JP2008076439A - Filter device and electronic musical instrument - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To attain a timbre variation equivalent to that of a natural musical instrument. <P>SOLUTION: In the filter device, a filter coefficient calculation circuit 26 has a parameter table. The parameter table stores a plurality of sets of filter coefficients associated with a first parameter based on a frequency and a second parameter based on respective plurality of levels representing a degree of attenuation or enhancement of a gain of a filter in filter characteristics. The filter coefficient calculation circuit 26 extracts a set of filter coefficients from the parameter table 35 with the use of the first parameter and the second parameter determined according to a frequency and a strength of a musical tone signal, and outputs the extracted set of filter coefficients to a filter means 22. A filter circuit 22 performs filter processing for a musical tone signal, based on the filter characteristics determined by the set of filter coefficients. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a filter device and an electronic musical instrument equipped with the filter device.

  The timbre change of a natural musical instrument is caused by the change in the amplitude ratio of the harmonic overtone. In particular, in the timbre change due to the strength of performance, the amplitude ratio of the higher harmonics increases as the performance is stronger (for example, the key is strongly pressed on a piano). On the other hand, the amplitude ratio of the higher harmonics decreases as the performance is weaker (the key is weaker on the piano).

  12 (a) and 12 (b) are graphs showing the waveform of an acoustic piano. FIG. 12 (a) is a waveform when a key is played with fortissimo, and FIG. 12 (b) is a waveform when a key is played with a meso piano. is there. FIGS. 13A and 13B are graphs showing spectra calculated from the waveforms of FIGS. 12A and 12B. 13A and 13B, the horizontal axis (frequency axis) is linear, and the vertical axis is dB. These are calculated based on 4096 samples from the vicinity of the start of sound generation of the waveforms shown in FIGS. 12A and 12B, which are cut out by the Blackmann window function.

FIG. 14 is a diagram showing a spectrum envelope extracted from the spectra shown in FIGS. 13 (a) and 13 (b). As shown in FIG. 14, it can be seen that the spectrum envelope has a substantially constant slope from the fundamental wave toward the harmonic, and the slope varies uniformly depending on the strength of the performance. In the example of FIG. 14, the inclination increases (that is, approaches 0) as the performance strength increases. Therefore, even when the timbre is changed by the filter circuit in the electronic musical instrument, it is desirable that the filter circuit has a filter characteristic having a spectrum envelope as shown in FIG.
JP-A-4-78213

  The transfer function of the conventional filter is shown in equation (1).

Ω ₀ = 2πfc / fs (0 <ω ₀ <1)
Here, fc is the cutoff frequency, fs is the sampling frequency, ω ₀ is the cutoff angular frequency, and Q is the selectivity.

In the conventional filter circuit, since the filter characteristic is changed by ω ₀ (or fc) and Q, the filter circuit accepts ω ₀ (or fc) and Q as parameters, and changes the characteristic based on this. It is configured to let you.

  FIG. 15 is a graph showing filter characteristics when fc is changed as a parameter in the second-order IIR filter having the characteristics shown in the equation (1). Also in FIG. 15, the horizontal axis (frequency axis) is linear (0 to 10 kHz), and the vertical axis is dB. As shown in FIG. 15, even if fc is changed, the changed filter characteristics are far from the characteristics shown in FIGS. 13 (a), (b) and FIG. 14, and therefore the cutoff frequency is controlled. Therefore, there is a problem that it is very difficult to realize a change equivalent to a change in piano tone.

  Currently, many electronic musical instruments adopt the PCM method. In this PCM method, the waveform at the time of the occurrence of a musical sound (at the time of attack) and the waveform data of the waveform for repetition thereafter are stored in the waveform memory, and after the time of attack, the waveform of the waveform for repetition is stored. By repeatedly reading the waveform data, the capacity of the waveform memory is reduced. However, since the timbre is fixed and monotonous only by repeating the same waveform, the timbre is changed over time by the filter circuit. In particular, an acoustic piano attenuates faster with higher harmonics over time. Such a change cannot be realized in a conventional filter circuit.

  For example, Patent Document 1 proposes a filter that uses, as parameters, a specific frequency f0 at which the transfer characteristic starts to change, and its rate of change, instead of the cutoff frequency fc. However, even if the filter disclosed in Patent Document 1 is used, it has been particularly difficult to realize a change equivalent to a change in piano timbre. In particular, there is a problem in that a frequency (hereinafter referred to as “transition frequency” in this specification) having a predetermined ratio from the maximum level of attenuation or enhancement at a predetermined frequency moves due to a change in parameters.

