JPH029360B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a two-signal waveform similarity measurement device that quantitatively measures waveform similarity from two signals. Conventionally, in order to determine the similarity of waveforms from two signals,
The similarity was understood from the cross-correlation function of the signals. However, it has not been possible to quantitatively determine the degree of similarity only from the cross-correlation function. This invention has been made in view of the above-mentioned problems, and is based on the digital signal x _i at time T and the digital signal y _i at time T when the phase difference between the digital signal x i and the digital signal x _i is zero.
Find the waveform areas S _{x and} S _y from which the DC component has been removed, calculate the autocorrelation function of the digital signal x _i , find the maximum peak value P _xx of the function, _and Calculate the cross-correlation function and find the maximum peak value P _xy of the function, K=
By calculating P _xy /P _xx・S _x /S _y , the signal waveform of the digital signal x _i (corresponding to the signal x inputted before the transmission under test) becomes irregular like the audio signal waveform. The present invention provides a two-signal waveform similarity measurement device that can measure the waveform similarity using the value K even in regular cases. This invention will be explained below. FIG. 1 is a block diagram showing an embodiment of the present invention. In this figure, the first memory circuit 1
is composed of a RAM, etc., and has a storage capacity of 1024 to store digital signals x _i sampled and digitized every 20 μs at time T, and stores the digital signals x _i from 0th to 1023rd. are sequentially stored in storage locations. The second storage circuit 2 is composed of a RAM or the like, and has a storage capacity of 1049 pieces for storing the digital signal y _i at a time T + ΔT delayed by ΔT = 500 μs with respect to the digital signal x _i . The digital signal y _i sampled and digitized at the same time as the signal x _i is applied to the digital signal x _i for 500 μs.
The data is sequentially stored in memory locations from 0th to 1048th while remaining delayed. The delay correction device 3 includes a delay calculation device 3a, a counting circuit 3b, and a third storage circuit 3.
c and a fourth memory circuit 3d.
The delay calculation device 3a receives the digital signals stored in the first and second storage circuits 1 and 2, respectively.
Calculate the cross-correlation function of both digital signals by receiving x _i and y _i , and calculate the cross-correlation function of both digital signals from the position of its maximum peak.
Digital delay value d = 25 (1) corresponding to 500μs
20μs). The counting circuit 3b counts the digital delay value d=25 and stores it in the second storage circuit 2.
The digital signals y _i+25 (i=0, 1, 2, . . . , 1023) stored from the 25th to the 1048th are read out.
The third storage circuit 3c stores the digital signal y _i+25 read from the second storage circuit 2 in the 0th to 1023rd storage locations. First memory circuit 1
The digital signal x _i stored in is stored in the 0th to 1023rd storage locations of the fourth storage circuit 3d. Therefore, the digital signals x _i and
The phase difference at y _i+25 becomes zero, meaning that the delay has been corrected. Note that the next digital signal at time T is stored in the first and second storage circuits 1 and 2. The arithmetic device 4 includes third and fourth storage circuits 3c,
Digital signals x _i stored respectively from 3d
and the waveform area after removing the DC component of the delay-corrected digital signal y _i S _x = _N-1 〓 ⁱ⁼⁰ (｜x _i ｜-1/N _N-1 〓 ⁱ⁼⁰ x _i ) (x _i : i-th sampled peak value of the waveform of digital signal x _i , i = 0, 1, 2, ..., N-1)
and S _y = _N-1 〓 ⁱ⁼⁰ (｜y _i ｜-1/N _N-1 〓 ⁱ⁼⁰ y _i ) (y _i : i-th sampled peak value of the waveform of digital signal y _i ) Calculate. Furthermore, the autocorrelation function of the digital signal x _i is calculated to obtain its maximum peak value P _xx . Furthermore, the digital signal x _i
Then, the cross-correlation function of the delay-corrected digital signal y _i is calculated to obtain its maximum peak value P _xy . Using the waveform areas S _x and S _y and the maximum peak values P _xx and P _xy obtained above, K=P _xy /P _xx ·S _x /S _y is calculated. Next, the operation will be explained. The analog input signal (audio signal) x of the transmission system under test applied to the input terminal 5 (channel 1) is sampled at N=1024 points every 20 μs by the sampling circuit 6, and converted into a digital signal x _i by the A-D converter 7. (i=0, 1, 2, ..., 1023). This digital signal x _i has a time T (20μs×1024)
1024 points corresponding to 1024 points are stored in the 0th to 1023rd storage locations of the first storage circuit 1, as shown in FIG. 2a. In addition, the analog output signal y of the transmission system under test, which is delayed by ΔT=500 μs with respect to the analog input signal x, is added to the input terminal 8 (channel 2), and the sampling circuit 9 simultaneously inputs the input signal x at 1024+25 points (25 points). (corresponds to a delay of 500 μs) After being sampled, it is converted into a digital signal y _i (i=0, 1, 2, . . . , 1048) by the A/D converter 10. This digital signal y _i is a digital signal
Time T+ with a delay of ΔT=500μs relative to x _i
1049 points corresponding to ΔT are stored in the 0th to 1048th memory locations of the second memory circuit 2, as shown in FIG. 2b. Digital signal x _i , y _i
is added to the delay arithmetic unit 3a, where first, for the digital signal x _i,

