Dual-track audio encryption method and device

Abstract

The invention relates to a dual-channel audio encryption method and device, and the method employs a stream cipher system to design an encryption algorithm of dual-channel audio data based on a 1 + 1-dimensional nonlinear fractional order pseudo-parabolic equation encryption and decryption model. The operations performed on the audio data include Euclidean subtraction (addition), key matrix averaging, encryption and decryption model numerical calculation, numerical constraint (release) and the like. In the application, the signal time domain feature and the signal spectrum feature of the encrypted audio data can be effectively masked, the encryption algorithm effectively changes the audio data, and the peak signal-to-noise ratio statistical attack can be effectively resisted.

Description

Technical Field

[0001] The present application belongs to the technical field of data encryption and decryption, and specifically relates to a dual-channel audio encryption method and device. Background Art

[0002] Cryptography is a widely used data security technology. When data is threatened by various attacks, this technology is an important measure for application security. The core of encryption and decryption technology is the encryption and decryption algorithm. Traditional symmetric encryption algorithms include AES encryption, DES encryption, etc., and traditional asymmetric encryption algorithms include RSA encryption, etc. Common voice and image encryption algorithms include chaotic encryption and permutation encryption. Chaotic encryption uses the operation of chaotic equations to generate keys. The uncertainty of the chaotic model after entering the chaotic state is equivalent to the security of chaotic encryption. Its advantage is that the chaotic state cannot be predicted and is difficult to crack. The disadvantage is that the initial value sensitivity is extremely strong, and a small initial value difference can lead to a large result error; the principle of the permutation algorithm is to change the position of the original digital signal to achieve the encryption effect. The key is the information of the relative position before and after the change. Its advantage is that the encryption and decryption cost is low and the overhead is low, and the disadvantage is that the security is relatively low. In summary, the traditional encryption and decryption algorithms have the characteristics of high cost, large algorithm complexity, and unstable security performance. In today's network environment with large traffic, many users, wide range, and high requirements for information authenticity, it is particularly important to study an encryption and decryption algorithm with smaller errors, stronger reliability and stability. Summary of the invention

[0003] The present invention provides a dual-channel audio encryption and decryption method, which can achieve high-reliability encryption and maintain encryption stability.