  An object of the present invention is to provide a filter device and an electronic musical instrument that can realize a timbre change equivalent to that of a natural musical instrument, in particular, an acoustic piano.

An object of the present invention is to provide filter coefficient output means for outputting a set of filter coefficients, and a set of filter coefficients output from the filter coefficient output means for a musical sound signal having a predetermined frequency supplied from the outside. Filter means for performing a filtering process based on a specified filter characteristic,
For each of a plurality of frequencies, the filter coefficient output means uses a first parameter based on the frequency and a second parameter based on each of a plurality of levels representing the degree of attenuation or enhancement of the filter gain in the filter characteristics. By the associated parameter table storing a plurality of types of filter coefficient sets in the filter means, and the first parameter and the second parameter determined corresponding to the frequency and intensity of the supplied musical sound signal, This is achieved by a filter device comprising: filter coefficient generation means for extracting a corresponding filter coefficient set from the parameter table and outputting it to the filter means.

In a preferred embodiment, the parameter table is defined by the frequency of the supplied musical sound signal, and rises and falls with the level of the frequency, and from the reference frequency from which attenuation or enhancement of the filter gain is started. The filter characteristic is attenuated or enhanced with a substantially constant inclination, and the inclination changes so as to increase as the intensity of the supplied musical sound signal increases.
Transition at a predetermined frequency with a maximum level of gain attenuation or enhancement from the reference frequency as a second parameter and a frequency at which the gain level is a predetermined ratio from the maximum level For each of a plurality of first parameters, a set of filter coefficients based on the first parameter and each of a plurality of second parameters having different maximum levels is stored with the frequency as the first parameter.

  In another preferred embodiment, the filter coefficient generating means further comprises an envelope generating means for changing the second parameter with the passage of time.

  In a more preferred embodiment, the envelope generating means adds or subtracts a value that changes over time with respect to the second parameter.

  In another preferred embodiment, the parameter table receives a plurality of bits of data consisting of a first upper bit of the first parameter and a second upper bit of the second parameter as input, and corresponds to the parameter table. A set of filter coefficients, and the filter coefficient generation means further outputs the filter coefficient output from the parameter table for each of the filter coefficients included in the set of filter coefficients, as the first parameter of the first parameter. Interpolating means for interpolating with predetermined first lower bits excluding the higher-order bits and predetermined second lower-order bits excluding the second higher-order bits of the second parameter.

  In a more preferred embodiment, the interpolation means includes a filter coefficient obtained by changing the first upper bit, a difference value thereof, a first interpolation value obtained by the second lower bit, and a second The filter coefficient obtained by changing the first higher order bit, the difference value thereof, the second interpolation value obtained by the second lower order bit, and the first higher order bit, And the third interpolation value obtained from the difference value between the second interpolation value and the second interpolation value and the first lower order bit, and the third interpolation value is output as an interpolated filter coefficient.

  In another preferred embodiment, the apparatus further comprises control means for generating the second parameter based on the intensity of the supplied musical tone signal.

Further, the object of the present invention is to provide sound generation instruction means for instructing the frequency and intensity of a musical sound to be generated by an external operation,
Control means for generating a first parameter based on the frequency and calculating a second parameter based on the intensity;
A parameter table storing a plurality of sets of filter coefficients associated with a plurality of levels representing the degree of attenuation or enhancement of the gain of the filter in the frequency and filter characteristics for each of the plurality of frequencies;
Filter coefficient generation means for extracting from the parameter table a set of filter coefficients corresponding to the frequency and level determined based on the first and second parameters calculated by the control means;
Filter means for performing a filtering process based on a filter characteristic defined by a set of filter coefficients from the filter coefficient generating means and outputting the musical sound signal from the music sound generating means;
An electronic musical instrument characterized by comprising:

  In a preferred embodiment, the filter coefficient generation means further includes envelope generation means for changing the second parameter with the passage of time.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to provide the filter apparatus and electronic musical instrument which can implement | achieve the timbre change equivalent to the timbre change of a natural musical instrument, especially an acoustic piano.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a block diagram showing an outline of an electronic musical instrument according to an embodiment of the present invention. As shown in FIG. 1, the electronic musical instrument includes a microcomputer 1, a ROM (Read Only Memory) 2, a RAM (Random Access Memory) 3, switches 4, a touch detection circuit 5, a keyboard 6, a tone generation circuit 7, and a waveform ROM 8. , A D / A converter (DAC) 9, an amplifier circuit 10, and speakers 11 and 12. The microcomputer 1, the ROM 2, the RAM 3 and the tone generation circuit 7 are connected to the data bus 13, and the touch detection circuit 5 is connected to the microcomputer 1.

