Analog audio multi-resolution lossless reversible privacy protection method and system

Abstract

The application provides an analog audio multi-resolution lossless reversible privacy protection method and system, and the method comprises the following steps: sequentially performing reading and coding operations on original digital audio to obtain single-precision floating-point bit strings; arranging all single-precision floating-point bit strings in sequence and calculating numerical contributions; calculating base matrix elements by using the numerical contributions; and arranging the base matrix elements in sequence to obtain a base matrix; dividing the single-precision floating-point bit strings into a third bit block, a second bit block and a first bit block by using three preset bit numbers; restoring all bit blocks under a third authorization level; and taking the bit string matrix dimension being the same and the difference bit number being 0 as a judgment condition to realize verifiable bit-level lossless restoration.

Description

Technical Field

[0001] This invention belongs to the field of audio data encryption technology, and specifically relates to a method and system for simulating multi-resolution lossless and reversible privacy protection of audio. Background Technology

[0002] The core objective of privacy protection is to ensure that sensitive personal information is not misused or leaked. When the data being processed is audio data such as voice, music, or ambient sounds, audio privacy becomes an issue. Audio privacy protection requires ensuring the usability of audio information while restricting unauthorized access to the complete audio content and its underlying sample information; therefore, a balance must be struck between privacy security and authorized usability.

[0003] A common method for protecting audio privacy is to encrypt the entire audio data. This method converts the raw audio into ciphertext that only users with the full key can decrypt, thus ensuring the security of the audio during storage and transmission. However, overall encryption also prevents users without the full key from obtaining any usable audio information. In scenarios involving cloud storage, multi-party sharing, and tiered access, it is difficult to provide recovery results with varying degrees of accuracy to different authorized users.

[0004] When an audio file needs to be shared and accessed by multiple users, one of the main challenges is how to control the audio information that each user can recover based on their different authorization levels. One approach is to divide the underlying bit representation of the audio sample into multiple bit blocks and set interrelated encryption parameters and access starting points for different bit blocks. This allows lower-authority users to recover only a portion of the bit blocks, while higher-authority users can continue to recover more bit blocks, thus creating different levels of audio availability.

[0005] On the other hand, audio samples can be represented in computers using floating-point numbers. When converting between sample values, floating-point bit strings, encrypted data, and audio files, quantization, truncation, normalization, transcoding, or other processes that alter the underlying numerical representation can cause irreversible differences between the recovered sample and the original sample. In scenarios where the highest-authority user needs complete access to the original data, it is difficult to prove, based solely on subjective listening impressions or waveform similarity, that each recovered audio sample is completely identical to the original sample in its underlying bit representation.

[0006] Another challenge is how to protect audio privacy without sacrificing authorized users' ability to use audio information. If encryption and access control methods are too simplistic, users can only choose between complete unavailability and complete public access; if the processing structure is too complex, it may increase storage, transmission, and recovery overhead. Therefore, an audio privacy protection scheme is needed that can directly provide hierarchical protection based on the bit representation of audio samples and achieve verifiable lossless recovery at the highest authorization level. Summary of the Invention

[0007] In view of the above, the main objective of this invention is to propose a method and system for lossless and reversible privacy protection of analog audio multi-resolution, in order to solve the above-mentioned technical problems.

[0008] This invention proposes a method for lossless and reversible privacy protection of simulated audio multi-resolution, the method comprising the following steps: Step 1: Read and encode the original digital audio sequentially to obtain single-precision floating-point bit strings. Arrange all single-precision floating-point bit strings in order and calculate their numerical contributions. Calculate the elements of the basis matrix using the numerical contributions, and then arrange them in order to obtain the basis matrix. Divide the single-precision floating-point bit strings into the third, second, and first bit blocks using three preset bit lengths. Step 2: Encrypt the third bit block using the preset first key and first initialization vector to obtain the ciphertext of the third bit block, and derive the second key and second initialization vector. Then encrypt the second bit block and the first bit block in sequence to obtain the ciphertext of the second bit block and the ciphertext of the first bit block, and derive the third key and the third initialization vector. Reassemble the ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block in their original order and calculate the ciphertext value contribution. Then combine it with the basis matrix to obtain the ciphertext audio sample. Step 3: Associate and save the ciphertext audio sample, three bit block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors to obtain three different sets of ciphertext data. Step 4: Distribute three different sets of encrypted data to the receiving end according to three preset authorization levels. Decrypt the three sets of encrypted data and reassemble them in order to obtain the recovery bit string of the corresponding authorization level. Calculate the recovery value contribution of the recovery bit string of the corresponding authorization level. Then combine the recovery value contribution with the basis matrix to obtain the recovery audio sample of the corresponding authorization level. Use the recovery audio sample of the corresponding authorization level to determine whether bit-level lossless recovery has been achieved, so as to obtain the final lossless privacy protection result.

[0009] This invention also proposes a simulated audio multi-resolution lossless reversible privacy protection system, the system comprising: The audio sample processing module is used for: The original digital audio is read and encoded sequentially to obtain single-precision floating-point bit strings. All single-precision floating-point bit strings are arranged in order and their numerical contributions are calculated. The elements of the basis matrix are calculated using the numerical contributions and then arranged in order to obtain the basis matrix. The single-precision floating-point bit strings are divided into the third bit block, the second bit block, and the first bit block using three preset bit lengths. The encrypted audio processing module is used for: The third bit block is encrypted using the preset first key and first initialization vector to obtain the ciphertext of the third bit block. The second key and second initialization vector are then derived. The second bit block and the first bit block are then encrypted in sequence to obtain the ciphertext of the second bit block and the ciphertext of the first bit block. The third key and the third initialization vector are then derived. The ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block are then recombined in their original order and the ciphertext numerical contribution is calculated. Finally, the ciphertext is combined with the basis matrix to obtain the ciphertext audio sample. The key processing module is used for: The encrypted audio sample, three bit-block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors are associated and saved separately to obtain three different sets of ciphertext data. The audio recovery module is used for: Three different sets of encrypted data are distributed to the receiving end according to three different preset authorization levels. The three sets of encrypted data are decrypted and reassembled in order to obtain the recovery bit string of the corresponding authorization level. The recovery value contribution of the recovery bit string of the corresponding authorization level is calculated. The recovery value contribution is then combined with the basis matrix to obtain the recovery audio sample of the corresponding authorization level. Using the recovery audio sample of the corresponding authorization level, it is determined whether bit-level lossless recovery has been achieved to obtain the final lossless privacy protection result.

[0010] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Compared with existing technologies, this invention directly targets the IEEE 754 standard 32-bit single-precision floating-point bit string for block protection, and maintains the reconstructability of audio samples through a combination of basis matrix and numerical contribution. 2. By using three bit blocks, the corresponding key and the corresponding initialization vector derivation, and the distribution of authorization start parameters, the same ciphertext data can be encrypted once, and different authorization levels can obtain audio data with different bit block recovery ranges, thereby achieving a balance between privacy and availability. 3. Under the third authorization level, all bit blocks are restored, and the same bit string matrix dimension and zero difference in bit bits are used as the judgment condition to achieve verifiable bit-level lossless recovery.