[0004] The technical solution of this application is as follows: A dual-channel audio encryption method comprises the following steps: Step 1: Get the two-channel audio data matrix Audio , whose dimensions are L ×2, which means as follows: ; In the formula, the first column represents the left channel data from top to bottom, and the second column represents the right channel data from top to bottom; Step 2: Audio Perform matrix transformation operation of Euclidean subtraction to obtain Audio 1, which means as follows: ; Step 3: Create a dimension L ×2 key matrix Rand 1. Rand 1 and Audio The elements at the corresponding positions in 1 are averaged and then established Audio 2; Step 4: Audio The elements in 2 are grouped and expanded by columns to form Audio 3. In Audio In 3, each group contains 16 elements in sequence and when the last group has less than 16 elements, it is padded with 0; Audio The specific form of 3 is as follows: ; Step 5: Use the encrypted model of the 1+1 dimensional nonlinear fractional-order pseudo-parabolic equation to Audio 3. The encryption model is: ; In the formula, u and t is a parameter variable; express right Find the second-order partial derivatives of Find the first-order partial derivative; express right Find the first-order partial derivative; for t is the function at time zero; is the boundary condition; f ( x , t ) is a given function; T is a constant; Indicates about x Caputo fractional terms; During encryption, the nonlinear fractional-order pseudo-parabolic equation is directly linearized and iteratively solved by calculating the initial moment, and the encryption of the remaining moments is completed, which specifically includes the following steps: Step 1: Solve Audio 3 corresponds to the function value at the initial time ,include: The spatial variable x and time t Discrete into: ; Then the discrete space variable at the initial moment is x The nodes are: ; Then the boundary condition The corresponding discrete nodes are: ; Step 2: Discrete space variables at the initial moment xNode substitution function In the initial moment, we get the function value matrix , which is expressed as follows: ; Step 3: Construct one and Audio 3. A calculation matrix with the same dimensions, and in the calculation matrix, wrap it with 0 Audio 3 is the non-zero initial time function value; the calculation matrix is ​​in the following form: ; Step 4: Take Audio Each column of data in 3 is filled with the calculation matrix to obtain the encryption matrix at the initial moment: ; In the formula, the lower left corner subscript of the element means that the value is calculated at the i-th (i=2,3,4,5,6)th order in the matrix; Step 5: Calculate the matrix form of the encryption model based on the direct linearization iterative format of the nonlinear fractional-order pseudo-parabolic equation T Function value matrix at time , which is expressed as follows: ; Based on the calculation matrix construction principle in step three and step four, we get T The encryption matrix of the moment: ; Step 6: Repeat step 5 to Audio 3. Each data in the encryption is encrypted. Audio The original data in 3 is arranged in order and the encrypted data is obtained to obtain the final encryption matrix Audio 4, expressed as: ; Step 6: Construct Dimensions and Audio 4 Same numerical label matrix G , the numerical label matrix G The initial values ​​of the elements in are all 0; Will Audio Remove the negative sign from the negative elements in 4. When you remove the negative sign of an element, that is, in the matrix G Add a value to the element position at the corresponding position; after all is completed, write Audio 4 for ;right The values ​​of the elements in the matrix are constrained. G Mark the value; after the constraint Expand to a column, delete some data at the end of the expanded column, the number of deleted data is the same as the number of 0s added in step 4; after the constraint operation is completed, record the numerical labeling matrix G is the key matrix; The front of the expanded column data L Data and L The data are respectively used as the first and second columns to form a dual-channel ciphertext audio data matrix Audio 5.

[0005] Furthermore, in step 5, the matrix form of the direct linearization iterative format encryption model of the nonlinear fractional-order pseudo-parabolic equation is: ; ; In the formula, R 1 represents the coefficient matrix, and They are x , t of p 1. p The 2nd order differential matrix, is the integral matrix, is the gamma function, and They are m +1st level, l +1st-order identity matrix; f For a given function.

[0006] Furthermore, in the matrix of the direct linearization iterative format encryption model of the nonlinear fractional pseudoparabolic equation, the permutation method is used to impose boundary conditions and initial adjustments, and the accuracy of the iterative calculation is set to e=10 -9 .

[0007] Furthermore, in step six, The values ​​of the elements in the matrix are constrained and G The method for numerical marking is: When the element value is in the interval [0,1), keep and matrix G Original data; When the element value is in the interval [1,10), divide the element value by 10 in the matrix G The corresponding element position of plus 10; When the element value is in the interval [10,20), divide the element value by 40 and add it to the matrix G The corresponding element position of plus 100; When the element value is in the interval [20,40), divide the element value by 80 and add it to the matrix G The corresponding element position of plus 1000; When the element value is in the interval [40,80), divide the element value by 160 and add it to the matrix G The corresponding element position of plus 10000; When the element value is in the interval [80,200), divide the element value by 400 and add it to the matrix G The corresponding element position of plus 100000; When the element value is greater than or equal to 200, in the matrix G The corresponding element position of is reset to zero.

[0008] Furthermore, the processor is configured to execute a dual-channel audio encryption method as described in any one of claims 1 to 4 when running the program instructions.