  The microcomputer 1 controls the entire electronic musical instrument, reads the program and data from the ROM 2 storing the program and data, and executes the program. Data generated by executing the program is stored in the RAM 3 which is a work area. The switches 4 are arranged on the console panel of the electronic musical instrument. The microcomputer 1 detects the operation of the switches 4 by the performer. The touch detection circuit 5 sends a scanning signal to the keyboard 6 at a predetermined timing, and in response to turning on the two switches arranged on the keys of each keyboard 6, performance operation data (pitch and touch response data). ) And output to the microcomputer 1. In the present embodiment, the keyboard 6 has 88 keys, and each key is provided with two switches in the longitudinal direction. When the key is pressed, the first switch is first turned on. When pressed, the second switch is turned on.

  The microcomputer 1 controls the musical tone generation circuit 7 based on the tone color designated by the operation of the switches 4 and the performance operation data including the touch response data and the pitch output from the touch detection circuit 5. A predetermined musical tone is generated. The musical tone generation circuit 7 reads the waveform data of the tone color designated from the waveform ROM 8, generates a musical tone having a pitch and volume (velocity) according to the performance operation data, and outputs it to the DAC 9. The DAC 9 converts the digital data output from the tone generation circuit 7 into an analog signal. The analog signal is emitted from the speakers 11 and 12 via the amplifier circuit 10.

  FIG. 2 is a block diagram showing the musical tone generating circuit according to the present embodiment in more detail. As shown in FIG. 2, the tone generation circuit 7 includes an interface 20, a waveform generation circuit 21, a filter circuit 22, a multiplication circuit 23, a mixer 24, a first envelope generation circuit 25, a filter coefficient calculation circuit 26, and a second envelope. A generation circuit 27 is provided. The interface 20, waveform generation circuit 21, filter circuit 22, mixer 24, first envelope generation circuit 25, filter coefficient calculation circuit 26 and second envelope generation circuit 27 are connected to the internal bus 28.

  The interface 20 is connected to the data bus 13 shown in FIG. 1, and writes setting data and the like to each arithmetic unit such as the waveform generation circuit 21 and the filter circuit 22 in the musical tone generation circuit 7 via the internal bus 28. The waveform generation circuit 21 is connected to the waveform ROM 8, reads PCM data from a predetermined address of the waveform ROM 8, and generates musical tone waveform data having a frequency corresponding to the pitch in the performance operation data.

  The filter coefficient circuit 26 is a filter coefficient according to a parameter based on a performance operation signal given from the microcomputer 1 via the interface 20 and a first envelope signal output from the first envelope generation circuit 25 and time-varying. Is calculated. The frequency characteristic is changed by the first envelope signal, and the filter coefficient changes with time. The filter circuit 22 performs a filter process on the musical sound waveform data according to the filter coefficient.

  The multiplier 23 multiplies the musical sound waveform data and the second envelope signal that is output from the second envelope generation circuit 27 and changes with time. Thereby, the rising and falling of the musical sound according to the on / off of the key of the keyboard 6 and the volume according to the touch response data are controlled. Note that the waveform generation circuit 21, the filter circuit 22, the first envelope generation circuit 25, the filter coefficient calculation circuit 26, the multiplier 23, and the second envelope generation circuit 27 perform processing of the maximum simultaneous sound number in a time division manner. Thus, it is possible to generate musical sounds with a sufficient number of channels for keyboard performance. The mixer 24 accumulates the generated musical tones of the maximum number of simultaneous pronunciations with predetermined weights, respectively, and finally distributes the musical tones of the two left and right channels. The output of the mixer 24 is output to the DAC 9.

  The outline of the filter circuit 22 and the filter coefficient calculation circuit 26 according to the present invention will be described below. FIG. 3 is a diagram for explaining ideal filter characteristics for realizing a timbre filter of an acoustic piano depending on performance strength. In FIG. 3, the vertical axis represents gain, and the horizontal axis represents linear frequency.

