CN106024000B - End-to-end voice encryption and decryption method based on frequency spectrum mapping - Google Patents
End-to-end voice encryption and decryption method based on frequency spectrum mapping Download PDFInfo
- Publication number
- CN106024000B CN106024000B CN201610343431.7A CN201610343431A CN106024000B CN 106024000 B CN106024000 B CN 106024000B CN 201610343431 A CN201610343431 A CN 201610343431A CN 106024000 B CN106024000 B CN 106024000B
- Authority
- CN
- China
- Prior art keywords
- coefficient
- lpc
- mapping
- lsf
- speech
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013507 mapping Methods 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000001228 spectrum Methods 0.000 title claims abstract description 19
- 230000015572 biosynthetic process Effects 0.000 claims abstract description 20
- 238000003786 synthesis reaction Methods 0.000 claims abstract description 20
- 230000009466 transformation Effects 0.000 claims abstract description 14
- 230000003595 spectral effect Effects 0.000 claims abstract description 13
- 230000008569 process Effects 0.000 claims description 13
- 238000009795 derivation Methods 0.000 claims description 7
- 238000013519 translation Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims 1
- 230000006854 communication Effects 0.000 description 8
- 238000006243 chemical reaction Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 238000009432 framing Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005311 autocorrelation function Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/08—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
- G10L19/12—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
- G10L19/13—Residual excited linear prediction [RELP]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
- H04L9/06—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols the encryption apparatus using shift registers or memories for block-wise or stream coding, e.g. DES systems or RC4; Hash functions; Pseudorandom sequence generators
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The invention discloses an end-to-end voice encryption method based on frequency spectrum mapping, which comprises the following steps: carrying out Linear Prediction (LPC) analysis on the digital voice signal to obtain an LPC coefficient; the LPC coefficient is converted into a Linear Spectral Frequency (LSF) coefficient, the LSF coefficient is mapped and transformed according to a given key, the mapped LSF coefficient is converted into the LPC coefficient, and a synthesis filter is constructed; and finally, the prediction residual signal passes through a synthesis filter constructed by the LPC coefficient after mapping transformation to obtain an encrypted voice signal. The voice characteristics of the encrypted voice signal are ensured, and effective voice encryption is realized.
Description
Technical Field
The invention relates to a voice encryption and decryption method, in particular to a voice encryption and decryption method based on linear predictive coding and frequency spectrum mapping.
Background
Voice is an important means for people to obtain information, and voice communication is one of the most effective and convenient means in modern communication. With the development of communication technology, various voice communications are occurring in people's lives. But voice communication in real life is inevitably subjected to security threats such as eavesdropping, telephone tracking, telephone hijacking, denial of service and the like. Therefore, voice encryption is important to secure voice. The importance of voice encryption is manifested in two aspects: on one hand, as the consciousness of protecting the privacy of people gradually rises, the attention on voice encryption becomes higher and higher; on the other hand, in special applications such as military communication, business negotiation, political negotiation, etc., once information is revealed, it will cause huge loss, where voice encryption is necessary.
However, the traditional mobile communication lacks an end-to-end encryption mechanism, and the plaintext information after analog-to-digital conversion is transmitted between mobile network node devices, so that the risk of eavesdropping is extremely high. The existing mobile communication process generally goes through several processes of analog-to-digital conversion, coding transmission, base station decoding and recoding technology of mobile phone end voice signals, mobile phone end decoding and digital-to-analog conversion into voice signals, and the like. The existing encryption means are mostly used for encrypting after the encoding process, and then encrypting after decryption and decoding of the base station, and the encryption mode is established on the premise that a core network part is safe and credible, because only the safety of a wireless channel part is considered, plaintext information of the wireless channel part is visible for the base station, and the system cannot provide end-to-end safe communication for users. The encryption mode after coding has the advantages of less encrypted data, small occupied channel and the like which can not be repudiated, but the encryption mode has the risk of being intercepted due to the characteristic of transparency to the base station.
Chinese patent document CN201210055857 discloses a speech encryption system, which proposes a speech encryption algorithm, first performing time domain to frequency domain conversion on speech signals according to a set length segment, then scrambling speech signal frequency packets in the frequency domain, and finally converting the frequency domain into the time domain to form encrypted speech signals. Although the risk of eavesdropping can be reduced to a certain extent, theoretical analysis and experiments prove that the encryption voice signal synthesized by the method has high synchronization requirements on the encryption and decryption ends, and the method is poor in practicability and cannot be widely applied.
