KR20020031654A - Method and apparatus for embedding watermarks using fast fourier transformed data - Google Patents

Abstract

PURPOSE: A method and a system for inserting and extracting a watermark using Fourier transform are provided to additionally insert lots of information into a digital audio signal to offer various information items to a listener. CONSTITUTION: A digital audio watermark inserting system includes a watermark generator, a frequency converter(712), a critical frequency measuring unit(713), and a watermark summing unit(714). The watermark generator converts a watermark to be inserted into a digital audio signal into a bit stream. The frequency converter Fourier-transforms the digital audio signal. The critical frequency measuring unit calculates a critical frequency of the digital audio spectrum transformed by the frequency converter. The watermark summing unit calculates an intermediate value for a frequency band higher than the critical frequency, and controls the spectrum value upward or downward on the basis of the intermediate value according as the data value of the watermark bit stream is "0" or "1".

Description

Watermark Insertion and Extraction Method using Fourier Transform {METHOD AND APPARATUS FOR EMBEDDING WATERMARKS USING FAST FOURIER TRANSFORMED DATA}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and method for embedding and extracting a watermark into a digital audio signal. In particular, the present invention relates to a watermark after performing a frequency transformation through Fourier transform of digital audio data into a frequency domain. It is about a technique to insert.

Recently, with the development of digital signal technology and network technology, a world that easily receives multimedia information is coming. In particular, since digital audio is compressed by a technology such as MPEG and easily propagated through the Internet, cases of illegal distribution on a network may frequently occur.

As described above, in order to protect copyright of digital audio, a technique of embedding a watermark in the data or encrypting the data using a public key algorithm or the like is used. The watermarking method of digital audio can be classified into four types such as a lower bit encoding method, a phase encoding method, a spread spectrum method, and an echo concealment method.

The lower bit manipulation technique according to the prior art replaces a relatively insignificant bit stream with a watermark at a sampling point of digital audio. For example, when the sampling rate is 8 KHz, the watermark information may be up to 8 Kbit / sec.

In addition, the phase encoding technique is a technique for changing the phase of digital audio data, and is a technique for inserting watermarking into the phase data by using a sensitivity that is generally sensitive to sound volume but insensitive to a change in phase. That is, the phase coding technique is a technique of converting music data in the time domain into data in the frequency domain, inserting and changing the watermarking, and then changing the time data back to the time domain.

However, phase changes are very sensitive and can easily be lost to lossy compression. On the other hand, the spread spectrum technique according to the prior art is a technique for widening and spreading the watermark information so as not to be biased anywhere, and is more resistant to lossy compression, resampling, and other data manipulation than the conventional frequency domain method.

However, the spread spectrum technique has a limitation in that the watermark must be weakened to a size within the 0.5% range of the audio signal so as to maintain the noise level and not be heard. Meanwhile, the echo concealment technique according to the related art is a technique capable of inserting a strong watermark without losing a signal of digital audio, and may hide information using an echo. However, the echo concealment technique has the disadvantage that the mark can be easily broken in the third region, such as echo addition or linear prediction cepstrum.

Accordingly, it is a first object of the present invention to provide a watermarking method and apparatus applicable to digital audio.

It is a second object of the present invention to provide a watermarking method and apparatus which can provide various information to a listener by additionally inserting a lot of information into digital audio in addition to the first object.

It is a third object of the present invention to provide a method and apparatus capable of performing watermarking without a digital audio source in addition to the first object.

A fourth object of the present invention is to provide a method and apparatus capable of performing watermarking in the frequency domain in addition to the first object.

1 is a diagram showing the configuration of a digital audio watermark embedding and extraction system according to the present invention;

2 is a diagram showing the configuration of a watermark generating apparatus according to the present invention;

3 is a diagram showing the configuration of a watermark embedding apparatus according to the present invention;

4A and 4B show the results of performing a Fast Fourier Transform of a digital audio signal into a spectrum in the frequency domain through a frequency converter in accordance with the present invention.

Fig. 5 is a diagram comparing the distribution of the frequency spectrum in the case of large and small kutosis values.