[0011] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by means of embodiments of the invention. Attached Figure Description

[0012] Figure 1 This is a flowchart of the steps of the simulated audio multi-resolution lossless reversible privacy protection method proposed in this invention; Figure 2This is a comparison diagram of the original audio, encrypted audio, and difference signals of the analog audio multi-resolution lossless and reversible privacy protection method proposed in this invention. Figure 3 This is a comparison chart showing the difference between the recovered audio under different authorization levels and the recovered audio under the third authorization level compared with the original audio using the simulated audio multi-resolution lossless reversible privacy protection method proposed in this invention. Figure 4 This is a schematic diagram of the module composition of the analog audio multi-resolution lossless reversible privacy protection system proposed in this invention. Detailed Implementation

[0013] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0014] These and other aspects of the embodiments of the present invention will become clear from the following description and accompanying drawings. In these descriptions and drawings, some specific embodiments of the present invention are specifically disclosed to provide some ways of implementing the principles of the embodiments of the present invention; however, it should be understood that the scope of the embodiments of the present invention is not limited thereto.

[0015] refer to Figure 1 This invention proposes a method for lossless and reversible privacy protection of simulated audio multi-resolution, which includes the following steps: Step 1: Read and encode the original digital audio sequentially to obtain single-precision floating-point bit strings. Arrange all single-precision floating-point bit strings in order and calculate their numerical contributions. Calculate the elements of the basis matrix using the numerical contributions, and then arrange them in order to obtain the basis matrix. Divide the single-precision floating-point bit strings into the third, second, and first bit blocks using three preset bit lengths.

[0016] In step 1, the original digital audio is read and encoded sequentially to obtain single-precision floating-point bit strings. All single-precision floating-point bit strings are arranged in order and their numerical contributions are calculated. The elements of the basis matrix are calculated using the numerical contributions and then arranged in order to obtain the basis matrix. The single-precision floating-point bit strings are then divided into a third bit block, a second bit block, and a first bit block using three preset bit lengths. The specific steps include the following: Read the raw digital audio and construct a column vector of audio sample values ​​according to the sampling order; Each sample in the audio sample value column vector is encoded using single-precision floating-point encoding to obtain a single-precision floating-point bit string; the single-precision floating-point bit string includes a sign bit, an exponent field, and a mantissa field; Arrange all the single-precision floating-point bit strings in order to obtain the original bit string matrix; Convert the exponent field in the original bit string matrix into an unsigned decimal value to determine the exponent weight; The numerical contribution under the corresponding exponential weight is calculated using the last digit segment of the single-precision floating-point bit string. The sign coefficients of the samples are determined based on the original audio sample values ​​in the column vector of audio sample values. Subtract the absolute value of the original audio sample value from the numerical contribution, and then multiply it with the sample sign coefficient to obtain the basis matrix elements; Arrange all the elements of the basis matrix in order to obtain the basis matrix; Each single-precision floating-point bit string is divided into the first segment, the second segment, and the third segment in sequence from the high bit to the middle bit and then to the low bit using three preset bit numbers, respectively. The first segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the first bit block; the second segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the second bit block; the third segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the third bit block.

[0017] The process involves reading the original digital audio and constructing a column vector of audio sample values ​​according to the sampling order. The corresponding relationship in this process is as follows: ; in, Represents a column vector of audio sample values. This represents the value of the first original audio sample. This represents the value of the second original audio sample. This represents the last original audio sample value. This represents the total number of original audio sample values. Indicates transpose; A single-precision floating-point bit string includes a sign bit, an exponent field, and a mantissa field. The corresponding relationship in the process is as follows: ; in, Indicates the first A single-precision floating-point bit string of an audio sample. Indicates the first The sign bit of each audio sample This indicates a bit string concatenation operation. Indicates the first The index field of each audio sample, Indicates the first The last digit segment of each audio sample; Arrange all the single-precision floating-point bit strings in order to obtain the original bit string matrix. The corresponding relationship in this process is as follows: ; in, Represents the original bit string matrix. This represents the first single-precision floating-point bit string. This represents the second single-precision floating-point bit string. This represents the last single-precision floating-point bit string. This represents the total number of single-precision floating-point bits. The exponent field in the original bit string matrix is ​​converted into an unsigned decimal value to determine the exponent weight. The corresponding relationship in this process is as follows: ; in, Indicates the first The unsigned decimal value of each audio sample. Indicates the positional order within the index field. Indicates the first The index field of the audio samples, arranged from highest to lowest, is the [number]th [sample]. Bit, Indicates the first The exponential weights of each audio sample; The numerical contribution under the corresponding exponential weight is calculated using the last digit segment of the single-precision floating-point bit string. The corresponding relationship in this process is as follows: ; in, Indicates the first The numerical contribution of each audio sample. This represents the positional index within the last digit field. Indicates the first The last digit segment of the audio sample The bit value at each bit; Based on the original audio sample values ​​in the audio sample value column vector, the sample sign coefficients are determined. The corresponding relationship in this process is as follows: ; in, Indicates the first The sample sign coefficients of each audio sample. Indicates the first The original audio sample values ​​of each audio sample; The absolute value of the original audio sample value is subtracted from the numerical contribution, and then multiplied by the sample sign coefficient to obtain the basis matrix elements. The corresponding relationship in this process is as follows: ; in, Indicates the first The basis matrix elements of each audio sample, Indicates taking the absolute value; The first segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the first bit block; the second segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the second bit block; the third segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the third bit block. The corresponding relationship in this process is as follows: ; in, Indicates the first Units block This represents the first single-precision floating-point bit string. Units segment This represents the second single-precision floating-point bit string. Units segment This indicates the last single-precision floating-point bit string. The units digit segment, among which and The values ​​are equal.

[0018] It should be noted that in step 1, the method of separating the basis values ​​from the numerical contributions and corresponding the index numbers of the same sample is adopted to ensure that the elements of the basis matrix correspond one-to-one with the three different bit blocks, thus providing a numerical basis for subsequent recovery at different precisions.

[0019] Step 2: Encrypt the third bit block using the preset first key and first initialization vector to obtain the ciphertext of the third bit block, and derive the second key and second initialization vector. Then encrypt the second bit block and the first bit block in sequence to obtain the ciphertext of the second bit block and the ciphertext of the first bit block, and derive the third key and the third initialization vector. Reassemble the ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block in their original order and calculate the ciphertext value contribution. Then combine it with the basis matrix to obtain the ciphertext audio sample.