Disclosure of Invention
In order to solve the problems in the prior art, the invention aims to: on the basis of linear predictive coding and spectral mapping, carrying out Linear Predictive (LPC) analysis on a digital voice signal at an encryption end to obtain an LPC coefficient, then converting the LPC coefficient into a Linear Spectral Frequency (LSF) coefficient, carrying out mapping transformation on the LSF coefficient according to a given secret key, then converting the mapped LSF coefficient into the LPC coefficient, and constructing a synthesis filter; on the other hand, the original voice is subjected to linear prediction to obtain a prediction residual signal, and finally the prediction residual signal is subjected to synthesis filter constructed by the LPC coefficient after mapping transformation to obtain an encrypted voice signal, so that the voice characteristics of the encrypted voice signal are ensured, and effective voice encryption is realized.
The technical scheme of the invention is as follows:
an end-to-end voice encryption method based on spectrum mapping is characterized by comprising the following steps:
s01: performing linear predictive LPC analysis on the digital voice signal to obtain an LPC coefficient;
s02: the LPC coefficient is converted into a linear spectrum frequency LSF coefficient, the LSF coefficient is mapped and transformed according to a given key, the mapped LSF coefficient is converted into the LPC coefficient, and a synthesis filter is constructed;
s03: and finally, the prediction residual signal passes through a synthesis filter constructed by the LPC coefficient after mapping transformation to obtain an encrypted voice signal.
Preferably, the LPC coefficients in step S01 are obtained by:
linear prediction LPC is a method of predicting present or future sample values using past p sample values s (n)The prediction error ε (n) is:
wherein, aiIs a linear prediction coefficient, and n is a natural number;
Yule-Walker equation for linear prediction of LPC:
the Levinson-Durbin algorithm recursion formula is as follows:
aki＝ak-1,i+akkak-1,k-i,i＝1,2,…,k-1
thereby, the prediction coefficient a of the two-order linear prediction can be solvedi,i＝1,2,…,p。
Preferably, the converting the LPC coefficients into linear spectral frequency LSF coefficients comprises the steps of:
the p-order linear prediction filter function is:
definition, p (z) ═ a (z) + z-(p+1)A(z-1)，Q(z)＝A(z)-z-(p+1)A(z-1)
Then the process of the first step is carried out,
when the order p is an even number:
when the order p is odd, the following are present:
P′(z)＝P(z)
p '(z) and Q' (z) are symmetric even polynomials, the root is a complex conjugate pair, and only the root positioned in the upper semicircle needs to be determined, and the roots arranged in the upper semicircle P '(z) and Q' (z) arei 1,2, …, p, with line spectrum frequency at the root of angular frequency 0<ωi<π；
When the order p is even number
M1＝p/2,M2＝p/2
When the order p is odd
M1＝(p+1)/2,M2＝(p-1)/2
Respectively taking the conjugate zero point logarithms as M by using the Taylor series expansion principle1And M2Corresponding P '(z) and Q' (z) spreads with M1+M2P, with z ejwSubstituting and converting by using the cosine theorem, the following steps are included:
P″(ω)＝2cosM1ω+2P′(1)cos(M1-1)ω+…+2P′(M1-1)cosω+P′(M1)
Q″(ω)＝2cosM1ω+2Q′(1)cos(M1-1)ω+…+2Q′(M1-1)cosω+Q′(M1)
let x be cos omega, and use the above formula with Chebyshev polynomial Tm(x) Cos (mx) is developed with:
chebyshev polynomial T of mth order xm(x) Satisfy recurrence Tm(x)＝2xTk-1(x)-Tk-2(x) Initial value T0(x)＝1，T1(x) X is determined to be [1, -1 ═ x]Within the interval, root { x) of P ″ (x) ═ 0 and Q ″ (x) ═ 0 is searched foriAnd the corresponding parameter value ω of the LSFiFrom omegai＝arccosxiTo be determined.
Preferably, the mapping transformation of step S02 includes linear mapping and nonlinear mapping, where the linear mapping includes translation mapping, rotation mapping, similarity mapping, and inversion mapping, and the nonlinear mapping is implemented by using various nonlinear operators.
