the frequency discrete sample points, Indicates the position of the time window; then extracts key speech features including spectral features, fundamental frequency and formant frequency from the spectrogram, wherein the spectral features are the Amplitude , the fundamental frequency is calculated by the speech signal The autocorrelation function ,in is the delay time sample point, and according to the formula It is calculated from the delay time corresponding to the peak of the autocorrelation function, where is the sampling frequency of the speech signal, and the formant frequency is determined by performing peak detection on the spectrum graph as the frequency point corresponding to the peak with a large amplitude and a prominent peak and marked as ,in are different resonance peak numbers.

3. A sound interference method for confidential places according to claim 1, characterized in that: The S2 specifically includes: generating a narrowband interference signal based on the speech formant, for each formant frequency , with its center frequency , construct the narrowband interference signal component , whose expression is ,in Determined based on the energy of the target speech at the corresponding formant and the expected interference intensity, Initially set to a random value, Select based on speech characteristics and expected interference frequency range; then superimpose each narrowband interference signal component, according to the formula Get the spectrum of the entire interference signal .

4. A sound interference method for confidential places according to claim 1, characterized in that: The S2 also includes: adjusting the interference signal spectrum based on information entropy, calculating the interference signal spectrum Information entropy , the calculation formula is ,in At the same time, the spectrum information entropy of the target speech is calculated in a similar way to the above-mentioned calculation of the interference signal information entropy. ; Set the information entropy of the interference signal Information entropy of target speech The ratio is , by adjusting the amplitude of each component of the interference signal spectrum Parameters, recalculate information entropy and adjust iteratively until it satisfies , to obtain the adjusted interference signal spectrum .

5. A sound interference method for confidential places according to claim 1, characterized in that: The S3 specifically includes: Hilbert transform application, for discrete interference signals, analogy continuous signal , according to the Hilbert transform definition The discretization method is used to calculate the Hilbert transform ,in is a discrete time sample point, and then the analytical signal is constructed , and then through the formula Calculate the phase of the analytic signal , obtain the phase information corresponding to each discrete time point of the interference signal, and prepare for subsequent phase adjustment.

6. A sound interference method for confidential places according to claim 1, characterized in that: The S3 also includes: adjusting the interference signal phase based on phase interference, and adjusting the interference signal spectrum The components of the target speech at each frequency Phase , according to the formula Adjust its phase ,in is a small phase adjustment amount, and its value range is determined according to the actual interference experiment; for the entire interference signal spectrum After the phase adjustment of each component is completed, the frequency domain interference signal after phase adjustment is converted back to the time domain interference signal through inverse Fourier transform. .

7. A sound interference method for confidential places according to claim 1, characterized in that: The S4 specifically includes: deploying an adaptive filtering algorithm to filter a mixed input signal containing target speech and ambient noise and the expected output signal As input, set the coefficient vector of the adaptive filter ,in is the filter order, according to the input signal , according to the formula Calculate filter output ; Define the error signal , and based on the update formula of the least mean square adaptive filtering algorithm To update the filter coefficients, so as to dynamically adjust the filter coefficients according to the input signal characteristics and the expected convergence speed, where is the step size factor.

8. A sound interference system for confidential places, characterized by: A method for sound interference in confidential places according to any one of claims 1 to 7, the system comprising: A speech feature extraction module is used to extract speech features from target speech signals collected from confidential places; an interference signal spectrum design module, connected to the speech feature extraction module, for designing an interference signal spectrum based on the extracted speech features; A phase modulation module, connected to the interference signal spectrum design module, for performing phase modulation on the interference signal spectrum; A time-varying characteristic design module, connected to the phase modulation module, for performing time-varying characteristic design on the phase-modulated interference signal; The sound-generating unit is connected to the time-varying characteristic design module and is used to transmit the interference signal finally generated to the confidential place space to interfere with the target voice.

9. A sound jamming device for a confidential place, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for sound interference in confidential places as described in any one of claims 1 to 7 is implemented.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for sound interference in confidential places as described in any one of claims 1 to 7 is implemented.

Images