CN101114450A - Speech encoding selectivity encipher method - Google Patents
Speech encoding selectivity encipher method Download PDFInfo
- Publication number
- CN101114450A CN101114450A CNA2007100527999A CN200710052799A CN101114450A CN 101114450 A CN101114450 A CN 101114450A CN A2007100527999 A CNA2007100527999 A CN A2007100527999A CN 200710052799 A CN200710052799 A CN 200710052799A CN 101114450 A CN101114450 A CN 101114450A
- Authority
- CN
- China
- Prior art keywords
- safe class
- bit
- parameter
- speech frame
- stream
- 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.)
- Granted
Links
Images
Abstract
The invention discloses a speech coding selective encryption method, which comprises a step A, first of all, performing the initialization, selecting the safe level the encryption needs, and confirming the adjusting parameter and initial value of the chaotic stream encryption method; a step B, choosing and using the encryption bit position table according to the safe level, and performing the encryption operation using the chaotic stream encryption method to encrypt a part of the bit. The invention also comprises a step C, the decryption part first processes the initialization after receiving the encryption data, and then performs the decryption operation. The invention selects the bit of the speech coding frame which needs encrypting according to the safe level, increases the real time of the speech talking, decreases the system energy consumption, and satisfies the business requirement of different levels.
Description
Technical field
The present invention relates to the speech coding technology field.Especially relate to a kind of audio coding selective cryptographic method, particularly relate to the G.729 selective cryptographic method of voice coding.
Background technology
Along with the growth at full speed of communication service demand, in order to guarantee the transmission of communication service data, people carry out the research work of various data compression techniques energetically.And the compressed digital transmission of voice signal is the direction of people's effort always for many years.
Prior art generally adopts the low rate speech coding technology to carry out voice transfer.The low rate voice coding is under the situation that as far as possible reduces distortion, reduces code rate, so that shared bandwidth when reducing to transmit.It is compared to analog transmission, can save bandwidth, is convenient to realize and internet (Internet Protocol, fusion IP).
G.729 be International Telecommunications Union's (International Telecommunication Union, ITU) voice coding standard of formulating for low code check multimedia communication.This voice coding scheme is the H.324 ingredient of standard series of ITU-T, can be with other audio signal components of low-down code check compressed voice or multimedia equipment, and G.729 the sample frequency of voice signal is 8kHz, frame length 10ms.G.729 adopted conjugate structure algebraic code excited linear prediction (Conjugate-SstructureAlgebraic-Code-Excited Linear-Prediction, CS-ACELP) algorithm, it is based on Code Excited Linear Prediction (Code-Excited Linear-Predictive, CELP) coding techniques, per 80 sampling points are a speech frame, mathematical model according to voice signal, comprehensive utilization linear predictive coding (Linear Predictive Coding, LPC), analysis-by-synthesis method (Analysis-By-Synthesis, ABS), vector quantization (Vector Quantizer, VQ) and perceptual weighting technology such as (Perceptual Weighting), to speech signal analysis and extract various parameter (coefficient of linear prediction wave filter, code book index in adaptive codebook (Adaptive codebook) and the fixed codebook (Fixed codebook), adaptive codebook vector gain and fixed codebook vector gain), these parameter codings are sent.Decoding end reverts to parameter coding to the bit stream of receiving, decoding obtains each parameter, from adaptive codebook, obtain the adaptive code vector with the adaptive codebook index, from fixed codebook, be fixed code vector with fixed codebook indices, multiply by their gain respectively, by constituting activation sequence after the addition.Constitute composite filter with coefficient of linear prediction wave filter, realize when long or the fundamental tone synthetic filtering with adaptive code this method, calculate synthetic speech after, further strengthen tonequality with postfilter.
As shown in Figure 1, be coding principle figure G.729.
Input signal is earlier through the high-pass filtering pre-service, and the voice signal that forms after handling is as " input signal " of subsequent analysis.Signal estimation is analyzed every frame and is carried out once linear prediction (Linear Predictive, LP) analyze, calculate the LP filter coefficient, these Parameters Transformation be line spectrum pair (Line Spectrum Pairs, LSP) because it is frequency domain parameter, so and there is more closely contact at the peak of voice signal spectrum envelope, constitute composite filter with the LSP parameter simultaneously and guarantee stability easily, and LSP has the error relativity, promptly the LSP deviation on certain Frequency point only exerts an influence near the voice this frequency.To the quantification of LSP parameter, adopt two-stage prediction division formula vector quantization, (MovingAverage, MA) the LSP coefficient calculations value of predictor predicts present frame and predicted value difference quantize with two-stage VQ with one 4 rank running mean.First order LSP vector dimension is 10 dimensions, uses 7 bits of encoded; Partial quantification input is quantification output and its input poor of first order quantizer, and its output is split into two 5 dimension code book L2, L3, each code book 5 bits of encoded.Pumping signal is searched for for estimating so that the error perceptual weighting of raw tone and synthetic speech is minimum with analysis-by-synthesis method (ABS) search.This process is that error signal is passed through a perceptual weighting wave filter, and perceptual weighting filter forms with non-quantized LP coefficients to construct.
Excitation parameters comprises fixed codebook (Fixed codebook) and adaptive codebook (Adaptivecodebook), and every subframe (5ms, 40 sampling points) is calculated once.What quantize is used for second subframe with non-quantized LP filter coefficient, and uses the LP coefficient of interpolation (Interpolated) in first subframe, estimates an open-loop pitch time-delay according to the every frame of perceptual weighting voice signal (10ms, 80 sampling points).Each subframe all repeats operation.Echo signal obtains by the LP residual computations that weighted synthesis filter filters.Revise the original state of these wave filters by error filtering between the LP residual sum excitation, this is equivalent to the zero input response that deducts weighted synthesis filter from weighted speech signal.Calculate the impulse response of weighted synthesis filter, do closed loop pitch analysis (promptly seeking adaptive codebook time-delay and gain) with near the value echo signal and the time-delay of impulse response search open-loop pitch then, mark fundamental tone time-delay (Fractional Pitch Delay, FPD) resolution 1/3 sampling point at interval, encode with 8bits fundamental tone time delay in first subframe, encode with 5bits in second subframe, from echo signal, deduct the contribution of adaptive codebook.New echo signal is used for fixing codebook search and seeks Optimum Excitation, and the algebraic codebook of 17bits is used for fixing the code book coding.
The gain of adaptive codebook and fixed codebook quantizes with the 7bits combined vectors.Adaptive codebook and fixed codebook gain are carried out vector quantization with the code book of two-stage conjugated structure.Fixed codebook gain can be expressed as: gc=γ gc ', and wherein gc ' is based on the prediction gain of former fixed codebook energy, and γ is a correction coefficient.Adaptive codebook gain gp and correction coefficient γ are to use the vector of two-stage conjugated structure codebook quantification.It is synthetic that " conjugation " is meant that each input vector is quantified as the linearity of two code books.The first order comprises the two-dimentional code book GA of one 3 bit, and the second level comprises the two-dimentional code book GB of one 4 bit.First element of each code book, representative quantizes the part of adaptive codebook gain, and second element represented the part of fixed codebook gain correction coefficient.