Preferably, the LSF coefficients are converted into LPC coefficients in step S02, which includes the following steps:
from the LSF parameter ωiInverse derivation of the Chebyshev polynomial to solve for the LSF parameter value ωkLet xk＝cosωkK is 1,2, …, p is an intermediate formula
According to the corresponding relation of the original derivation process, P '(omega) and P' (z) are reversely deduced from P '(x), Q (omega) and Q (z) are also reversely deduced from Q' (x), and then P (z) and Q (z) are obtained by reverse derivation:
P(z)＝P′(z)*(1+z-1)
Q(z)＝Q′(z)*(1-z-1)
then there are:
A(z)＝(P(z)+Q(z))/2
from A (z) the LPC parameter a is obtainedi。
Preferably, before step S01, the input digital speech signal is windowed and framed, i.e. S (n) is multiplied by a window function w (n), such that the windowed speech signal Sw＝s(n)*w(n)。
The invention also discloses a voice decryption method, which is characterized by comprising the following steps:
s11: carrying out linear prediction LPC analysis on the encrypted digital voice signal to obtain an LPC coefficient;
s12: the LPC coefficient is converted into a linear spectrum frequency LSF coefficient, the LSF coefficient is subjected to inverse mapping transformation according to a key, the LSF coefficient subjected to inverse mapping is converted into the LPC coefficient, and a decryption speech synthesis filter is constructed;
s13: and finally, the prediction residual signal passes through a decryption speech synthesis filter constructed by the LPC coefficient after inverse mapping transformation to obtain an original speech signal.
Compared with the prior art, the invention has the advantages that:
1. the invention is a voice encryption and decryption method based on linear predictive coding and spectrum mapping, which comprises the steps of carrying out Linear Predictive (LPC) analysis on a digital voice signal at an encryption end to obtain an LPC coefficient, then converting the LPC coefficient into a Linear Spectral Frequency (LSF) coefficient, carrying out mapping transformation on the LSF coefficient according to a given secret key, then converting the mapped LSF coefficient into the LPC coefficient, and constructing a synthesis filter; on the other hand, the original voice is subjected to linear prediction to obtain a prediction residual signal, and finally the prediction residual signal is subjected to synthesis filter constructed by the LPC coefficient after mapping transformation to obtain an encrypted voice signal, so that the voice characteristics of the encrypted voice signal are ensured, and effective voice encryption is realized.
2. The decryption process is the reverse of the encryption process. The decryption end performs Linear Prediction (LPC) on the encrypted digital voice signal to obtain an LPC coefficient of the encrypted signal, then converts the LPC coefficient into a Linear Spectrum Frequency (LSF) coefficient, inversely maps the LSF coefficient according to a key, converts the inversely mapped LSF coefficient into the LPC coefficient, and constructs a voice synthesis filter; meanwhile, the encrypted voice obtains a predicted residual signal through linear prediction, and finally, the predicted residual signal passes through a synthesis filter constructed by the LPC coefficient after inverse mapping change to obtain an original voice signal, so that the original signal is simply and efficiently recovered. The method has low synchronization requirements on the encryption and decryption ends, has better practicability and has wide application prospect.
Drawings
The invention is further described with reference to the following figures and examples:
FIG. 1 is a block diagram of a voice encryption apparatus according to the present invention;
FIG. 2 is a block diagram of the voice decryption apparatus according to the present invention;
FIG. 3 is a waveform diagram of an original speech signal;
FIG. 4 is a waveform diagram of an encrypted speech signal;
fig. 5 is a waveform diagram of a decrypted speech signal.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
Example (b):
as shown in fig. 1, the signal to be encrypted is a digital speech signal, and as shown in fig. 3, according to the short-time stationary performance of the speech signal, in order to facilitate the analysis of the characteristic parameters of the speech signal, it is necessary to perform framing on the speech signal, where the framing method adopted here is a windowing framing technique. Meanwhile, the frames are in smooth transition to ensure the continuity of the voice signals, and the signals after framing are s (n).