[0020] In step 2, the third bit block is encrypted using a preset first key and first initialization vector to obtain the ciphertext of the third bit block. A second key and second initialization vector are then derived, and the second and first bit blocks are encrypted sequentially to obtain the ciphertext of the second and first bit blocks. A third key and third initialization vector are then derived. The ciphertexts of the third, second, and first bit blocks are then reassembled in their original order, and the ciphertext numerical contribution is calculated. Finally, this is combined with the basis matrix to obtain the ciphertext audio sample. The specific steps include: The third bit block is encrypted using AES-CFB block encryption with the preset first key and the preset first initialization vector to obtain the ciphertext of the third bit block. Extract the first sub-block of the third bit block and the first sub-block of the ciphertext of the third bit block respectively to obtain the first plaintext sub-block and the first ciphertext sub-block of the third bit block; concatenate the first plaintext sub-block and the first ciphertext sub-block of the third bit block to obtain the first concatenation result; perform SHA-256 hash operation on the first concatenation result to obtain the second key; The first initial vector is encrypted and concatenated sequentially using the first preset key to obtain the second concatenation result. The second concatenation result is then hashed using SHA-256 to obtain the first hash result. A preset 128 bits are extracted from the first hash result to obtain the second initial vector. The second bit block is encrypted using AES-CFB block encryption with the second key and the second initialization vector to obtain the ciphertext of the second bit block. Extract the first sub-block of the second bit block and the first sub-block of the ciphertext of the second bit block respectively to obtain the first plaintext sub-block and the first ciphertext sub-block of the second bit block; concatenate the first plaintext sub-block and the first ciphertext sub-block of the second bit block to obtain the third concatenation result; perform SHA-256 hash operation on the third concatenation result to obtain the third key; The second initial vector is encrypted and concatenated using the second key to obtain the fourth concatenation result. The fourth concatenation result is then hashed using SHA-256 to obtain the second hash result. A preset 128 bits are extracted from the second hash result to obtain the third initial vector. The first bit block is encrypted using AES-CFB block encryption with the third key and the third initialization vector to obtain the ciphertext of the first bit block. The ciphertext bit segments in the first bit block ciphertext, the ciphertext bit segments in the second bit block ciphertext, and the ciphertext bit segments in the third bit block ciphertext are reassembled in order to obtain the ciphertext reassembled bit string corresponding to each audio sample; Extract the exponent field and tail number field from the ciphertext recombinant bit string corresponding to each audio sample, and calculate the ciphertext numerical contribution of each audio sample; combine the ciphertext numerical contribution of each audio sample with the corresponding element in the basis matrix to obtain the ciphertext audio sample.

[0021] Specifically, the third bit block is encrypted using AES-CFB block encryption with a preset first key and a preset first initialization vector to obtain the ciphertext of the third bit block. The corresponding relationship in this process is as follows: ; in, This indicates the ciphertext of the third bit block. This indicates AES-CFB block encryption operation. This indicates the third bit block. This represents the first preset key. This represents the first preset initial vector; The first plaintext sub-block and the first ciphertext sub-block of the third bit block are concatenated to obtain the first concatenation result. The first concatenation result is then subjected to a SHA-256 hash operation to obtain the second key. The corresponding relationship in this process is as follows: ; in, This represents the second key. This indicates that a conversion to a fixed-length 256-bit hash value will be performed. This indicates the first plaintext sub-block of the third bit block. This indicates the first ciphertext subblock of the third bit block; Using the preset first key, the preset first initial vector is sequentially encrypted with AES and then concatenated to obtain the second concatenated result. The second concatenated result is then subjected to SHA-256 hashing to obtain the first hash result. A preset 128 bits are extracted from the first hash result to obtain the second initial vector. The corresponding relationship in this process is as follows: ; in, This represents the second initial vector. This indicates that a preset 128-bit operation will be performed to use as the initial vector for the next layer. This represents the second concatenation result obtained by performing AES encryption on a preset first initialization vector using a preset first key; The second bit block is encrypted using AES-CFB block encryption with the second key and the second initialization vector to obtain the ciphertext of the second bit block. The corresponding relationship in the process is as follows: ; in, This indicates the ciphertext of the second bit block. Indicates the second bit block; The first plaintext sub-block and the first ciphertext sub-block of the second bit block are concatenated to obtain the third concatenation result. The third concatenation result is then subjected to a SHA-256 hash operation to obtain the third key. The corresponding relationship in this process is as follows: ; in, This represents the third key. This indicates the first plaintext sub-block of the second bit block. This indicates the first ciphertext subblock of the second bit block; The second initial vector is encrypted and concatenated using the second key to obtain the fourth concatenated result. A SHA-256 hash operation is then performed on the fourth concatenated result to obtain the second hash result. A preset 128 bits are extracted from the second hash result to obtain the third initial vector. The corresponding relationship in this process is as follows: ; in, This represents the third initial vector. This represents the fourth concatenation result obtained by performing AES encryption on the second initial vector using the second key; The first bit block is encrypted using AES-CFB block encryption with the third key and the third initialization vector to obtain the ciphertext of the first bit block. The corresponding relationship in the process is as follows: ; in, This indicates the ciphertext of the first bit block. Indicates the first bit block; The ciphertext segments in the first, second, and third bit blocks are reassembled sequentially to obtain the reassembled ciphertext bit string corresponding to each audio sample. The relationship in this process is as follows: ; in, Indicates the first The ciphertext reconstructed bit string corresponding to each audio sample Indicates the first The ciphertext bit field in the first bit block ciphertext corresponding to each audio sample. Indicates the first The ciphertext bit segment in the second bit block ciphertext corresponding to each audio sample. Indicates the first The ciphertext bit segment in the third bit block ciphertext corresponding to each audio sample; Extract the exponent field and tail number segment from the ciphertext reassembled bit string corresponding to each audio sample, and calculate the ciphertext numerical contribution of each audio sample; combine the ciphertext numerical contribution of each audio sample with the corresponding element in the basis matrix to obtain the ciphertext audio sample. The relationship in the correspondence process is as follows: ; in, Indicates the first One audio sample ciphertext audio sample, Indicates from the first The ciphertext numerical contribution of each audio sample.

[0022] It should be noted that: Step 2 uses the corresponding key and the corresponding initial vector to derive layer by layer and encrypt bit blocks in sequence, so that the encryption level of the three bit blocks corresponds to the parameter derivation level; at the same time, the numerical contribution of the bit block ciphertext is combined with the basis matrix to construct the ciphertext audio sample.

[0023] It needs to be further explained that: This embodiment performs bit-level difference statistics on encrypted audio samples under different combinations of bit block lengths; The test audio sampling rate was 16000Hz, the duration was 5s, and the number of samples was 80000. The original audio, the encrypted audio, and the difference signal between the two are as follows: Figure 2 As shown; Please see Figure 2 ,exist Figure 2 In Figure 2 In the image (a), the original audio time-domain waveform is shown. Figure 2 (b) in the image represents the time-domain waveform of the encrypted audio. Figure 2 (c) in the figure represents the difference signal between the original audio and the encrypted audio; The original audio samples and the encrypted audio samples were converted into IEEE 754 standard 32-bit single-precision floating-point strings respectively. The number of different samples and the number of different bits were counted. The results are shown in Table 1.