As shown in Figure 2, be the schematic diagram of G.729 decoding, at first, from the bit stream that receives, extract parameter reference, these parameter references of decoding obtain 10ms speech frame corresponding codes parameter, comprise LSP coefficient, 2 mark fundamental tone time delays, 2 fixed codebook vectors, 2 self-adaptations and fixed codebook gain.The LSP coefficient of each subframe is by interpolation, and is transformed to the LP filter coefficient.To each 5ms subframe, carry out following operation then:
1, rebuilds pumping signal: with adaptive codebook and fixed codebook vector (respectively and gain separately multiply each other) addition.
2, reconstructed speech signal: pumping signal is passed through the filtering of LP composite filter.
3, aftertreatment: the reconstructed speech signal is successively by self-adaptation postfilter (comprise when long and composite filter) in short-term, Hi-pass filter, calibration at last.
By above process as can be known, needed transmission parameters to have by the frame signal after the scrambler processing: channel parameters---LSP parameter, in order to construct the LPC composite filter in decoding end; Excitation parameters---pitch period and long-term prediction gain, the pulse position of random code book and gain.
Traditional voice coding encryption method all is that the stream of compressed voice G.729 integral body is encrypted.
Yet compare with text message, the voice flow data volume is big, real-time requires high, directly it is carried out bulk encryption with conventional cryptography algorithm such as AES (Advanced Encryption Standard), 3DES (3 Data EncryptionStandard) algorithm etc., can cause tangible time delay, reduce its real-time performance.Moreover, it also can significantly increase the computational load of system, takies more resources, consumes more multipotency source.This is in many occasions, under the very overcritical situation of low energy consumption, is intolerable such as mobile communication etc.In addition, audio data stream is carried out bulk encryption, be inconvenient to satisfy the business demand of different brackets.
Summary of the invention
Problem to be solved by this invention is to provide a kind of audio coding selective cryptographic method, and it needs the bit of the G.729 vocoder frames of encryption by the safe class selection.
A kind of audio coding selective cryptographic method for realizing that the present invention provides comprises the following steps:
Steps A is at first carried out initialization, selects to encrypt required safe class, determines the initial value of stream encryption method and regulates parameter;
Step B determines the bit position table selected for use according to safe class, promptly determines the concrete bit that will encrypt to utilize described stream encryption method to carry out cryptographic operation.
Described audio coding selective cryptographic method can also comprise the following steps:
Step C after deciphering side receives enciphered data, at first carries out initialization, carries out decryption oprerations then.
Described steps A comprises the following steps:
Described step B comprises the following steps:
Step B1 determines the bit position table of selecting for use according to safe class;
Step B2 after the bit position table of determining to select for use according to safe class, carries out the cryptographic operation of present frame;
Step B3, finish the cryptographic operation of present frame after, if key stream uses, utilize the adjusting parameter in the described chaotic stream encryption method that current chaos sequence value is carried out iteration, obtain new key stream, upgrade the key stream table; If also have speech frame, then go back to step B2, then handle next frame, if there is not speech frame, encrypts and finish.
Described step C comprises the following steps:
Step C1 receives controlled variable and obtains the encryption safe grade, and obtains the initial value of chaotic stream encryption method and regulate parameter;
Step C2, effective binary digit of getting initial value is converted to byte type data, deposits data converted in byte type array as the key stream table, and this key stream table will be in order to carry out XOR with the position of audio data stream ciphertext encryption, to obtain expressly;
Step C3 determines the bit position table of selecting for use according to safe class;
Step C4 carries out the decryption oprerations of present frame;
Step C5, finish the decryption oprerations of present frame after, if key stream uses, utilize to regulate parameter current chaos sequence value carried out iteration, obtain new key stream, upgrade the key stream table; If also have speech frame, then go back to step C4, then handle next frame, if there is not speech frame, deciphering finishes.
Described stream encryption method is the chaotic stream encryption method.
Described chaotic stream encryption method is the chaotic stream encryption method that improved Logistics mapped one-dimensional nonlinear iteration method characterizes, and described alternative manner is shown below:
G(x)＝(β+1)(1+1/β)
βx(1-x)
β
Wherein, β ∈ (1,4), x
0∈ (0,1), the initial value of x are x
0, can obtain x by this formula iteration
1, x
2, x
3... x
n....
Among the described step B, determine the bit position table selected for use by encrypting required safe class, for:
To the effect of reconstructed speech and the difference of importance, bit is divided into different classes of according to each parameter in the received pronunciation frame G.729, encrypts different classes of bit and obtain different safety class or effect.
The classification of the bit of parameter has following two kinds of patterns in the described speech frame:
The first, be primary goal with confidentiality and efficient, obtain confidentiality as well as possible to encrypt the least possible position, be intended to reduce voice intellingibility.Be divided into six safe classes.
Safe class one, 20 bits in the encrypted speech frame;
Safe class two, 21 bits in the encrypted speech frame;
Safe class three, 22 bits in the encrypted speech frame;
Safe class four, 23 bits in the encrypted speech frame;
Safe class five, 24 bits in the encrypted speech frame;
Safe class six, 36 bits in the encrypted speech frame;
The second, be that target is considered with the business demand that may occur.Following scheme is proposed:
Safe class seven, fuzzy voice sex (promptly by listening voice can't accurately judge former speaker's sex, can understand semanteme, be male voice or female voice but can not listen) need to select the bit of encryption;
Safe class eight, reduce to converse/difference of non-conversation phase need to select the bit encrypted;
In the practical application, can be used in combination above safe class according to concrete needs.
The invention has the beneficial effects as follows: audio coding selective cryptographic method of the present invention, can improve the real-time of voice call on the one hand.It carries out selective encryption to speech frame, in frame, by the needs of safe class, bit to those susceptibility height (promptly big to the reconstructed speech influence) is encrypted, the voice medium content is carried out bulk encryption compare with traditional, this accelerates enciphering rate undoubtedly, raises the efficiency; On the other hand, it can reduce system energy consumption, reduces taking resource.Cutting down the consumption of energy in the very nervous applied environment of some energy resources, carry finite energy as its mobile device in the wireless communication field, is very crucial problem, reduces the taking of resource, and also can alleviate the load of network processing node.Further, it can be convenient to satisfy the business demand of different brackets.Selective encryption not only can satisfy better demand technically, and it also has actual using value in the business of many internets.Such as, the website of all kinds of audio service is provided, its phonetic material can offer the user with different Cipher Strengths, if provide free audition to the user, can take certain low intensive encryption to voice, to carry out Fuzzy Processing, for another example, be appreciated that under the situation of voice content, the voice sex is beyond recognition.Suchlike many application, in a word, selective encryption provides feasibility and good technical support for the business demand of different brackets.