First, Linear Prediction (LPC) analysis is performed on the speech signal frame s (n), the p-order linear prediction coefficient aiThe autocorrelation function is used to calculate the autocorrelation by the Levinson Durbin algorithm. Given that the linear spectral frequency LSF coefficients have good quantization and interpolation properties and that the linear spectral frequency LSF may correspond well to the position and bandwidth of the spectrally preserved formants, the LSF is often directly controlled and processed in speech processing, and hence is herein directly mapped to the LSF coefficients. The LPC coefficients found are herein converted to line spectral pair LSF parameters using the Chebyshev method. The LSF coefficients are mapped according to a given key by selecting a suitable mapping method, which is a main implementation part of voice encryption. And finally, converting the mapped LSF into an LPC coefficient to form a voice encryption synthesis filter for synthesizing the encrypted voice signal.
On the other hand, the original voice is subjected to linear prediction to obtain a prediction residual signal e (n), and finally, the prediction residual signal e (n) is subjected to voice encryption synthesis filter constructed by mapping the transformed LPC coefficients to obtain an encrypted voice signal. Thus, the encryption of the original voice signal is realized, and the waveform diagram of the encrypted voice is shown in fig. 4.
As shown in fig. 2, the decryption process is the reverse of the encryption process. The decryption end analyzes the encrypted digital voice signal Linear Prediction (LPC) to obtain an LPC coefficient of the encrypted signal, then converts the LPC coefficient into a Linear Spectrum Frequency (LSF) coefficient, and reversely maps the LSF coefficient according to a provided secret key to recover the LSF coefficient of the original voice signal. In order to synthesize an original voice signal, the LSF coefficient needs to be converted into an LPC coefficient, and a voice synthesis filter is constructed; meanwhile, the encrypted speech is subjected to linear prediction to obtain a predicted residual signal, and finally, the predicted residual signal is subjected to a synthesis filter constructed by inverse mapping of the changed LPC coefficients to obtain an original speech signal, so that the original signal is simply and efficiently restored, and a waveform diagram is shown in fig. 5.
The specific implementation method comprises the following steps:
due to the short-term stationarity of the speech signal, in order to avoid the truncation effect, it is usually necessary to window and frame the input speech signal, and it is also necessary to ensure the continuity between frames of the speech signal.
Framing is performed by weighting with a window of finite length, which is movable, i.e. multiplying s (n) by a window function w (n) such that the windowed speech signal sw＝s(n)*w(n).
The common window functions in speech signal digital processing are rectangular window and Hamming window, and the expression is as follows:
rectangular window:
hamming window:
wherein, N is the frame length, and N is a natural number.
(1) Linear predictive analysis
The basic idea of linear prediction is to use the past p sample values s (n) to predict the present or future sample values
The prediction error ε (n) is:
wherein, aiAnd i is 1,2, …, p is a linear prediction coefficient and can be obtained by a Levinson-Durbin algorithm, and the Levinson-Durbin algorithm provides an efficient recursion algorithm by utilizing the symmetry and Toepltz property of an autocorrelation matrix.
Yule-Walker equation for linear prediction
As can be seen from the equations, it has p +1 equations, when k is 0,1,2, …, pWhen known, can be solved topk[k＝1,2,…,p]Andsample p +1 unknowns. The Levinson-Durbin algorithm recursion formula is as follows:
aki＝ak-1,i+akkak-1,k-i,i＝1,2,…,k-1 (7)
thereby, the prediction coefficient a of the two-order linear prediction can be solvediI is 1,2, …, p, where l is an intermediate variable.
(2) Conversion of LPC coefficients to LSF coefficients
The idea of the method is to convert an LPC coefficient into an LSF coefficient, encrypt the LSF coefficient, convert the encrypted LSF coefficient into an LPC coefficient, synthesize an encrypted voice signal, and ensure that the synthesized encrypted voice signal contains voice characteristics, so that the LPC coefficient needs to be converted into the LSF coefficient.
The recursive relation of the transfer function of the ith order linear prediction error filter is as follows:
Ai(z)＝Ai-1(z)-kiz-1Ai-1(z-1) (9)
then there is a p-order linear prediction filter function as:
defining:
P(z)＝A(z)+z-(p+1)A(z-1) (11)Q(z)＝A(z)-z-(p+1)A(z-1)
therefore:
it can be shown that when the zero point of a (Z) is within the Z-plane unit circle, the zero points of p (Z) and q (Z) are both on the unit circle, and that p (Z) and q (Z) have conjugate complex roots and zero points alternating with increasing ω along the unit circle. P (z) must have one root, z-1 (ω ═ pi), q (z) must have one root, z-1.