[0024] Table 1 shows the bit-level differences in ciphertext audio under different bit block parameters.

[0025] As shown in Table 1, under the six sets of bit block parameters mentioned above, all 80,000 ciphertext audio samples showed bit-level differences compared to their corresponding original audio samples, with a difference rate of 100% and a bit difference rate ranging from 43.4586% to 43.5844%. These data indicate that different bit block division methods can cause significant bit-level changes in the ciphertext bit string compared to the original bit string.

[0026] Step 3: Associate and save the ciphertext audio sample, three bit-block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors to obtain three different sets of ciphertext data.

[0027] In step 3, the encrypted audio sample, three bit-block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initialization vectors are associated and saved respectively to obtain three different sets of ciphertext data. The specific steps include the following: The encrypted audio sample, the first bit block ciphertext, the second bit block ciphertext, the third bit block ciphertext, the basis matrix, three preset bit lengths, the third key, and the third initial vector are associated and saved to obtain the first ciphertext data set; The encrypted audio sample, the first bit block ciphertext, the second bit block ciphertext, the third bit block ciphertext, the basis matrix, three preset bit lengths, the second key, and the second initial vector are associated and saved to obtain the second ciphertext data set; The encrypted audio sample, the first bit block encrypted text, the second bit block encrypted text, the third bit block encrypted text, the basis matrix, three preset bit lengths, the preset first key, and the preset first initial vector are associated and saved to obtain the third encrypted data set.

[0028] Specifically, the encrypted audio sample, the first bit block of encrypted text, the second bit block of encrypted text, the third bit block of encrypted text, the basis matrix, three preset bit lengths, the third key, and the third initial vector are associated and stored to obtain the first encrypted data set. The corresponding relationship in this process is as follows: ; in, This represents the first set of encrypted data. Represents the basis matrix, This represents the first preset bit position out of three preset bit positions. This represents the second preset number of bits out of three preset number of bits. This indicates the third preset number of bits out of three preset number of bits.

[0029] The encrypted audio sample, the first bit block of encrypted text, the second bit block of encrypted text, the third bit block of encrypted text, the basis matrix, three preset bit lengths, the second key, and the second initialization vector are associated and saved to obtain the second encrypted data set. The corresponding relationship in this process is as follows: ; in, This represents the second set of encrypted data; The encrypted audio sample, the first bit block of encrypted text, the second bit block of encrypted text, the third bit block of encrypted text, the basis matrix, three preset bit lengths, the preset first key, and the preset first initialization vector are associated and saved to obtain the third encrypted data set. The corresponding relationship in this process is as follows: ; in, This represents the third set of encrypted data.

[0030] It should be further explained that step 3 uses a method that maps different authorization levels to different starting positions in the key derivation chain, so that different authorization levels correspond to different numbers of recoverable bit blocks, enabling the hierarchical use of the same ciphertext data. The ciphertext data set is used to represent the associated data required to complete the recovery of the corresponding authorization level. The encryption end distributes the corresponding ciphertext data and decryption start parameters according to users with different authorization levels.

[0031] Step 4: Distribute three different sets of encrypted data to the receiving end according to three preset authorization levels. Decrypt the three sets of encrypted data and reassemble them in order to obtain the recovery bit string of the corresponding authorization level. Calculate the recovery value contribution of the recovery bit string of the corresponding authorization level. Then combine the recovery value contribution with the basis matrix to obtain the recovery audio sample of the corresponding authorization level. Use the recovery audio sample of the corresponding authorization level to determine whether bit-level lossless recovery has been achieved, so as to obtain the final lossless privacy protection result.

[0032] In step 4, three different sets of encrypted data are distributed to the receiving end according to three preset authorization levels. Each of the three sets of encrypted data is decrypted and reassembled in sequence to obtain the recovery bit string for the corresponding authorization level. The recovery value contribution of the recovery bit string for the corresponding authorization level is calculated, and then combined with the basis matrix to obtain the recovery audio sample for the corresponding authorization level. Using the recovery audio sample for the corresponding authorization level, it is determined whether bit-level lossless recovery has been achieved to obtain the final lossless privacy protection result. Specifically, this includes the following steps: The system distributes a first set of encrypted data to the receiving end according to a preset first authorization level, and decrypts it to obtain a first authorization level recovery audio sample; it distributes a second set of encrypted data to the receiving end according to a preset second authorization level, and decrypts it to obtain a second authorization level recovery audio sample; it distributes a third set of encrypted data to the receiving end according to a preset third authorization level, and decrypts it to obtain a third authorization level recovery audio sample. The decryption process for the first authorized level is as follows: Using the third key and the third initialization vector, the first bit block ciphertext is decrypted using AES-CFB to obtain the first authorized level recovery bit block; The plaintext bit field in the first authorized level recovery bit block, the ciphertext bit field in the second bit block ciphertext, and the ciphertext bit field in the third bit block ciphertext are reassembled according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the first authorized level. Extract the exponent field and tail number segment from the recovered bit string of the first authorization level, and calculate the recovery numerical contribution of the first authorization level; combine the recovery numerical contribution of the first authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the first authorization level; The pre-defined decryption process for the second authorization level is as follows: The second bit block ciphertext is decrypted using AES-CFB using the second key and the second initialization vector to obtain the second authorization level recovery bit block; Extract the first sub-block of the second authorized level recovery bit block and the first sub-block of the second bit block ciphertext respectively to obtain the first plaintext sub-block of the second recovery bit block and the first ciphertext sub-block of the second bit block ciphertext; concatenate the first plaintext sub-block of the second recovery bit block and the first ciphertext sub-block of the second bit block ciphertext, and perform SHA-256 hash operation to obtain the third recovery key; use the second key to perform AES encryption operation, concatenation operation, SHA-256 hash operation and truncate the second initial vector to obtain the third initial vector; Using the third recovery key and the third initialization vector, the first bit block ciphertext is decrypted using AES-CFB to obtain the first authorized level recovery bit block; The plaintext bit field in the first authorized level recovery bit block, the plaintext bit field in the second authorized level recovery bit block, and the ciphertext bit field in the third bit block are reassembled according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the second authorized level. Extract the exponent field and tail number segment from the recovered bit string of the second authorization level, and calculate the recovery numerical contribution of the second authorization level; combine the recovery numerical contribution of the second authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the second authorization level; The decryption process for the preset third-level authorization is as follows: According to the decryption method of the preset second authorization level decryption process, the third bit block ciphertext is decrypted by AES-CFB using the preset first key and the preset first initialization vector to obtain the third authorization level recovery bit block. Then, the second key and the second initialization vector are recovered using the preset first key, the preset first initialization vector and the third authorization level recovery bit block. The second bit block ciphertext is then decrypted by AES-CFB using the second key and the second initialization vector to obtain the second authorization level recovery bit block. Then, the third key and the third initialization vector are recovered using the second key, the second initialization vector and the second authorization level recovery bit block. Finally, the first bit block ciphertext is decrypted by AES-CFB using the third key and the third initialization vector to obtain the first authorization level recovery bit block. The plaintext bit fields in the third authorization level recovery bit block, the plaintext bit fields in the second authorization level recovery bit block, and the plaintext bit fields in the first authorization level recovery bit block are recombined according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the third authorization level. Extract the exponent field and tail number segment from the recovered bit string of the third authorization level, and calculate the recovery numerical contribution of the third authorization level; combine the recovery numerical contribution of the third authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the third authorization level; The original audio sample and the third authorized level recovered audio sample are converted into the original bit string matrix and the lossless recovered bit string matrix, respectively. The corresponding bits of the original bit string matrix and the lossless recovered bit string matrix are compared bit by bit. The number of different corresponding bits is counted to obtain the number of difference bits. It is also determined whether the dimensions of the original bit string matrix and the lossless recovered bit string matrix are the same when the dimensions are the same. When the dimensions of the original bit string matrix and the lossless recovered bit string matrix are the same and the number of difference bits is 0, it is determined that bit-level lossless recovery has been achieved, so as to obtain the final lossless privacy protection result.

[0033] Specifically, the plaintext bit fields in the third authorization level recovery bit block, the second authorization level recovery bit block, and the first authorization level recovery bit block are reassembled according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the third authorization level. The corresponding relationship in this process is as follows: ; in, Indicates the first The recovered bit string corresponding to the third license level for each audio sample Indicates the first The first plaintext bit segment in the third license level recovery bit block corresponding to each audio sample. Indicates the first The second plaintext bit segment in the third license level recovery bit block corresponding to each audio sample. Indicates the first The third plaintext bit segment in the third license level recovery bit block corresponding to each audio sample; Extract the exponent field and tail number segment from the recovered bit string of the third authorization level, and calculate the recovery numerical contribution of the third authorization level; combine the recovery numerical contribution of the third authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the third authorization level. The relationship in the corresponding process is as follows: ; in, Indicates the first The third authorization level recovers the audio sample corresponding to each audio sample. Indicates the first The recovered sample sign coefficients of each audio sample. Indicates the third level of authorization. The numerical contribution of the recovered audio sample; By comparing corresponding bits of the original bit string matrix with the lossless recovered bit string matrix bit by bit, and counting the number of different corresponding bits, the number of differing bits is obtained. The relationship in the correspondence process is as follows: ; in, Indicates the first The audio sample and the recovered audio sample obtained from the third-level authorization level The audio sample is the first one in the single-precision floating-point mantissa segment. Difference marker bit on the bit Indicates the first The single-precision floating-point bit tail digital segment in the original bit string matrix of the audio sample Bit, Indicates the third level of authorization restored to the first The lossless recovery bit string tail segment in the lossless recovery bit string matrix of the audio sample Bit, Indicates the number of difference bits.

[0034] It should be noted that step 4 uses the method of restoring the bit blocks in situ and combining the restored numerical contribution with the basis matrix to transform the bit information restored by different license levels into audio samples of different precision; under the third license level, all bit blocks are restored and bit-by-bit verification is performed, which achieves the effect of coordinating hierarchical restoration and bit-level lossless verification.

[0035] It should be further explained that, to verify the bit-level lossless restoration under the third authorization level, the original audio samples and the restored audio samples under the third authorization level were converted into the original bit string matrix and the lossless restored bit string matrix, respectively, in the same sample order. With the two bit string matrices having the same dimension, an XOR comparison was performed on the difference in the number of bits; 0 for the same bits and 1 for different bits. When the original bit string matrix and the lossless recovery bit string matrix have the same dimension, it indicates that each bit of the lossless recovery bit string matrix is ​​the same as the corresponding bit in the original bit string matrix. It is determined that the recovered audio sample of the third authorization level is consistent with the original audio sample in the IEEE 754 standard 32-bit single-precision floating-point representation sample by sample and bit by bit, thus achieving bit-level lossless recovery. When the matrix dimensions are different or the sample-by-sample and bit-by-bit are not completely consistent, it is determined that bit-level lossless recovery has not been achieved.

[0036] It should be further explained that: in order to illustrate the graded restoration effect and the bit-level lossless restoration effect under the third authorized level, the parameter combination of the first preset bit depth of 22, the second preset bit depth of 4, and the third preset bit depth of 6 was selected for testing. The test audio sampling rate was 16000 Hz and the number of samples was 80000. Please see Figure 3 The recovery effects of different authorization levels are as follows: Figure 3 As shown, in Figure 3 In Figure 3 (a) in the text represents the first authorized level of audio restoration. Figure 3 (b) in the text refers to the second authorization level for audio restoration. Figure 3 (c) in the diagram represents the result of audio recovery at the third authorization level. Figure 3 In the table, (d) represents the difference between the original audio and the audio recovered at the third license level; the recovery quality and bit-level consistency data for different license levels are shown in Table 2.

[0037] Table 2 shows the recovery quality and bit-level consistency data for different authorization levels.

[0038] As shown in Table 2, as the authorization level increases, the number of bits recovered increases from 22 bits to 32 bits, while the mean square error and bit difference rate gradually decrease. At the third authorization level, the mean squared error and the number of difference bits are both 0, and the accuracy is 100%, indicating that the lossless recovery bit string matrix is ​​consistent with the original bit string matrix sample by sample and bit by bit, achieving bit-level lossless recovery. The signal-to-noise ratio of the third authorization level in the table is denoted as ∞, indicating that the recovery error is 0. Please see Figure 4 The present invention also proposes a simulated audio multi-resolution lossless reversible privacy protection system, the system comprising: The audio sample processing module is used for: The original digital audio is read and encoded sequentially to obtain single-precision floating-point bit strings. All single-precision floating-point bit strings are arranged in order and their numerical contributions are calculated. The elements of the basis matrix are calculated using the numerical contributions and then arranged in order to obtain the basis matrix. The single-precision floating-point bit strings are divided into the third bit block, the second bit block, and the first bit block using three preset bit lengths. The encrypted audio processing module is used for: The third bit block is encrypted using the preset first key and first initialization vector to obtain the ciphertext of the third bit block. The second key and second initialization vector are then derived. The second bit block and the first bit block are then encrypted in sequence to obtain the ciphertext of the second bit block and the ciphertext of the first bit block. The third key and the third initialization vector are then derived. The ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block are then recombined in their original order and the ciphertext numerical contribution is calculated. Finally, the ciphertext is combined with the basis matrix to obtain the ciphertext audio sample. The key processing module is used for: The encrypted audio sample, three bit-block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors are associated and saved separately to obtain three different sets of ciphertext data. The audio recovery module is used for: Three different sets of encrypted data are distributed to the receiving end according to three different preset authorization levels. The three sets of encrypted data are decrypted and reassembled in order to obtain the recovery bit string of the corresponding authorization level. The recovery value contribution of the recovery bit string of the corresponding authorization level is calculated. The recovery value contribution is then combined with the basis matrix to obtain the recovery audio sample of the corresponding authorization level. Using the recovery audio sample of the corresponding authorization level, it is determined whether bit-level lossless recovery has been achieved to obtain the final lossless privacy protection result.