Description of drawings
Fig. 1 is the G.729 speech coding principles figure of received pronunciation frame of prior art;
Fig. 2 is the G.729 tone decoding schematic diagram of received pronunciation frame of prior art;
20 bits of Fig. 3 for needing in the speech frame in the safe class one of the present invention to encrypt;
21 bits of Fig. 4 for needing in the speech frame in the safe class two of the present invention to encrypt;
22 bits of Fig. 5 for needing in the speech frame in the safe class three of the present invention to encrypt;
23 bits of Fig. 6 for needing in the speech frame in the safe class four of the present invention to encrypt;
24 bits of Fig. 7 for needing in the speech frame in the safe class five of the present invention to encrypt;
36 bits of Fig. 8 for needing in the speech frame in the safe class six of the present invention to encrypt;
The bit that Fig. 9 need encrypt for fuzzy voice sex in the safe class seven of the present invention;
Figure 10 is for reducing in the safe class eight of the present invention to converse/bit that the difference of non-conversation phase need be encrypted;
Figure 11 is an audio coding selective encipher-decipher method process flow diagram of the present invention;
Figure 12 is step S100 initial method process flow diagram in the audio coding selective encryption of the present invention;
Figure 13 is the concrete encryption operation method process flow diagram of step S200 in the audio coding selective encryption of the present invention;
Figure 14 is step S300 decryption method process flow diagram in the audio coding selective encryption of the present invention;
G.729 received pronunciation frame the 1st byte ciphering process synoptic diagram when Figure 15 is safe class 7;
G.729 received pronunciation frame the 8th byte ciphering process synoptic diagram when Figure 16 is safe class 7;
G.729 received pronunciation frame the 9th byte ciphering process synoptic diagram when Figure 17 is safe class 7;
G.729 received pronunciation frame the 10th byte ciphering process synoptic diagram when Figure 18 is safe class 7.
Embodiment
In order to make purpose of the present invention, technical scheme and advantage clearer,, a kind of audio coding selective cryptographic method of the present invention is further elaborated below in conjunction with drawings and Examples.Should be appreciated that specific embodiment described herein only in order to explanation the present invention, and be not used in qualification the present invention.
Usually, in to voice encryption measure of effectiveness standard, comprise objective and subjective evaluation and test two aspects, but arrowband compressed voice signal for low rate, the parameter of objective evaluating such as signal to noise ratio (S/N ratio) etc., be difficult to its performance performance of real reflection, and subjectivity is evaluated and tested more closing to reality situation, so more pay attention to subjective evaluation and test usually.Several main criterion in the subjective evaluation and test promptly comprise following some:
1, voice intelligibility (semanteme)
2, the sex of voice is judged
3, plaintext attack
4, quiet/non-quiet differentiation
This just wishes to that is to say standard of reaching in the evaluation and test of the subjectivity of voice, in the embodiment of the invention, should be by encrypting, and the people can not be made after listening voice about above some effective resolution behavior again.
Need to prove, more than four standards be equivalent to the highest requirement of encryption safe intensity.The cipher round results that different safe classes obtains in the embodiment of the invention is different, be not to satisfy above four standards, and the method among the present invention is to the efficient of encrypting and a kind of compromise of security intensity, seeks an equilibrium point between the two.Above standard only is our a total standard and principle of wishing to reach, and does not have direct correlation to get in touch with safe class of the present invention or the like, also not ordering successively.
Digital speech, image and video compression data stream all have a characteristic, be the unbalanced importance (nonuniform perceptual importance) on the sense organ, that is to say that the mistake that part position upward takes place in the data stream is more significantly than the influence that the mistake that takes place is brought on other.
Therefore; be different from and encrypt whole multimedia stream; audio coding selective cryptographic method of the present invention; adopt a kind of selective cryptographic method; different with the way of encrypting all data in the conventional cryptography method; selective cryptographic method is only encrypted the more great bit of perception influence a part; rest parts does not then directly add transmitting of protection in channel; promptly by keeping encryption back data format information and control information constant; encryption section real data; thereby keep encrypting the compatibility of back data stream, it is being guaranteed under the situation of required security intensity, reduces computational load; also just saved more resources; energy consumption; transmission speed is greatly improved, and better performance has also just been arranged aspect real-time, can requirement of real time.Selective cryptographic method requires to combine with concrete data layout, and according to the difference that level of security requires, select to encrypt different sensitive datas, thereby meet the different needs,
The embodiment of the invention illustrates audio coding selective cryptographic method of the present invention with the selective encryption of voice coding G.729, but the present invention is not limited to only be applicable to G.729 voice coding, and it goes for the voice coding of other standard equally.
As shown in figure 11, wherein, Figure 11 is an audio coding selective encipher-decipher method process flow diagram of the present invention.Describe a kind of audio coding selective cryptographic method of the present invention below in detail:
Step S100 at first carries out initialization, selects to encrypt required safe class, determines to encrypt the initial value of alternative manner and regulates parameter;
As shown in figure 12, it specifically comprises the steps:
Step S110, user select to encrypt required safe class;
Step S120 determines that improved Logistics mapped one-dimensional nonlinear iteration method characterizes the initial value x of chaotic stream encryption method
0With the adjusting parameter beta;
The user selects this two parameter x
0And β, these two parameters promptly are equivalent to key.x
0Value between 0～1, β is value between 1...～4, and the figure place behind the radix point is not required, and preferably, in view of the accessible figure place of computing machine, two parameters are got behind radix point and are no more than 10.The embodiment of the invention illustrates, and described two parameters are value x
0=0.31234567, β=2.5.Simultaneously, need to prove, can be in Windows Xp, the VC programmed environment in operating system, these two parameters is stored as the double-precision floating points type, so that choose 64 effective binary digits.
Step S130 gets initial value x
0Radix point after 64 effective binary digits, be converted to 8 byte types (byte) data, deposit the array of conversion in an array as the key stream table, this key stream table will carry out XOR in order to the position that expressly will encrypt with audio data stream, to obtain ciphertext.
The described prior art that is converted to, it is divided into 8 groups with 64 binary digits, and 8 bits of each group just form byte type data;
Described key stream table is 8 byte type (byte) type array keyList[for data converted being deposited in a length] obtain, it carries out XOR in order to the position that expressly will encrypt with audio data stream, to obtain ciphertext.
Step S200 determines the encryption bit position table selected for use according to safe class, carries out cryptographic operation.
In the voice coding encrypted process, the user will select the bit of the definite voice coding that need encrypt according to the format characteristic of concrete cryptographic object voice coding G.729 after selecting to encrypt required safe class.
As shown in figure 13, specifically comprise the steps:
Step S210 at first determines the encryption bit position table selected for use according to safe class.
Step S220 after the bit position table of determining to select for use according to safe class, carries out the concrete cryptographic operation of present frame.