As shown in FIG. 3, the ideal filter characteristic is inferred from the spectrum envelope of the acoustic piano shown in FIG. 13, and the harmonic component of the fundamental wave is almost linear from a certain point based on the fundamental wave. It is like decaying. In addition, the inclination is small when the volume is low (that is, the degree of right-down is large: the inclination is small), and increases as the volume increases (that is, the degree of right-down is small and becomes more horizontal). : Increasing the inclination).
In the following, the point (frequency) at which linear attenuation or enhancement starts based on the fundamental wave is the reference frequency, the attenuation (dB) at a frequency half the sampling frequency fs is the gain, and the gain is a predetermined ratio. Such a frequency (in this embodiment, a frequency that is ½ of the attenuation at fs / 2) is referred to as a transition frequency. The degree of linear attenuation or enhancement from the reference frequency is referred to as inclination. For the slope, the frequency axis is the x-axis and the attenuation is the y-axis. Therefore, the larger the negative attenuation amount, the smaller the slope in the frequency characteristic. The closer the negative attenuation amount is to 0, the larger the slope becomes and approaches 0. The greater the positive current guess, the greater the slope in the frequency characteristic.

  In the present embodiment, the reference frequency changes as the fundamental wave changes. In addition, the transition frequency changes according to a certain rule as the reference frequency changes.

  In this embodiment, as will be described later, the filter circuit is controlled by a transition frequency that changes based on the fundamental wave and a gain that indicates the depth of the filter. This realizes control of harmonic components with respect to the fundamental wave, which is almost the same as the actual timbre change of an acoustic piano.

In the present embodiment, it is considered that not only the harmonic components are attenuated but also enhanced. This is to change the timbre from the PCM data of the waveform ROM 8 over a wider range. An FIR (Finite Impulse Response) filter is suitable for realizing such characteristics. However, the FIR filter has a large calculation amount and requires large-scale hardware. Therefore, in the present embodiment, a characteristic approximate to the ideal filter characteristic shown in FIG.
Impulse Resonance (infinite impulse response) filter is used, and the filter coefficient is controlled to be almost equal to the ideal filter characteristics.

  In order to obtain a filter having the above filter characteristics, various methods are conceivable. In order to minimize the amount of calculation, an approximation with a first-order IIR filter is attempted, and the ideal is obtained by extending the approximation. A frequency specification close to is obtained. First, the transfer function of the first-order IIR filter is considered as equation (2).

In equation (2), if the maximum gain (that is, the maximum gain at a frequency that is ½ of the sampling frequency fs) is A, A can be expressed as equation (3).

If the gain is ½ of the maximum gain at the transition frequency f0, the equation (4) is derived.

When the above equations (3) and (4) are solved, the coefficients b and c are as shown in equations (5) and (6).

Therefore, the coefficients b and c of the first-order IIR filter shown in the equation (2) can be obtained from the maximum gain A and the transition frequency f0 as shown in the equations (5) and (6). Although there are two sets of filter coefficients from the above equation (5), it is only necessary to select one that is easier to control from the coefficient range.

  The first-order IIR filter has too slow attenuation characteristics. Therefore, in the present embodiment, two primary filters IIR having different transition frequencies are connected in series to form a secondary IIR filter, thereby obtaining a filter characteristic more approximate to an ideal filter. Equation (7) is a diagram showing a transfer function of a secondary IIR filter in which primary IIR filters are connected in series.

FIG. 4 shows the characteristics of the second-order IIR filter when the gain is changed at a certain transition frequency. It can be understood that the characteristics shown in FIG. 4 approximate the ideal characteristics shown in FIG. In the present embodiment, the ideal filter is approximated by a secondary IIR filter. However, by increasing the order of the filter, the characteristic can be further approximated to the characteristic of the ideal filter.

  FIG. 5 is a graph showing second-order filter coefficients for realizing the filter characteristics shown in FIG. The horizontal axis represents the negative gain level. Level 1 represents a gain having the largest negative level (that is, the slope is minimum), and level 0 represents a gain having the smallest negative level (that is, the slope is 0). Thus, in the present embodiment, the filter coefficients b, c, d, e, and f are determined at levels 1 to 9, respectively.

  The filter coefficient for the positive gain can be calculated by exchanging the denominator and numerator of the above equation (7) (see equation (8)).

Therefore, the negative gain coefficients b to f are
b: e / d, c: f / d, d: 1 / d, e: b / d, f: c / d
Can be obtained to obtain a filter coefficient having a positive gain characteristic.

  In this embodiment, this filter coefficient is not calculated by calculation, but a table (parameter table) with transition frequency and gain as addresses is provided in the filter coefficient calculation circuit 26, and a data value is extracted from the table. An appropriate filter coefficient is calculated by interpolating the data value.