[0039] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0040] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0041] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for lossless and reversible privacy protection of analog audio at multiple resolutions, characterized in that, The method includes the following steps: Step 1: Read and encode the original digital audio sequentially to obtain single-precision floating-point bit strings. Arrange all single-precision floating-point bit strings in order and calculate their numerical contributions. Calculate the elements of the basis matrix using the numerical contributions, and then arrange them in order to obtain the basis matrix. Divide the single-precision floating-point bit strings into the third, second, and first bit blocks using three preset bit lengths. Step 2: Encrypt the third bit block using the preset first key and first initialization vector to obtain the ciphertext of the third bit block, and derive the second key and second initialization vector. Then encrypt the second bit block and the first bit block in sequence to obtain the ciphertext of the second bit block and the ciphertext of the first bit block, and derive the third key and the third initialization vector. Reassemble the ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block in their original order and calculate the ciphertext value contribution. Then combine it with the basis matrix to obtain the ciphertext audio sample. Step 3: Associate and save the ciphertext audio sample, three bit block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors to obtain three different sets of ciphertext data. Step 4: Distribute three different sets of encrypted data to the receiving end according to three different preset authorization levels. Decrypt the three different sets of encrypted data and reassemble them in order to obtain the recovery bit string of the corresponding authorization level. Calculate the recovery value contribution of the recovery bit string of the corresponding authorization level. Then combine the recovery value contribution with the basis matrix to obtain the recovery audio sample of the corresponding authorization level. By using the recovered audio samples of the corresponding authorization level, it is determined whether bit-level lossless recovery has been achieved, so as to obtain the final lossless privacy protection result.

2. The analog audio multi-resolution lossless reversible privacy protection method according to claim 1, characterized in that, In step 1, the original digital audio is read and encoded sequentially to obtain single-precision floating-point bit strings. All single-precision floating-point bit strings are arranged in order and their numerical contributions are calculated. The elements of the basis matrix are then calculated using these numerical contributions and arranged in order to obtain the basis matrix. The single-precision floating-point bit strings are then divided into a third bit block, a second bit block, and a first bit block using three preset bit lengths. Specifically, this includes the following steps: Read the raw digital audio and construct a column vector of audio sample values ​​according to the sampling order; Each sample in the audio sample value column vector is encoded using single-precision floating-point encoding to obtain a single-precision floating-point bit string; Arrange all the single-precision floating-point bit strings in order to obtain the original bit string matrix; Convert the exponent field in the original bit string matrix into an unsigned decimal value to determine the exponent weight; The numerical contribution under the corresponding exponential weight is calculated using the last digit segment of the single-precision floating-point bit string. The sign coefficients of the samples are determined based on the original audio sample values ​​in the column vector of audio sample values. Subtract the absolute value of the original audio sample value from the numerical contribution, and then multiply it with the sample sign coefficient to obtain the basis matrix elements; Arrange all the elements of the basis matrix in order to obtain the basis matrix; Each single-precision floating-point bit string is divided into the first segment, the second segment, and the third segment in sequence from the high bit to the middle bit and then to the low bit using three preset bit numbers, respectively. The first segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the first bit block; the second segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the second bit block; the third segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the third bit block.

3. The analog audio multi-resolution lossless reversible privacy protection method according to claim 2, characterized in that, The original digital audio is read, and an audio sample value column vector is constructed according to the sampling order. The corresponding relationship in this process is as follows: ； in, Represents a column vector of audio sample values. This represents the value of the first original audio sample. This represents the value of the second original audio sample. This represents the last original audio sample value. This represents the total number of original audio sample values. Indicates transpose; Arrange all the single-precision floating-point bit strings in order to obtain the original bit string matrix. The corresponding relationship in this process is as follows: ； in, Represents the original bit string matrix. This represents the first single-precision floating-point bit string. This represents the second single-precision floating-point bit string. This represents the last single-precision floating-point bit string. This represents the total number of single-precision floating-point bits. The exponent field in the original bit string matrix is ​​converted into an unsigned decimal value to determine the exponent weight. The corresponding relationship in this process is as follows: ； in, Indicates the first The unsigned decimal value of each audio sample. Indicates the positional order within the index field. Indicates the first The index field of the audio samples, arranged from highest to lowest, is the [number]th [sample]. Bit, Indicates the first The exponential weights of each audio sample; The numerical contribution under the corresponding exponential weight is calculated using the last digit segment of the single-precision floating-point bit string. The corresponding relationship in this process is as follows: ； in, Indicates the first The numerical contribution of each audio sample. This represents the positional index within the last digit field. Indicates the first The last digit segment of the audio sample The bit value at each bit; Based on the original audio sample values ​​in the audio sample value column vector, the sample sign coefficients are determined. The corresponding relationship in this process is as follows: ； in, Indicates the first The sample sign coefficients of each audio sample. Indicates the first The original audio sample values ​​of each audio sample; The absolute value of the original audio sample value is subtracted from the numerical contribution, and then multiplied by the sample sign coefficient to obtain the basis matrix elements. The corresponding relationship in this process is as follows: ； in, Indicates the first The basis matrix elements of each audio sample, Indicates taking the absolute value; The first segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the first bit block; the second segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the second bit block; the third segment of each single-precision floating-point bit string is concatenated sequentially according to the original digital audio sequence to obtain the third bit block. The corresponding relationship in this process is as follows: ； in, Indicates the first Units block This represents the first single-precision floating-point bit string. Units segment This represents the second single-precision floating-point bit string. Units segment This indicates the last single-precision floating-point bit string. Units segment.

4. The analog audio multi-resolution lossless reversible privacy protection method according to claim 3, characterized in that, In step 2, the third bit block is encrypted using a preset first key and first initialization vector to obtain the ciphertext of the third bit block. A second key and second initialization vector are then derived. The second bit block and the first bit block are then encrypted sequentially to obtain the ciphertext of the second bit block and the ciphertext of the first bit block. A third key and a third initialization vector are then derived. The ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block are then reassembled in their original order, and the ciphertext numerical contribution is calculated. Finally, this is combined with the basis matrix to obtain the ciphertext audio sample. Specifically, the steps include the following: The third bit block is encrypted using AES-CFB block encryption with the preset first key and the preset first initialization vector to obtain the ciphertext of the third bit block. Extract the first sub-block of the third bit block and the first sub-block of the ciphertext of the third bit block respectively to obtain the first plaintext sub-block and the first ciphertext sub-block of the third bit block; The first plaintext sub-block of the third bit block and the first ciphertext sub-block of the third bit block are concatenated to obtain the first concatenation result. The first concatenation result is then subjected to SHA-256 hash operation to obtain the second key. The first initial vector is encrypted and concatenated sequentially using the first preset key to obtain the second concatenation result. The second concatenation result is then hashed using SHA-256 to obtain the first hash result. A preset 128 bits are extracted from the first hash result to obtain the second initial vector. The second bit block is encrypted using AES-CFB block encryption with the second key and the second initialization vector to obtain the ciphertext of the second bit block. Extract the first sub-block of the second bit block and the first sub-block of the ciphertext of the second bit block respectively to obtain the first plaintext sub-block of the second bit block and the first ciphertext sub-block of the second bit block; The first plaintext sub-block of the second bit block and the first ciphertext sub-block of the second bit block are concatenated to obtain the third concatenation result. The third concatenation result is then subjected to SHA-256 hash operation to obtain the third key. The second initial vector is encrypted and concatenated using the second key to obtain the fourth concatenation result. The fourth concatenation result is then hashed using SHA-256 to obtain the second hash result. A preset 128 bits are extracted from the second hash result to obtain the third initial vector. The first bit block is encrypted using AES-CFB block encryption with the third key and the third initialization vector to obtain the ciphertext of the first bit block. The ciphertext bit segments in the first bit block ciphertext, the ciphertext bit segments in the second bit block ciphertext, and the ciphertext bit segments in the third bit block ciphertext are reassembled in order to obtain the ciphertext reassembled bit string corresponding to each audio sample; Extract the exponent field and tail number field from the ciphertext reassembled bit string corresponding to each audio sample, and calculate the ciphertext numerical contribution of each audio sample. The ciphertext audio sample is obtained by combining the ciphertext numerical contribution of each audio sample with the corresponding element in the basis matrix.

5. The analog audio multi-resolution lossless reversible privacy protection method according to claim 4, characterized in that, The third bit block is encrypted using AES-CFB block encryption with a preset first key and a preset first initialization vector to obtain the ciphertext of the third bit block. The corresponding relationship in the process is as follows: ； in, This indicates the ciphertext of the third bit block. This indicates AES-CFB block encryption operation. This indicates the third bit block. This represents the first preset key. This represents the first preset initial vector; The first plaintext sub-block and the first ciphertext sub-block of the third bit block are concatenated to obtain the first concatenation result. The first concatenation result is then subjected to a SHA-256 hash operation to obtain the second key. The corresponding relationship in this process is as follows: ； in, This represents the second key. This indicates that a conversion to a fixed-length 256-bit hash value will be performed. This indicates the first plaintext sub-block of the third bit block. This indicates the first ciphertext subblock of the third bit block; Using the preset first key, the preset first initial vector is sequentially encrypted with AES and then concatenated to obtain the second concatenated result. The second concatenated result is then subjected to SHA-256 hashing to obtain the first hash result. A preset 128 bits are extracted from the first hash result to obtain the second initial vector. The corresponding relationship in this process is as follows: ； in, This represents the second initial vector. This indicates a 128-bit truncation operation. This represents the second concatenation result obtained by performing AES encryption on a preset first initialization vector using a preset first key; The ciphertext segments in the first, second, and third bit blocks are reassembled sequentially to obtain the reassembled ciphertext bit string corresponding to each audio sample. The relationship in this process is as follows: ； in, Indicates the first The ciphertext reconstructed bit string corresponding to each audio sample Indicates the first The ciphertext bit field in the first bit block ciphertext corresponding to each audio sample. Indicates the first The ciphertext bit segment in the second bit block ciphertext corresponding to each audio sample. Indicates the first The ciphertext bit segment in the third bit block ciphertext corresponding to each audio sample; Extract the exponent field and tail number segment from the ciphertext reassembled bit string corresponding to each audio sample, and calculate the ciphertext numerical contribution of each audio sample; combine the ciphertext numerical contribution of each audio sample with the corresponding element in the basis matrix to obtain the ciphertext audio sample. The relationship in the correspondence process is as follows: ； in, Indicates the first One audio sample ciphertext audio sample, Indicates from the first The ciphertext numerical contribution of each audio sample.

6. The analog audio multi-resolution lossless reversible privacy protection method according to claim 5, characterized in that, In step 3, the encrypted audio sample, three bit-block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors are associated and saved to obtain three different sets of ciphertext data. Specifically, this includes the following steps: The encrypted audio sample, the first bit block ciphertext, the second bit block ciphertext, the third bit block ciphertext, the basis matrix, three preset bit lengths, the third key, and the third initial vector are associated and saved to obtain the first ciphertext data set; The encrypted audio sample, the first bit block ciphertext, the second bit block ciphertext, the third bit block ciphertext, the basis matrix, three preset bit lengths, the second key, and the second initial vector are associated and saved to obtain the second ciphertext data set; The encrypted audio sample, the first bit block encrypted text, the second bit block encrypted text, the third bit block encrypted text, the basis matrix, three preset bit lengths, the preset first key, and the preset first initial vector are associated and saved to obtain the third encrypted data set.

7. The analog audio multi-resolution lossless reversible privacy protection method according to claim 6, characterized in that, The encrypted audio sample, the first bit block of encrypted text, the second bit block of encrypted text, the third bit block of encrypted text, the basis matrix, three preset bit lengths, the third key, and the third initialization vector are associated and saved to obtain the first encrypted data set. The corresponding relationship in this process is as follows: ； in, This represents the first set of encrypted data. This indicates the ciphertext of the first bit block. This indicates the ciphertext of the second bit block. Represents the basis matrix, This represents the first preset bit position out of three preset bit positions. This represents the second preset number of bits out of three preset number of bits. This indicates the third preset bit position out of three preset bit positions. The third key, The third initial vector.