Step 230, finish the cryptographic operation of present frame after, if key stream uses, utilize the adjusting parameter in the described chaotic stream encryption method that current chaos sequence value is carried out iteration, obtain new key stream, upgrade the key stream table; If also have speech frame, then go back to step 220, then handle next frame, if there is not speech frame, encrypts and finish.
Further, can also comprise the following steps:
Step S300 after deciphering side receives enciphered data, at first carries out initialization, carries out decryption oprerations then.
As shown in figure 14, specifically comprise the steps:
Step S310 receives controlled variable, obtains the encryption safe grade, and obtains the initial value x of chaotic stream encryption method
0With the adjusting parameter beta;
Step S320 gets x
0Radix point after 64 effective binary digits, be converted to 8 byte type data, depositing data converted in a length is 8 byte type array keyList[] as the key stream table, in order to the audio data stream ciphertext in the position of encrypting carry out XOR, to obtain expressly;
Step S330 determines the encryption bit position table of selecting for use according to safe class.
Step S340 after the bit position table of determining to select for use according to safe class, carries out the concrete decryption oprerations of present frame;
Step S350, finish the decryption oprerations of present frame after, if key stream uses, utilize to regulate parameter current chaos sequence value carried out iteration, obtain new key stream, upgrade the key stream table; If also have speech frame, then go back to step S340, then handle next frame, if there is not speech frame, deciphering finishes.
Carry out the concrete deciphering flow process of selective cryptographic method of the present invention, the cryptographic operation flow process is corresponding with carrying out, and those skilled in the art can finish decrypting process according to the described ciphering process of the specific embodiment of the invention, therefore, in embodiments of the present invention, be described in detail no longer one by one.
Describe the adjusting parameter x of determining the chaotic stream encryption method that improved Logistics mapped one-dimensional nonlinear iteration method characterizes among the step S120 below in detail
0Concrete grammar process with β:
Preferably, take all factors into consideration characteristics, efficient and the security of selective encryption, adopt the chaotic stream encryption algorithm in the embodiment of the invention, but xor operation is not a unit with the bit, but is unit with the byte, and the use of chaotic key stream is improved, in conjunction with to selected characteristics that need the encryption section position of frame structure G.729, can suitably reuse, and extremely not influence its security performance, thereby raise the efficiency chaotic key stream.
The advantage of stream encryption is that wrong expansion is little, real-time is high, and its secrecy depends on the randomness of the key sequence that key generator produces, and when key sequence during very near random series, its security is very high.But in traditional stream encryption method, the pseudo-random sequence generator of employing is linear congruence generator or linear feedback shift register, and their security is all relatively poor, easily is cracked.Therefore, the embodiment of the invention adopts improved Logistics Map chaotic stream encryption method.
As a kind of enforceable mode, a kind of simple and autonomous one-dimensional discrete dynamic system that be widely studied is the Logistics mapping method, and it is shown below with the chaotic stream encryption method that the one-dimensional nonlinear alternative manner characterizes.
F(x
n)＝λx
n(1-x
n)
Wherein, n=0,1,2 ..., x
0With λ is to regulate parameter, works as x
0∈ (0,1), during λ ∈ (3.5699456..., 4), the Logistics mappings work is in chaos state, promptly by starting condition x
0Sequence { the x that under the effect of Logistics mapping method, is produced
k, k=0,1,2,3...} is non-periodic, convergent not, and very responsive to initial value and parameter.
This chaotic stream encryption method has two problems, one is the point of fixity of stream encryption method, and promptly repeatedly iteration levels off to some fixed values, and another is " stability window ", promptly assemble at the point in certain interval, the sequence of iterations that produces in the window can not provide as the necessary security of key stream.
The embodiment of the invention adopts improved Logistics map (Logistics map) one-dimensional nonlinear alternative manner to avoid prior art problems.The alternative manner of the embodiment of the invention is shown below:
G(x)＝(β+1)(1+1/β)
βx(1-x)
β
Wherein, β ∈ (1,4), x
0∈ (0,1), the initial value of x are x
0, can obtain x by this formula iteration
1, x
2, x
3... x
n....
In theory, the sequence that following formula produces is non-periodic, but owing to be subjected to the restriction of machine word length, the actual chaos sequence that is obtained by Computer Simulation all is approaching objective chaos, therefore " circulation window " problem is arranged, pass through the several times iteration exactly, iterative value occurs periodically.According to the result of test, when adopting the floating-point operation of double precision, available iteration average time＞=2*10
7
Describe in detail below among the step S200 by encrypting required safe class and determine the concrete grammar process of the bit position table selected for use:
For being described, the required safe class of encryption of the present invention determines the bit position table of selecting for use, G.729 standard frame at first is described, and analyze its parameter role and importance according to its coding principle, select the bit that to encrypt by safe class or actual demand according to its parameter role then, bit is divided into different classes of, encrypts different classes of bit and obtain different safety class or effect.
As shown in table 1 is the bit allocation table of G.729 vocoder frames.
The bit allocation table of table 1G.729 vocoder frames
| Parameter | Code word | First subframe | Second subframe | Every frame amounts to (bit) | |
| Line spectrum pair (LSP) | L0，L1，L2，L3 | 18 | |||
| The adaptive codebook time delay | P1， |
8 | 5 | 13 | |
| The parity checking of fundamental tone | P0 | 1 | 1 | ||
| Fixed codebook indices | C1，C2 | 13 | 13 | 26 |
| The fixed codebook symbol | S1， |
4 | 4 | 8 |
| Code book gain (stage one) | GA1， |
3 | 3 | 6 |
| Code book gain (stage two) | GB1， |
4 | 4 | 8 |
| Amount to | - | - | - | 80 |
By analysis to G.729 cataloged procedure, the contrast frame structure, can determine the effect of each parameter in the frame:
1) the LSP parameter L 0 in the frame G.729, L1, L2, L3 is the parameter of linear prediction filter (composite filter in short-term), in order to constitute the LPC composite filter in decoding end, very crucial, the intelligibility (i.e. semanteme) of LPC coefficient and last voice has close association;
2) G.729 the time-delay of the adaptive codebook in the frame (gene time-delay) parameter P1, P2 is long time base sound composition in the total excitation of decoding end, also is the recurrent pulse composition, and this composition influence the sex judgement of people's ear to voice;
3) G.729 code book gain G A, the GB in the frame is the combined coding of the gain of adaptive codebook and fixed codebook, and gain parameter affects people's ear to the conversation of voice and the resolution characteristic of quiet phase;
4) fixed codebook indices in the frame and fixed codebook symbol symbol and the position of representing the coded pulse of fixed codebook (the non-periodic pulse composition in the voice-activated) respectively G.729.Fixed codebook is in the cataloged procedure to pumping signal, to approaching of the difference of target vector and long time base sound contribution, being the residual signal of total excitation, also is that it is the non-periodic pulse composition in total excitation to the approaching of the difference of target vector and the contribution of long time base sound.
Therefore, by to voice encryption measure of effectiveness standard and G.729 each parameter role unite consideration, clear and definite LSP parameter, code book gain and three parameters of adaptive codebook time-delay are extremely important, need when selecting encryption parameter at first to consider that the encryption of fixed codebook parameters can be considered in postpone.
The present invention (is the encryption of the safe class six among the present invention with the G.729 speech frame significant bits position 36bit that people such as K.Swaninathan propose, see Fig. 8) minimal set for the basis, three parameters having selected in the above-mentioned analysis of main consideration, in conjunction with the more detailed separately cataloged procedure in its inside, further carry out the selection of bit, in 36 bits, to filter out bit still less once more three parameter inside.
1, the inner further selection of LSP coefficient (LPC coefficient).Can see by Fig. 8, two parameters of L1, L2 have only been comprised in 36 bits, do not comprise L0, L3, this is because during G.729 to the LSP coefficient coding, at first use 4 rank running mean (MA, Moving Average) fallout predictor is predicted the line spectral frequencies of present frame, and the method that the two-stage division formula predictive vector that difference calculating and line spectral frequencies prediction adopts quantizes quantizes.The first order is that 10 n dimensional vector ns quantize, and L1 represents code book index, encodes with 7bit; The second level is the code book that 10 dimensions is split into two 5 dimensions, and L2 represents 5 lower dimension code book indexes, and L3 represents 5 higher dimension code book indexes, encodes with 5bits respectively.L0 only be used to select two kind of 4 rank MA fallout predictor G.729 providing one of them.Quantization encoding process by research LSP coefficient knows that L1 is the index of first order vector quantization, and is the most important, and it can have influence on the decoding of decoding end place to L2, L3, therefore should preferentially encrypt in L1, L2, L3; The first part of coding when L2 is second level vector quantization can influence at the L3 of coding and decoding thereafter, therefore has precedence over L3 it is encrypted.In addition, because L1, L2, L3 are code book indexes, be not code book itself, to its part bit map, can greatly influence the code book of searching for out by them, therefore on the above-mentioned basis that two parameters of L1, L2 are encrypted, further, only this L1, L2 part position is encrypted;
2, the further selection of adaptive codebook delay parameter inside.Each subframe of adaptive codebook search (5ms) is carried out once.Decide the optimum delay of each subframe with closed-Loop Analysis minimizing Weighted square error, the criterion of closed loop pitch searcher is to make all square perceptual weighting error minimum between raw tone and the reconstruct voice.Determine after the fundamental tone time delay, carry out the code word of adaptive codebook time delay and calculate.The first subframe fundamental tone time delay T1 encodes with 8bits, and the second subframe fundamental tone time delay T2 encodes with 5bits.The value coding of fundamental tone time delay T2 is relevant with the value of T1, is equivalent to differential coding.Therefore, in decoding end, the importance of P1 is higher than P2, because it is also influential to the decoding of T2.So, to P1, when P2 encrypts, should be to the bit that P1 encrypts more than P2.We are to P1, P2 encryption section position;
3, the further selection of code book gain parameter inside.Adaptive codebook gain and fixed codebook gain are the vector quantizations with 7bits, search for the gain code book with the error weighted mean square minimum between raw tone and the reconstruct voice.Same, because GA, GB are code book indexes, be not code book itself, to its part bit map, can greatly influence the code book of searching for out by them, therefore on the above-mentioned basis that two parameters of GA, GB are encrypted, further, only this GA, GB are carried out the part bit encryption;
On the basis of above analysis, after G.729 speech frame carries out the encryption of different bits, encrypted speech is carried out audition repeatedly and analyzes relatively, at last speech frame selective encryption G.729 is divided into 8 safe classes according to the audition result, the first six safe class is intended to obtain high as far as possible confidentiality with minimum encrypted bits, and latter two safe class is intended to provide possibility and technical support for all kinds of business demands.
8 safe classes propose respectively according to following two kinds of targets:
The first, be primary goal with the confidentiality, obtain confidentiality as well as possible to encrypt the least possible position, be intended to reduce voice intellingibility.Be divided into six safe classes:
Safe class one, 20 bits in the encrypted speech frame are as Fig. 3;
Safe class two, 21 bits in the encrypted speech frame are as Fig. 4;
Safe class three, 22 bits in the encrypted speech frame are as Fig. 5;
Safe class four, 23 bits in the encrypted speech frame are as Fig. 6;
Safe class five, 24 bits in the encrypted speech frame are as Fig. 7;
Safe class six, 36 bits in the encrypted speech frame are as Fig. 8;
The second, be that target is considered with the business demand that may occur, following method is proposed:
Safe class seven, fuzzy voice sex (promptly by listening voice can't accurately judge former speaker's sex, can understand, be male voice or female voice but can not listen) need to select the bit of encryption, as Fig. 9;
Safe class eight, reduce to converse/difference of non-conversation phase need to select the bit encrypted, as Figure 10;
In embodiments of the present invention, provided the selective encryption of 8 class safe classes, without limits certain concrete application background and demand for security.If in certain practical application, bit is carried out the application requirements that the encryption safe effect does not reach concrete environment, can between 8 kinds of safe classes of the present invention, carry out independent assortment, up to obtaining required effect.
As a kind of enforceable mode, the data structure of the bit position table that each byte need be encrypted in described safe class and selective encryption one frame can be expressed as follows:
/ * safe class */
#define SECURITY_LEVEL1 1 // safe class one
#define SECURITY_LEVEL2 2 // safe class two
#define SECURITY_LEVEL3 3 // safe class three
#define SECURITY_LEVEL4 4 // safe class four
#define SECURITY_LEVEL4 5 // safe class five
#define SECURITY_LEVEL4 6 // safe class six
#define SECURITY_LEVEL4 7 // safe class seven
#define SECURITY_LEVEL4 8 // safe class eight
The bit position table * that each byte need be encrypted in/* selective encryption one frame/
byte?level1_Bitpos_List[]＝{0x1e，0x07，0x7c，0x00，0x00，0x90，0x19，0x00，0x00，0x32}；
During // safe class 1, the bit position table of selecting for use
byte?level2_Bitpos_List[]＝{0x1e，0x07，0x7c，0x00，0x00，0x90，0x39，0x00，0x00，0x32}；
During // safe class 2, the bit position table of selecting for use
byte?level3_Bitpos_List[]＝{0x1e，0x07，0x7c，0x00，0x00，0xb0，0x19，0x00，0x00，0x36}；
During // safe class 3, the bit position table of selecting for use
byte?level4_Bitpos_List[]＝{0xle，0x07，0x7c，0x00，0x00，0xb0，0x39，0x00，0x00，0x36}；
During // safe class 4, the bit position table of selecting for use
byte?level5_Bitpos_List[]＝{0x3e，0x07，0x7c，0x00，0x00，0xb0，0x39，0x00，0x00，0x36}；
During // safe class 5, the bit position table of selecting for use
byte?level6_Bitpos_List[]＝{0xfe，0x1f，0xfc，0x01，0x00，0xf0，0x3f，0x00，0x00，0xfe}；
During // safe class 6, the bit position table of selecting for use
byte?leve17_Bitpos_List[]＝{0x00，0x00，0xfc，0x7f，0x00，0x00，0xf8，0x0f，0x00，0x00}；
During // safe class 7, the bit position table of selecting for use
byte?level8_Bitpos_List[]＝{0x00，0x00，0x00，0x00，0x00，0xf0，0x07，0x00，0x00，0xfe}；
During // safe class 8, the bit position table of selecting for use
Describe the concrete grammar process that step S220 of the present invention carries out cryptographic operation below in detail:
In embodiments of the present invention, be that 7 situation is that example is analyzed explanation with safe class, the bit position table that encrypt this moment is selected level7_Bitpos_List[for use].