[0009] Due to the adoption of the above technical solution, the beneficial effects of this application are as follows: This application adopts a stream cipher system, which is a non-traditional cryptographic encryption and decryption system with a class of differential equations as the core. The key factors of this cryptographic system are the numerical calculation method and the selection of differential equations. This application designs an encryption algorithm for dual-channel audio data based on the (1+1)-dimensional nonlinear fractional-order pseudo-parabolic equation encryption and decryption model. The operations performed on the audio data include Euclidean subtraction (addition), key matrix averaging, encryption and decryption model numerical calculation, numerical constraints (release), etc. In this application, the signal time domain characteristics and signal spectrum characteristics of the encrypted audio data can be effectively concealed, and the histogram of the ciphertext audio data can be greatly different from the histogram of the original plaintext audio data, which shows that the audio encryption algorithm effectively changes the frequency distribution range of the audio data. The correlation coefficients between the ciphertext audio and the left and right channel data of the original plaintext audio are close to 0, indicating that the encryption algorithm effectively changes the audio data and can effectively resist peak signal-to-noise ratio statistical attacks. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 A flow chart of a dual-channel audio encryption method provided by this application; Figure 2 A flowchart of a decryption method corresponding to a dual-channel audio encryption method provided in this application; Figure 3 A time domain comparison diagram of the original audio, the encrypted audio signal, and the decrypted audio signal; Figure 4 A comparison chart of the spectrograms of the original audio, the encrypted audio signal, and the decrypted audio signal; Figure 5is the distribution histogram of the original audio signal; Figure 6 is the distribution histogram of the encrypted audio signal. DETAILED DESCRIPTION

[0011] Based on the background technology, as shown in the attached Figure 1 As shown, the present application provides a dual-channel audio encryption method, the basic unit of audio data is the signal value, the audio can be a mono file or a dual-channel file, and the principle is the same. The present application is an encryption method applied to dual-channel audio, specifically including the following steps: Step 1: Get the two-channel audio data matrix Audio , whose dimensions are L ×2, which means as follows: ; In the formula, the first column represents the left channel data from top to bottom, and the second column represents the right channel data from top to bottom.

[0012] exist Audio In the figure, the number in the lower right corner of the element indicates the number of data.

[0013] Step 2: Audio Perform matrix transformation operation of Euclidean subtraction to obtain Audio 1, which means as follows: .

[0014] The matrix transformation operation of Euclidean subtraction in step 2 is to first subtract the second column from the first column of the matrix to replace the first column of the original matrix, and then subtract the replaced first column data from the second column data. That is: .

[0015] Step 3: Create a dimension L ×2 key matrix Rand 1. Rand 1 and Audio The elements at the corresponding positions in 1 are averaged and then established Audio 2. That is: .

[0016] Step 4: Audio The elements in 2 are grouped and expanded by columns to form Audio 3. In Audio In 3, each group contains 16 elements in sequence and when the last group has less than 16 elements, it is padded with 0; Audio The specific form of 3 is as follows: .

[0017] Will Audio The data in 2 are expanded by columns, with a total of 2 L There are 16 data in each group. If the last group is less than 16, it will be padded with 0. Group data.

[0018] Step 5: Use the encrypted model of the 1+1 dimensional nonlinear fractional-order pseudo-parabolic equation to Audio 3. The encryption model is: ; In the formula, u and t is a parameter variable; express right Find the second-order partial derivatives of Find the first-order partial derivative; express right t Find the first-order partial derivative; for t is the function at time zero; f ( x , t ) is a given function; T is a constant; Indicates about x The Caputo fractional order term. is the boundary condition.

[0019] During encryption, the nonlinear fractional-order pseudo-parabolic equation is directly linearized and iteratively solved by calculating the initial moment, and the encryption of the remaining moments is completed, which specifically includes the following steps: Step 1: Solve Audio 3 corresponds to the function value at the initial time ,include: The spatial variable x and time t Discrete into: ; Then the discrete space variable at the initial moment is x The nodes are: ; Then the boundary condition The corresponding discrete nodes are: ; Step 2: Discrete space variables at the initial moment x Node substitution function In the initial moment, we get the function value matrix , which is expressed as follows: ; Step 3: Construct one and Audio 3. A calculation matrix with the same dimensions, and in the calculation matrix, wrap it with 0 Audio 3 is the non-zero initial time function value; the calculation matrix is ​​in the following form: .

