JP4326282B2 - Filter device - Google Patents

Description

  The present invention includes a chamberin type filter unit that filters an input digital musical sound signal with a filtering characteristic according to two set filter coefficients, and a coefficient calculation for obtaining the two filter coefficients and setting the filter unit in the filter unit A filter device.

  In the timbre setting filter of an electronic musical instrument, the cutoff frequency needs to be continuously changed in order to change the cutoff frequency in accordance with the envelope and LFO. However, if a filter coefficient that can be heard so that the cut-off frequency continuously changes is tabulated, the table capacity becomes enormous. Therefore, in the prior art, a known interpolation calculation, that is, a one-dimensional interpolation calculation is performed, and for example, only the filter coefficient corresponding to the cutoff frequency is interpolated to provide the filter coefficient to the digital filter. (For example, refer to Patent Document 1).

  That is, referring to the block configuration diagram of FIG. 5, the CPU 10 executes the operation program recorded in the ROM 12, and Fc (cutoff frequency value: a parameter indicating a frequency at which the amplification degree decreases by 3 dB), which is two parameters, When Q (resonance value: parameter indicating the degree of the peaking state of the frequency characteristic) is given, each of the coefficient F interpolation calculation unit 50 and the coefficient W interpolation calculation unit 60 sets a set of filter coefficients corresponding to Fc and Q to the coefficient table. Thus, the filter coefficient F and the filter coefficient W are obtained by interpolation processing, and are set in the filter unit 30 to adjust the LPF (low-pass filter) characteristics and the like of the filter unit 30.

  The filter unit 30 is a state variable multi-mode filter, and can output an LPF (low-pass filter) signal, an HPF (high-pass filter) signal, a BPF (band-pass filter) signal, and the like with one secondary filter. The chamber filter is set with two filter coefficients F and W. However, according to such a conventional technique, a large-scale coefficient table is required to adjust the filter coefficients F and W to the required values according to the input parameters, and it is also necessary to perform an interpolation operation. For this reason, there is a problem that the scale of the filter device is increased.

  Therefore, it has been proposed to obtain the two coefficients F and W of the chamber filter by an arithmetic expression (for example, see Non-Patent Document 1). According to this, for example, “F = 2π · (Fc / Fs), W = 1 / Q: where Fc is a cutoff frequency, Fs is a sampling frequency, Q is a parameter indicating a resonance value, and when Q = 1, resonance The two coefficients F and W are obtained by a simple expression “minimum value, maximum resonance value when Q is infinite”.

Japanese Patent Laid-Open No. 6-195079 (page 3-4, FIG. 1) Jon Dattorro, "Effect Design", Journal of the AES, (USA), Audio Engineering Society, Inc., September 1997, Volume 45 (Vol. 45), No. 9 (No, 9), p660-p684

  However, the frequency characteristics obtained by such calculation cannot be used. FIG. 8 shows frequency (unit: Hz) on the horizontal axis and amplitude (unit: dB) on the vertical axis. According to the conventional method, the frequency characteristics of LPF (low-pass filter) output as the cutoff frequency is increased. However, for example, the frequency characteristics are raised as indicated by symbols A to D, and it is difficult to use the frequency characteristics as they are as the LPF output frequency characteristics. In addition, in the case of using this calculation formula, if it is in the range of about half or less of the Nyquist frequency, it has characteristics that can be used as a low-pass filter. A method of increasing the sampling frequency, filtering with this filter, and then returning to the original sampling frequency has been adopted. However, even in this case, there is a problem that originally unnecessary resonance remains or the amount of calculation increases because the sampling frequency is temporarily increased.

  Therefore, the present invention has been made to solve such a conventional problem, and provides a filter device capable of obtaining a chambering filter coefficient having a good frequency characteristic with a simple configuration. Objective.

In order to achieve the above object, the present invention includes a chamberin type filter unit for filtering an input digital musical sound signal with a filtering characteristic according to two set filter coefficients (F, W), and the two A coefficient calculation unit that obtains a filter coefficient (F, W) and sets the filter coefficient in the filter unit,
The chamberin type filter unit is:
A first adder (90), a second adder (91), a first multiplier (92) having a filter coefficient F, a third adder (93), and a filter coefficient F; A second multiplier (94) and a fourth adder (95) are connected in series, and the filter output is output by the first delay element (98) to the second adder (91) and the fourth adder (95). The second delay element (97) feeds back the output of the third adder (93) to the adder 93 and has a filter coefficient W. Configured to feed back to the first adder (90) through a multiplier (3) of 3;
The coefficient calculator is
The cutoff frequency value (Fc) that is the first parameter and the resonance value (Q) that is the second parameter are received, and “F = (1 + Q) · Fc, W = (1−Q) · (2− Fc) "directly determines the two filter coefficients (F, W) ,
The obtained filter coefficient F is given to the first multiplier (92) and the second multiplier (94), and the obtained filter coefficient W is given to the third multiplier (96). It was made to feature.