Array keyList[8] deposit chaotic key stream.Get 64 binary digits behind the Xi radix point, generate byte type (byte) the type data of 8 bytes, deposit array keyList[8 in] in.Sign amount times is set, to write down the number of times that same key uses in this method.Its initial value is set to 0.
Specifically comprise the steps:
Step S221, at first receive first frame of speech frame, array level7_Bitpos_List[] the bit that need encrypt of the 1st byte of the 1st record speech frame, take out its value 00011110, bit is that the corresponding bits position of the corresponding speech frame byte of 1 explanation needs to encrypt, otherwise need not;
Step S222, with 00011110 with array keyList[] ((1+times) %8) Xiang Xiangyu of (depositing 64 key streams of generation), the 1st byte of result who obtains and speech frame is carried out XOR, the result sends speech frame the 1st byte back to, as shown in figure 15, diagram is an example during with times=0, therefore array keyList[among Figure 15 this moment] promptly be to use the 1st;
Step S223 searches array level7_Bitpos_List[again] the 2nd, the bit that need encrypt of the 2nd byte of its record speech frame takes out its value 000001111;
Step S224, with 00011110 with array keyList[] ((2+times) %8) Xiang Xiangyu, the 2nd byte of result who obtains and speech frame is carried out XOR, the result sends speech frame the 2nd byte back to;
By that analogy, up to being encrypted, the 8th byte of speech frame finishes, as shown in figure 16.
Step S225 when the 9th byte of speech frame need be encrypted, searches array level7_Bitpos_List[] the 9th, with keyList[] ((9+times) %8) Xiang Xiangyu of table, carry out XOR encryption with this result and speech frame the 9th byte.
As shown in figure 17, when proceeding to step S225, keyList[] 8 bytes all used one to take turns, when the 9th byte of speech frame need be encrypted, keyList[] table gets back to gauge outfit ((9+times) %8), diagram is an example during with times=0, and therefore promptly be to get back to the 1st of gauge outfit this moment among Figure 17.Search array level7_Bitpos_List[] the 9th, with this byte and array keyList[] ((9+times) %8) Xiang Xiangyu, result who obtains and speech frame the 9th byte are carried out XOR and are encrypted.Because the bit that each byte of speech frame will be encrypted is also inequality, even keyList[] to reuse, the bit that wherein really uses each time also is not quite similar.
Step S226 slides backward and continues to encrypt the 10th byte, simultaneously, and keyList[] show also to slide backward level7_Bitpos_List[] also move one backward;
As shown in figure 18, after speech frame the 9th byte has been encrypted, slide backward and continue to encrypt the 10th byte, simultaneously, keyList[] table also slide rearwardly to ((10+times) %8), level7_Bitpos_List[] also move one backward, search array array level7_Bitpos_List[] the 10th, with this byte and array keyList[] ((10+times) %8) Xiang Xiangyu, result who obtains and speech frame the 10th byte are carried out XOR and are encrypted, diagram is an example during with times=0, therefore array keyList[among Figure 18 this moment] what use is the 2nd.
Arrive this, this speech frame is encrypted and is finished, and sign amount times increases 1.
Describe step S230 of the present invention below in detail and carry out the more concrete grammar process of new key:
Step S231, the current speech frame encrypt finish after, whether judgement symbol amount times is 8, does not upgrade key stream if not 8, goes to step S234; If 8, key stream had been used 8 times, needed to upgrade key stream, continued next step;
Step S232, sign amount times assignment again are 0;
Step S233, keyList[] the table calcellation, need to upgrade.Utilize following formula
G(x)＝(β+1)(1+1/β)
βx(1-x)
β，
To x
i(currency of x) carries out iteration, obtains x
I+1, get x
I+1Radix point after 64 effective binary digits, be converted to 8 byte type data, deposit data converted in array keyList[], realized keyList[] renewal.
Step S234 if the speech frame that will encrypt in addition then continues to get next frame, goes back to step S220; If no, encrypt end.
Audio coding selective cryptographic method of the present invention uses selective encryption that encoded voice is G.729 carried out chaotic stream encryption, has following advantage:
1) real-time of raising voice call
Audio coding selective cryptographic method improves encryption efficiency from following several respects:
A1) speech frame is carried out selective encryption, in frame, by the needs of safe class, the bit high to those susceptibility (promptly big to the reconstructed speech influence) encrypted.The voice medium content is carried out bulk encryption compare with traditional, this accelerates enciphering rate undoubtedly.
A2) use stream cipher algorithm is incited somebody to action expressly direct and key stream carries out XOR, compares grouping algorithm, and this method speed is faster.
A3) in order to strengthen security, the key stream that uses in the stream cipher algorithm is chaos sequence, is obtained by improved Logistics map mapping.This method is in iterative process, and computation complexity is bigger, in order to reduce the calculated amount of key stream,, suitably chaotic key stream is reused in conjunction with selected characteristics that need the encryption section position of frame structure G.729, and its security performance of not appreciable impact.
2) reduce system energy consumption, reduce taking resource
Speech frame is carried out selective encryption, use stream cipher algorithm, chaotic key stream is carried out suitable recycling, these measures also can reduce system energy consumption when raising the efficiency, and reduce the resource quantity that takies.Cutting down the consumption of energy in the very nervous applied environment of some energy resources, carry finite energy as its mobile device in the wireless communication field, is very crucial problem.Minimizing takies resource, also can alleviate the load of network processing node.
3) be convenient to satisfy the business demand of different brackets
Selective encryption not only can satisfy better demand technically, and it also has actual using value in the business of many internets.In speech frame, press the needs of practical business, specific partial bit position is encrypted, can obtain various effects, as the sex of fuzzy voice, make the people can't the identification men and women, for another example, the website of all kinds of audio service is provided, its phonetic material can offer the user with different Cipher Strengths, if provide free the audition to the user, can take certain low Cipher Strength to voice, to carry out Fuzzy Processing, reduce tonequality.Suchlike many application, in a word, selective encryption provides feasibility and good technical support for the business demand of different brackets.
More than specific embodiments of the invention are described and illustrate it is exemplary that these embodiment should be considered to it, and be not used in and limit the invention, the present invention should make an explanation according to appended claim.
Claims (12)
1. an audio coding selective cryptographic method is characterized in that, comprises the following steps:
Steps A is carried out initialization, selects required safe class, determines the initial value of stream encryption method and regulates parameter;
Step B determines the bit position table selected for use according to safe class, utilizes described stream encryption method to carry out cryptographic operation.
2. audio coding selective cryptographic method according to claim 1 is characterized in that, also comprises the following steps:
Step C after deciphering side receives enciphered data, at first carries out initialization, carries out decryption oprerations then.
3. audio coding selective cryptographic method according to claim 1 and 2 is characterized in that described steps A comprises the following steps:
Steps A 1, the user selects to encrypt required safe class;
Steps A 2 is determined the initial value of stream encryption method and is regulated parameter;
Steps A 3 is got effective binary digit of regulating the initial value in the parameter, and hand-over word nodal pattern data deposit data converted in an array as the key stream table, and this key stream table will carry out XOR in order to the position that expressly will encrypt with audio data stream, to obtain ciphertext.
4. audio coding selective cryptographic method according to claim 1 and 2 is characterized in that described step B comprises the following steps:
Step B1 determines the bit position table of selecting for use according to safe class;
Step B2 after the bit position table of determining to select for use according to safe class, carries out the cryptographic operation of present frame;
Step B3, finish the cryptographic operation of present frame after, if key stream uses, utilize the adjusting parameter in the described chaotic stream encryption method that current chaos sequence value is carried out iteration, obtain new key stream, upgrade the key stream table; If also have speech frame, then go back to step B2, then handle next frame, if there is not speech frame, encrypts and finish.
5. described audio coding selective cryptographic method according to claim 1 and 2 is characterized in that described step C comprises the following steps:
Step C1 receives controlled variable, obtains the encryption safe grade, and obtains the initial value of chaotic stream encryption method and regulate parameter;
Step C2 gets effective binary digit of regulating parameter, is converted to byte type data, deposits data converted in byte type array as the key stream table, this key stream table will in order to the audio data stream ciphertext in the position of encrypting carry out XOR, to obtain expressly;
Step C3 determines the bit position table of selecting for use according to safe class;
Step C4 carries out the decryption oprerations of present frame;
Step C5, finish the decryption oprerations of present frame after, if key stream uses, utilize to regulate parameter current chaos sequence value carried out iteration, obtain new key stream, upgrade the key stream table; If also have speech frame, then go back to step C4, then handle next frame, if there is not speech frame, deciphering finishes.
6. audio coding selective cryptographic method according to claim 1 and 2 is characterized in that, described stream encryption method is the chaotic stream encryption method.
7. audio coding selective cryptographic method according to claim 6 is characterized in that, described chaotic stream encryption method is the chaotic stream encryption method that improved Logistics mapped one-dimensional nonlinear iteration method characterizes.
8. audio coding selective cryptographic method according to claim 7 is characterized in that, described alternative manner is shown below:
G(x)＝(β+1)(1+1/β)
β×(1-x)
β
Wherein, β ∈ (1,4), x
0∈ (0,1), the initial value of x are x
0, can obtain x by this formula iteration
1, x
2, x
3... x
n....
9. audio coding selective cryptographic method according to claim 1 and 2 is characterized in that, described voice coding is the coding of received pronunciation G.729.
10. audio coding selective cryptographic method according to claim 9 is characterized in that, among the described step B, determines the bit position table selected for use by encrypting required safe class, for:
To the effect of reconstructed speech and the difference of importance, bit is divided into different classes of according to each parameter in the received pronunciation frame G.729, encrypts different classes of bit and obtain different safety class or effect.
11. audio coding selective cryptographic method according to claim 10 is characterized in that, the classification of the bit of parameter comprises one of following grade or its combination in the described speech frame:
Safe class one, 20 bits in the encrypted speech frame;
Safe class two, 21 bits in the encrypted speech frame;
Safe class three, 22 bits in the encrypted speech frame;
Safe class four, 23 bits in the encrypted speech frame;
Safe class five, 24 bits in the encrypted speech frame;
Safe class six, 36 bits in the encrypted speech frame.
12. audio coding selective cryptographic method according to claim 11 is characterized in that, the classification of the bit of parameter also comprises one of following grade or its combination in the described speech frame:
Safe class seven, the bit that fuzzy voice sex selection need be encrypted;
Safe class eight, reduce to converse/difference of non-conversation phase need to select the bit encrypted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2007100527999A CN101114450B (en) | 2007-07-20 | 2007-07-20 | Speech encoding selectivity encipher method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2007100527999A CN101114450B (en) | 2007-07-20 | 2007-07-20 | Speech encoding selectivity encipher method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101114450A true CN101114450A (en) | 2008-01-30 |
| CN101114450B CN101114450B (en) | 2011-07-27 |
Family
ID=39022758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2007100527999A Expired - Fee Related CN101114450B (en) | 2007-07-20 | 2007-07-20 | Speech encoding selectivity encipher method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101114450B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102036232A (en) * | 2010-12-17 | 2011-04-27 | 中兴通讯股份有限公司 | Method and device for transmitting and receiving base station data |
| CN102307089A (en) * | 2011-08-12 | 2012-01-04 | 黑龙江大学 | Chaotic encryption method based on variable initial value and pseudo decryption |
| CN102307075A (en) * | 2011-08-09 | 2012-01-04 | 深圳科立讯电子有限公司 | Voice transmission encryption method of DMR (digital mobile radio) communication terminal |
| CN104113625A (en) * | 2014-07-28 | 2014-10-22 | 深圳市中兴移动通信有限公司 | Conversation recording system, method and device and mobile terminals |
| CN105871842A (en) * | 2016-03-31 | 2016-08-17 | 宇龙计算机通信科技(深圳)有限公司 | Voice encryption and decryption methods, voice encryption and decryption devices and terminals |
| CN106098073A (en) * | 2016-05-23 | 2016-11-09 | 苏州大学 | A kind of end-to-end speech encrypting and deciphering system mapping based on frequency spectrum |
| CN106486129A (en) * | 2014-06-27 | 2017-03-08 | 华为技术有限公司 | A kind of audio coding method and device |
| CN106935243A (en) * | 2015-12-29 | 2017-07-07 | 航天信息股份有限公司 | A kind of low bit digital speech vector quantization method and system based on MELP |
| CN107508796A (en) * | 2017-07-28 | 2017-12-22 | 北京明朝万达科技股份有限公司 | A kind of data communications method and device |
| CN107864129A (en) * | 2017-10-31 | 2018-03-30 | 江苏神州信源系统工程有限公司 | A kind of method and apparatus for ensureing network data security |
| CN110601814A (en) * | 2019-09-24 | 2019-12-20 | 深圳前海微众银行股份有限公司 | Federal learning data encryption method, device, equipment and readable storage medium |
| CN111818519A (en) * | 2020-07-16 | 2020-10-23 | 郑州信大捷安信息技术股份有限公司 | End-to-end voice encryption and decryption method and system |
| CN113132418A (en) * | 2021-06-17 | 2021-07-16 | 北京电信易通信息技术股份有限公司 | Variable-grade encryption method, system and device |
-
2007
- 2007-07-20 CN CN2007100527999A patent/CN101114450B/en not_active Expired - Fee Related
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012079332A1 (en) * | 2010-12-17 | 2012-06-21 | 中兴通讯股份有限公司 | Method and apparatus for transmitting or receiving base station data |
| CN102036232B (en) * | 2010-12-17 | 2015-12-09 | 中兴通讯股份有限公司 | A kind of base station data sending, receiving method and device |
| CN102036232A (en) * | 2010-12-17 | 2011-04-27 | 中兴通讯股份有限公司 | Method and device for transmitting and receiving base station data |
| CN102307075A (en) * | 2011-08-09 | 2012-01-04 | 深圳科立讯电子有限公司 | Voice transmission encryption method of DMR (digital mobile radio) communication terminal |
| CN102307089A (en) * | 2011-08-12 | 2012-01-04 | 黑龙江大学 | Chaotic encryption method based on variable initial value and pseudo decryption |
| CN102307089B (en) * | 2011-08-12 | 2014-02-26 | 黑龙江大学 | Chaotic encryption method based on variable initial value and pseudo decryption |
| CN106486129B (en) * | 2014-06-27 | 2019-10-25 | 华为技术有限公司 | A kind of audio coding method and device |
| US11133016B2 (en) | 2014-06-27 | 2021-09-28 | Huawei Technologies Co., Ltd. | Audio coding method and apparatus |
| CN106486129A (en) * | 2014-06-27 | 2017-03-08 | 华为技术有限公司 | A kind of audio coding method and device |
| US10460741B2 (en) | 2014-06-27 | 2019-10-29 | Huawei Technologies Co., Ltd. | Audio coding method and apparatus |
| CN104113625A (en) * | 2014-07-28 | 2014-10-22 | 深圳市中兴移动通信有限公司 | Conversation recording system, method and device and mobile terminals |
| CN106935243A (en) * | 2015-12-29 | 2017-07-07 | 航天信息股份有限公司 | A kind of low bit digital speech vector quantization method and system based on MELP |
| CN105871842B (en) * | 2016-03-31 | 2020-04-07 | 宇龙计算机通信科技(深圳)有限公司 | Voice encryption and decryption method, encryption and decryption device and terminal |
| WO2017166556A1 (en) * | 2016-03-31 | 2017-10-05 | 宇龙计算机通信科技(深圳)有限公司 | Voice encryption and decryption method, encryption and decryption device, and terminal |
| CN105871842A (en) * | 2016-03-31 | 2016-08-17 | 宇龙计算机通信科技(深圳)有限公司 | Voice encryption and decryption methods, voice encryption and decryption devices and terminals |
| CN106098073A (en) * | 2016-05-23 | 2016-11-09 | 苏州大学 | A kind of end-to-end speech encrypting and deciphering system mapping based on frequency spectrum |
| CN107508796B (en) * | 2017-07-28 | 2019-01-04 | 北京明朝万达科技股份有限公司 | A kind of data communications method and device |
| CN107508796A (en) * | 2017-07-28 | 2017-12-22 | 北京明朝万达科技股份有限公司 | A kind of data communications method and device |
| CN107864129A (en) * | 2017-10-31 | 2018-03-30 | 江苏神州信源系统工程有限公司 | A kind of method and apparatus for ensureing network data security |
| CN110601814A (en) * | 2019-09-24 | 2019-12-20 | 深圳前海微众银行股份有限公司 | Federal learning data encryption method, device, equipment and readable storage medium |
| CN110601814B (en) * | 2019-09-24 | 2021-08-27 | 深圳前海微众银行股份有限公司 | Federal learning data encryption method, device, equipment and readable storage medium |
| CN111818519A (en) * | 2020-07-16 | 2020-10-23 | 郑州信大捷安信息技术股份有限公司 | End-to-end voice encryption and decryption method and system |
| CN111818519B (en) * | 2020-07-16 | 2022-02-11 | 郑州信大捷安信息技术股份有限公司 | End-to-end voice encryption and decryption method and system |
| CN113132418A (en) * | 2021-06-17 | 2021-07-16 | 北京电信易通信息技术股份有限公司 | Variable-grade encryption method, system and device |
| CN113132418B (en) * | 2021-06-17 | 2021-08-27 | 北京电信易通信息技术股份有限公司 | Variable-grade encryption method, system and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101114450B (en) | 2011-07-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101114450B (en) | Speech encoding selectivity encipher method | |
| CN101059957B (en) | An audio coding selective cryptographic method | |
| CN101925950B (en) | Audio encoder and decoder | |
| CN102119414B (en) | Device and method for quantizing and inverse quantizing LPC filters in a super-frame | |
| CN100362568C (en) | Method and apparatus for predictively quantizing voiced speech | |
| CN101996636B (en) | Sub-band voice codec with multi-stage codebooks and redundant coding | |
| CN105244034B (en) | For the quantization method and coding/decoding method and equipment of voice signal or audio signal | |
| CN101023471B (en) | Scalable encoding apparatus, scalable decoding apparatus, scalable encoding method, scalable decoding method, communication terminal apparatus, and base station apparatus | |
| CN101615396B (en) | Voice encoding device and voice decoding device | |
| CN101903945B (en) | Encoder, decoder, and encoding method | |
| CN100409308C (en) | Voice coding method and device and voice decoding method and device | |
| CN1947174B (en) | Scalable encoding device, scalable decoding device, method thereof, and scalable coding device | |
| CN101076853B (en) | Wide-band encoding device, wide-band lsp prediction device, band scalable encoding device, wide-band encoding method | |
| CN1890714B (en) | Optimized multiple coding method | |
| CN104123946A (en) | Systemand method for including identifier with packet associated with speech signal | |
| KR20000010994A (en) | Audio signal coding and decoding methods and audio signal coder and decoder | |
| CN104934036A (en) | Audio Encoding Device, Method And Program, And Audio Decoding Device, Method And Program | |
| Wu et al. | An approach of steganography in G. 729 bitstream based on matrix coding and interleaving | |
| JP6812504B2 (en) | Voice coding method and related equipment | |
| CN101399043A (en) | Self-adapting multi-speed narrowband coding method and coder | |
| CN1751338B (en) | Method and apparatus for speech coding | |
| CN101090312B (en) | Improved speech code flow enciphering Improved speech code flow enciphering method | |
| CN101208741B (en) | Method for adapting for an interoperability between short-term correlation models of digital signals | |
| CN100585700C (en) | Sound encoding device and method thereof | |
| CN102568484A (en) | Warped spectral and fine estimate audio encoding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110727 Termination date: 20120720 |