FIG. 11 is a diagram for explaining filter characteristics and a parameter table in the present embodiment. FIG. 11A is a diagram illustrating filter characteristics at the transition frequency F ₁ based on a certain reference frequency f ₁ . As described above, the reference frequency f ₁ is determined according to the frequency of the fundamental wave. For example, the reference frequency may be substantially the same as the fundamental wave, or may be a frequency that is a predetermined multiple of the fundamental wave. As shown in FIG. 11 (a), the single transition frequencies F _1, is considered a filter characteristic having a plurality of tilt. In the example of FIG. 11A, gain levels G ₁₁ , G ₁₂ ,... G _1i (= attenuation “0”) from the smallest negative characteristic (attenuating characteristic), and G ₁ as the positive characteristic. _{(I + 1)} ... G1 _(2i-1) is shown. Further, considering the case where the reference frequency is changed from f ₁ to f ₂ (f ₁ <f ₂ ) as the fundamental wave is different, as shown in FIG. 11B, the reference frequency is changed. Te, the transition frequency _{F 2} is changed. In addition, gain levels G ₂₁ , G ₂₂ ,..., G _2i ,..., G _{2 (2i−1)} can also be changed.

For example, in FIG. 11 (a), the the single transition frequencies _{F 1,} 2i-1 single level exists. Therefore, in this example, there is one set of filter coefficients (b, c, d, e, f) for a combination of the transition frequency F _{1 and a} certain level. That is, if there are 2i-1 levels, there are 2i-1 filter coefficient pairs. The same applies to FIG. 11B. Therefore, if there are n types of transition frequencies, there may be a set of n × (number of levels ((2i−1) in the example of FIG. 11)) filter coefficients.

  If a set of filter coefficients for each gain is held for each transition frequency, it is necessary to prepare a table of enormous size (number of transition frequency settings × number of gain settings × type of filter coefficient (5)). Therefore, by limiting the number of transition frequency settings and the number of gain settings, prepare only a table that stores a series of filter coefficients according to these combinations, and interpolate both the transition frequency and gain (two-dimensional interpolation). By realizing the above, the size of the table is reduced and the filter coefficient is optimized.

  Hereinafter, the configurations of the filter coefficient calculation circuit 26 and the filter circuit 22 according to the present embodiment will be described in more detail.

  FIG. 6 is a block diagram showing the configuration of the filter coefficient calculation circuit according to this embodiment. As shown in FIG. 6, the filter coefficient calculation circuit 26 includes an interface 31, adders 32 to 34, a parameter table 35, an interpolation circuit 36, and registers 37 to 41. Although not shown in FIG. 6, the filter coefficient calculation circuit 26 controls to output address signal increment signals X1 and Y1 to the parameter table 35 and the selection signal SEL of the parameter table 35 at a predetermined timing. It has a circuit.

  The interface 31 is connected to the internal bus 28 of the tone generation circuit 7 and is transmitted from the microcomputer 1 via the interface 20, that is, transition frequency data F [15: 0] and gain data G [15 : 0] is received, held, and output to the filter calculation circuit 26. The transition frequency data F corresponds to the transition frequency described above, and the gain data G corresponds to the gain. The gain data G [15: 0] is given to the adder 32. The adder 32 is supplied with the first envelope signal that changes over time from the first envelope generation circuit 25 as the other input. Therefore, the value of the gain data G [15: 0] output from the adder changes with time based on the first envelope signal.

  Of the gain data G [15: 0], G [15:12] is supplied to the adder 33. Further, F [15:13] of the transition frequency data F [15: 0] is supplied to the adder 45. In the adders 33 and 34, the increment signal “1” is added at a predetermined timing as necessary, and is output as an address of the parameter table 35. The adder 33 can output a continuous address (G [15:12] +1) by adding the increment signal X1 at the timing next to the output of G [15:12]. The adder 34 can also output a continuous address (F [15:13] +1) by adding the increment signal Y1 at the next timing of the output of F [15:13].

  A signal from the adder 33 is given to the lower address A [4: 0] of the parameter table 35, and a signal from the adder 34 is given to the next higher address A [8: 5]. Further, a selection signal SEL from a control circuit (not shown) is given to the highest address A [11: 9].

  In the present embodiment, the parameter table 35 includes 9 types of transition frequencies × 17 types of gains (8 types of negative gain, gain “0”, and 9 types of positive gains) × 5 types of filter coefficients. Remember. Therefore, the parameter table 35 stores the number of filters of 9 × 17 × 5 = 765 words.

  Regarding the transition frequency, nine types of addresses are prepared in order to interpolate the upper 3 bits (8 types) with the lower 13 bits. As for the gain, 17 kinds of addresses are prepared in order to complement the upper 4 bits (16 kinds) with the lower 12 bits. The selection signal SEL is used to select one of the five types of coefficients b to f.