  According to the present invention, the filter unit filters the input digital musical sound signal with a filtering characteristic corresponding to the set filter coefficient (F, W), and the coefficient calculation unit obtains the filter coefficient to the filter unit. The coefficient calculation unit accepts a cutoff frequency value (Fc) that is a first parameter and a resonance value (Q) that is a second parameter, and “F = (1 + Q) · Fc, W = (1-Q) · (2-Fc) "is used to directly obtain the two filter coefficients (F, W) and set the obtained filter coefficients (F, W) in the filter unit. It is possible to realize a filter device that can obtain a filter coefficient of a chambering filter with a simple configuration that is unnecessary and does not require an interpolation operation. Here, the cut-off frequency value Fc, which is the first parameter, is a parameter indicating the frequency at which the degree of amplification decreases by 3 dB, and the resonance value Q, which is the second parameter, indicates the degree of the peaking state of the frequency characteristic. It is a parameter to show.

  The reason why the two filter coefficients (F, W) of the chamber filter can be obtained with this simple formula is as follows. LPF (low-pass filter) using a simple first-order IIR (Infinite Impulse Response) filter makes it easy to control the cutoff frequency by a coefficient, and a chamber filter (described in detail in an embodiment later). Constitutes a second-order IIR filter by itself, so that the transfer function of the chamber filter is equivalent to the case where the first-order IIR is cascaded in two stages, the coefficients F, W of the chamber filter Thought. FIG. 6 shows two-stage cascaded primary IIR filters. In the first stage, a multiplier 400 having a coefficient k and an adder 402 are connected, and the output of the adder 402 is connected to a delay element 404 and a coefficient “ The first-order IIR filter is configured to feed back via a 1-k multiplier 406, and the second stage is also a first-order IIR filter having a similar configuration. The multiplier 410 and the adder 412 have a coefficient k. And the output of the adder 412 is fed back via a delay element 414 and a multiplier 416 having a coefficient “1-k”.

  The transfer function of a filter of the type in which the first-order IIR filters are cascade-connected is shown at the top of FIG. On the other hand, the transfer function of the LPF (low-pass filter) of the chamber filter is the second from the top in FIG. Therefore, first, assuming that no resonance is applied, the cutoff frequency of the first-order IIR filter is easily controlled only by the coefficient k, and the frequency characteristics at that time are also good. Consider associating coefficients F, W and coefficient k of the chamber filter so that they match. First, assuming that the numerators of both transfer functions are the same, “F = k” is derived from the fact that the square of k and the square of F are equal. In the denominator, the first-order terms are equal and “−2 (1 −k) = − 2 + F · W + F∧2 (hereinafter, ∧ represents a power. In this case, F∧2 represents the square of F) and “F = k”, “W = 2−2. k "is obtained. The quadratic terms of the denominators of both transfer functions are equal when F and W are substituted.

  From these facts, by setting “F = k, W = 2−k”, the same transfer function as that obtained by cascading the first-order IIR filters in two stages can be obtained from the LPF output of the chamber filter. The chamber characteristic filter coefficient of the frequency characteristic can be easily obtained. The above is a description of a state where no resonance is applied. In the following, it will be considered that the simple filter coefficient arithmetic expression is modified so that the resonance can be applied. When the resonance Q is maximized, it is known that the coefficient W of the chamber filter becomes “0”, and the following calculation is added to realize this.

  The transfer function when the cutoff frequency is minimized and the resonance is maximized, that is, when the peak in the DC component DC rises, is a filter in which the first-order IIR filters are cascaded in two stages. It will be third. At this time, since the filter coefficient W is “0” in the chamber filter, the coefficient F is considered. However, since the transfer function is not exactly the same as that of the first-order IIR filter, paying attention to only the denominator term of the transfer function and assuming that the filter poles match, the first-order term of the denominator is equal to “−2 = −2 + F · W + F∧2 ”, that is,“ F∧2 + F · W = 0 ”. From this and“ W = 0 ”,“ F = 0 ”. In the second-order term of the denominator, “W = 0” results in “F · W = 0”, so the filter coefficient F cannot be obtained.

  Next, the filter coefficient F of the chamber filter is obtained from the transfer function when the resonance is maximized by maximizing the cutoff frequency and the transfer function of the chamber filter using the two-stage cascade connection of the primary IIR filter. . The transfer function when the cut-off frequency and resonance are both maximum in the two-stage cascade connection of the primary IIR filter is as shown in the lowermost stage (fourth from the uppermost stage) in FIG. Here, only the denominator of the transfer function is considered, and the filter poles are made to coincide. Assuming that the first-order terms of the denominators of both transfer functions are equal, “2 = −2 + F · W + F∧2”, that is, “F∧2 + F · W = 4”. From this and “W = 0”, the square of F is 4, and “F = 2” is obtained.

  Therefore, when “the cutoff frequency is minimum and the resonance is maximum, F = 0, W = 0”, and when “the cutoff frequency is maximum and the resonance is maximum, F = 2, W = 0”, the resonance is maximum. The filter coefficients F and W for the chamber filter are obtained. These, and a coefficient calculation expression “F = k, W = 2−k” using k when there is no resonance, and a resonance parameter Q (here, when Q is “0”, the resonance is minimum, and Q is “1”) In this case, the resonance is made maximum.

  For F, “(1) When resonance is minimum (Q = 0), F = k regardless of cutoff frequency”, “(2) When resonance is maximum (Q = 1), cutoff is minimum (k = 0) ), F = 0 ”,“ (3) Resonance maximum (Q = 1), and cutoff maximum (k = 1), F = 2 ”. = (1 + Q) · k ”. On the other hand, for W, regardless of the cut-off frequency, “(1) When resonance is minimum (Q = 0), W = 2−k”, “(2) When resonance is maximum (Q = 1), W = Therefore, “W = (1−Q) · (2−k)”.

  Therefore, when these are put together, “F = (1 + Q) · Fc, Q = (1−Q) · (2−Fc)” is obtained, and the filter coefficients W and F of the chamber filter can be obtained by a simple calculation, and a large coefficient table is obtained. Can be made unnecessary. When this filter coefficient is used in a chamber filter, the frequency characteristics are equal to those obtained by cascading two stages of primary IIR filters when resonance is not applied, and up to the maximum Nyquist frequency when resonance is applied. Since it is possible to oscillate, it is possible to eliminate the disadvantages of the conventional chamber filter.

In addition, a chamber type filter unit for filtering an input digital musical sound signal with a filtering characteristic corresponding to two set filter coefficients (F, W) and the two filter coefficients (F, W) are obtained. There are a plurality of sets of coefficient calculation units set in the filter unit, a cutoff frequency value (Fc) that is a first parameter, and a power value obtained by multiplying the number by a predetermined value (Fc raised to a predetermined value: Fcq). A storage unit for storing the set,
The chamberin type filter unit is:
A first adder (90), a second adder (91), a first multiplier (92) having a filter coefficient F, a third adder (93), and a filter coefficient F; A second multiplier (94) and a fourth adder (95) are connected in series, and the filter output is output by the first delay element (98) to the second adder (91) and the fourth adder (95). The second delay element (97) feeds back the output of the third adder (93) to the adder 93 and has a filter coefficient W. Configured to feed back to the first adder (90) through a multiplier (3) of 3;
The coefficient calculator is
Means for obtaining a power value (Fcq) corresponding to the cutoff frequency value (Fc) as the given first parameter with reference to the storage content of the storage unit;
The resonance value (Q) which is the given second parameter and the obtained power value are substituted into the formula “F = (1 + Q) · Fcq, W = (1−Q) · (2−Fcq)”. look including a means for determining the filter coefficients (F, W) Te,
The obtained filter coefficient F is given to the first multiplier (92) and the second multiplier (94), and the obtained filter coefficient W is given to the third multiplier (96). A filter device is also provided. Also in the present invention, it is possible to realize a filter device having a simple configuration that does not require a coefficient table and does not require an interpolation calculation.

  According to the present invention, there is an effect that it is possible to realize a filter device capable of obtaining chamber chamber filter coefficients having a good frequency characteristic with a simple configuration.

  The best mode for carrying out the present invention will be described below with reference to the drawings. In the present invention, the term “chamberline filter” generally refers to one second-order IIR (Infinite Impulse Response) filter to LPF (low-pass filter), HPF (high-pass filter), BPF (band-pass filter). These filters are also called state variable filters, and are designed by digitally replacing analog state variable filters.

  FIG. 1 is a block diagram of a filter device 1 according to an embodiment of the present invention. This filter device 1 calculates the chamber coefficient filter unit 300 for filtering the input digital musical sound signal with the filtering characteristics corresponding to the two filter coefficients F and W to be set, and the filter coefficients F and W. And a coefficient calculation unit 200 that sets the obtained filter coefficients F and W in the filter unit 300.

  Then, the CPU 10 determines the cutoff frequency value of the low-pass filter (Fc: the frequency at which the amplification degree decreases by 3 dB) and the resonance value (Q: the peaking state of the frequency characteristics) by an operation according to a program stored in the ROM 12. (Shown parameter) is supplied as two parameters. At this time, the cutoff frequency value Fc and the resonance value Q are directly supplied to the coefficient calculation unit 200.

  The coefficient calculation unit 200 calculates the filter coefficients F and W of the filter unit 300 using the simple expression “F = (1 + Q) · Fc, W = (1−Q) · (2−Fc)” described above. The filter unit 300 is obtained and set. Therefore, since the equations for obtaining the filter coefficients F and W are simple, it is possible to obtain the filter coefficients F and W with an extremely simple arithmetic expression without requiring a coefficient table or the like.

  FIG. 4 shows a configuration example of the filter unit 300 (chamberline filter), which includes adders 90, 91, 93, and 95, delay elements 97 and 98, multipliers 92 and 94 having a filter coefficient F, and a filter coefficient W. And a multiplier 96. More specifically, the chamber filter includes an adder 90 that is a first adder, an adder 91 that is a second adder, and a first filter coefficient F (first filter coefficient). A multiplier 92 as a multiplier, an adder 93 as a third adder, a multiplier 94 as a second multiplier having a filter coefficient F (first filter coefficient), and a fourth adder Are connected in series, and the filter output is fed back to the adder 91 as the second adder and the adder 95 as the fourth adder by the delay element 98 as the first delay element. Further, the output of the adder 93 as the third adder is fed back to the adder 93 by the delay element 97 as the second delay element, and the filter coefficient W (second filter coefficient) is obtained. A multiplier 96 which is a third multiplier having Is configured to be fed back to the adder 90 is a first adder. The filter coefficients F and W of the multiplier 92, the multiplier 94, and the multiplier 96 are obtained and set by the coefficient calculation unit 200 by a simple expression.

  Then, as shown in FIG. 4, the filtered signal is output to the outside of the apparatus as a BRF (band reject filter) signal, an HPF (high pass filter) signal, a BPF (band pass filter) signal, and an LPF (low pass filter) signal. Thus, by setting the filter coefficients F and W to required values, a desired filtering signal of the tone signal is output.

(Operation)
The overall operation of the filter device 1 will be described as follows. The CPU 10 supplies two parameters Fc and Q. At this time, the parameter Fc and the parameter Q are directly supplied to the coefficient calculation unit 200. Next, when the parameter Fc and the parameter Q are given, the coefficient calculation unit 200 performs filtering using a simple filter coefficient calculation expression “F = (1 + Q) · Fc, W = (1-Q) · (2-Fc)”. The filter coefficients F and W of the unit 300 are obtained and set in the filter unit 300. The coefficient-adjusted filter unit 300 outputs a BRF signal, an HPF signal, a BPF signal, and an LPF signal as musical tone signals with the adjusted frequency characteristics.

  Therefore, according to this embodiment, the filter unit 300 filters the input digital musical sound signal with a filtering characteristic according to the set filter coefficients F and W, and the coefficient calculation unit 200 calculates the filter coefficients F and W. The coefficient calculation unit 200 directly receives the cut-off frequency value (Fc) that is the first parameter and the resonance value (Q) that is the second parameter, and obtains “F”. = (1 + Q) · Fc, W = (1−Q) · (2−Fc) ”, the two filter coefficients F and W are directly obtained, and the obtained filter coefficients F and W are set in the filter unit 300. Therefore, it is possible to realize a filter device that can obtain the filter coefficient of the chamber filter with a simple configuration that does not require a coefficient table and does not require an interpolation operation. .

  FIG. 9 is an explanatory diagram of the frequency characteristics of the LPF (low pass filter) output of the filter device 1 to which the present invention is applied. The horizontal axis represents frequency (unit Hz) and the vertical axis represents amplitude (unit dB). Compared to the above method (see FIG. 8), even if the cut-off frequency is increased, the frequency characteristic of the LPF (low-pass filter) output does not increase in the characteristic curve, and the filter coefficient can be adjusted. It can be seen that a low-pass filter characteristic can be obtained.

(Other embodiments)
FIG. 7 is a block diagram of a filter device 2 according to another embodiment. The filter device 2 is characterized in that the Fc index conversion table 100 is provided between the CPU 10 and the coefficient calculation unit 200 in the filter device 1 of FIG. 1, and the other structural points are the filter device of FIG. The same components as those of the filter device 1 are denoted by the same reference numerals. In this filter device 2, it is assumed that the present invention is applied to a musical instrument device or the like in which Fc changes exponentially according to a user's key input operation.

  That is, the filter device 2 calculates and obtains the chamber coefficient filter unit 300 that filters the input digital musical sound signal with the filtering characteristics according to the set filter coefficients F and W, and calculates the filter coefficient. A coefficient calculation unit 200 that sets filter coefficients (F, W) in the filter unit 300. The filter device 2 includes an Fc index conversion table 100 between the CPU 10 and the coefficient calculation unit 200. As shown in FIG. 2, this table 100 stores a plurality of sets of parameter Fc and a value of Fc raised to 2 (Fcq: a power value obtained by multiplying 2 (predetermined value) by the number indicated by Fc). . In the example illustrated in FIG. 2, “2 to the power of a”, “2 to the power of b”, and so on are stored in pairs for Fc “a”, “b”,.

  When Fc is given, the coefficient calculation unit 200 is configured to obtain a value of 2 Fc corresponding to this Fc with reference to the stored contents of the table 100, and can calculate a power value every time. It becomes unnecessary.

  Then, the CPU 10 determines the cutoff frequency value of the low-pass filter (Fc: the frequency at which the amplification degree decreases by 3 dB) and the resonance value (Q: the peaking state of the frequency characteristics) by an operation according to a program stored in the ROM 12. (Shown parameter) is supplied as two parameters. At this time, the resonance value Q is directly supplied to the coefficient calculation unit 200, while the cutoff frequency Fc is supplied to the Fc index conversion table 100. As described above, the coefficient calculation unit 200 obtains the value of 2 to the power of Fc corresponding to Fc with reference to the stored contents of the table 100.

  Then, the filter coefficients F and W are obtained by substituting the cutoff frequency value and the 2nd power of Fc (Fcq) into an expression defined using two parameters. Specifically, the value of Q in the expression “F = (1 + Q) · Fc, W = (1−Q) · (2−Fc)” described above is substituted with the value obtained directly from the CPU 10, By substituting “Fcq”, which is an Fc power value of 2 obtained from the Fc index conversion table 100, into the Fc value, two filter coefficients F and W are obtained. Therefore, even in this filter device 2, the formulas for obtaining the respective filter coefficients F and W are simple formulas, and therefore the filter coefficients F and W can be obtained with extremely simple arithmetic formulas without requiring a coefficient table or the like. It becomes possible.

(Operation)
The overall operation of the filter device 2 will be described as follows. The CPU 10 supplies two parameters Fc and Q. At this time, the parameter Q is directly supplied to the coefficient calculation unit 200, while the parameter Fc is supplied to the Fc index conversion table 100.

  Next, when Fc is given, the coefficient calculation unit 200 obtains a value of Fc raised to the power of 2 (Fcq) corresponding to this Fc with reference to the stored contents of the table 100. Then, the coefficient calculation unit 200 obtains the filter coefficients F and W of the filter unit 300 with a simple filter coefficient calculation expression “F = (1 + Q) · Fcq, W = (1−Q) · (2−Fcq)”. This is set in the filter unit 300. The coefficient-adjusted filter unit 300 outputs a BRF signal, an HPF signal, a BPF signal, and an LPF signal as musical tone signals with the adjusted frequency characteristics.

  Therefore, according to the present embodiment, the filter unit 300 that is a chamberin type filter filters the input digital musical tone signal with the filtering characteristics according to the set filter coefficients F and W, and the coefficient calculation unit 200 The filter coefficient is obtained and set in the filter unit 300. The coefficient calculation unit 200 refers to the stored contents of the Fc index conversion table 100 (storage unit), and is given a cutoff frequency value as the first parameter. A power value (Fcq) corresponding to (Fc) is obtained, and the resonance value (Q), which is the second parameter directly given, and the power value are expressed as “F = (1 + Q) · Fcq, W = (1− Q) · (2-Fcq) ”is substituted into the equation to obtain the filter coefficients F and W, so that the filter table with a simple configuration that eliminates the need for a coefficient table and no interpolation calculation It is possible to achieve. Moreover, it is extremely easy to apply the present invention to a musical instrument device or the like in which Fc changes exponentially according to a user's key input operation.

  Although the embodiments of the present invention have been described above, various modifications and changes can be made to the above-described embodiments without departing from the gist of the present invention. Further, the configuration of the filter device 1 can be realized by hardware such as a dedicated LSI as much as possible, and the function can be realized by the CPU (or DSP) executing the operation program as much as possible. It is. Further, in the Fc index conversion table 100 shown in FIG. 2, the Fc power of 2 is set, but a Fc power of a predetermined value other than 2 may be set.

  As described above, a chamber device filter coefficient can be obtained with a simple configuration, and a filter device suitable for processing musical tone signals and the like can be provided.

It is a lineblock diagram of filter device 1 of an embodiment of the invention. It is a block diagram of an Fc index conversion table. It is explanatory drawing of various filters and its transfer function. 2 is a configuration diagram of a filter unit 200. FIG. It is explanatory drawing of a prior art. It is explanatory drawing of the filter of the type which cascade-connected the 2nd stage of the primary IIR filter. It is a block diagram of the filter apparatus 2 of other embodiment. It is explanatory drawing of the frequency characteristic (LPF output) of a conventional system. It is explanatory drawing of the frequency characteristic (LPF output) of this system.

Explanation of symbols

1 Filter Device 2 Filter Device 10 CPU
12 ROM
90 Adder 91 Adder 93 Adder 94 Multiplier 95 Adder 96 Multiplier 97 Delay Element 98 Delay Element 100 Fc Exponential Conversion Table 200 Coefficient Operation Unit 300 Filter Unit

Claims (2)

A chamber type filter unit for filtering an input digital musical sound signal with a filtering characteristic according to two set filter coefficients (F, W), and the two filter coefficients (F, W) to obtain the filter A coefficient calculation unit to be set in the unit,
The chamberin type filter unit is:
A first adder (90), a second adder (91), a first multiplier (92) having a filter coefficient F, a third adder (93), and a filter coefficient F; A second multiplier (94) and a fourth adder (95) are connected in series, and the filter output is output by the first delay element (98) to the second adder (91) and the fourth adder (95). The second delay element (97) feeds back the output of the third adder (93) to the adder 93 and has a filter coefficient W. Configured to feed back to the first adder (90) through a multiplier (3) of 3;
The coefficient calculator is
The cutoff frequency value (Fc) that is the first parameter and the resonance value (Q) that is the second parameter are received, and “F = (1 + Q) · Fc, W = (1−Q) · (2− Fc) "directly determines the two filter coefficients (F, W) ,
The obtained filter coefficient F is given to the first multiplier (92) and the second multiplier (94), and the obtained filter coefficient W is given to the third multiplier (96). A filter device. A chamber type filter unit for filtering an input digital musical sound signal with a filtering characteristic according to two set filter coefficients (F, W), and the two filter coefficients (F, W) to obtain the filter A plurality of sets of coefficient calculation units to be set in the unit, a cutoff frequency value (Fc) that is the first parameter, and a power value obtained by multiplying that number by a predetermined value (Fc raised to a predetermined value: Fcq) A storage unit,
The chamberin type filter unit is:
A first adder (90), a second adder (91), a first multiplier (92) having a filter coefficient F, a third adder (93), and a filter coefficient F; A second multiplier (94) and a fourth adder (95) are connected in series, and the filter output is output by the first delay element (98) to the second adder (91) and the fourth adder (95). The second delay element (97) feeds back the output of the third adder (93) to the adder 93 and has a filter coefficient W. Configured to feed back to the first adder (90) through a multiplier (3) of 3;
The coefficient calculator is
Means for obtaining a power value (Fcq) corresponding to the cutoff frequency value (Fc) as the given first parameter with reference to the storage content of the storage unit;
The resonance value (Q) which is the given second parameter and the obtained power value are substituted into the formula “F = (1 + Q) · Fcq, W = (1−Q) · (2−Fcq)”. look including a means for determining the filter coefficients (F, W) Te,
The obtained filter coefficient F is given to the first multiplier (92) and the second multiplier (94), and the obtained filter coefficient W is given to the third multiplier (96). A filter device.

Images