Let zero of P (z) beZero of Q (z) isSince the zeros of p (z) and q (z) are both on the unit circle, these zeros can be directly expressed in frequency:
0<ωi<θi<…<ωp/2<θp/2<π (13)
ωi，θioccurring in pairs, reflecting spectral characteristics, and are called Line Spectral Pairs (LSPs) since the LSF parameters areThe frequency domain parameter, so it is more closely related to the peak of the speech signal spectral envelope.
When the order p is an even number, there are
(14)
When the order p is odd, there are
P′(z)＝P(z)
(15)
P '(z) and Q' (z) are symmetrical even polynomials, the root is a complex conjugate pair, and the root positioned in the upper semicircle only needs to be determined. The root of the upper semi-circle P '(z) and Q' (z) isi 1,2, …, p, with line spectrum frequency at the root of angular frequency 0<ωi<π。
When the order p is even number
M1＝p/2,M2＝p/2 (16)
When the order p is odd
M1＝(p+1)/2,M2＝(p-1)/2 (17)
Respectively taking the conjugate zero point logarithms as M by using the Taylor series expansion principle1And M2Corresponding P '(z) and Q' (z) spreads with M1+M2P. With z ═ ejwSubstituted and converted by the cosine theorem
P″(ω)＝2cosM1ω+2P′(1)cos(M1-1)ω+…+2P′(M1-1)cosω+P′(M1)
Q″(ω)＝2cosM1ω+2Q′(1)cos(M1-1)ω+…+2Q′(M1-1)cosω+Q′(M1)
(18)
Let x be cos omega, and use the above formula with Chebyshev polynomial Tm(x) Is developed with cos (mx)
Chebyshev polynomial T of mth order xm(x) Satisfy recurrence Tm(x)＝2xTk-1(x)-Tk-2(x) Initial value T0(x)＝1,
T1(x) The LSF parameter value is found by a search method. The above Chebyshev polynomial solution is substantially that x is calculated to be in the range of [1, -1]Within the interval, root { x) of P ″ (x) ═ 0 and Q ″ (x) ═ 0 is searched foriAnd the corresponding LSF value can be represented by ωi＝arccosxiTo be determined.
LSF analysis is a method of representing the spectral characteristics of a speech signal by using P discrete frequencies, and the LSF coefficient deviation only affects the speech spectrum around the frequency, but not the LSF speech spectrum at other frequencies.
(3) LSF mapping
Mapping and transforming the LSF coefficients according to the provided key is the main process of encryption. The data mapping includes linear mapping and non-linear mapping, wherein the linear mapping can be classified into translation mapping, rotation mapping, similarity mapping, inversion mapping, and the like.
This text is mainly based on the comparison of the LSF coefficient ωiThe mapping of (2) implements a mapping process of line spectrum frequencies, wherein the key is a mapping factor. Due to the LSF coefficient omegaiIn the range of 0 to pi, so that the rotating mapping is adopted herein, the mapping transformation of the LSF coefficients is realized according to the provided key.
(4) Conversion of LSF coefficients into LPC coefficients
For conversion of LSF to LPC, the LSF parameter ωiAnd thetaiReverse-pushing Chebyshev duoSolving by a polynomial, LSF parameter value omegakLet xk＝cosωkK is 1,2, …, p has the intermediate formula:
according to the corresponding relation of the original derivation process, P '(ω) and P' (z) can be reversely deduced from P '(x), Q (ω) and Q (z) can be reversely deduced from Q' (x), and P (z) and Q (z) can be reversely deduced according to equation (11), and the following steps are provided:
P(z)＝P′(z)*(1+z-1)
Q(z)＝Q′(z)*(1-z-1) (22)
then there are:
A(z)＝(P(z)+Q(z))/2 (23)
then the LPC parameter a is obtained from A (z)iAnd ki。
It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.
Claims (7)
1. An end-to-end voice encryption method based on spectrum mapping is characterized by comprising the following steps:
s01: performing linear predictive LPC analysis on the digital voice signal to obtain an LPC coefficient;
s02: the LPC coefficient is converted into a linear spectrum frequency LSF coefficient, the LSF coefficient is mapped and transformed according to a given key, the mapped LSF coefficient is converted into the LPC coefficient, and an encrypted voice synthesis filter is constructed;
s03: the original digital voice signal is filtered by LPC, and a prediction residual signal epsilon (n) is obtained as follows:
wherein, aiIs a linear prediction coefficient, n is a natural number, and s (n) is past p sample values; and finally, the prediction residual signal passes through an encrypted voice synthesis filter constructed by the LPC coefficient after mapping transformation to obtain an encrypted voice signal.
