JPH0728203B2 - Transversal filter coefficient calculator - Google Patents

Description

The present invention relates to a transversal filter (hereinafter, referred to as a transversal filter) that realizes an arbitrary frequency characteristic by deviating a convolution integral of a finite number of coefficients (hereinafter, FIR coefficient) and a delayed signal. , A FIR filter) and a transversal filter coefficient calculation device for obtaining an FIR coefficient to be set.

2. Description of the Related Art A sound quality adjustment device is a typical device that uses a FIR filter. The desired frequency characteristic realized by this is usually
It is indicated by the amplitude frequency characteristic (power spectrum), which does not contain phase information. Therefore, it is not possible to obtain the FIR coefficient which is a time function by subjecting this directly to the inverse Fourier transform. Conventionally, a conversion method assuming linear phase has been used as a method for obtaining phase information from a power spectrum, and an FIR coefficient that realizes a desired frequency characteristic has been obtained from this (Reference 1: Basics of Digital Signal Processing Maeda Wata Ohmsha).

Problems to be Solved by the Invention According to Reference Document 1, when linear phase conversion is discussed in a discrete system, its conversion formulas are expressed by Formulas (1) and (2).

Hi (k) = Hr (k) * tan {-(N-1) / N * πk} ...
(2) However, Hr (k) and Hi (k) are original real functions and imaginary functions, and A (k) is an amplitude frequency characteristic.

In the equation (1), even if tan (.........) is used as table data, N points can be converted into N-phase linear phase conversion.
* Minimum of 2 multiplications required. N * 2 multiplications will determine the conversion time of the linear phase conversion. When the frequency band and the frequency resolution are decided, the linear phase conversion calculation time is naturally decided and cannot be shortened.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a transversal filter coefficient calculation device that requires a small number of multiplications for linear phase conversion and has a short conversion time.

Means for Solving the Problems The present invention divides a desired frequency characteristic into a plurality of bands in order to reduce the time required for linear phase conversion for obtaining an FIR coefficient that realizes an arbitrary frequency characteristic, and the characteristics of the divided bands The characteristics of each band are discretely expressed with a sampling frequency and frequency resolution that can express, and linear phase conversion and inverse Fourier transform are performed for each of these bands.
The FIR coefficient is sought. Furthermore, it is necessary to match the sampling frequencies in order to find the FIR coefficient that can express the entire desired frequency characteristic band.
Insert a zero point between data of R coefficient, or linear interpolation,
Alternatively, sampling frequency conversion is performed, and this is passed through a low-pass filter, band-pass filter, or high-pass filter that passes only the divided bands, and then the FIR coefficients are all added to obtain the desired frequency characteristics of all bands. The FIR coefficient that realizes is obtained.

Effect The present invention is to obtain the FIR coefficient for each band without excessively requiring the number of data points representing the characteristics by the above configuration.

EXAMPLES Examples of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a first embodiment of the FIR coefficient calculation device of the present invention. In FIG. 1, reference numeral 1 is an input means of a desired frequency characteristic to be realized by an FIR filter, 2 is a band dividing means for dividing the inputted characteristic into a plurality of bands, and 201, 202 and 203 are the dividing means. This is a sampling point thinning-out means for thinning out the points input by the desired frequency characteristic inputting means 1 for each predetermined number of points in order to represent the characteristics of each of the selected bands with fewer points. FIG. 3 (a) shows a desired frequency characteristic, which is discretely indicated by points on a curve, and the frequency interval is determined by the frequency resolution of the desired frequency characteristic in the low temperature range. FIGS. 3 (b), (c) and (d) show three bands divided by the band dividing means 2 in FIG.

In FIG. 3, f _N1 , f shown on the frequency axis, which is the horizontal axis,
_N2, f _N is the maximum frequency of the divided bands, the digital signal theory is 1/2 of and sampling frequency which is called the Nyquist frequency.

In FIG. 1, 3 is a linear phase conversion means for linearly converting the bands shown in FIGS. 3 (b), (c) and (d) respectively, and 4 is for obtaining an FIR coefficient for realizing a desired frequency characteristic. It is an inverse Fourier transforming means for performing an inverse Fourier transform using the real function and the imaginary function obtained by the linear phase transforming means 3. Reference numeral 5 is a sampling point interpolation means, 6 is a delay means for adjusting the centers of the linear phase coefficients of the respective bands, 701, 702 and 703 are low pass filter means, band pass filter means and high pass filter means (hereinafter referred to as filter means). , It has a filter configuration that passes signals in each band only. FIG. 4 (a) shows an example of sample point interpolation. A zero value is inserted at the time indicated by x between the values on the time axis (round points) obtained by the inverse Fourier transform. By passing this through a bandpass filter, the time axis waveform shown in FIG. 4 (b) is obtained, and in this example, the sampling frequency is tripled. 5 (a), (b) and (c) show a filter means 701, a filter means 702 and a filter means 70.
3 shows the filter characteristic of 3.

Reference numeral 8 is a FIR coefficient adding means for adding the FIR coefficients obtained for each band. FIGS. 6 (a), (b), and (c) are FIR coefficients for realizing the desired frequency characteristics of each band obtained by passing the sampling frequency-converted time waveform through the filter means 701, 702, and 703. FIR that realizes the desired frequency characteristics in the previous band by adding them all together
The coefficient is obtained.