  The output of the parameter table 35, that is, the set of filter coefficients before interpolation is given to the interpolation circuit 36. Further, the lower-order data F [12: 0] of the transition frequency data and the lower-order data G [11: 0] of the gain data are also supplied to the interpolation circuit 36. In the interpolation circuit 36, five types of post-interpolation filter coefficients b to f are calculated and output. The registers 37 to 41 hold the filter coefficients b to f after interpolation that are sequentially output, respectively.

  FIG. 7 is a block diagram showing the interpolation circuit according to this embodiment in more detail. As shown in FIG. 7, the interpolation circuit 36 includes registers 51 to 54, a subtractor 55, a multiplier 56, an adder 57, a register 58, a subtractor 65, a multiplier 66, an adder 67, a register 68, and a subtractor 70. , A multiplier 71, and an adder 72.

  The subtractor 55 calculates a difference value obtained by subtracting the output of the register 51 from the output of the register 51, and the multiplier 56 multiplies the difference value by the lower data G [11: 0] of the gain data. The value is output to the adder 57. The adder 57 adds the multiplication value and the output from the register 51, and the added value is stored in the register 58.

  Similarly, the subtracter 65 calculates a difference value obtained by subtracting the output of the register 54 from the output of the register 53, and the multiplier 66 multiplies the difference value by the lower data G [11: 0] of the gain data. The multiplied value is output to the adder 67. The adder 67 adds the multiplication value and the output from the register 53, and the added value is stored in the register 68.

  Further, the subtractor 70 calculates a difference value obtained by subtracting the output of the register 68 from the output of the register 58, and the multiplier 71 multiplies the difference value by the lower-order data F [12: 0] of the transition frequency data. The multiplied value is output to the adder 72. The adder 72 adds the multiplication value and the output from the register 58. The output from the adder 72 becomes the interpolated filter coefficient.

  FIG. 8 is a block diagram showing an outline of the filter circuit according to the present embodiment. As shown in FIG. 8, the filter circuit 22 includes adders 80 and 88, multipliers 81, 82, 85, 86 and 87, and delay circuits 83 and 84. Filter coefficients b, c, d, e, and f calculated by the filter coefficient calculation circuit 22 are given to the multipliers 81, 82, 85, 86, and 87, respectively, and applied to respective inputs of the multipliers. The signal and the filter coefficient are multiplied.

  Hereinafter, two-dimensional interpolation according to the present embodiment will be described. FIG. 9 is a diagram for explaining the two-dimensional interpolation according to the present embodiment. For convenience of explanation, it is assumed that x = G [15:12], dx = G [11: 0], y = F [15:13], and dy = F [12: 0]. The following values are stored in the registers 51 to 54 of FIG.

Register 51: o [x, y] (where o [x, y] represents the output of the parameter table 35 at the address xy)
Register 52: o [x + 1, y]
Register 53: o [x, y + 1]
Register 54: o [x + 1, y + 1]
The subtracter 55 subtracts the output o [x, y] of the register 51 from the output o [x + 1, y] of the register 52, and the multiplier 56 multiplies the subtracted value by dx. Thereafter, the adder 57 adds the multiplication value and o [x, y] and stores them in the register 58. Therefore, the register 58 stores the following values.

Register 58: o [x, y] + (o [x + 1, y] −o [x, y]) * dx
= O [x + dx, y]
As shown in FIG. 9, this corresponds to linear interpolation between o [x, y] and o [x + 1, y] based on dx.

  The subtracter 65 subtracts the output o [x, y + 1] of the register 53 from the output o [x + 1, y + 1] of the register 54, and the multiplier 66 multiplies the subtracted value by dx. Thereafter, the adder 67 adds the multiplication value and o [x, y + 1] and stores them in the register 58. Therefore, the register 68 stores the following values.

Register 68: o [x, y + 1] +
(O [x + 1, y + 1] −o [x, y + 1]) * dx
= O [x + dx, y + 1]
As shown in FIG. 9, this corresponds to linear interpolation between o [x, y + 1] and o [x + 1, y + 1] based on dx.

  Further, the subtracter 70 subtracts the output o [x + dx, y] of the register 58 from the output o [x + dx, y + 1] of the register 68, and the multiplier 71 multiplies the subtracted value by dy. Thereafter, the adder 72 adds the multiplication value and o [x + dx, y] and outputs the result. Therefore, the output value is as follows.

Output value: o [x + dx, y] +
(O [x + dx, y + 1] −o [x + dx, y]) * dy
= O [x + dx, y + dy]
As shown in FIG. 9, the output value corresponds to an output value interpolated between o [x + dx, y] and o [x + dx, y + 1] based on dy. Therefore, the obtained output value is a two-dimensional interpolation value based on dx and dy.