[0020] because Audio Each group of data in 3 has 16, so if the encrypted data is wrapped with 0, the dimension of the calculation matrix is ​​6 × 6. It should be noted that wrapping with 0 in this step means wrapping with a layer of 0.

[0021] In the calculation matrix, the number in the lower left corner of the element indicates the number of times this column of data is obtained in the numerical calculation in this matrix.

[0022] Step 4: Take Audio Each column of data in 3 is filled with the calculation matrix to obtain the encryption matrix at the initial moment: .

[0023] Step 5: Calculate the matrix form of the encryption model based on the direct linearization iterative format of the nonlinear fractional-order pseudo-parabolic equation T Function value matrix at time , which is expressed as follows: ; Based on the calculation matrix construction principle in step three and step four, we get T The encryption matrix at this moment: .

[0024] The calculation matrix is ​​in the form of: .

[0025] The matrix form of the direct linearization iterative format encryption model of the nonlinear fractional-order pseudo-parabolic equation is: ; ; In the formula, R 1 represents the coefficient matrix, and They are x , t of p 1. p The 2nd order differential matrix, is the integral matrix, is the gamma function, and They are m+1st level, l +1-order unit matrix. In the matrix of the direct linearization iterative format encryption model of the nonlinear fractional pseudoparabolic equation, the permutation method is used to impose boundary conditions and initial adjustments, and the accuracy of the iterative calculation is set to e=10 -9 .

[0026] Step 6: Repeat step 5 to Audio 3. Each data in the encryption is encrypted. Audio The original data in 3 is arranged in order and the encrypted data is obtained to obtain the final encryption matrix Audio 4, expressed as: ; Step 6: Construct Dimensions and Audio 4 Same numerical label matrix G , the numerical label matrix G The initial values ​​of the elements in the matrix are all 0; G It is expressed as: .

[0027] Will Audio Remove the negative sign from the negative elements in 4. When you remove the negative sign of an element, that is, in the matrix G Add a value to the element position at the corresponding position; after all is completed, write Audio 4 for ;right The values ​​of the elements in the matrix are constrained. G Mark the value; after the constraint Expand to a column, delete some data at the end of the expanded column, the number of deleted data is the same as the number of 0s added in step 4; after the constraint operation is completed, record the numerical labeling matrix G is the key matrix; The values ​​of the elements in the matrix are constrained and G The method for numerical marking is: When the element value is in the interval [0,1), keep and matrix G Original data; When the element value is in the interval [1,10), divide the element value by 10 in the matrix G The corresponding element position of plus 10; When the element value is in the interval [10,20), divide the element value by 40 and add it to the matrix G The corresponding element position of plus 100; When the element value is in the interval [20,40), divide the element value by 80 and add it to the matrix G The corresponding element position of plus 1000; When the element value is in the interval [40,80), divide the element value by 160 and add it to the matrix G The corresponding element position of plus 10000; When the element value is in the interval [80,200), divide the element value by 400 and add it to the matrix G The corresponding element position of plus 100000; When the element value is greater than or equal to 200, in the matrix G The corresponding element position of is reset to zero.

[0028] Finally, the front of the expanded column data L Data and L The data are respectively used as the first and second columns to form a dual-channel ciphertext audio data matrix Audio 5.

[0029] The decryption method of the above encryption method is the inverse operation of the algorithm.

[0030] As attached Figure 2 As shown, for the dual-channel ciphertext audio data matrix Audio 5, whose dimensions are L ×2, that is L Row 2 column, where Left 1 means the first data of the left channel (first column), Right 1 represents the first data of the right channel (the second column), and so on.

[0031] .

[0032] Will Audio 5 data are expanded by column, with a total of 2 L There are 16 data in each group. If the last group is less than 16, it will be padded with 0. Group data. After writing into a matrix: .

[0033] According to the matrix G and binding rules release .

[0034] The specific expression is: .

[0035] , , , , , Represents the numerical label matrix G exist The units, tens, hundreds, thousands, ten thousand, and hundred thousand digits of the element value are recorded as follows: .

[0036] The data in the above matrix is ​​numerically calculated using the decryption model. The decryption model of the heat flow cipher system in this paper is: ; Assume the boundary condition is 0, the data matrix to be decrypted Corresponding to the function value at the terminal time . Time variable x and spatial variables t are discretized into Discretized spatial variables at terminal time x The nodes are: ; Boundary conditions The corresponding discrete nodes are: ; Substitute all interpolation nodes into the function In the figure, it is the function value matrix at the terminal value moment : .

[0037] Write the data to be decrypted into the middle 4 of the 6 final value interpolation nodes, and perform a numerical calculation to decrypt the 4 data. There are 16 data in each group, and the following 6×6 dimension matrix is ​​designed: ; The data to be decrypted is written in the middle 16-bit non-zero element position. Each numerical calculation takes 1 column of data. The subscript in the lower left corner of the element means the number of numerical calculations in this matrix to take this column of data. Filling the 6×6 dimensional matrix with each set of data in can be obtained: One numerical calculation decrypts 16 data.

[0038] The matrix form of the direct linearization iterative format decryption model of the nonlinear fractional-order pseudo-parabolic equation is: Use the substitution method to impose boundary conditions and final value conditions, and set the accuracy of the iterative calculation to e=10 -9 , numerically calculate the function value matrix at the initial time for: The 6×6 dimensional matrix corresponding to the function value at the final value moment has the following matrix form at the initial moment: The result at the initial moment corresponds to the decrypted matrix: .

[0039] Will The 16 data in each column are numerically calculated according to this method, and then rearranged in order to obtain the data matrix decrypted by the equation decryption model. , the expression is Will Audio 3 Expand by column and take the first 2 L data, and discard the rest of the data. L The data is used as the left channel signal of the audio data. L The data is used as the right channel signal of the audio data and recombined into a dimension of The data matrix Audio 2, Audio 2 and key matrix Rand 1 Perform the following operations: Get the data matrix Audio 1. Assume Audio The expression of 1 is in and They respectively represent the first signal data of the left channel and the first signal data of the right channel of the current audio data, and the rest are similar.

[0040] for Audio 1. Perform Euclidean addition on the two columns of the data matrix. The data in the second column plus the data in the first column replaces the original data in the second column. The data in the first column plus the replaced data in the second column replaces the original data in the first column. The process of Euclidean addition is equivalent to matrix transformation: Audio That is, the decrypted audio data: in and They respectively represent the first signal data of the left channel and the first signal data of the right channel of the current audio data, and the rest are similar.

[0041] In this embodiment, the feasibility and security of the audio encryption and decryption algorithm are verified by simulation experiments. ,The simulation software is MATLAB (R2020b), the processor is Intel Core i7-11800H, the 64-bit operating system is Windows 11, the running memory is 16GB. The information of the encrypted and decrypted audio in the simulation ,experiment is shown in Table 1.

[0042] Table 1 Simulation experiment audio information As attached Figure 3 As shown in the figure, it can be seen from the time domain diagram of the ciphertext audio signal that the encryption algorithm of the present application effectively conceals the time domain characteristics of the audio signal, and no valid audio signal information can be obtained from the figure. It can be seen from the time domain diagram of the decrypted plaintext audio signal that the time domain characteristics of the decrypted audio signal are highly consistent with the time domain characteristics of the original plaintext audio signal, that is, the decryption algorithm can effectively restore the audio data.

[0043] As attached Figure 4 As shown, the signal spectrum is the result of statistics on speech signals with different frequency band amplitudes. It can be seen from the signal spectrum diagram of the ciphertext audio data that the audio encryption algorithm of the present application effectively masks the spectral characteristics of the audio signal, and the frequency amplitude of the signal is concentrated in several specific values: 0, 0.5, and 1.5, and it is impossible to obtain valuable information based on the frequency distribution. It can be seen from the spectrum diagram of the decrypted plaintext audio signal that the spectral characteristics of the decrypted audio signal are highly consistent with the spectral characteristics of the original plaintext audio signal, that is, the decryption algorithm can effectively restore the audio plaintext data.

[0044] As attached Figure 5 and attached Figure 6 As shown in the figure, a histogram is drawn based on the distribution of the audio signal to verify whether the encryption algorithm has effectively changed the signal distribution of the original plaintext audio data. The horizontal axis of the histogram represents the amplitude of the signal, and the vertical axis represents the proportion of the signal at this amplitude to the total number of signals. Figure 5 is the signal distribution histogram of the original plaintext two-channel data, which is generally normally distributed with the axis of symmetry being x=0. Figure 6 This is the signal distribution histogram of the two channel data of the ciphertext. It can be seen that its amplitude range is 0~1, which is different from the interval symmetric about x=0 of the original plaintext. Its frequency value and frequency statistical range have changed significantly, indicating that the encryption algorithm of this application has effectively changed the distribution law of the audio signal.

[0045] Table 2 Audio correlation data Table 3 Sample PSNR averages Tables 2 and 3 respectively list the correlation degree of the left and right channel data before and after encryption from the perspective of audio data correlation, and the peak signal-to-noise ratio (PSNR) of the left and right channels before and after audio encryption to evaluate the encryption effect of audio. As can be seen from Tables 2 and 3, the correlation coefficients of the two channel data before and after encryption are very close to 0, indicating that the audio encryption algorithm in this paper has effectively changed the audio data; the PSNR values ​​of the left and right channel data are both small, indicating that the encrypted audio is significantly different from the audio before encryption, effectively hiding the information in the audio data, and can better resist cracking attacks based on peak signal-to-noise ratio statistics.

[0046] The disclosed embodiment also provides a dual-channel audio encryption device, including a processor and a memory. Optionally, the device may also include a communication interface and a bus. The processor, the communication interface, and the memory may communicate with each other through the bus. The communication interface may be used for information transmission. The processor may call the logic instructions in the memory to execute a dual-channel audio encryption method of the above embodiment.

[0047] Anything not described in this application can be achieved by adopting or drawing on existing technologies. The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A dual-channel audio encryption method, characterized in that: The following steps are involved: Step 1: Get the two-channel audio data matrix Audio , whose dimensions are L ×2, which means as follows: ； In the formula, the first column represents the left channel data from top to bottom, and the second column represents the right channel data from top to bottom; Step 2: Audio Perform matrix transformation operation of Euclidean subtraction to obtain Audio 1, which means as follows: ； Step 3: Create a dimension L ×2 key matrix Rand 1. Rand 1 and Audio The elements at the corresponding positions in 1 are averaged and then established Audio 2; Step 4: Audio The elements in 2 are grouped and expanded by columns to form Audio 3. In Audio In 3, each group contains 16 elements in sequence and when the last group has less than 16 elements, it is padded with 0; Audio The specific form of 3 is as follows: ； Step 5: Use the encrypted model of the 1+1 dimensional nonlinear fractional-order pseudo-parabolic equation to Audio 3. The encryption model is: ； In the formula, u and t is a parameter variable; express right Find the second-order partial derivatives of Find the first-order partial derivative; express right Find the first-order partial derivative; for t is the function at time zero; is the boundary condition; f ( x , t ) is a given function; T is a constant; Indicates about x Caputo fractional terms; During encryption, the nonlinear fractional-order pseudo-parabolic equation is directly linearized and iteratively solved by calculating the initial moment, and the encryption of the remaining moments is completed, which specifically includes the following steps: Step 1: Solve Audio 3 corresponds to the function value at the initial time ,include: The spatial variable x and time t Discrete into: ； Then the discrete space variable at the initial moment is x The nodes are: ； Then the boundary condition The corresponding discrete nodes are: ； Step 2: Discrete space variables at the initial moment x Node substitution function In the initial moment, we get the function value matrix , which is expressed as follows: ； Step 3: Construct one and Audio 3. A calculation matrix with the same dimensions, and in the calculation matrix, wrap it with 0 Audio 3 is the non-zero initial time function value; the calculation matrix is ​​in the following form: ； Step 4: Take Audio Each column of data in 3 is filled with the calculation matrix to obtain the encryption matrix at the initial moment: ； In the formula, the lower left corner subscript of the element means the i ( i =2,3,4,5,6) times are calculated; Step 5: Calculate the matrix form of the encryption model based on the direct linearization iterative format of the nonlinear fractional-order pseudo-parabolic equation T Function value matrix at time , which is expressed as follows: ； Based on the calculation matrix construction principle in step three and step four, we get T The encryption matrix of the moment: ； Step 6: Repeat step 5 to Audio 3. Each data in the encryption is encrypted. Audio The original data in 3 is arranged in order and the encrypted data is obtained to obtain the final encryption matrix Audio 4, expressed as: ； Step 6: Construct Dimensions and Audio 4 Same numerical label matrix G , the numerical label matrix G The initial values ​​of the elements in are all 0; Will Audio Remove the negative sign from the negative elements in 4. When you remove the negative sign of an element, that is, in the matrix G Add a value to the element position at the corresponding position; after all is completed, write Audio 4 for ;right The values ​​of the elements in the matrix are constrained. G Mark the value; after the constraint Expand to a column, delete some data at the end of the expanded column, the number of deleted data is the same as the number of 0s added in step 4; after the constraint operation is completed, record the numerical labeling matrix G is the key matrix; The front of the expanded column data L Data and L The data are respectively used as the first and second columns to form a dual-channel ciphertext audio data matrix Audio 5.

2. A dual-channel audio encryption method according to claim 1, characterized in that: In step 5, the matrix form of the direct linearization iterative format encryption model of the nonlinear fractional-order pseudo-parabolic equation is: ； ； In the formula, R 1 represents the coefficient matrix, and They are x , t of p 1. p The second-order differential matrix, is the integral matrix, is the gamma function, and They are m +1st level, l +1st-order identity matrix; f For a given function.

3. A dual-channel audio encryption method according to claim 2, characterized in that: In the matrix of the direct linearization iterative format encryption model of the nonlinear fractional pseudo-parabolic equation, the permutation method is used to impose boundary conditions and initial adjustments, and the accuracy of the iterative calculation is set to e=10 -9 .

4. A dual-channel audio encryption method according to claim 3, characterized in that: In step six, The values ​​of the elements in the matrix are constrained and G The method for numerical marking is: When the element value is in the interval [0,1), keep and matrix G Original data; When the element value is in the interval [1,10), divide the element value by 10 in the matrix G The corresponding element position of plus 10; When the element value is in the interval [10,20), divide the element value by 40 and add it to the matrix G The corresponding element position of plus 100; When the element value is in the interval [20,40), divide the element value by 80 and add it to the matrix G The corresponding element position of plus 1000; When the element value is in the interval [40,80), divide the element value by 160 and add it to the matrix G The corresponding element position of plus 10000; When the element value is in the interval [80,200), divide the element value by 400 and add it to the matrix G The corresponding element position of plus 100000; When the element value is greater than or equal to 200, in the matrix G The corresponding element position of is reset to zero.

5. A dual-channel audio encryption device, comprising a processor and a memory storing program instructions, characterized in that: The processor is configured to execute a dual-channel audio encryption method as described in any one of claims 1 to 4 when running the program instructions.

Images