Audio scrambling method based on tornado transform

Abstract

At present, the activities of transmitting audio files and communicating with voice on the Internet are becoming more and more frequent. In order to protect the security of audio content, a tornado transform is designed, and an audio scrambling method based on tornado transform is proposed. By adopting tornado transform, the pixel value of each layer of tornado is converted into a vector in a certain way, and these vectors are randomly cyclically shifted respectively; then the elements in the new vector after cyclic shift are placed in the pixel position of each layer of tornado in a certain way to realize the scrambling operation. The experimental results and method analysis show that this method has strong security and good performance, and improves the security and efficiency of the scrambling process.

Description

Technical Field

[0001] The invention relates to an audio scrambling method, which mainly performs scrambling on a single audio file. Background Art

[0002] In recent years, with the development of big data technology and the popularization of 5G, people are increasingly relying on mobile phones, the Internet and other tools for voice communication and exchange. At the same time, there are great risks in the network. Problems such as information leakage and malicious tampering are emerging in an endless stream. Audio encryption uses the matrix characteristics of digital audio to scramble or diffuse according to specific transformation rules, making the original audio information "chaotic".

[0003] Audio encryption mainly includes two methods: phoneme scrambling and phoneme diffusion. The purpose of phoneme scrambling is to scramble the positions of phonemes; the purpose of phoneme diffusion is to change the values ​​of phonemes. In order to improve the security and efficiency of audio encryption and ensure the safe and efficient transmission of audio, an audio scrambling method based on tornado transformation is designed using chaos theory. This method combines chaos and tornado transformation to effectively protect the security of audio file network transmission and storage, lays the foundation for further audio encryption, and improves the security and efficiency of the scrambling process. Summary of the invention

[0004] The purpose of the present invention is to protect audio information from being stolen and to improve the security and efficiency of encryption during network transmission. An audio scrambling method based on tornado transformation is proposed.

[0005] Technical solution of the present invention: To achieve the above-mentioned purpose of the invention, the technical solution adopted is an audio scrambling method based on tornado transformation, and the scrambling steps are described in detail as follows:

[0006] Step 1: Upscaling audio data: Let the interactive audio file be A 1 , whose size is m ×1, m is the number of phonemes, 1 means the audio file is mono; A 1 Convert to a size of s × t The matrix A 2 ,in, s, t is a positive integer, m = s × t ;

[0007] Step 2: Calculation A 2 Number of layers: A 2 The rectangle formed by the outermost phoneme valuesC 1 , called the first layer of a tornado, A 2 The rectangle formed by the phoneme values ​​of the second outer circle C 2 , called the second layer of the tornado, ..., layer by layer, until A 2 Central location C k , called the tornado k Layer; the total number of layers in a tornado k for:

[0008] k = min ( ceil (0.5× s ), ceil (0.5× t )), (1)

[0009] in, min (•) is the minimum function, ceil (•) is the ceiling function;

[0010] Step 3: Generate chaotic sequence: Logistic mapping is:

[0011] x n+1 = rx n (1+ x n ), (2)

[0012] in, r ∈[3.57, 4) is the control parameter, x n ∈(0, 1); randomly selected r and initial value x 0, iterate formula (2) k times, we can get a length of k Chaotic sequence L 1 ;

[0013] Step 4: Integerization of chaotic sequence: Use formula (3) to L 1 Integerization,

[0014] L 2 = floor ( mod ( L 1 ×10 16 , 2×(s + t ))), (3)

[0015] in, mod (•) represents the modulo operation, floor (•) is the floor function, L 2 For a length of k Integer chaotic sequence of ;

[0016] Step 5: Tornado transformation: According to a certain method, the phoneme value of each layer of the tornado is converted into a vector, and these vectors are randomly cyclically shifted; then the elements in the new vector after cyclic shift are placed in the phoneme position of each layer of the tornado according to a certain method;

[0017] Step 6: Generate a scrambled matrix: Use the tornado transformation to C i All phoneme values ​​are converted into vectors H i ,use L 2 And formula (4) H i Perform random circular shift operations,

[0018] P i = circshift ( H i , L 2 ( i )), i =1, 2, …, k , (4)

[0019] in, L 2 ( i )∈ L 2 , circshift ( H i , L 2 ( i )) is correct H i Cyclic shift L 2 ( i ) bit function, P i is the new vector after circular shift; P i All the phoneme values ​​in are placed in the tornado i In the phoneme position of the layer,A 2 The scrambling operation can obtain a size of s × t The scrambled phoneme matrix A 3 ;

[0020] Step 7: Audio data dimensionality reduction: A 3 Convert to a length of m Vector A 4 , which is scrambled audio.

[0021] In the process of audio decryption, the scrambled audio is operated using the same chaotic sequence to restore the original audio. The inverse scrambling process is the inverse process of audio scrambling.

[0022] Beneficial effects: In order to protect audio information from being stolen and improve the security and efficiency of encryption during network transmission, an audio scrambling method based on tornado transform is proposed. The main contributions are: (1) Inspired by tornado, a tornado transform is designed for audio scrambling; (2) Using tornado transform, an audio scrambling method based on tornado transform is proposed to realize the scrambling operation of the original audio; (3) Experimental results show that the new method has achieved good results in terms of visual effect and scrambling degree. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 : Scrambling flow chart;

[0024] Figure 2 : Time domain waveform of the original audio;

[0025] Figure 3 : Time domain waveform of scrambled audio. DETAILED DESCRIPTION

[0026] The implementation process of the present invention is further described in detail below with reference to specific drawings and examples.

[0027] Figure 1 This is the scrambling flow chart of this method.

[0028] The programming software used is Matlab R2016a. An audio file with a size of 73710×1 is selected as the original audio. The time domain waveform of the original audio is as follows: Figure 2 As shown; using this method, the detailed process of scrambling the original audio is described as follows.

[0029] Step 1: Upscaling audio data: Let the interactive audio file be A 1, whose size is 73710×1, 73710 is the number of phonemes, and 1 means the audio file is mono. A 1 Converted into a matrix of size 270×273 A 2 .

[0030] Step 2: Calculation A 2 Number of layers: Use formula (1) to calculate the total number of layers of a tornado k is 135.

[0031] Step 3: Generate chaotic sequence: randomly select r and initial value x 0, iterate formula (2) 135 times to get a chaotic sequence with a length of 135 L 1 .

[0032] Step 4: Integerization of chaotic sequence: Use formula (3) to convert L 1 Integerization gives an integer chaotic sequence with a length of 135. L 2 .

[0033] Step 5: Tornado transformation: According to a certain method, the phoneme value of each layer of the tornado is converted into a vector, and these vectors are randomly cyclically shifted; then the elements in the new vector after the cyclic shift are placed in the phoneme position of each layer of the tornado according to a certain method.

[0034] Step 6: Generate a scrambled matrix: Use the tornado transformation to C i All phoneme values ​​are converted into vectors H i ,use L 2 And formula (4) H i Perform random circular shift operations, P i is the new vector after circular shift; P i All the phoneme values ​​in are placed in the tornado i In the phoneme position of the layer, A 2 The scrambling operation can obtain a scrambled phoneme matrix of size 270×273 A 3 .

[0035] Step 7: Audio data dimensionality reduction: A3 Converted into a vector of length 73710 A 4 , which is the scrambled audio, the time domain waveform of the scrambled audio, such as Figure 3 shown.

[0036] In the audio decryption process, the scrambled audio is decrypted using the same chaotic sequence to restore the original audio. The inverse scrambling process is the inverse process of scrambling.

Claims

1. An audio scrambling method based on tornado transformation, characterized in that: The scrambling process includes the following steps: Step 1: Upscaling audio data: Let the interactive audio file be A 1 , whose size is m ×1, m is the number of phonemes, 1 means the audio file is mono; A 1 Convert to a size of s × t The matrix A 2 ,in, s, t is a positive integer, m = s × t ; Step 2: Calculation A 2 Number of layers: A 2 The rectangle formed by the outermost phoneme values C 1 , called the first layer of a tornado, A 2 The rectangle formed by the phoneme values ​​of the second outer circle C 2 , called the second layer of the tornado, ..., layer by layer, until A 2 Central location C k , called the tornado k Layer; the total number of layers in a tornado k for: k = min ( ceil (0.5× s ), ceil (0.5× t ))， （1） in, min (•) is the minimum function, ceil (•) is the ceiling function; Step 3: Generate chaotic sequence: Logistic mapping is: x n+1 = rx n (1+ x n ）， （2） in, r ∈[3.57, 4) is the control parameter, x n ∈(0, 1); randomly selected r and initial value x 0, iterate formula (2) k times, we can get a length of k Chaotic sequence L 1 ; Step 4: Integerization of chaotic sequence: Use formula (3) to L 1 Integerization, L 2 = floor ( mod ( L 1 ×10 16 , 2×( s + t )))， （3） in, mod (•) represents the modulo operation, floor (•) is the floor function, L 2 For a length of k Integer chaotic sequence of ; Step 5: Tornado transformation: According to a certain method, the phoneme value of each layer of the tornado is converted into a vector, and these vectors are randomly cyclically shifted; then the elements in the new vector after cyclic shift are placed in the phoneme position of each layer of the tornado according to a certain method; Step 6: Generate a scrambled matrix: Use the tornado transformation to C i All phoneme values ​​are converted into vectors H i ,use L 2 And formula (4) H i Perform random circular shift operations, P i = circshift ( H i , L 2 ( i ))， i =1, 2, …, k ， （4） in, L 2 ( i )∈ L 2 , circshift ( H i , L 2 ( i )) is correct H i Cyclic shift L 2 ( i ) bit function, P i is the new vector after circular shift; P i All the phoneme values ​​in are placed in the tornado i In the phoneme position of the layer, A 2 The scrambling operation can obtain a size of s × t The scrambled phoneme matrix A 3 ; Step 7: Audio data dimensionality reduction: A 3 Convert to a length of m Vector A 4 , which is scrambled audio.

Images