2. The speech encryption method according to claim 1, wherein the LPC coefficients in the step S01 are obtained by:
linear prediction LPC is a method of predicting present or future sample values using past p sample values s (n)
Yule-Walker equation for linear prediction of LPC:
the Levinson-Durbin algorithm recursion formula is as follows:
aki＝ak-1，i+akkak-1，k-i，i＝1，2，…，k-1
thereby, the prediction coefficient a of the two-order linear prediction can be solvedi，i＝1，2，...，p。
3. The speech encryption method according to claim 1, wherein said converting LPC coefficients to linear spectral frequency LSF coefficients comprises the steps of:
the p-order linear prediction filter function is:
definition, p (z) ═ a (z) + z-(p+1)A(z-1)，Q(z)＝A(z)-z-(p+1)A(z-1)
Then the process of the first step is carried out,
when the order p is an even number:
when the order p is odd, the following are present:
P′(z)＝P(z)
p '(z) and Q' (z) are symmetric even polynomials, the root is a complex conjugate pair, and only the root positioned in the upper semicircle needs to be determined, and the roots arranged in the upper semicircle P '(z) and Q' (z) arei 1,2, …, p, with line spectrum frequency of 0 < ωi＜π；
When the order p is even number
M1＝p/2，M2＝p/2
When the order p is odd
M1＝(p+1)/2，M2＝(p-1)/2
Respectively taking the conjugate zero point logarithms as M by using the Taylor series expansion principle1And M2Corresponding P '(z) and Q' (z) spreads with M1+M2P, with z ejwSubstituting and converting by using the cosine theorem, the following steps are included:
P″(ω)＝2cosM1ω+2P′(1)cos(M1-1)ω+…+2P′(M1-1)cosω+P′(M1)
Q″(ω)＝2cosM1ω+2Q′(1)cos(M1-1)ω+…+2Q′(M1-1)cosω+Q′(M1)
let x be cos omega, and use the above formula with Chebyshev polynomial Tm(x) Cos (mx) is developed with:
chebyshev polynomial T of mth order xm(x) Satisfy recurrence Tm(x)＝2xTk-1(x)-Tk-2(x) Initial value T0(x)＝1，T1(x) X is determined to be [1, -1 ═ x]Within the interval, root { x) of P ″ (x) ═ 0 and Q ″ (x) ═ 0 is searched foriAnd the corresponding parameter value ω of the LSFiFrom omegai＝arccosxiTo be determined.
4. The speech encryption method according to claim 1, wherein the mapping transformation of step S02 includes linear mapping and nonlinear mapping, wherein the linear mapping includes translation mapping, rotation mapping, similarity mapping, and inversion mapping, and the nonlinear mapping is mapping implemented by using various types of nonlinear operators.
