JP4028676B2 - Acoustic signal transmission apparatus and acoustic signal transmission method, and data extraction apparatus and data extraction method for extracting embedded data of an acoustic signal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily and safely embed a large amount of information without entailing any sense of imcompatibility. SOLUTION: Digital sound signals of L and R channels are divided into frames of constant time length and the sound signal of the L channel is outputted as it is. According to the embedded data, an impulse response function is generated (8). Convolutional operation (10) between the L-channel sound signal and the impulse response function generated from the embedded data is carried out and the result is added to the R-channel sound signal (12) to generate an R-channel sound signal having the embedded data. The frames of the L and R channels are each put together and sent.

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an acoustic signal transmission device and an acoustic signal transmission method, and a data extraction device and a data extraction method for extracting embedded data of an acoustic signal.
[0002]
More specifically, the present invention relates to an acoustic signal transmission apparatus and an acoustic signal transmission method for transmitting specific information data embedded in a two-channel stereo digital acoustic signal. Furthermore, the present invention relates to a data extraction apparatus and a data extraction method for inputting a first channel digital acoustic signal and a second channel digital acoustic signal in which specific information data is embedded and extracting the specific information data. It is.
[0003]
[Prior art]
The following techniques are known as conventional techniques for transmitting predetermined information embedded in a digital acoustic signal.
[0004]
(1) A method of embedding predetermined information in lower bits of digital audio data (see Japanese Patent Laid-Open No. 9-214636).
[0005]
(2) A method of embedding predetermined information in a phase component after frequency conversion of an acoustic signal (see Nikkei Electronics 1997 No. 684, “Data Hiding Technology that Supports Digital Watermarking”).
[0006]
(3) A method of embedding predetermined information in an acoustic signal that has been spectrum-spread and frequency-converted (see Nikkei Electronics 1997 No. 684, “Data Hiding Technology that Supports Digital Watermarking”).
[0007]
(4) A method of embedding predetermined information as an echo by changing the time delay parameter of the impulse response (see Nikkei Electronics 1997 No. 684, “Data Hiding Technology that Supports Digital Watermarking”).
[0008]
[Problems to be solved by the invention]
However, the conventional techniques described in the above items (1) to (4) have the following problems.
[0009]
The prior art of item (1) has the disadvantage that it is easy to specify the position where the information is embedded, so that the resistance to manipulation such as tampering is weak and it is easily affected by noise in the transmission path.
[0010]
In the prior art of item (2), the amount of distortion caused by changing the phase depends on the phase of the original acoustic signal, so that it is difficult to embed predetermined information in an audible manner. Is included.
[0011]
In the prior art of item (3), when embedding information is extracted on the receiving side, it is difficult to specify the position where the data corresponding to the predetermined information is embedded and to match the reference timing, and this is possible. Even in this case, there is a disadvantage that the apparatus becomes large-scale.
[0012]
In the prior art of item (4), only one bit is embedded in one frame (a part of the original signal divided for performing a series of data embedding processes, which usually has a length of several seconds to several tens of seconds). Since it cannot be performed, it has a drawback that the amount of data to be embedded cannot be increased.
[0013]
Therefore, in view of the above-described points, an object of the present invention is to provide an acoustic signal transmission device and an acoustic signal transmission method capable of easily and safely embedding a large amount of information without causing a sense of incongruity in hearing, and An object of the present invention is to provide a data extraction apparatus and a data extraction method for extracting embedded data of an acoustic signal.
[0014]
More specifically, an object of the present invention is to solve all the problems listed below at the same time.
(1) Excellent resistance to noise contamination and operation due to transmission.
(2) A difference in audibility is unlikely to occur between the acoustic signal in which the data is embedded and the original acoustic signal.
(3) The specific information can be embedded and the information can be extracted by a relatively simple process.
(4) The amount of data that can be embedded is large.
(5) The original voice is not required when extracting the embedded data.
[0015]
[Means for Solving the Problems]
In order to achieve the above object, the present invention according to claim 1 is an acoustic signal transmission apparatus that embeds specific information data in a two-channel stereo digital acoustic signal and sends it out, and the first channel digital acoustic signal is inputted. An embedding unit that embeds specific information data in a signal, and a synthesis that forms a new second channel digital acoustic signal by synthesizing the input second channel digital acoustic signal and the digital acoustic signal output from the embedding unit. Means, and an output means for sending out the input first channel digital sound signal and the new second channel digital sound signal as a two-channel stereo digital sound signal.
[0016]
According to a second aspect of the present invention, in the acoustic signal transmission device according to the first aspect, the embedding means creates an impulse response function based on the specific information data and predetermined key data, and Convolution operation means for creating a digital sound signal in which the specific information data is embedded by performing a convolution operation on the input digital acoustic signal of the first channel and the impulse response function.
[0017]
The present invention according to claim 3 includes a first step of embedding the specific information data in the input digital audio signal of the first channel when the specific information data is embedded in the 2-channel stereo digital audio signal and transmitted. A second step of synthesizing the digital audio signal of the second channel and the digital audio signal obtained in the first step to form a new digital audio signal of the second channel; and the input digital of the first channel An acoustic signal transmission method comprising: a third step of transmitting an acoustic signal and the new second channel digital acoustic signal as a two-channel stereo digital acoustic signal.
[0018]
According to a fourth aspect of the present invention, in the acoustic signal transmission method according to the third aspect, the first step of embedding specific information data in the input digital audio signal of the first channel includes the specific information data and a predetermined key. An impulse response creating step for creating an impulse response function based on the data, and a digital acoustic signal in which the specific information data is embedded by performing a convolution operation on the input digital acoustic signal of the first channel and the impulse response function A convolution calculation step to be created.
[0019]
According to a fifth aspect of the present invention, there is provided a data extracting apparatus for inputting the first channel digital acoustic signal and the second channel digital acoustic signal in which the specific information data is embedded, and extracting the specific information data, The first channel digital audio signal and the second channel digital audio signal are separately determined in advance while synchronizing the 1 channel digital audio signal and the second channel digital audio signal in which the specific information data is embedded. Separating means for sequentially dividing the data sequence into time lengths, and the second channel digital acoustic signal in which the first channel digital acoustic signal and the specific information data corresponding to the first channel digital acoustic signal are embedded, An impulse response function corresponding to the specific information data is extracted using a cross spectrum method. To, those provided with the extraction means for inverse transforming the impulse response function with the extraction based on the correspondence information of the specific information data and said impulse response function at the time of embedding specific information data to said specific information data.
[0020]
According to a sixth aspect of the present invention, when the first channel digital sound signal and the second channel digital sound signal embedded with the specific information data are input and the specific information data is extracted, the first channel digital sound signal is extracted. The first channel digital audio signal and the second channel digital audio signal have a predetermined length of time separately in synchronization with the second channel digital audio signal in which the specific information data is embedded. A first step of sequentially dividing the data sequence; and a cross-spectral method for the second channel digital acoustic signal in which the first channel digital acoustic signal and the specific information data corresponding to the first channel digital acoustic signal are embedded The impulse response function corresponding to the specific information data is extracted using A data extraction method and a second step of inversely transforming the impulse response function with the extraction based on the correspondence information of the specific information data and said impulse response function at the time of embedding specific information data to said specific information data.
[0021]
DETAILED DESCRIPTION OF THE INVENTION
Outline of the embodiment In this embodiment, an impulse response function generated in accordance with "embedded data" is generated from information (hereinafter referred to as "embedded data") necessary for acoustic signal identification and copy protection. It is embedded in the acoustic signal by convolution processing with the acoustic signal so as not to cause a change in audibility. The “embedded data” can be extracted from an acoustic signal in which the data is embedded by using a cross spectrum method, which will be described later, and can be used as digital watermark or database index information.
[0022]
In other words, the present embodiment provides a technique for embedding information necessary for identification and copy protection of a sound signal in a digital sound signal and a specific technique for extracting “embedded data” from an embedded sound signal with high accuracy. Disclosure.
[0023]
Detailed description of the embodiments <br/> below with reference to the accompanying drawings, this embodiment will be described in detail.
[0024]
FIG. 1 shows a block configuration of an “embedded data” transmission system to which the present invention is applied. In this figure, 4 is an “embedded data” input unit, 6 is a key information input unit, 8 is an impulse response creation unit, 10 is a convolution operation unit, 12 is a synthesis unit, and 16 is a storage program for embedding processing of “embedded data” It is a control part controlled according to.
[0025]
FIG. 2 is a flowchart showing the “embedded data” embedding processing procedure executed by each block shown in FIG.
[0026]
The process on the embedding side will be described with reference to FIG. 1 and FIG.
[0027]
(Processing in step S1)
(A) The L and R channel digital audio signals are set at a fixed time length (for example, set so that a sufficient pulse train can be estimated at the time of extraction) by an embedding frame dividing unit (not shown). Divide sequentially into frames.
(B) “Embedded data” and key information are input to the impulse response creating unit 8 to create an impulse response function. This impulse response function is composed of pulses corresponding to “embedded data” having a predetermined amplitude and a predetermined random number sequence with a predetermined seed number with respect to the first frame as a reference. . This seed number is stored as key information for use during extraction.
(C) The convolution calculation unit 10 performs a convolution operation on the impulse response function created by encoding “embedded data” and the L-channel acoustic signal for each frame divided in (a) above, and embeds the data. Create an acoustic signal.
(D) The processes (a) to (c) are performed for each frame.
[0028]
(Process in step S2)
The embedded acoustic signal created for each frame is sequentially synthesized in the time axis direction by the synthesis unit 12 to create an acoustic signal in which data is embedded.
[0029]
The process of embedding “embedded data” may be repeated to improve the probability of extracting “embedded data” and to improve the digital watermark resistance against unauthorized use of only a part of the acoustic signal. desirable.
[0030]
(Processing in step S3)
The L channel sound signal is transmitted as it is as the L channel sound signal, and the sound signal in which the “embedded data” is embedded is transmitted as the R channel sound signal.
[0031]
FIG. 3 shows a block configuration of an “embedded data” extraction system. In this figure, 20 is a synchronization unit, 22 and 24 are frame division units, 26 is an extraction unit, and 28 is a control unit that controls extraction processing of “embedded data” according to a storage program.
[0032]
FIG. 4 is a flowchart showing the “embedded data” extraction processing procedure executed by each block shown in FIG.
[0033]
With reference to FIG. 3 and FIG. 4, processing on the extraction side will be described.
[0034]
(Processing in step S11)
An L channel acoustic signal and an R channel acoustic signal in which “embedded data” is embedded are input.
[0035]
(Processing in Step S12)
The synchronization unit 20 obtains the cross-correlation between the R-channel embedded sound signal and the L-channel sound signal, compares the waveforms, takes a rough timing, and further examines the result of changing the time difference of the cross-correlation in detail. Find the timing that maximizes the cross-correlation.
[0036]
(Processing in step S13)
Based on the frame division information at the time of data embedding, the frame dividing units 22 and 24 set the R channel so as to be the same as the frame divided on the embedding side while synchronizing both signals based on the timing found in step S12. The embedded acoustic signal and the L channel acoustic signal are divided into frames.
[0037]
(Processing in step S14)
Using a known cross spectrum method, the extraction unit 26 extracts a composite pulse train (“embedded data”) in the frame. The processing of the cross spectrum method will be described in detail below.
[0038]
Cross-spectrum method: An L channel acoustic signal x (n) in a frame and an R channel embedded acoustic signal y (n) obtained by convolving a synthetic pulse train (“embedded data”) d (n) with this signal. By using it, the synthetic pulse train is estimated statistically.
[0039]
[Expression 1]
y (n) = d (n) * x (n) (1)
x _i (n) = x ((i-1) L + n), n = 1,2, ..., M (2)
y _i (n) = y ((i-1) L + n), n = 1,2, ..., M (3)
Xi (k) = FFT [w (n) x _i (n)] (4)
Yi (k) = FFT [w (n) y _i (n)] (5)
Sxx (k) = Σ Xi (k) Xi (k) (6)
Sxy (k) = Σ Xi (k) Yi (k) (7)
h (n) = IFFT [Sxy (k) / Sxx (k)] (8)
In the above equation, * represents a convolution integral, and _ represents a complex conjugate.
[0040]
Define block signals x (n) and y (n) consisting of 1 block M points (block length M) in the frame, and 2 of the i-th block obtained by shifting the block signals L points from the beginning of the frame. Assume that two signals are x _i (n) and y _i (n) as shown in equations (2) and (3). The i-th block signal is subjected to weighting w (n) such as a Hamming window as shown in equations (4) and (5) in order to reduce the wraparound phenomenon at both ends of the block.
[0041]
Sxx and Sxy represent the average power spectrum of the L channel acoustic signal and the average cross power spectrum of the L channel acoustic signal and the R channel embedded acoustic signal, respectively. In equation (8), h (n) is an estimated pulse train that approximates the original synthesized pulse train d (n). When the average number of times is P, the number of data necessary for the estimation of the equation (8) is ((P-1) L + M) points.
[0042]
The detection of “embedded data” is based on the time of the first frame of the extracted impulse response acoustic signal, and the delay time corresponding to the “embedded data” set corresponding to the uniform random number with a known seed number This is done by examining whether there are any other pulses.
[0043]
The processing on the embedding side and the extracting side described above is summarized as follows.
[0044]
Processing on the embedding side:
Divide the L and R channel digital audio signals into frames of a certain length of time.
[0045]
・ The sound signal of L channel is output as it is.
[0046]
Create an impulse response function based on “embedded data”.
[0047]
-Convolution of the L channel acoustic signal and the impulse response function created from "embedded data", and by adding and synthesizing this with the R channel acoustic signal, the R channel acoustic signal in which "embedded data" is embedded is created. To do.
[0048]
・ Synthesize the L and R channel frames.
[0049]
Processing on the extraction side:
The L channel sound signal and the R channel sound signal are synchronized and divided into the same frame as that at the time of embedding using the frame division information at the time of embedding.
[0050]
The respective impulse response functions are estimated using a known cross spectrum method for the L channel acoustic signal and the corresponding R channel acoustic signal frame, and “embedded data” is extracted.
[0051]
Effects according to the embodiment According to the present embodiment, the following special effects can be obtained.
[0052]
(I) By constructing a composite pulse train based on uniform random numbers, the “embedded data” can be concealed, and at the same time, amplitude peaks and dips occurring at specific frequencies can be reduced. it can.
[0053]
(Ii) The L channel acoustic signal and the created impulse response are convoluted to generate an embedded acoustic signal, which is added to the R channel acoustic signal and synthesized, and the R channel acoustic signal and the L channel acoustic signal are heard simultaneously. In this case, it is possible to make it difficult to produce a difference in audibility with the original sound signal. In addition, since human auditory sense has integral characteristics in the time axis direction, an R channel created by adding and synthesizing an embedded acoustic signal created by convolution of the created impulse response and an L channel acoustic signal and an R channel acoustic signal. In an acoustic signal, it is difficult to hear an echo for the sake of hearing, and a difference from an R channel acoustic signal before embedding is difficult to detect.
[0054]
(Iii) Since it is considered that acoustic signals of the L and R channels often pass through the same transmission system, the present embodiment using the correlation of the L and R channels is mixed with noise generated randomly during transmission, etc. Has great resistance to.
[0055]
【The invention's effect】
As described above, according to the present invention, an acoustic signal transmission device, an acoustic signal transmission method, and an acoustic signal that can embed a large amount of information easily and safely without causing a sense of incongruity in hearing. A data extraction apparatus and a data extraction method for extracting embedded data can be realized.
[Brief description of the drawings]
FIG. 1 is a block diagram of an “embedded data” transmission system to which the present invention is applied.
FIG. 2 is a flowchart showing an “embedded data” embedding process procedure executed by each block shown in FIG. 1;
FIG. 3 is a block configuration diagram of an “embedded data” extraction system;
4 is a flowchart showing an “embedded data” extraction processing procedure executed by each block shown in FIG. 3;
[Explanation of symbols]
4 “embedded data” input unit 6 key information input unit 8 impulse response creation unit 10 convolution operation unit 12 synthesis unit 16 control unit 20 synchronization unit 22, 24 frame division unit 26 extraction unit 28 control unit

Claims (6)

An acoustic signal transmission apparatus that embeds specific information data in a two-channel stereo digital acoustic signal and sends it out.
An embedding means for embedding specific information data in the input digital audio signal of the first channel;
A synthesizing unit for synthesizing the input second channel digital audio signal and the digital audio signal output from the embedding unit to form a new second channel digital audio signal;
An acoustic signal transmission apparatus comprising: output means for transmitting the input first channel digital acoustic signal and the new second channel digital acoustic signal as a two-channel stereo digital acoustic signal. The acoustic signal transmission device according to claim 1,
The embedding means is
Impulse response creating means for creating an impulse response function based on the specific information data and predetermined key data;
And a convolution operation means for generating a digital sound signal in which the specific information data is embedded by performing a convolution operation on the input digital audio signal of the first channel and the impulse response function. Signal transmission device. When embedding specific information data in a 2-channel stereo digital sound signal,
A first step of embedding specific information data in the input digital audio signal of the first channel;
A second step of synthesizing the input second channel digital audio signal and the digital audio signal obtained in the first step to form a new second channel digital audio signal;
An acoustic signal transmission method comprising: a third step of transmitting the input first channel digital acoustic signal and the new second channel digital acoustic signal as a two-channel stereo digital acoustic signal. The acoustic signal transmission method according to claim 3,
The first step of embedding specific information data in the input digital audio signal of the first channel includes:
An impulse response creating step for creating an impulse response function based on the specific information data and predetermined key data;
A convolution operation step of generating a digital audio signal in which the specific information data is embedded by performing a convolution operation on the input digital audio signal of the first channel and the impulse response function. Signal transmission method. A data extraction device that inputs a first channel digital sound signal and a second channel digital sound signal in which specific information data is embedded, and extracts the specific information data,
The first channel digital sound signal and the second channel digital sound signal are predetermined separately while synchronizing the first channel digital sound signal and the second channel digital sound signal in which the specific information data is embedded. Separating means for sequentially dividing the data string into a predetermined length of time,
An impulse corresponding to the specific information data using a cross spectrum method for the first channel digital audio signal and a second channel digital audio signal in which the specific information data corresponding to the first channel digital audio signal is embedded. Extraction means for extracting a response function and reversely converting the extracted impulse response function into the specific information data based on correspondence information between the specific information data and the impulse response function when the specific information data is embedded. Feature data extraction device. When the first channel digital sound signal and the second channel digital sound signal embedded with the specific information data are input and the specific information data is extracted,
The first channel digital sound signal and the second channel digital sound signal are predetermined separately while synchronizing the first channel digital sound signal and the second channel digital sound signal in which the specific information data is embedded. A first step of sequentially dividing the data sequence into a predetermined length of time,
Impulse corresponding to the specific information data using a cross spectrum method for the first channel digital audio signal and the second channel digital audio signal in which the specific information data corresponding to the first channel digital audio signal is embedded A second step of extracting a response function and inversely converting the extracted impulse response function into the specific information data based on correspondence information between the specific information data and the impulse response function when the specific information data is embedded. A featured data extraction method.

Images