8. The analog audio multi-resolution lossless reversible privacy protection method according to claim 7, characterized in that, In step 4, three different sets of encrypted data are distributed to the receiving end according to three preset authorization levels. Each of the three sets of encrypted data is decrypted and reassembled in sequence to obtain the recovery bit string for the corresponding authorization level. The recovery value contribution of the recovery bit string for the corresponding authorization level is calculated, and then combined with the basis matrix to obtain the recovery audio sample for the corresponding authorization level. Using the recovery audio sample for the corresponding authorization level, it is determined whether bit-level lossless recovery has been achieved to obtain the final lossless privacy protection result. Specifically, this includes the following steps: The system distributes a first set of encrypted data to the receiving end according to a preset first authorization level, and decrypts it to obtain a first authorization level recovered audio sample; it distributes a second set of encrypted data to the receiving end according to a preset second authorization level, and decrypts it to obtain a second authorization level recovered audio sample; it distributes a third set of encrypted data to the receiving end according to a preset third authorization level, and decrypts it to obtain a third authorization level recovered audio sample. The decryption process for the first authorized level is as follows: Using the third key and the third initialization vector, the first bit block ciphertext is decrypted using AES-CFB to obtain the first authorized level recovery bit block; The plaintext bit field in the first authorized level recovery bit block, the ciphertext bit field in the second bit block ciphertext, and the ciphertext bit field in the third bit block ciphertext are reassembled according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the first authorized level. Extract the exponent field and tail number segment from the recovered bit string of the first authorization level, and calculate the recovery numerical contribution of the first authorization level; combine the recovery numerical contribution of the first authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the first authorization level; The pre-defined decryption process for the second authorization level is as follows: The second bit block ciphertext is decrypted using AES-CFB using the second key and the second initialization vector to obtain the second authorized level recovery bit block; Extract the first sub-block of the second authorized level recovery bit block and the first sub-block of the second bit block ciphertext respectively to obtain the first plaintext sub-block of the second recovery bit block and the first ciphertext sub-block of the second bit block ciphertext; concatenate the first plaintext sub-block of the second recovery bit block and the first ciphertext sub-block of the second bit block ciphertext, and perform SHA-256 hash operation to obtain the third recovery key; use the second key to perform AES encryption operation, concatenation operation, SHA-256 hash operation and truncate the second initial vector to obtain the third initial vector; Using the third recovery key and the third initialization vector, the first bit block ciphertext is decrypted using AES-CFB to obtain the first authorized level recovery bit block; The plaintext bit field in the first authorized level recovery bit block, the plaintext bit field in the second authorized level recovery bit block, and the ciphertext bit field in the third bit block are reassembled according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the second authorized level. Extract the exponent field and tail number segment from the recovered bit string of the second authorization level, and calculate the recovery numerical contribution of the second authorization level; combine the recovery numerical contribution of the second authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the second authorization level; The decryption process for the preset third-level authorization is as follows: Using the preset first key and preset first initialization vector, the third bit block ciphertext is decrypted using AES-CFB to obtain the third authorized level recovery bit block. Then, the preset first key, preset first initialization vector, and third authorized level recovery bit block are used to recover the second key and second initialization vector. The second bit block ciphertext is then decrypted using AES-CFB using the second key and second initialization vector to obtain the second authorized level recovery bit block. The second key, second initialization vector, and second authorized level recovery bit block are then used to recover the third key and third initialization vector. Finally, the first bit block ciphertext is decrypted using AES-CFB using the third key and third initialization vector to obtain the first authorized level recovery bit block. The plaintext bit fields in the third authorization level recovery bit block, the plaintext bit fields in the second authorization level recovery bit block, and the plaintext bit fields in the first authorization level recovery bit block are recombined according to the original arrangement order of the single-precision floating-point bit string from high bit to middle bit to low bit to obtain the recovery bit string of the third authorization level. Extract the exponent field and tail number segment from the recovered bit string of the third authorization level, and calculate the recovery numerical contribution of the third authorization level; combine the recovery numerical contribution of the third authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the third authorization level; The original audio sample and the third authorized level recovered audio sample are converted into the original bit string matrix and the lossless recovered bit string matrix, respectively. The corresponding bits of the original bit string matrix and the lossless recovered bit string matrix are compared bit by bit. The number of different corresponding bits is counted to obtain the number of difference bits. It is also determined whether the dimensions of the original bit string matrix and the lossless recovered bit string matrix are the same when the dimensions are the same. When the dimensions of the original bit string matrix and the lossless recovered bit string matrix are the same and the number of difference bits is 0, it is determined that bit-level lossless recovery has been achieved, so as to obtain the final lossless privacy protection result.

9. The analog audio multi-resolution lossless reversible privacy protection method according to claim 8, characterized in that: The plaintext bit fields in the recovery bit blocks of the third authorization level, the second authorization level, and the first authorization level are reassembled according to the original arrangement of the single-precision floating-point bit strings from the high-order bit to the middle-order bit and then to the low-order bit, to obtain the recovery bit string of the third authorization level. The corresponding relationship in this process is as follows: ； in, Indicates the first The recovered bit string corresponding to the third license level for each audio sample Indicates the first The first plaintext bit segment in the third license level recovery bit block corresponding to each audio sample. Indicates the first The second plaintext bit segment in the third license level recovery bit block corresponding to each audio sample. Indicates the first The third plaintext bit segment in the third license level recovery bit block corresponding to each audio sample; Extract the exponent field and tail number segment from the recovered bit string of the third authorization level, and calculate the recovery numerical contribution of the third authorization level; combine the recovery numerical contribution of the third authorization level with the corresponding element in the basis matrix to obtain the recovered audio sample of the third authorization level. The relationship in the corresponding process is as follows: ； in, Indicates the first The third authorization level recovers the audio sample corresponding to each audio sample. Indicates the first The recovered sample sign coefficients of each audio sample. Indicates the third level of authorization. The numerical contribution of the recovered audio sample; By comparing corresponding bits of the original bit string matrix with the lossless recovered bit string matrix bit by bit, and counting the number of different corresponding bits, the number of differing bits is obtained. The relationship in the correspondence process is as follows: ； in, Indicates the first The audio sample and the recovered audio sample obtained from the third-level authorization level The audio sample is the first one in the single-precision floating-point mantissa segment. Difference marker bit on the bit Indicates the first The single-precision floating-point bit tail digital segment in the original bit string matrix of the audio sample Bit, Indicates the third level of authorization restored to the first The lossless recovery bit string tail segment in the lossless recovery bit string matrix of the audio sample Bit, Indicates the number of difference bits.

10. A simulated audio multi-resolution lossless reversible privacy protection system, characterized in that, The system employs the analog audio multi-resolution lossless reversible privacy protection method as described in any one of claims 1-9, and the system comprises: The audio sample processing module is used for: The original digital audio is read and encoded sequentially to obtain single-precision floating-point bit strings. All single-precision floating-point bit strings are arranged in order and their numerical contributions are calculated. The elements of the basis matrix are calculated using the numerical contributions and then arranged in order to obtain the basis matrix. The single-precision floating-point bit strings are divided into the third bit block, the second bit block, and the first bit block using three preset bit lengths. The encrypted audio processing module is used for: The third bit block is encrypted using the preset first key and first initialization vector to obtain the ciphertext of the third bit block. The second key and second initialization vector are then derived. The second bit block and the first bit block are then encrypted in sequence to obtain the ciphertext of the second bit block and the ciphertext of the first bit block. The third key and the third initialization vector are then derived. The ciphertext of the third bit block, the ciphertext of the second bit block, and the ciphertext of the first bit block are then recombined in their original order and the ciphertext numerical contribution is calculated. Finally, the ciphertext is combined with the basis matrix to obtain the ciphertext audio sample. The key processing module is used for: The encrypted audio sample, three bit-block ciphertexts, the basis matrix, three preset bit lengths, three keys, and three initial vectors are associated and saved separately to obtain three different sets of ciphertext data. The audio recovery module is used for: Three different sets of encrypted data are distributed to the receiving end according to three different preset authorization levels. The three sets of encrypted data are decrypted and reassembled in order to obtain the recovery bit string of the corresponding authorization level. The recovery value contribution of the recovery bit string of the corresponding authorization level is calculated. The recovery value contribution is then combined with the basis matrix to obtain the recovery audio sample of the corresponding authorization level. Using the recovery audio sample of the corresponding authorization level, it is determined whether bit-level lossless recovery has been achieved to obtain the final lossless privacy protection result.

Images