  In this embodiment, by sequentially changing the selection signal SEL and outputting the pre-interpolated values of the filter coefficients b, c, d, e, and f from the parameter table 35, the interpolation circuit 36 Interpolated values of c, d, e, and f are generated and output.

  FIG. 10 is a timing chart for calculating filter coefficients. 10, R0 to R5 correspond to the registers 51 to 54, 58, and 68 in FIG. 7, respectively, and B to F correspond to the registers 37 to 41 in FIG. 6, respectively.

  In FIG. 10, the address A [4: 0] is switched to “x” or “x + 1” from the signal X1, and the address A [8: 5] is changed to “y” or “y + 1” by the signal Y1. Can be switched to. Further, the address A [9:11] is changed by the selection signal SEL. In one cycle of the first Y1 (see reference numeral 1001), the selection signal SEL indicates “b”. Therefore, the filter coefficient b is determined by four types of addresses given to the parameter table 35 in the cycle of the first Y1. O [x, y], o [x + 1, y], o [x, y + 1] and o [x + 1, y + 1] are output and stored in R0 to R3.

  Based on o [x, y] and o [x + 1, y] stored in R0 and R1, respectively, o [x + dx, y] is calculated and stored in R4. Further, o [x + dx, y + 1] is calculated based on o [x, y + 1] and o [x + 1, y + 1] stored in R2 and R3, respectively, and stored in R5. Based on o [x + dx, y] and o [x + dx, y + 1] stored in R4 and R5, an interpolated value b [x + dx, y + dy] for the filter coefficient b is calculated and stored in B. .

  Similarly, when the selection signal SEL indicates “c”, “d”, “e”, and “f”, finally, the interpolated values c [x + dx, y + dy] of the filter coefficients c to f are finally obtained. ], D [x + dx, y + dy], e [x + dx, y + dy] and f [x + dx, y + dy] are calculated, and the calculated values are stored in C to F, respectively.

  According to this embodiment, the parameter table for outputting the filter coefficient based on the transition frequency and the gain is provided, and the filter circuit is controlled by the filter coefficient using the parameter table. The transition frequency is based on the frequency of the fundamental wave of the musical sound to be generated, and the gain indicates the degree of attenuation (or the degree of enhancement). Therefore, it is possible to realize a digital filter having a filter characteristic very close to the filter characteristic of an acoustic instrument, in particular, an acoustic piano, with parameters that are easy to understand intuitively and without performing complicated calculations. .

  Further, according to the present embodiment, the envelope data that changes with time is output from the first envelope generation circuit, and this envelope data is added to the gain data by the filter coefficient calculation circuit. Therefore, it is possible to realize a change in time of gain attenuation (or enhancement) by a filter and a desired change in tone color with a simple circuit configuration.

  Furthermore, according to the present embodiment, the upper bits of the transition frequency data and the upper bits of the gain data are used as addresses, the parameter table outputs a coarse filter coefficient, and the lower bits of the transition frequency data and the lower bits of the gain data Thus, the coarse filter coefficient is two-dimensionally corrected to obtain the corrected filter coefficient. As a result, it is possible to obtain an appropriate filter coefficient with high accuracy without increasing the number of data (coefficients) stored in the parameter table.

  The present invention is not limited to the above embodiments, and various modifications can be made within the scope of the invention described in the claims, and these are also included in the scope of the present invention. Needless to say.

FIG. 1 is a block diagram showing an outline of an electronic musical instrument according to an embodiment of the present invention. FIG. 2 is a block diagram showing the musical tone generating circuit according to the present embodiment in more detail. FIG. 3 is a diagram for explaining ideal filter characteristics for realizing a timbre filter of an acoustic piano depending on performance strength. FIG. 4 shows the characteristics of the second-order IIR filter when the gain is changed at a certain transition frequency. FIG. 5 is a graph showing second-order filter coefficients for realizing the filter characteristics shown in FIG. FIG. 6 is a block diagram showing the configuration of the filter coefficient calculation circuit according to this embodiment. FIG. 7 is a block diagram showing the interpolation circuit according to this embodiment in more detail. FIG. 8 is a block diagram showing an outline of the filter circuit according to the present embodiment. FIG. 9 is a diagram for explaining the two-dimensional interpolation according to the present embodiment. FIG. 10 is a timing chart when calculating the filter coefficient according to the present embodiment. FIG. 11 is a diagram for explaining filter characteristics and a parameter table in the present embodiment. 12A and 12B are graphs showing waveforms of an acoustic piano. FIGS. 13A and 13B are graphs showing spectra calculated from the waveforms of FIGS. 12A and 12B. FIG. 14 is a diagram showing a spectrum envelope extracted from the spectra shown in FIGS. 13 (a) and 13 (b). FIG. 15 is a graph showing filter characteristics when fc is changed as a parameter in a conventional feedback secondary filter.

Explanation of symbols

1 Microcomputer 2 ROM
3 RAM
4 Switches 5 Touch detection circuit 6 Keyboard 7 Musical sound generation circuit 8 Waveform ROM
9 DAC
DESCRIPTION OF SYMBOLS 10 Amplification circuit 11, 12 Speaker 20 Interface 21 Waveform generation circuit 22 Filter circuit 23 Multiplier 24 Mixer 25 1st envelope generation circuit 26 Filter coefficient calculation circuit 27 2nd envelope generation circuit

Claims (9)

Filter coefficient output means for outputting a set of filter coefficients;
Filter means for applying a filtering process based on a filter characteristic defined by a set of filter coefficients output from the filter coefficient output means to a musical sound signal having a predetermined frequency supplied from the outside A device,
The filter coefficient output means comprises:
For each of a plurality of frequencies, the filter means associated with a first parameter based on the frequency and a second parameter based on each of a plurality of levels representing the degree of attenuation or enhancement of the filter gain in the filter characteristics A parameter table storing a plurality of sets of filter coefficients in
Based on the first parameter and the second parameter determined in accordance with the frequency and intensity of the supplied tone signal, a corresponding filter coefficient set is extracted from the parameter table and output to the filter means. Filter coefficient generation means;
A filter device comprising: The parameter table is
Specified by the frequency of the musical sound signal to be supplied, increased or decreased with the level of the frequency, and attenuated or enhanced with a substantially constant slope from the reference frequency from which attenuation or enhancement of the filter gain is started, And so that the inclination becomes a filter characteristic that changes so as to increase as the intensity of the supplied tone signal increases.
Transition at a predetermined frequency with a maximum level of gain attenuation or enhancement from the reference frequency as a second parameter and a frequency at which the gain level is a predetermined ratio from the maximum level With frequency as the first parameter,
The set of filter coefficients based on the first parameter and each of a plurality of second parameters having different maximum levels is stored for each of the plurality of first parameters. Filter device.   3. The filter device according to claim 1, wherein the filter coefficient generation unit further includes an envelope generation unit that changes the second parameter as time elapses.   4. The filter device according to claim 3, wherein the envelope generating means adds or subtracts a value that changes with the passage of time with respect to the second parameter. The parameter table receives a plurality of bits of data consisting of a first upper bit of the first parameter and a second upper bit of a second parameter, and outputs the corresponding set of filter coefficients.
The filter coefficient generation means further includes, for each of the filter coefficients included in the set of filter coefficients.
The filter coefficient output from the parameter table has a predetermined first low-order bit excluding the first high-order bit of the first parameter and a predetermined high-order bit excluding the second high-order bit of the second parameter. 5. The filter device according to claim 1, further comprising an interpolation unit configured to interpolate using the second lower-order bit. 6.   The interpolation means changes a filter coefficient obtained by changing the first higher order bit, a difference value thereof, a first interpolation value obtained by the second lower order bit, and a second higher order bit. In addition, the filter coefficient obtained by changing the first upper bit, the difference value thereof, the second interpolation value obtained by the second lower bit, the first interpolation value, and the second 6. The difference value between the interpolation values of the first interpolation value and the third interpolation value obtained by the first lower-order bit is calculated, and the third interpolation value is output as an interpolated filter coefficient. The filter device according to 1.   7. The filter device according to claim 1, further comprising a control unit that generates the second parameter based on the intensity of the supplied musical sound signal. Pronunciation instruction means for instructing the frequency and intensity of a musical tone to be generated by an external operation;
Control means for generating a first parameter based on the frequency and calculating a second parameter based on the intensity;
A parameter table storing a plurality of sets of filter coefficients associated with a plurality of levels representing the degree of attenuation or enhancement of the gain of the filter in the frequency and filter characteristics for each of the plurality of frequencies;
Filter coefficient generation means for extracting from the parameter table a set of filter coefficients corresponding to the frequency and level determined based on the first and second parameters calculated by the control means;
Filter means for performing a filtering process based on a filter characteristic defined by a set of filter coefficients from the filter coefficient generating means and outputting the musical sound signal from the music sound generating means;
An electronic musical instrument characterized by comprising:   9. The electronic musical instrument according to claim 8, wherein the filter coefficient generation unit further includes an envelope generation unit that changes the second parameter as time elapses.