Reference numeral 9 denotes FIR coefficient transfer means for transferring the FIR coefficient to the FIR filter.

Seventh Example of FIR Filter Used in First Embodiment
Shown in the figure. In FIG. 7, 10 is a digital signal input means, 11 is a digital signal storage delay means for storing and delaying an input signal for realizing an FIR filter, and 12 is a transversal filter coefficient computing device of this embodiment. FIR coefficient holding means for holding the FIR coefficient to be transferred, 121
Is a coefficient transfer line to the FIR coefficient holding means 12, 13 is a multiplication means 131
And a sum-of-products means having an addition means 132, and 14 is a digital signal output means for outputting a signal subjected to FIR filtering.

FIG. 2 shows a second embodiment of the present invention. The characteristic of this embodiment is to obtain the FIR coefficient of each band and directly transfer it to the FIR filter.

FIG. 8 shows an example of the FIR filter used in this embodiment. The feature of this filter is digital signal input means.
The digital signal input to 10 is passed through the low pass filter 151,
After passing through the band pass filter 152 and the high pass filter 153, the low band sampling point thinning means 161 and the middle band sampling point thinning means 162 are set so that the upper frequencies of the low band and the band become sampling frequencies at which the Nyquist frequency becomes the Nyquist frequency. The sampling points are thinned out, and the product sum processing similar to that shown in FIG. 7 is performed in the blocks surrounded by the respective dotted lines. FIR coefficient holding means 12
The FIR coefficient corresponding to each band is sent to the receiver from the transfer means 9 of the second embodiment. Further, the same low-pass sampling point interpolating means 171, middle-passage sampling point interpolating means 172, low-pass filter 151, band-pass filter 152, and high-pass filter 153 as shown in the first embodiment are passed. After that, the digital signals are added by the digital signal adding means 18 and then output from the digital signal output means 14.

EFFECTS OF THE INVENTION As described above, according to the present invention, it is not necessary to use an excessive number of data to represent the characteristics of each band.
The arithmetic processing shown in the equation (2) can be performed in a short time with a small circuit scale, which is very useful in practice.

[Brief description of drawings]

FIG. 1 is a block diagram showing an FIR coefficient calculation device in the first embodiment of the present invention, and FIG. 2 is a block diagram of the second embodiment of the present invention.
FIG. 3 is a block diagram showing an FIR coefficient computing device, FIG. 3 is a conceptual diagram showing that the input desired frequency characteristic according to the present invention is divided into a plurality of bands to reduce the number of points representing the characteristic, and FIG. 4 is a zero point. Characteristic diagram showing sampling frequency conversion by insertion and band filter, Fig. 5 is a frequency characteristic diagram of each filter,
FIG. 6 is a characteristic diagram showing the output of each filter of FIG. 1 and the output obtained by adding them, FIG. 7 is a block diagram showing an example of the FIR filter used in the first embodiment, and FIG. It is a block diagram which shows an example of the FIR filter used by the Example of 2nd. 1 ... Desired frequency characteristic input means, 2 ... Band division means,
201, 202, 203 ... Sample point thinning-out means, 3 ... Linear phase converting means, 4 ... Inverse Fourier transforming means, 5 ... Sample point interpolating means, 6 ... Delaying means, 8 ... FIR coefficient adding means, 9
...... FIR coefficient transfer means, 10 …… Digital signal input means, 1
2 ... FIR coefficient holding means, 13 ... sum of products means, 14 ... digital signal output means, 131 ... multiplication means, 132 ... adding means,
151 …… Low-pass filter, 152 …… Band-pass filter, 153 …… High-pass filter, 171 …… Low-pass sampling point interpolating means, 172 …… Middle-pass sampling point interpolating means, 701… Low-pass Pass filter means, 702 ... Band pass filter means, 703 ... High pass filter means.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Akihisa Kawamura 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (56) Reference JP 63-26112 (JP, A) JP 63-234617 (JP, A) JP 63-244924 (JP, A) JP 63-278410 (JP, A) JP-B 6-91416 (JP, B2)

Claims (2)

[Claims] 1. Input means for inputting data having a desired amplitude frequency characteristic, band dividing means for dividing a band of data having the input frequency characteristic into a predetermined number, and each of the divided portions. And sampling point thinning-out means for thinning out sampling points of frequency characteristics sampled for each band of a predetermined number of points, and for frequency characteristics of the divided bands,
Each of the linear phase conversion means for performing a linear phase conversion, an inverse Fourier transform means for obtaining a filter coefficient by subjecting the output of the linear phase conversion means to an inverse Fourier transform, and a filter coefficient obtained by the inverse Fourier transform means A transversal filter coefficient calculation device having a transfer means for transferring directly or indirectly. 2. Sampling frequency converting means for making the sampling frequencies of the filter coefficients obtained by the inverse Fourier transforming means in each of the band-divided bands all the same, and low for inputting the output of the sampling frequency converting means. A band-pass filter unit for each of the band, the middle band, and the high band, and an adding unit for adding the filter coefficients found in each band, and the added filter coefficients are transferred by the transfer unit. The transversal filter coefficient calculation device according to the first item of the above.