[Formula] (X _k is the spectrum at each frequency point, k = 0, 1, 2,
..., N-1, j=√-1) using fast Fourier transform (FFT). Then for the digital signal y _i

Execute the calculation in [Formula] using FFT. Next, after calculating Z _k = X ^* _k・Y _k (* is the complex conjugate), for Z _k

The calculation in [Formula] is performed using inverse fast Fourier transform (IFFT) with Z ₀ (DC component) set to zero. As a result, the cross-correlation function of the digital signals x _i and y _i , that is, the input and output signals x and y, is determined as shown in FIG. Figure 3 shows the first and second
The digital signal y _i stored in the memory circuits 1 and 2 of
is delayed with respect to the digital signal x _i by a digital delay value d=25, that is, 500 μs (delay value 1=20 μs). The digital delay value d=25 obtained in this way is added to the counting circuit 3b for counting, and as shown in FIG.
Read out y _i+25 (i=0, 1, 2,..., 1024).
The read digital signal y _i+25 is stored in the 0th to 1023rd storage locations of the third storage circuit 3c, as shown in FIG. 2d. Further, the digital signal x _i stored in the first storage circuit 1 is read out and stored at 0 in the fourth storage circuit 3d as shown in FIG. 2c.
It is stored in memory locations from 1st to 1023rd.
In this way, the third and fourth memory circuits 3c, 3d
The digital signals x _i and y _i+25 stored in are outputted to the arithmetic unit 4 with a phase difference of zero. Digital signal x _i and delay-corrected digital signal stored in the third and fourth storage circuits
Add y _i+25 to arithmetic unit 4. Here, the waveform area S x of the digital signal x _i and the delay-corrected digital signal y _{i+25 is} S _x = _N-1 〓 ⁱ⁼⁰ (|x _i |-1/N _N-1 〓 ⁱ⁼⁰ x _i ) and S _y = _N-1 〓 ⁱ⁼⁰ (|y _i+25 |-1/N _N-1 〓 ⁱ⁼⁰ y _i+25 ). Next, for digital signal x _i

Execute [formula] using FFT. Then, after calculating A _k =X _k・X〓 _k , A _k
against

Execute [Equation] using IFF with A ₀ (DC component) set to zero, calculate the autocorrelation function of digital signal x _i , and find the peak value P _xx of a _i . Also, for digital signal y _i+25