6 is a view showing music in which a watermark is inserted in a frequency domain according to the present invention;

7 is a diagram showing the configuration of a watermark extraction apparatus according to the present invention;

<Explanation of symbols for the main parts of the drawings>

50: Watermark information

51: Watermark data

100: watermark generator

200: watermark insertion device

400: watermark extraction device

710: bit conversion unit

711: intensity control unit

712: frequency converter

713: critical frequency measurement unit

714: Watermark totaling department

715: inverse frequency converter

716,716 ': An example of the spectrum of a fast Fourier transformed digital audio

717,717 ': Example spectrum of fast Fourier transformed digital audio

718, 719: Area to insert watermark

720: middle value measuring unit

721: bit information converter

780: Middle value

783: Watermark insertion area

In order to achieve the above object, the present invention uses a fast Fourier transform (FFT) as a technique for inserting a watermark after performing a frequency conversion through the Fourier Transform (Digital Audio data) The speed of watermarking is fast and a large amount of information can be inserted into the digital audio data by using a critical frequency meter and an auditory psychological model.

The present invention makes it possible to insert a large amount of information, thereby enabling the insertion of various additional information as well as a detailed description of the copyright holder. In other words, the maximum frequency that an average person can listen sensitively is about 6 KHz. Given that digital audio has a frequency bandwidth of 44.1 KHz, it is possible to watermark about 38.1 Kbits of data per second.

In the present invention, a threshold frequency meter is used to insert a watermark, and it can be used not only for music copyright information but also for song lyrics information or for data communication.

Hereinafter, a watermarking technique according to the present invention will be described in detail with reference to FIGS. 1 to 7.

1 is a diagram showing the configuration of a digital audio watermark embedding and extraction system according to the present invention. Referring to FIG. 1, the watermark information 50 is converted into a bit stream signal W 51 through the watermark generating apparatus 100. The converted watermark bit signal (W) 51 is inserted into the original audio OM (x) digital audio signal 40 through the watermark embedding apparatus 200, and the music 210 WM (x) having the watermark embedded therein is The watermark extraction apparatus 400 performs watermark authentication in real time on the embedded audio player. At this time, music is not played in the audio player of the illegal user, and the music can be played only when legitimate user authentication information 51 is extracted.

2 is a diagram showing the configuration of a watermark generating apparatus 100 according to the present invention. Referring to FIG. 2, the watermark information 50 is transmitted through the bit converter 710, for example, 10101110. It is converted into a binary bit stream and the like, and is input to the intensity controller 711. The strength adjusting unit 711 according to the present invention adjusts the strength of the binary bit stream watermark information according to the value S generated by the threshold frequency measuring unit 713.

3 is a diagram showing the configuration of a watermark embedding apparatus 200 according to the present invention. Referring to FIG. 3, the watermark embedding apparatus 200 may include a frequency converter 712, a threshold frequency measurer 713, a watermark adder 714, a frequency inverse converter 715, and the like. The frequency converter 712 according to the present invention receives the original music 40 OM (x), and is converted into frequency components through a Fourier transform.

In a preferred embodiment of the present invention, the frequency converter 712 may be implemented as a fast Fourier transform (FFT) or a fast Fourier transform program module. For example, for CD music with a band bandwidth of 44.1 KHz, fast Fourier transforms on 2048-bit data results in symmetric frequency spectrum.

4A and 4B illustrate a result of performing fast Fourier transform of a digital audio signal into a spectrum of a frequency domain through a frequency converter according to the present invention. Referring to FIG. 4A, the original music 40 data is distributed over a wide frequency band 716 and 716 ', in which case the water inserted when the watermark is inserted into the rectangular area 718 shown in FIG. 4A. The mark amount is small while the watermark intensity is large.

On the other hand, referring to Fig. 4B, since the original music data is distributed over a relatively narrow frequency band 717, 717 ', even if the watermark is inserted into the band 719 with a low frequency component, the listener inserts the watermark. It is more likely to recognize that it is. Therefore, in this case, although the amount of watermark to be inserted is large, the intensity of the watermark is inserted through the intensity adjusting unit 711 to be inserted. At this time, the intensity control adjusted by the intensity control unit 711 may be determined by the following equation.

if F (t) ≤ Threshold_Freq 1, Scale = α1

if Threshold_Freq 1 <F (t) ≤ Threshold_Freq 2, Scale = α2

if Threshold_Freq 2 <F (t), Scale = α3

Here, F (t) refers to the frequency measured by the threshold frequency measuring unit 713, and Scale refers to the intensity of the watermark. As a preferred embodiment of the present invention, the scale intensity can set any value.

F (t) designates the corner frequency of the portion where the frequency component spectral values start to become significantly smaller in order to define the rectangular regions 718 and 719 shown in Figs. 4A and 4B. At this time, the threshold frequency measuring unit 713 may use the auditory psychological model for detecting the portion where the frequency magnitude is small. As a preferred embodiment according to the present invention, it is possible to calculate F (t) on the basis of 10 KHz or more, which is not easy for human recognition.

The threshold frequency measuring unit 713 according to the present invention may measure kutosis to calculate the watermark embedding regions 718 and 719. Mathematically, the larger the value of the Kutosis, the more the data distribution is concentrated in the center. If the value is zero or less, the distribution is smooth.

FIG. 5 is a diagram showing a comparison of distributions of frequency spectra in the case of large and small kutosis values. FIG. Referring to FIG. 5, since the larger the data distribution is centralized, the higher the values of the kuthosis, the higher the frequency values are smaller than the low frequency values, so that the listeners can easily recognize the small changes. Therefore, in this case, the watermarks 718 and 719 need to be moved to a higher frequency side, so that the watermark can be inserted into an area that the listener cannot recognize.

On the contrary, since the smaller the kutosis value means that the value of the high frequency component and the low frequency component are not large, it is not easy for the listener to sense a small change in the high frequency component. Therefore, in this case, a large amount of watermark information can be inserted by setting a threshold frequency toward a lower frequency side than in the previous case.

On the other hand, the calculation formula for obtaining the Kutosis is as follows.

here, And N represents the number of data.

According to an embodiment of the present invention, when sampling 2,048 bits of audio data having a bandwidth of 44.1 KHz, a total of about 1,024 bits of the total frequency corresponding to a frequency of 7 KHz to 22.05 KHz, which is calculated from the Kutosis equation, is approximately 2 bits. Watermark data can be inserted in the / 3 area (approximately 650 bits).

The watermark embedding technology according to the present invention enables to insert not only simple information for copyright protection of copyright holders, but also various kinds of other information including song lyrics in digital audio under the level that the listener cannot detect in the frequency domain. Let's do it.

Referring back to FIG. 3, the signal output from the threshold frequency measuring device 713 is input to the watermark adding unit 714, which is added with the watermark bit stream, and is time domaind by the frequency inverse transform unit 715. Is converted to digital audio. The feature of the watermark embedding apparatus according to the present invention is that when the watermark is added to the original signal, the watermark embedding apparatus is inserted after preprocessing all the frequency component spectrum values of the watermark embedding regions 718 and 719 into an average value. Is that you lose. As described above, in the process of summing the watermark bit streams by preprocessing the intermediate values, information is loaded by raising or lowering the spectral value according to whether the value is "0" or "1".

6 is a view showing music in which a watermark is inserted in a frequency domain according to the present invention. Referring to Fig. 6, the watermark embedding area 783 is defined through the above-described kuthosis calculation and smoothed to the intermediate value 780 through the line processing process. Subsequently, according to the data value of the watermark data 51 coded into the bit stream, if the value is "1", the spectral value is raised 781, and if it is "0", it is downward 778. By doing so, binary information is loaded.

7 is a diagram showing the configuration of a watermark extraction apparatus 400 according to the present invention. Referring to FIG. 7, the watermark extractor 400 may include a frequency converter 712, a threshold frequency measurer 713, an intermediate value measurer 720, a bit information converter 721, and the like. When the digital audio 210 WM (x) having the watermark inserted therein is inputted, it is converted into a frequency component spectrum through the frequency converter 712 and then the position where the watermark is inserted through the threshold frequency measurement unit 713. Figure out.

Subsequently, the median measuring unit 720 calculates the median value 780 for the area 783 in which the watermark is inserted, and considers the signal "1" 781 if it is larger than the value for each data point. If small, the bit signal 782 of "0" is extracted. The extracted bit signal is converted into copyright information through a bit information converter.

The foregoing has outlined rather broadly the features and technical advantages of the present invention to better understand the claims of the invention which will be described later. Additional features and advantages that make up the claims of the present invention will be described below. It should be appreciated by those skilled in the art that the conception and specific embodiments of the disclosed subject matter can be used immediately as a basis for designing or modifying other structures for carrying out similar purposes to the present invention.

In addition, the inventive concepts and embodiments disclosed herein may be used by those skilled in the art as a basis for modifying or designing other structures for carrying out the same purposes of the present invention. In addition, such modifications or altered equivalent structures by those skilled in the art may be variously changed, substituted, and changed without departing from the spirit or scope of the invention described in the claims.

As described above, according to the present invention, a large amount of watermark information can be inserted by inserting a watermark into a frequency region insensitive to listening to the FFT converted frequency spectrum of the digital audio signal.

As a result, it is possible to prevent the reproduction of illegally distributed digital contents and to insert not only copyright information but also additional information such as song lyrics.

Claims (10)

Generating a watermark to be inserted into the digital audio signal as a bit stream; Fourier transforming the digital audio signal into a frequency domain; Calculating a threshold frequency with respect to a frequency domain spectrum of the Fourier transformed digital audio signal; And Inserting the watermark bit stream in a frequency domain for a band having a frequency greater than the threshold frequency Digital audio watermark embedding method comprising a. The digital audio watermark embedding method of claim 1, wherein the measuring of the threshold frequency sets a threshold frequency by calculating kutosis for a Fourier transform spectrum of the digital audio signal. The method of claim 1, wherein the embedding of the watermark bit stream in a frequency band greater than the threshold frequency is performed. Adjusting the strength of the watermark bit stream Digital audio watermark embedding method further comprising. The method of claim 1, wherein the inserting of the watermark bit stream in a frequency region greater than the threshold frequency is performed. Measuring an intermediate value for a frequency region greater than the threshold frequency; For a frequency domain larger than the threshold frequency, the Fourier transform value at each data point is greater or less than the median, depending on whether the data value of the watermark bit stream is "1" or "0". Inserting the watermark bit stream by setting to a value Digital audio watermark embedding method comprising a. The method of claim 1, wherein the digital audio watermark embedding method Fourier inverse transforming the watermarked frequency spectrum subsequently after inserting the watermark bit stream in a frequency domain Digital audio watermark embedding method comprising a. Fourier transforming the watermark-embedded digital audio signal into a frequency domain; Calculating a threshold frequency with respect to a frequency domain spectrum of the Fourier transformed digital audio signal; Measuring an intermediate value for a frequency band greater than the threshold frequency; Extracting watermark bitstream information by determining a zero crossing based on the intermediate value at each data pointer for a frequency band larger than the threshold frequency; Digital audio watermark extraction method comprising a. A watermark generator for converting the watermark to be inserted into the digital audio signal into a bit stream in binary form; A frequency converter for Fourier transforming the digital audio signal; A threshold frequency measuring unit for calculating a threshold frequency for the spectrum of digital audio converted by the frequency converter; And The watermark is calculated by calculating an intermediate value for a frequency band greater than the threshold frequency and adjusting the spectrum value up or down based on the intermediate value according to whether the data value of the watermark bit stream is "0" or "1". Watermark adder to insert Digital audio watermark insertion device comprising a. The digital audio watermark embedding apparatus according to claim 7, wherein the watermark generating unit further comprises a strength adjusting unit for adjusting the strength of the watermark bit stream. 8. The digital audio watermark embedding apparatus of claim 7, wherein the digital audio watermark embedding apparatus comprises an inverse frequency converting unit for performing Fourier inverse transform of the digital audio spectrum into which the watermark is inserted following the watermark adding unit. A frequency converter for performing Fourier transform on the watermark-embedded digital audio signal; A threshold frequency measuring unit for calculating a threshold frequency for the spectrum of digital audio converted by the frequency converter; An intermediate value measuring unit for calculating an intermediate value for a frequency region greater than the threshold frequency; And A bit information converter configured to extract watermark bitstream information by determining zero crossing for a frequency band greater than the threshold frequency based on the intermediate value calculated by the intermediate value measurer. Digital audio watermark extraction apparatus comprising a.