JP2003174437A - Transmission path error detecting device by two system voice signal - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a transmission part error detecting device capable of effectively and highly reliably detecting a transmission path error in communication in redundant configuration using two system transmission paths. <P>SOLUTION: A difference signal z(t) of voice signals x(t) and y(t) transmitted from two system transmission paths is compared with the signal x(t) by an error judging part 34, and when a difference between the signals z(t) and x(t) exceeds a certain threshold, it is judged that a transmission path error is present. Then, which of those two system transmission paths generates the error is judged by calculating wavelet conversion coefficients X(k), Y(k), and Z(k) from the signals x(t), y(t) and the difference signal z(t) by a featured value deriving part 35, and calculating the correlation coefficient CK<SB>(</SB>X, Z<SB>)</SB>and C=CK<SB>(</SB>Y, Z<SB>)</SB>of the wavelet conversion coefficients X(k) and Z(k), and Y(k) and Z(k) by correlating parts 36 and 37, and comparing the correlation coefficients CK<SB>(</SB>X, Z<SB>)</SB>and CK<SB>(</SB>Y, Z<SB>)</SB>with a preliminarily learned threshold CT by a threshold processing part 38. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a transmission path error detecting apparatus for a two-system audio signal, and more particularly to a transmission path error detecting apparatus for a two-system audio signal capable of detecting a transmission path error with high reliability.

[0002]

2. Description of the Related Art Conventionally, a system has been used in which the same image information is divided into two routes and transmitted, and a normal route is selected on the output side to ensure high reliability of communication. ing. The purpose of this system is to monitor the difference between image information of two systems, detect an error in the transmission path, and switch the program without interruption when an error is detected. As a conventional technique of this kind, for example, Japanese Patent Laid-Open No. 2000-350 by the present applicant.
There is one described in Japanese Patent No. 238.

Conventionally, most of the methods for detecting a transmission line error using a voice signal are those for detecting only one line.

[0004]

When a slight transmission path error or the like occurs, it may be easier to detect the deterioration of the audio than the image. However, as in the conventional case, only one line is used. Since the one that detects a transmission line error is generally judged from the nature of the transmission line error, it is possible to judge that an originally error is not an error, or conversely it can be judged that an error is not an error. There was a lack of reliability.

The present invention has been made in view of the above-mentioned prior art, and an object thereof is to effectively and reliably improve a transmission line error in communication in a redundant configuration using two transmission lines. It is to provide a transmission path error detection device capable of detecting.

[0006]

[Means for Solving the Problems] To achieve the above-mentioned object.
For the purpose of the present invention, at least audio transmitted in two systems is used.
2 system for detecting transmission line error based on signal including signal
In the transmission line error detection device for the general voice signal,
The difference signal z (t) between the two-system audio signals x (t) and y (t) and the two-system system
The transmission path error is calculated from one of the audio signals x (t) and y (t).
Means for detecting the occurrence of an error, and the audio signal x
(t), y (t) and the wavelet transform of the difference signal z (t)
Et al., Wavelet transform coefficients X (k), Y (k), and Z
means for obtaining (k) and the wavelet transform coefficient X (k)
Correlation coefficient C between Z (k), Y (k) and Z (k) _{K (X, Z)}and
C = C_{K (Y, Z)}And a correlation coefficient C
_{K (X, Z)}And C_{K (Y, Z)}And request by learning in advance
Set threshold C_TAnd compare the belief that the error is
It is characterized in that it is provided with means for determining the number.

According to this feature, it becomes possible to accurately detect even a slight transmission line error and which of the two systems contains the error.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described in detail below with reference to the drawings. FIG. 1 is a conceptual diagram of a communication system having a redundant configuration using two transmission lines to which the present invention is applied.

On the transmitting side, an input signal 10 such as an audio signal or a video signal containing an audio signal is distributed to two systems,
The first encoder 11 of the first system and the second encoder 2 of the second system
Is encoded by the encoder 12. The signal encoded by the first encoder 11 is transmitted to the reception side via the first transmission path 13 and is decoded by the first decoder 15 on the reception side. Similarly, the signal encoded by the second encoder 12 is transmitted to the reception side via the second transmission path 14 and is decoded by the second decoder 16 on the reception side.

The signals x (t) and y (t) decoded by the first and second decoders 15 and 16 are input to the error detecting device 17,
A transmission line error (fault) is detected. Here, t represents time. If both of the systems are normal, the transmission line of the working system, for example, the first system is selected, and the decoded signal x (t) of the first system is output as the output signal 20. On the other hand, when it is determined that an error has occurred in the working line, the decoded signal y (t) of the second system spare line is selected and output as the output signal 20. This switching is performed without interruption.

In the present invention, when a slight error occurs in the transmission line, it is noted that the deterioration of the signal may be detected more easily in the audio signal than in the video signal. Error detection is performed based on the audio signal.
FIG. 2 is a block diagram showing the configuration of an embodiment of the error detection device 17.

The signals x (t) and y (t) decoded by the first and second decoders 15 and 16 are input to the error detection device 17. The windows 31 and 32 are the signals x (t) and y (t).
The voice signal of the same section is extracted from. The time lag between the two audio signals is compensated by a device (not shown), and signals without time lag are input to the windows 31 and 32. That is, the signal x (t), y
There is no time lag in (t).

The subtracting means 33 obtains a difference z (t) between the signals x (t) and y (t), and the difference z (t) is sent to the error judging section 34. In the error determination unit 34, the difference z (t) and the signal x
(t) (or y (t)) is compared, and if the absolute value of the difference is less than or equal to a predetermined threshold value, some failure occurs in either the first route or the second route. I judge that I did. If it is determined that there is no error, the signal x (t)
Is output from the error detection device 17. Note that the difference z
The error can be detected by comparing (t) with the signal x (t) (or y (t)), if there is no error in both transmission lines, z (t) = 0 or ≈0. , Z (t) and signal x (t)
(Or y (t)) becomes large, but if an error occurs in one of the transmission lines, z (t) does not become 0 and the signal x (t) (or y (t)) This is because the difference between

Next, when the error determination unit 34 determines that there is an error, the signals x (t) and y (t) are guided to the feature amount derivation unit 35, and which of the signals causes an error due to the subsequent processing. Is determined.

The following method can be considered for determining the error. First, simply, the difference z (t) and the signal x
A method of taking a correlation with (t) and y (t) can be considered. But,
In this method, if a large error occurs at one point, a normal signal and z (t), and a signal containing an error and z (t)
There is almost no difference in the correlation with and it is impossible to separate with this feature quantity.

Next, the difference z (t) and the signals x (t) and y
A method in which the power spectrum of (t) is used as a feature quantity is conceivable. According to this method, although it depends on the size of the window, as a result of the preliminary experiment, the difference z (t) and the signal x (t), y
No significant difference appeared in the correlation of the feature quantities of (t), and it was confirmed that separation was impossible, as in the above method. The power spectrum is the sum of squares of the magnitudes of the coefficients obtained by Fourier transforming a certain signal f (t).

Further, as another error determination method, a method in which a low-pass filter is applied in the time direction and a changed signal is regarded as having an error can be considered. However, according to this method, there is a possibility that the error cannot be detected or the high frequency component is erroneously determined to be an error when the error continues for several samples.

Therefore, in this embodiment, there are a learning phase and a detection phase as described below.
In the former phase, learning is performed in advance with an actual signal, and then in the latter detection phase, a threshold value C _T, which will be described later, obtained by the learning is used to calculate the signals x (t) and y (t). Determine which has an error.

First, the learning phase will be described. (1) As a preliminary experiment, a difference signal z (t) is obtained for a signal x (t) containing an error and a signal y (t) containing no error. Then, the three signals x (t), y (t), and z (t) are subjected to the conversions shown in the following documents, and X (k), Y (k), and Let Z (k). Reference: S. Mallat and Z. Zhang, 'Matching Pursuits Wit
h Time-Frequency Dictionaries'IEEE Trans. on Sign
al Processing, Dec. 1993. An outline of the conversion method described in the above document will be described with reference to FIG. This method is a kind of wavelet transform _, and decomposes a certain signal into a low frequency component l (t) and a high frequency coefficient d _{i, j} by a wavelet function P _{i, j} . For example, considering decomposition of a speech signal x (t) having a waveform as shown in FIG. 3, a wavelet function P _{i, j} is selected from a dictionary _{, and the coefficient} is calculated according to the following equation (1) based on the selected wavelet function P _{i, j.} The d _{i, j} and the low frequency component l (t) are obtained. That is, as shown in FIG. 3, the wavelet function P obtained from the dictionary is the original audio signal x (t).
_{Based on i and j} , it is decomposed as shown. As a result of this decomposition, the high-frequency coefficient d as shown in FIGS.
_{-3, j} , d- _{2, j} , d- _{1, j} and low-frequency component l (t)
Is obtained. The _{d i,} those enumerated a _{j (d - 3, j,} d -2, j, d -1, j) becomes the X (k). Here, k is the order. Similarly to the above, the audio signal y (t) and the difference signal z (t) are subjected to the conversion shown in the above document to obtain Y (k) and Z (k).

(2) Next, the correlation coefficient up to order k = K is calculated between X (k), Y (k) and Z (k). _{Let the} correlation coefficients be C _{K (X, Z)} and C _{K (Y, Z)} . (3) Take a plurality of samples of (1) and (2) above. (4) Signal x (t) containing a plurality of errors taken in (3) above
The correlation coefficient C _{K (X, Z) of} the signal y (t) and the correlation coefficient C _{K (Y, Z) of the} signal y (t) that does not include a plurality of errors are C = C _{K (X, Z).} And C = C _{K (Y, Z)} distribution is obtained. (5) Next, in the following method, the probability α of determining a signal containing an error as not containing an error and the probability β of determining a signal not containing an error as containing an error A threshold value C _T of the feature amount C that reduces both values is obtained.

The threshold value C _T can be obtained as follows. Now, assuming that Hypothesis H0: "either signal contains a channel error" or Hypothesis H1: "Neither signal contains a channel error", some feature extracted from the observed signal. When the quantity is C (for example, the feature quantities C _{K (X, Z)} and C _{K (Y, Z)} ), it can be considered as a random variable because it is a variable extracted from the uncertainties. Here, the hypotheses H0 and H1
The probability density function under p is p (C; H0), p (C; H
1), the probability of determining a signal containing an error as not containing an error (probability of detection failure) α, and the probability of determining a signal containing no error as containing an error ( The probability of erroneous detection) β can be expressed as in the following equations (2) and (3), where R is the rejection area. α = ∫ _R p (C; H0) dC ・ ・ ・ (2) β = ∫ _R p (C; H1) dC ・ ・ ・ (3) Here, the value of C that makes both α and β as small as possible Is a threshold value C _T. With the above, the learning phase is completed and the process proceeds to the detection phase.

Next, for the actual signals x (t) and y (t),
The phase (judgment phase) of applying the threshold value C _T to determine which of the signals x (t) and y (t) contains an error will be described with reference to FIG.

(1) An error is detected by the error determination unit 34.
When it is determined that the two systems
The signals x (t) and y (t) and their difference signal z (t) are described in the above document.
Conversion coefficient X (k), Y (k), and Z (k)
To get (2) The conversion coefficients X (k) and Z obtained by the feature amount deriving unit 35
(k) is a correlation unit 36 which calculates correlations up to order k = K as described above.
Thus asked. Further, the conversion coefficients Y (k) and Z (k) are
Correlation unit 37 similarly obtains correlations up to order k = K.
It The obtained correlations are respectively C_{K (X, Z)}And
And C_{K (Y, Z)}And (3) The threshold processing unit 38 uses the correlation C_{K (X, Z)}and
C_{K (Y, Z)}And the threshold C obtained in the learning_TAnd
Compare the threshold C_TIf it is greater than the
Is determined to be included. For example, the correlation C
_{K (X, Z)}Is the threshold C _TSignal x if greater
It is determined that (t) contains an error.

The present inventor conducted a simulation experiment in order to verify the effectiveness of this embodiment. In the simulation experiment, the AES / EBU digital audio format was targeted, and the error other than the payload part could be easily detected, so it was excluded from the target. It is also assumed that the time lag between the two systems has been compensated and that at least one of the two systems is operating normally.

Under the above conditions, the respective values α, β, and the threshold C _T of C with and without error are measured, and the result is as shown in FIG. The conventional method of FIG.
As described above, simply, the difference z (t) and the signal x (t), y
This is a method of detecting an error by correlating with (t).

As is apparent from FIG. 5, according to the method of the present invention, both the probability α of missed detection and the probability β of erroneous detection have extremely small values as compared with the conventional method. That is, the method of the present invention is 1
It can be seen 0 small ^four orders. This means that a signal containing an error is almost certainly detected and a signal containing no error is not extracted, and the effectiveness of this embodiment has been confirmed.

[0027]

As is apparent from the above description, according to the inventions of claims 1 to 3, in a transmission system having a dual transmission path, when a signal is deteriorated due to a transmission path error or the like, which signal Can be detected with high accuracy. Therefore, it is possible to greatly improve the reliability of the transmission path error detection in the two-system signal. In particular,
The detection accuracy is improved when a slight transmission line error occurs.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a dual transmission system to which the present invention is applied.

FIG. 2 is a block diagram showing the configuration of an embodiment of the error detection device of FIG.

FIG. 3 is an explanatory diagram of wavelet transform used in this embodiment.

FIG. 4 is an explanatory diagram of wavelet transform coefficients.

FIG. 5 is a diagram illustrating a comparison of effects of the present invention and a conventional method.

[Explanation of symbols]

17 ... Error detection device, 32 ... Subtraction means, 34 ... Error determination unit, 35 ... Feature amount derivation unit, 36, 37 ... Correlation unit, 38 ... Threshold processing unit.

Continued front page    (72) Inventor Masahiro Wada             2-15-1 Ohara, Kamifukuoka City, Saitama Stock             Company CAD Research Institute F-term (reference) 5J064 AA06 BA01 BA16 BB08 BC27                       BD02                 5J065 AA01 AB01 AC02 AE01 AH09                       AH15                 5K014 AA01 CA06 GA02

Claims (3)

[Claims] 1. A transmission path error detection device for a two-system audio signal for detecting a transmission path error based on a signal including at least a sound signal transmitted by two systems, wherein the two-system audio signal x (t), y ( t) difference signal z (t)
Means for detecting the occurrence of a transmission path error from one of the system audio signals x (t), y (t), the audio signals x (t), y (t), and the difference signal z From the wavelet transform of (t), wavelet transform coefficients X (k), Y
means for obtaining (k) and Z (k), and the wavelet transform coefficients X (k) and Z (k), Y (k) and Z (k)
Means for obtaining the correlation coefficients C _{K (X, Z)} and C = C _{K (Y, Z),} and the correlation coefficients C _{K (X, Z)} and C _{K (Y, Z)} obtained by learning in advance. was compared with the threshold C _T, transmission path error detection apparatus in the two systems audio signal, characterized by comprising a means for determining a signal that contains an error. 2. An absolute value of a difference between one of the two-system audio signals x (t) and y (t) and the difference signal is preset in the means for detecting that the transmission path error has occurred. The transmission path error detection device for a two-system audio signal according to claim 1, wherein it is determined that a transmission path error has occurred when the threshold value is equal to or more than the threshold value. 3. The means for determining a signal containing an error determines that a signal having a correlation coefficient larger than the threshold value C _T contains an error. 1 or 2 the transmission path error detection device for the two-system audio signal