After executing [Formula] with FFT and calculating C _k =X _k 〓・Y _k , for C _k

Execute [Formula] using IFFT with C ₀ (DC component) set to zero. Thus the digital signal x _i
Then, the cross-correlation function of the delay-corrected digital signal y _i+25 is calculated to obtain the peak value P _xy of C _i .
The waveform areas S _x , S _y and peak values obtained in this way
Using P _xx and P _xy , the calculation K=P _xy /P _xx ·S _x /S _y is performed. Using the value K obtained in this manner, the degree of similarity between the waveforms of the digital signals x _i and y _i+25 is measured. As explained above, the present invention includes a first storage circuit that stores a digital signal x _i at a time T, and a digital signal y _i at a time T+ΔT delayed by a time ΔT with respect to the digital signal x _{i .} a second storage circuit that stores the digital signal while being delayed by a time ΔT;
a delay correction device for correcting the delay of y _i ; and a waveform area S _x , S _y of the digital signal x _i and the delay-corrected digital signal y _i+ 〓 _i (Δi is the number of points corresponding to ΔT), and an arithmetic device that calculates K=P _xy /P _xx・S _x /S _y using the maximum peak value P _xy obtained from the cross-correlation function of the digital signal x _i and the digital signal y _i+ 〓 _i . Therefore, even if the two signals are irregular, such as audio signals, it is possible to know, for example, deterioration or abnormality in the transfer characteristics of the transmission system under test from the calculated value of K. In addition, it is possible to quantitatively determine the similarity of the waveforms of two signals, not only when the two signals have a causal relationship with each other, but even when the two signals have no causal relationship with each other. There is. Note that the present invention is not limited to the above-described embodiments, and can be implemented in various ways by specifically limiting the constituent elements without changing the gist of the invention.

[Brief explanation of drawings]

FIG. 1 is a block configuration diagram of an embodiment of the present invention, FIG. 2 is an explanatory diagram of the data storage state of digital signals x _i and y _i stored in a storage circuit, and FIG. 3 is a diagram of the data storage state of digital signals x _i and y i time-delayed digital signal
FIG. 3 is a diagram showing the peak position of the cross-correlation function of _yi . In the figure, 1 is a first storage circuit, 2 is a second storage circuit, 3 is a delay correction device, 3a is a delay calculation device, 3b is a counting circuit, 3c is a third storage circuit, 3
d represents a fourth storage circuit, and 4 represents an arithmetic unit.

Claims (1)

[Claims] 1. A first storage circuit 1 that stores a digital signal x _i sequentially from the m-th storage location to m, m+1, m ₊ 2, ...th locations; a second storage circuit 2 that receives the digital signal y _i delayed by a time ΔT and stores it sequentially from the m-th storage location to the m-th, m+1, m+2, . . .-th locations; from the first and second storage circuits; Receive the digital signals x _i and y _i stored respectively, calculate the cross-correlation function of both digital signals, calculate the digital delay value d corresponding to the unknown time ΔT from the position of the maximum peak, and use this. a delay correction device 3 for correcting the delay of the digital signal y _i ; determining the waveform areas S _x and S _y of the digital signal x _i and the digital signal y _i whose delay has been corrected by the delay correction device; In order to calculate the autocorrelation function of the digital signal x _i , calculate the maximum peak value P _xx of the digital signal x i.
Calculating the cross-correlation function of the digital signal x _i and the delay-corrected digital signal y _i to obtain its maximum peak value P _xy , and the waveform area S _x
and an arithmetic device 4 for calculating the following equations using S _y and the maximum peak values P _xx and P _xy , and converting the digital signal to the calculated value K.
A two-signal waveform similarity measuring device characterized in that it measures the similarity of waveforms of x _i and y _i . S _x = _N-1 〓 ⁱ⁼⁰ (｜x _i ｜-1/N _N-1 〓 ⁱ⁼⁰ x _i ) ... x _i : i-th sampled peak value i of the waveform of digital signal x _i =0, 1, 2, ..., N-1, S _y = _N-1 〓 ⁱ⁼⁰ (｜y _i ｜-1/N _N-1 〓 ⁱ⁼⁰ y _i )... y _i : Digital signal y The _i -th sampled peak value of the i waveform K=P _xy /P _xx・S _x /S _y ...... 2 The delay correction device is configured to convert the digital signal stored in the first and second storage circuits into x _i and y _i
a delay calculation device 3a that calculates the digital delay value d from the maximum peak position by calculating the cross-correlation function of A counting circuit 3b for reading out the stored digital signal y _i , and a counting circuit 3b for reading out the digital signal y _i read from the second storage circuit, sequentially from the pth storage location to p, p+1, p+2, ...th locations. a third memory circuit 3c for storing the digital signal x i read from the first memory circuit, and a fourth memory circuit for storing the digital signal x _i read from the first memory circuit in sequential locations p, p+1, p+2, . . . from the p-th memory location. 2. A waveform similarity measuring device for two signals according to claim 1, comprising a memory circuit 3d for delay correction.