5. The speech encryption method according to claim 3, wherein the LSF coefficients are converted into LPC coefficients in step S02, comprising the steps of:
from the LSF parameter ωiInverse derivation of the Chebyshev polynomial to solve for the LSF parameter value ωkLet xk＝cosωkK is 1,2, …, p is an intermediate formula
According to the corresponding relation of the original derivation process, P '(omega) and P' (z) are reversely deduced from P '(x), Q (omega) and Q (z) are also reversely deduced from Q' (x), and then P (z) and Q (z) are obtained by reverse derivation:
P(z)＝P′(z)*(1+z-1)
Q(z)＝Q′(z)*(1-z-1)
then there are:
A(z)＝(P(z)+Q(z))/2
from A (z) the LPC parameter a is obtainedi。
6. The speech encryption method according to claim 2, wherein the input digital speech signal is windowed and framed, i.e. is multiplied by a windowing function w (n) by S (n), before step S01, such that the windowed speech signal Sw＝s(n)*w(n)。
7. A method for decrypting speech, comprising the steps of:
s11: carrying out linear prediction LPC analysis on the encrypted digital voice signal to obtain an LPC coefficient;
s12: the LPC coefficient is converted into a linear spectrum frequency LSF coefficient, the LSF coefficient is subjected to inverse mapping transformation according to a key, the LSF coefficient subjected to inverse mapping is converted into the LPC coefficient, and a decryption speech synthesis filter is constructed;
s13: the encrypted digital speech signal is subjected to LPC filtering to obtain a prediction residual signal epsilon (n) which is:
wherein, aiIs a linear prediction coefficient, n is a natural number, and s (n) is past p sample values; and finally, passing the prediction residual signal through a decryption speech synthesis filter constructed by the LPC coefficient after inverse mapping transformation to obtain an original speech signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610343431.7A CN106024000B (en) | 2016-05-23 | 2016-05-23 | End-to-end voice encryption and decryption method based on frequency spectrum mapping |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610343431.7A CN106024000B (en) | 2016-05-23 | 2016-05-23 | End-to-end voice encryption and decryption method based on frequency spectrum mapping |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106024000A CN106024000A (en) | 2016-10-12 |
| CN106024000B true CN106024000B (en) | 2019-12-24 |
Family
ID=57096094
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610343431.7A Active CN106024000B (en) | 2016-05-23 | 2016-05-23 | End-to-end voice encryption and decryption method based on frequency spectrum mapping |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106024000B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107607783B (en) * | 2017-09-01 | 2019-09-20 | 广州辰创科技发展有限公司 | Efficient and flexible radar spectrum display method |
| CN116825117B (en) * | 2023-04-06 | 2024-06-21 | 浙江大学 | Microphone with privacy protection function and privacy protection method thereof |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001035395A1 (en) * | 1999-11-10 | 2001-05-17 | Koninklijke Philips Electronics N.V. | Wide band speech synthesis by means of a mapping matrix |
| CN101944362B (en) * | 2010-09-14 | 2012-05-30 | 北京大学 | Integer wavelet transform-based audio lossless compression encoding and decoding method |
| CN105471899A (en) * | 2015-12-30 | 2016-04-06 | 浙江德景电子科技有限公司 | End-to-end voice encryption communication method and device for mobile terminals |
-
2016
- 2016-05-23 CN CN201610343431.7A patent/CN106024000B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106024000A (en) | 2016-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107306124B (en) | Filter, transmitter and receiver and design method thereof | |
| CN106024000B (en) | End-to-end voice encryption and decryption method based on frequency spectrum mapping | |
| CN111526148B (en) | System and method for safely denoising encrypted audio in cloud computing environment | |
| Shi et al. | Generalized convolution theorem associated with fractional Fourier transform | |
| CN105279350B (en) | The design method of approximate Perfect Reconstruction Vertical Nonuniform Cosine modulated filter group | |
| Elshamy et al. | Efficient audio cryptosystem based on chaotic maps and double random phase encoding | |
| Abro et al. | Towards security of GSM voice communication | |
| KR20170143459A (en) | Receiver architecture for linear modulation based communication systems | |
| Krasnowski et al. | Introducing a novel data over voice technique for secure voice communication | |
| de Andrade et al. | Speech privacy for modern mobile communication systems | |
| CN112614498B (en) | Method for modulating low-speed data into voice-like signal | |
| Hermassi et al. | A joint encryption-compression codec for speech signals using the ITU-T G. 711 standard and chaotic map | |
| Nazarov | Signal constructions based on OFDM signals resistant to the influence of spectrum-concentrated interference | |
| CN106098073A (en) | A kind of end-to-end speech encrypting and deciphering system mapping based on frequency spectrum | |
| Özkan et al. | Data transmission via GSM voice channel for end to end security | |
| Ryan et al. | Communications and information systems | |
| Kumaresan | An inverse signal approach to computing the envelope of a real valued signal | |
| Biancucci et al. | Secure data and voice transmission over GSM voice channel: Applications for secure communications | |
| CN105788602A (en) | Voice encryption method and device for voice band compression system | |
| Chen et al. | End-to-end speech encryption algorithm based on speech scrambling in frequency domain | |
| Čubrilović et al. | Evaluation of improved classification of speech-like waveforms used for secure voice transmission | |
| Polak et al. | Secure voice transmission in the coalitional communication | |
| Qi et al. | Novel end-to-end voice encryption method in GSM system | |
| CN104994500A (en) | Voice secrecy transmission method and device used for mobile phone | |
| Siohan et al. | Design of optimal short OFDM-type windows |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |