CN105374360B - Intersect additivity spread spectrum audio frequency watermark embedding grammar, detection method and system - Google Patents
Intersect additivity spread spectrum audio frequency watermark embedding grammar, detection method and system Download PDFInfo
- Publication number
- CN105374360B CN105374360B CN201510833111.5A CN201510833111A CN105374360B CN 105374360 B CN105374360 B CN 105374360B CN 201510833111 A CN201510833111 A CN 201510833111A CN 105374360 B CN105374360 B CN 105374360B
- Authority
- CN
- China
- Prior art keywords
- watermark
- signal
- signal frame
- frequency
- frame
- 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
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; 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
- G10L19/018—Audio watermarking, i.e. embedding inaudible data in the audio signal
Abstract
The invention discloses a kind of intersection additivity spread spectrum audio frequency watermark embedding grammar, detection method and systems, embedding grammar includes: that will be transformed to frequency-region signal after time-domain audio signal framing, frequency-region signal is divided into odd-even interleaving frame, positive watermark bit and negative watermark bit are embedded in odd indexed signal frame and even number serial number signal frame respectively using additivity spread-spectrum watermark embedding inlay technique, the odd indexed signal frame for being embedded in watermark and even number serial number signal frame are combined into a frame signal.Detection method includes: that will receive signal to be divided into odd-even interleaving frame, coherent detection is carried out to odd indexed signal frame and even number serial number signal frame respectively, obtain odd decision statistic and even decision statistic, using the difference of odd decision statistic and even decision statistic as decision statistic, watermark is extracted according to decision statistic.The present invention can greatly reduce decision statistic variance, to reduce the error rate of watermark detection.
Description
Technical field
The present invention relates to Digital Audio Watermarking Techniques fields more particularly to a kind of intersection additivity to spread audio frequency watermark insertion side
Method, detection method and system.
Background technique
Digital audio frequency watermark is that certain digital watermark informations are added in audio signal to reach the identification of the file true and false, copyright
Protection, Information hiding the purpose of signal processing operations.Intersect the odd indexed point frame and even number referred in a frame audio sample point
Serial number point frame is cross arrangement distribution.The short-term stationarity of audio signal determines odd indexed point frame and even number serial number point frame
With stronger correlation.In traditional additivity spread spectrum audio frequency watermark system, due to there is no to utilize the height intersected between parity frame
Correlation is spent, the watermark signal after spread spectrum is directly added with main signal, when test side is using coherent detection, obtained decision
Statistic has biggish variance, causes the error rate of watermark detection high.
Summary of the invention
In view of the problems of the existing technology, the present invention provides the intersection additivitys that one kind can reduce watermark detection error rate
Spread audio frequency watermark embedding grammar, detection method and system.
In order to solve the above technical problems, the present invention adopts the following technical scheme that:
One, intersect additivity and spread audio frequency watermark embedding grammar, comprising steps of
S1 reads audio file and obtains time-domain audio signal xt and its sample rate f s1, will become after time-domain audio signal xt framing
It is changed to frequency-region signal xf；
S2 obtains the frequency domain sample of embeddable watermark according to the audio frequency range of sample rate f s1, frame length and human ear sensitivity
This point range, to obtain the frequency domain signal X F of the domain samples point range；
S3 generates the bipolarity puppet that length is (f max1-f min1+1) by rand function using random number seed seed
Random frequency expansion sequence u, f max1 and f min1 are the maximum value and minimum value of the domain samples point range of embeddable watermark；
S4 sequentially numbers signal frame in frequency domain signal X F, and the signal frame of odd indexed is combined into odd indexed signal frame
The signal frame of even number serial number is combined into even number serial number signal frame XFeven by XFodd；
S5 is according to bipolarity pseudorandom spreading sequence u, using additivity spread-spectrum watermark embedding inlay technique, to XFodd and XFeven points
It Qian Ru not watermark bit symbol and -1*symbol, odd indexed signal frame YFodd and even number serial number letter after watermark must be embedded in
Number frame YFeven, symbol are the watermark bit that should be embedded in；
Odd indexed signal frame YFodd and even number serial number signal frame YFeven are combined into signal frame YF by serial number by S6；
Signal frame YF is transformed to time-domain signal by S7, is generated audio file, that is, is embedded in the audio file of watermark.
S3 further comprises:
Rand function is initialized and is generated random number seed seed using RandStream function；
The random number sequence that length is (f max1-f min1+1) is generated using random number seed seed and rand function；
It rounds up respectively to random number each in random number sequence, obtains the binary sequence being made of 0 and 1；
2 are multiplied to binary number each in binary sequence and subtracts 1, obtains bipolarity pseudorandom spreading sequence u.
Two, intersect additivity and spread audio frequency watermark detection method, comprising steps of
The audio file that S1 reads insertion watermark must be embedded in the time-domain audio signal rt and its sample rate f s2 of watermark, by when
Frequency-region signal sf is transformed into after the audio signal rt framing of domain；
S2 obtains the frequency domain sample of embeddable watermark according to the audio frequency range of sample rate f s2, frame length and human ear sensitivity
This point range, to obtain the frequency-region signal SF of the domain samples point range；
S3 generates the bipolarity puppet that length is (f max1-f min1+1) by rand function using random number seed seed
Random frequency expansion sequence u, f max1 and f min1 are the maximum value and minimum of the domain samples point range of embeddable watermark；
S4 sequentially numbers signal frame in frequency-region signal SF, and the signal frame of odd indexed is combined into odd indexed signal frame
The signal frame of even number serial number is combined into even number serial number signal frame SFeven by SFodd；
S5 is based on bipolarity pseudorandom spreading sequence u, to odd indexed signal frame SFodd and even number serial number signal frame
SFeven carries out coherent detection operation respectively, obtains odd decision statistic rodd and even decision statistic reven；
S6 obtains decision statistic rall=rodd- according to odd decision statistic rodd and idol decision statistic reven
Reven, by the symbol extraction watermark for judging decision statistic rall.
S3 further comprises:
Rand function is initialized and is generated random number seed seed using RandStream function；
The random number sequence that length is (f max1-f min1+1) is generated using random number seed seed and rand function；
It rounds up respectively to random number each in random number sequence, obtains the binary sequence being made of 0 and 1；
2 are multiplied to binary number each in binary sequence and subtracts 1, obtains bipolarity pseudorandom spreading sequence u.
Three, intersect additivity and spread audio frequency watermark embedded system, comprising:
Time-frequency domain conversation module obtains time-domain audio signal xt and its sample rate f s1 for reading audio file, by time domain sound
Frequency-region signal xf is transformed to after frequency signal xt framing；
Frequency range selecting module, for obtaining according to the audio frequency range of sample rate f s1, frame length and human ear sensitivity
The domain samples point range for obtaining embeddable watermark, to obtain the frequency domain signal X F of the domain samples point range；
Frequency expansion sequence generation module, being used to generate length by rand function using random number seed seed is (f max1-
F min1+1) bipolarity pseudorandom spreading sequence u, f max1 and f min1 be embeddable watermark domain samples point range
Maximum value and minimum value；
Divide odd-even interleaving frame module, for sequentially numbering signal frame in frequency domain signal X F, by the signal frame of odd indexed
It is combined into odd indexed signal frame XFodd, the signal frame of even number serial number is combined into even number serial number signal frame XFeven；
Watermark intersects insertion module, is used to be embedded according to bipolarity pseudorandom spreading sequence u using additivity spread-spectrum watermark
Method is respectively embedded into watermark bit symbol and -1*symbol to XFodd and XFeven, the odd indexed letter after must being embedded in watermark
Number frame YFodd and even number serial number signal frame YFeven, symbol are the watermark bit that should be embedded in；
Odd-even interleaving frame composite module, for pressing odd indexed signal frame YFodd and even number serial number signal frame YFeven
Serial number is combined into signal frame YF；
Time-frequency domain inverse transform module generates audio file for signal frame YF is transformed to time-domain signal, i.e. insertion watermark
Audio file.
Four, intersect additivity and spread audio frequency watermark detection system, comprising:
Time-frequency domain conversation module, for read insertion watermark audio file must be embedded in watermark time-domain audio signal rt and
Its sample rate f s2 will be transformed into frequency-region signal sf after time-domain audio signal rt framing；
Frequency range selecting module, for obtaining according to the audio frequency range of sample rate f s2, frame length and human ear sensitivity
The domain samples point range for obtaining embeddable watermark, to obtain the frequency-region signal SF of the domain samples point range；
Frequency expansion sequence generation module, being used to generate length by rand function using random number seed seed is (f max1-
F min1+1) bipolarity pseudorandom spreading sequence u, f max1 and f min1 be embeddable watermark domain samples point range
Maximum value and minimum；
Divide odd-even interleaving frame module, for sequentially numbering signal frame in frequency-region signal SF, by the signal frame of odd indexed
It is combined into odd indexed signal frame SFodd, the signal frame of even number serial number is combined into even number serial number signal frame SFeven；
Coherent detection module is used to be based on bipolarity pseudorandom spreading sequence u, to odd indexed signal frame SFodd and idol
Number sequence signal frame SFeven carries out coherent detection operation respectively, obtains odd decision statistic rodd and even decision statistic reven；
Intersect judging module, for obtaining decision statistic according to odd decision statistic rodd and idol decision statistic reven
Rall=rodd-reven, by the symbol extraction watermark for judging decision statistic rall.
Compared to the prior art, the invention has the advantages that and the utility model has the advantages that
The present invention can greatly reduce decision statistic variance, to reduce the error rate of watermark detection.
Detailed description of the invention
Fig. 1 is the structural block diagram that the present invention intersects additivity spread spectrum audio frequency watermark embedded system；
Fig. 2 is the structural block diagram that the present invention intersects additivity spread spectrum audio frequency watermark detection system；
Fig. 3 is the specific flow chart that the present invention intersects additivity spread spectrum audio frequency watermark embedding grammar；
Fig. 4 is the specific flow chart that the present invention intersects additivity spread spectrum audio frequency watermark detection method.
Specific embodiment
Below in conjunction with attached drawing, specific embodiments of the present invention is described in detail.
Referring to Fig. 1, intersection additivity provided in this embodiment spreads audio frequency watermark embedded system, including time-frequency domain conversation module
1, frequency range selecting module 2, frequency expansion sequence generation module 3, divide odd-even interleaving frame module 4, watermark to intersect insertion module 5, odd
Even to intersect frame composite module 6 and time-frequency domain inverse transform module 7, when specific implementation, can realize each mould using software firming bechnology
Block.
The time-frequency domain conversation module 1 obtains time-domain audio signal xt and its sample rate f s1 for reading audio file,
Frequency-region signal xf will be transformed to after time-domain audio signal xt framing.Frequency-region signal xf is exported to frequency range selecting module 2.
The frequency range selecting module 2, for according to the sound frequency of sample rate f s1, frame length and human ear sensitivity
Range, obtains the domain samples point range of embeddable watermark, to obtain the frequency domain signal X F of the domain samples point range.It can
The maximum value and minimum value for being embedded in the domain samples point range of watermark are exported to frequency expansion sequence generation module 3, by frequency domain signal X F
It exports to dividing odd-even interleaving frame module 4.
The frequency expansion sequence generation module 3, for being by rand function generation length using random number seed seed
Bipolarity pseudorandom spreading sequence u, the f max1 and f min1 of (f max1-f min1+1) are the domain samples of embeddable watermark
The maximum value and minimum value of point range.Bipolarity pseudorandom spreading sequence u, which is exported, intersects insertion module 5 to watermark.
Described divides odd-even interleaving frame module 4, for sequentially numbering signal frame in frequency domain signal X F, by odd indexed
Signal frame combines odd indexed signal frame XFodd, and the signal frame of even number serial number is combined into even number serial number signal frame XFeven.It will
Odd indexed signal frame XFodd and even number serial number signal frame XFeven, which is exported, intersects insertion module 5 to watermark.
The watermark intersects insertion module 5, is used to according to bipolarity pseudorandom spreading sequence u, spreads water using additivity
Embedding inlay technique is printed, watermark bit symbol and -1*symbol are respectively embedded into XFodd and XFeven, the odd number after watermark must be embedded in
Serial number signal frame YFodd and even number serial number signal frame YFeven, symbol are the watermark bit that should be embedded in.By odd indexed signal
Frame YFodd and even number serial number frame frequency-region signal YFeven are exported to odd-even interleaving frame composite module 6.
The odd-even interleaving frame composite module 6 is used to odd indexed signal frame YFodd and even number serial number signal frame
YFeven is combined into signal frame YF by serial number.Signal frame YF is exported to time-frequency domain inverse transform module 7.
The time-frequency domain inverse transform module 7 generates audio file, i.e., for signal frame YF is transformed to time-domain signal
It is embedded in the audio file of watermark.
Referring to fig. 2, intersection additivity provided in this embodiment spreads audio frequency watermark embedded system, including time-frequency domain conversation module
8, frequency range selecting module 9, frequency expansion sequence generation module 10, divide odd-even interleaving frame module 11, coherent detection module 12 and hand over
Judging module 13 is pitched, each module can be realized using software firming bechnology when specific implementation.
Time-frequency domain conversation module 8, frequency expansion sequence generation module 10, divides odd-even interleaving frame module at frequency range selecting module 9
11 function respectively with time-frequency domain conversation module 1, frequency range selecting module 2, frequency expansion sequence generation module 3, divide odd-even interleaving
Frame module 4.
The time-frequency domain conversation module 8, for taking the audio file of insertion watermark that must be embedded in the time-domain audio letter of watermark
Number rt and its sample rate f s2, will be transformed into frequency-region signal sf after time-domain audio signal rt framing.Frequency-region signal xf is exported to frequency
Rate range selection module 9.
The frequency range selecting module 9, for according to the sound frequency of sample rate f s2, frame length and human ear sensitivity
Range, obtains the domain samples point range of embeddable watermark, to obtain the frequency-region signal SF of the domain samples point range.It can
The maximum value and minimum value for being embedded in the domain samples point range of watermark are exported to frequency expansion sequence generation module 10, by frequency-region signal SF
It exports to dividing odd-even interleaving frame module 11.
The frequency expansion sequence generation module 10, for being by rand function generation length using random number seed seed
Bipolarity pseudorandom spreading sequence u, the f max1 and f min1 of (f max1-f min1+1) are the domain samples of embeddable watermark
The maximum value and minimum of point range.Bipolarity pseudorandom spreading sequence u is exported to coherent detection module 12.
Described divides odd-even interleaving frame module 11, for sequentially numbering signal frame in frequency-region signal SF, by odd indexed
Signal frame be combined into odd indexed signal frame SFodd, the signal frame of even number serial number is combined into even number serial number signal frame
SFeven.Odd indexed signal frame SFodd and even number serial number signal frame SFeven are exported to coherent detection module 12.
The coherent detection module 12 is used to be based on bipolarity pseudorandom spreading sequence u, to odd indexed signal frame
SFodd and even number serial number signal frame SFeven carry out coherent detection operation respectively, obtain odd decision statistic rodd and even decision system
Measure reven.Odd decision statistic rodd and idol decision statistic reven are exported and give intersection judging module 13.
The intersection judging module 13, for that must be determined according to odd decision statistic rodd and idol decision statistic reven
Plan statistic rall=rodd-reven, by the symbol extraction watermark for judging decision statistic rall.
Referring to Fig. 3, intersection additivity provided in this embodiment spreads audio frequency watermark embedding grammar, can use computer software
It is automatic to carry out, specifically includes the following steps:
Step 1, it reads audio file and obtains time-domain audio signal xt and its sample rate f s1, by time-domain audio signal xt framing
After carry out time-frequency domain conversation, obtain frequency-region signal xf.
To the framing of time-domain audio signal xt elder generation, frame length frame.Time-frequency domain conversation can use Fast Fourier Transform (FFT)
Method (FFT), discrete cosine transform method (DCT), multiple modulation lapped transform method (MCLT) etc..
Step 2, according to sample rate f s1, frame length frame and the audio frequency range of human ear sensitivity, embeddable water is obtained
The domain samples point range of print, to obtain the frequency domain signal X F of the domain samples point range；
The routine techniques for being calculated as watermark embedding technique field of the domain samples point range of embeddable watermark, for convenient for benefit
In detailed calculation method is provided below.
The maximum value f max1 and most that formula (1) and (2) calculate the domain samples point range of embeddable watermark is respectively adopted
Small value f min1:
F max1=fix ((FMAX*2.0/fs1) * frame) (1)
F min1=fix ((FMIN*2.0/fs1) * frame) (2)
In formula (1)~(2), FMAX, FMIN respectively indicate the highest frequency and most of the audio frequency range of human ear sensitivity
Low frequency, fix () are bracket function.
Step 3, bipolarity pseudorandom spreading sequence u is generated by rand function using random number seed seed.
This step realizes that detailed process is as follows using function in MATLAB:
Firstly, rand function is initialized and is generated random number seed seed using RandStream function.Then,
The random number sequence that length is (f max1-f min1+1) is generated using random number seed seed and rand function.Due to rand
The random number that function generates is the number between 0~1, is rounded up to random number each in random number sequence, obtains being made of 0 and 1
Binary sequence, 2 are multiplied to binary number each in binary sequence and subtracts 1, i.e., acquisition bipolarity pseudorandom spreading sequence
u。
Step 4, signal frame in frequency domain signal X F is sequentially numbered, the signal frame of odd indexed is combined into odd indexed letter
Number frame XFodd, is combined into even number serial number signal frame XFeven for the signal frame of even number serial number.
Detailed process is as follows for this step:
If frequency domain signal X F has 2N frequency samples point, frequency domain signal X F can be written as follow form:
XF=[xf1odd,xf1even,xf2odd,xf2even,...,xfNodd,xfNeven] (3)
Odd indexed signal frame XFodd is as follows:
XFodd=[xf1odd,xf2odd,...,xfNodd] (4)
Even number serial number signal frame XFeven is as follows:
XFeven=[xf1even,xf2even,...,xfNeven] (5)
xfiodd、xfievenRespectively indicate 2i-1 and 2i signal frame in frequency domain signal X F, i=1,2 ... N.
Step 5, according to bipolarity pseudorandom spreading sequence u, using additivity spread-spectrum watermark embedding inlay technique, to XFodd and
XFeven is respectively embedded into watermark bit symbol and -1*symbol, odd indexed signal frame YFodd and idol after must being embedded in watermark
Number sequence signal frame YFeven, symbol are the watermark bit that should be embedded in.
The specific implementation process of this step is as follows:
Formula (6)~(7) are respectively adopted, watermark bit symbol and -1*symbol is embedded in odd indexed signal frame
XFodd and even number serial number signal frame XFeven:
YFodd=XFodd+smybol*u (6)
YFeven=XFeven-smybol*u (7)
Step 6, odd indexed signal frame YFodd and even number serial number signal frame YFeven are combined into signal frame by serial number
YF is shown in formula (8).
YF=[yf1odd,yf1even,yf2odd,yf2even,...,yfNodd,yfNeven] (8)
yfioddIndicate i-th of signal frame in odd indexed signal frame YFodd, yfievenIndicate even number serial number signal frame
I-th of signal frame in YFeven, i=1,2 ... N.
Step 7, time-frequency domain inverse transformation is carried out to signal frame YF, obtain time-domain signal, and generate audio file, i.e. insertion watermark
Audio file.
Referring to fig. 4, intersection additivity provided in this embodiment spreads audio frequency watermark detection method, can use computer software
It is automatic to carry out, specifically includes the following steps:
Step 8, the audio file for reading insertion watermark, must be embedded in the time-domain audio signal rt and its sample rate of watermark
Fs2, time-frequency domain conversation will be carried out after time-domain audio signal rt framing, obtains frequency-region signal sf.
To the framing of time-domain audio signal rt elder generation, frame length is the same as frame length frame in step 1.Time-frequency domain conversation method is synchronous
Rapid 1.
Step 9, according to sample rate f s2, frame length frame and the audio frequency range of human ear sensitivity, embeddable water is obtained
The domain samples point range of print, to obtain the frequency-region signal SF of the domain samples point range.
The maximum value f max1 and minimum value f min1 of the domain samples point range of embeddable watermark using formula (1) and
(2) it calculates and obtains.
Step 10, bipolarity pseudorandom spreading sequence u is generated by rand function using random number seed seed.
This step realized using function in MATLAB, detailed process with step 3, gained bipolarity pseudorandom spreading sequence and
Step 3 gained bipolarity pseudorandom spreading sequence is identical.
Step 11, signal frame in frequency-region signal SF is sequentially numbered, the signal frame of odd indexed is combined into odd indexed
The signal frame of even number serial number is combined into even number serial number signal frame SFeven by signal frame SFodd.
This step is repeated no more with step 4, specific implementation process.
Step 12, it is based on bipolarity pseudorandom spreading sequence u, to odd indexed signal frame SFodd and even number serial number signal
Frame SFeven carries out coherent detection operation respectively, obtains corresponding surprise decision statistic rodd and even decision statistic reven.
The calculation formula of odd decision statistic rodd and even decision statistic reven are as follows:
In formula (9)~(10), < > indicates coherent detection；SFoddi、SFeveni、uiRespectively SFodd, SFeven, u
In i-th of data point.
Step 12, decision statistic rall is obtained using odd decision statistic rodd and idol decision statistic reven, by sentencing
Disconnected decision statistic rall symbol extraction watermark.
Decision statistic rall is obtained using formula (11):
Rall=rodd-reven (11)
If rall >=0, then the watermark received_bit=1 extracted, otherwise, received_bit=0.
Claims (6)
1. intersecting additivity spreads audio frequency watermark embedding grammar, characterized in that comprising steps of
S1 reads audio file and obtains time-domain audio signal xt and its sample rate f s1, will be transformed to after time-domain audio signal xt framing
Frequency-region signal xf；
S2 obtains the domain samples point of embeddable watermark according to the audio frequency range of sample rate f s1, frame length and human ear sensitivity
Range, to obtain the frequency domain signal X F of the domain samples point range；
S3 generates the bipolarity pseudorandom that length is (f max1-f min1+1) by rand function using random number seed seed
Frequency expansion sequence u, f max1 and f min1 are the maximum value and minimum value of the domain samples point range of embeddable watermark；
S4 sequentially numbers signal frame in frequency domain signal X F, and the signal frame of odd indexed is combined into odd indexed signal frame
The signal frame of even number serial number is combined into even number serial number signal frame XFeven by XFodd；
S5 is embedding to XFodd and XFeven difference using additivity spread-spectrum watermark embedding inlay technique according to bipolarity pseudorandom spreading sequence u
Enter watermark bit symbol and -1*symbol, odd indexed signal frame YFodd and even number serial number signal frame after watermark must be embedded in
YFeven, symbol are the watermark bit that should be embedded in；
Odd indexed signal frame YFodd and even number serial number signal frame YFeven are combined into signal frame YF by serial number by S6；
Signal frame YF is transformed to time-domain signal by S7, is generated audio file, that is, is embedded in the audio file of watermark.
2. intersecting additivity as described in claim 1 spreads audio frequency watermark embedding grammar, it is characterized in that:
S3 further comprises:
Rand function is initialized and is generated random number seed seed using RandStream function；
The random number sequence that length is (f max1-f min1+1) is generated using random number seed seed and rand function；
It rounds up respectively to random number each in random number sequence, obtains the binary sequence being made of 0 and 1；
2 are multiplied to binary number each in binary sequence and subtracts 1, obtains bipolarity pseudorandom spreading sequence u.
3. intersecting additivity spreads audio frequency watermark detection method, characterized in that comprising steps of
The audio file that S1 reads insertion watermark must be embedded in the time-domain audio signal rt and its sample rate f s2 of watermark, by time domain sound
Frequency-region signal sf is transformed into after frequency signal rt framing；
S2 obtains the domain samples point of embeddable watermark according to the audio frequency range of sample rate f s2, frame length and human ear sensitivity
Range, to obtain the frequency-region signal SF of the domain samples point range；
S3 generates the bipolarity pseudorandom that length is (f max1-f min1+1) by rand function using random number seed seed
Frequency expansion sequence u, f max1 and f min1 are the maximum and minimum value of the domain samples point range of embeddable watermark；
S4 sequentially numbers signal frame in frequency-region signal SF, and the signal frame of odd indexed is combined into odd indexed signal frame
The signal frame of even number serial number is combined into even number serial number signal frame SFeven by SFodd；
S5 is based on bipolarity pseudorandom spreading sequence u, to odd indexed signal frame SFodd and even number serial number signal frame SFeven points
Not carry out coherent detection operation, obtain odd decision statistic rodd and even decision statistic reven；
S6 obtains decision statistic rall=rodd-reven according to odd decision statistic rodd and idol decision statistic reven, leads to
Cross the symbol extraction watermark for judging decision statistic rall.
4. intersecting additivity as claimed in claim 3 spreads audio frequency watermark detection method, it is characterized in that:
S3 further comprises:
Rand function is initialized and is generated random number seed seed using RandStream function；
The random number sequence that length is (f max1-f min1+1) is generated using random number seed seed and rand function；
It rounds up respectively to random number each in random number sequence, obtains the binary sequence being made of 0 and 1；
2 are multiplied to binary number each in binary sequence and subtracts 1, obtains bipolarity pseudorandom spreading sequence u.
5. intersecting additivity spreads audio frequency watermark embedded system, characterized in that include:
Time-frequency domain conversation module obtains time-domain audio signal xt and its sample rate f s1 for reading audio file, time-domain audio is believed
Frequency-region signal xf is transformed to after number xt framing；
Frequency range selecting module, for according to the audio frequency range of sample rate f s1, frame length and human ear sensitivity, acquisition can
It is embedded in the domain samples point range of watermark, to obtain the frequency domain signal X F of the domain samples point range；
Frequency expansion sequence generation module, being used to generate length by rand function using random number seed seed is (f max1-f
Min1+1 bipolarity pseudorandom spreading sequence u, f max1 and f min1) be embeddable watermark domain samples point range most
Big value and minimum value；
Divide odd-even interleaving frame module, for sequentially numbering signal frame in frequency domain signal X F, combines the signal frame of odd indexed
At odd indexed signal frame XFodd, the signal frame of even number serial number is combined into even number serial number signal frame XFeven；
Watermark intersects insertion module, is used to according to bipolarity pseudorandom spreading sequence u, right using additivity spread-spectrum watermark embedding inlay technique
XFodd and XFeven is respectively embedded into watermark bit symbol and -1*symbol, the odd indexed signal frame after must being embedded in watermark
YFodd and even number serial number signal frame YFeven, symbol are the watermark bit that should be embedded in；
Odd-even interleaving frame composite module is used to odd indexed signal frame YFodd and even number serial number signal frame YFeven pressing serial number
It is combined into signal frame YF；
Time-frequency domain inverse transform module generates audio file, that is, is embedded in the sound of watermark for signal frame YF is transformed to time-domain signal
Frequency file.
6. intersecting additivity spreads audio frequency watermark detection system, characterized in that include:
Time-frequency domain conversation module must be embedded in the time-domain audio signal rt of watermark and its adopt for reading the audio file of insertion watermark
Sample rate fs2 will be transformed into frequency-region signal sf after time-domain audio signal rt framing；
Frequency range selecting module, for according to the audio frequency range of sample rate f s2, frame length and human ear sensitivity, acquisition can
It is embedded in the domain samples point range of watermark, to obtain the frequency-region signal SF of the domain samples point range；
Frequency expansion sequence generation module, being used to generate length by rand function using random number seed seed is (f max1-f
Min1+1 bipolarity pseudorandom spreading sequence u, f max1 and f min1) be embeddable watermark domain samples point range most
Big value and minimum；
Divide odd-even interleaving frame module, for sequentially numbering signal frame in frequency-region signal SF, combines the signal frame of odd indexed
At odd indexed signal frame SFodd, the signal frame of even number serial number is combined into even number serial number signal frame SFeven；
Coherent detection module is used to be based on bipolarity pseudorandom spreading sequence u, to odd indexed signal frame SFodd and even number sequence
Number signal frame SFeven carries out coherent detection operation respectively, obtains odd decision statistic rodd and even decision statistic reven；
Intersect judging module, for obtaining decision statistic rall according to odd decision statistic rodd and idol decision statistic reven
=rodd-reven, by the symbol extraction watermark for judging decision statistic rall.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510833111.5A CN105374360B (en) | 2015-11-25 | 2015-11-25 | Intersect additivity spread spectrum audio frequency watermark embedding grammar, detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510833111.5A CN105374360B (en) | 2015-11-25 | 2015-11-25 | Intersect additivity spread spectrum audio frequency watermark embedding grammar, detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105374360A CN105374360A (en) | 2016-03-02 |
| CN105374360B true CN105374360B (en) | 2018-12-14 |
Family
ID=55376490
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510833111.5A Active CN105374360B (en) | 2015-11-25 | 2015-11-25 | Intersect additivity spread spectrum audio frequency watermark embedding grammar, detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105374360B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106409302B (en) * | 2016-06-22 | 2019-07-09 | 华中师范大学 | Audio-frequency water mark method and system based on insertion regional choice |
| CN106599630A (en) * | 2016-12-12 | 2017-04-26 | 中国矿业大学 | Improved additive spread spectrum watermarking method |
| CN109036442A (en) * | 2018-07-27 | 2018-12-18 | 北京感动无限科技有限公司 | Watermark recognition methods and device for broadcasting content |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1898396A1 (en) * | 2006-09-07 | 2008-03-12 | Deutsche Thomson-Brandt Gmbh | Method and apparatus for encoding/decoding symbols carrying payload data for watermarking of an audio or video signal |
| CN101185121A (en) * | 2005-06-02 | 2008-05-21 | 汤姆森许可贸易公司 | Method and apparatus for watermarking an audio or video signal with watermark data using a spread spectrum |
| CN101290773A (en) * | 2008-06-13 | 2008-10-22 | 清华大学 | Adaptive MP3 digital watermark embedding method |
| CN103957099A (en) * | 2014-05-16 | 2014-07-30 | 武汉大学 | Media encryption and decryption method combined with watermark |
| CN104217725A (en) * | 2014-09-29 | 2014-12-17 | 北京理工大学 | Audio watermarking method based on multi-echo core |
| CN104320719A (en) * | 2014-11-14 | 2015-01-28 | 武汉大学 | Television program interaction participating method and system based on audio watermarking |
| CN104658542A (en) * | 2015-03-16 | 2015-05-27 | 武汉大学 | Additive spread spectrum audio watermarking embedding method, additive spread spectrum audio watermarking detection method and additive spread spectrum audio watermarking embedding system based on orthogonality |
-
2015
- 2015-11-25 CN CN201510833111.5A patent/CN105374360B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101185121A (en) * | 2005-06-02 | 2008-05-21 | 汤姆森许可贸易公司 | Method and apparatus for watermarking an audio or video signal with watermark data using a spread spectrum |
| EP1898396A1 (en) * | 2006-09-07 | 2008-03-12 | Deutsche Thomson-Brandt Gmbh | Method and apparatus for encoding/decoding symbols carrying payload data for watermarking of an audio or video signal |
| CN101290773A (en) * | 2008-06-13 | 2008-10-22 | 清华大学 | Adaptive MP3 digital watermark embedding method |
| CN103957099A (en) * | 2014-05-16 | 2014-07-30 | 武汉大学 | Media encryption and decryption method combined with watermark |
| CN104217725A (en) * | 2014-09-29 | 2014-12-17 | 北京理工大学 | Audio watermarking method based on multi-echo core |
| CN104320719A (en) * | 2014-11-14 | 2015-01-28 | 武汉大学 | Television program interaction participating method and system based on audio watermarking |
| CN104658542A (en) * | 2015-03-16 | 2015-05-27 | 武汉大学 | Additive spread spectrum audio watermarking embedding method, additive spread spectrum audio watermarking detection method and additive spread spectrum audio watermarking embedding system based on orthogonality |
Non-Patent Citations (1)
| Title |
|---|
| "Improved multiplicative spread spectrum embedding for image data hiding";A Valizadeh，ZJ Wang;《IEEE International Conference on Image Processing》;20111231;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105374360A (en) | 2016-03-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104658542B (en) | Based on orthogonal additivity spread spectrum audio frequency watermark embedding grammar, detection method and system | |
| CN101101754B (en) | Steady audio-frequency water mark method based on Fourier discrete logarithmic coordinate transformation | |
| CN107461611B (en) | The leakage detection method and leak detecting device combined is reconstructed based on small echo and EMD | |
| CN105374360B (en) | Intersect additivity spread spectrum audio frequency watermark embedding grammar, detection method and system | |
| CN102157154B (en) | Audio-content-based non-uniform discrete cosine transform audio reliability authentication method | |
| CN113782041B (en) | Method for embedding and positioning watermark based on audio variable frequency domain | |
| CN110163787B (en) | Audio digital robust blind watermark embedding method based on dual-tree complex wavelet transform | |
| CN105976823A (en) | Adaptive audio watermarking method based on phase coding and system | |
| CN102664013A (en) | Audio digital watermark method of discrete cosine transform domain based on energy selection | |
| CN103455972A (en) | Multiple-bit spread spectrum watermark method with robustness enhanced by utilization of carrier | |
| CN105632506A (en) | Robust digital audio watermark embedding and detection method based on polar harmonic transform | |
| Maha et al. | DCT based blind audio watermarking scheme | |
| Lu et al. | RFPS: a robust feature points detection of audio watermarking for against desynchronization attacks in cyber security | |
| CN106409302B (en) | Audio-frequency water mark method and system based on insertion regional choice | |
| CN109377982B (en) | Effective voice obtaining method | |
| Wu et al. | Audio watermarking algorithm with a synchronization mechanism based on spectrum distribution | |
| Li et al. | Design of a general hardware in the loop underwater communication emulation system | |
| CN106205627B (en) | Digital audio reversible water mark algorithm based on side information prediction and histogram translation | |
| CN105895109B (en) | A kind of digital speech evidence obtaining and tamper recovery method based on DWT and DCT | |
| KR20070061285A (en) | Digital audio watermarking method using hybrid transform | |
| CN103354091B (en) | Based on audio feature extraction methods and the device of frequency domain conversion | |
| Zhao et al. | A robust audio sonic watermarking algorithm oriented air channel | |
| CN103685120B (en) | A kind of resynchronization method in LTE system and system | |
| CN101676992A (en) | Method for analyzing echo steganography | |
| CN108877819A (en) | A kind of voice content evidence collecting method based on coefficient correlation |
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 |