JP3551575B2 - Tone generator - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a tone generator that generates a tone based on a sounding mechanism of a wind instrument such as a clarinet.
[0002]
[Prior art]
In recent years, various devices have been developed to operate a model simulating the sounding mechanism of a natural musical instrument and thereby generate a musical sound of a natural musical instrument. -293898.
According to this type of device, a closed loop is formed by an excitation unit that simulates the non-linear operation of a portion consisting of a mouthpiece and a lead of a wind instrument, and a tube unit that simulates the transmission characteristics of a resonance tube that is a tube of the wind instrument. When the excitation signal corresponding to the blowing pressure is injected from the excitation section into the closed loop, signal circulation is excited in the closed loop, whereby a tone signal having a pitch (pitch) corresponding to the resonance operation of the resonance tube is obtained. Can be
[0003]
[Problems to be solved by the invention]
By the way, the above-described conventional tone generating device has a non-linear table for simulating a non-linear operation of a portion including a mouthpiece and a lead. This non-linear table converts the pressure change in the pipe into the above-mentioned excitation signal excited by the vibration of the reed, and uses the pressure change in the pipe due to the air flow passing through the gap between the mouthpiece and the reed as a read address. A pressure is generated that is excited by vibration corresponding to a large displacement.
In such a table conversion, conversion characteristics are stored in a ROM or the like. However, in the signal processing for tone generation, the input address length often exceeds 16 bits in order to improve processing accuracy. There is an adverse effect that the table has a storage capacity.
[0004]
By the way, in recent years, a self-excited vibration type sound source and a well-known waveform readout type sound source by the above-mentioned conventional musical sound generating device are mounted, and one of the sound sources is used depending on the type of musical sound to be generated. Devices are being devised. In such a case, both the self-excited vibration type sound source and the waveform readout type sound source require a configuration for reading a “table” or a “waveform memory”. In view of the above-mentioned problems, in order to reduce the table storage capacity or the waveform memory storage capacity, interpolation means that can be shared with a self-excited vibration type or waveform reading type sound source has been desired.
[0005]
Therefore, the present invention has been made in view of the above-described circumstances, and aims at reducing the table storage capacity or the waveform memory storage capacity, and further includes a tone generator having interpolation means that can be shared with a self-excited vibration type or waveform readout type sound source. It is intended to provide a generator.
[0006]
In order to achieve the above object, in the invention described in claim 1,Exciting means for generating an excitation signal based on the non-linear characteristics of the sounding body, and an exciting signal output from the exciting means for simulating the transmission characteristics of the resonator by circulating repeatedly while delaying the excitation signal for a predetermined time. A wave guide means for generating a musical tone having a pitch corresponding to the resonance operation of the musical instrument, wherein the exciting means comprises a signal corresponding to a blowing pressure supplied from the outside and a return path of the wave guide means. A first calculating means for calculating a pressure difference signal representing a difference from a signal corresponding to a reflected pressure wave fed back via the first calculating means, and a pressure difference signal supplied from the first calculating means in accordance with a blowing mode. Low-pass filter means for low-pass filtering based on the obtained resonance value and cutoff frequency, and a non-linear transformation of a signal from the low-pass filter means. A non-linear table means, and a second calculating means for multiplying a signal from the non-linear table means by the pressure difference signal and supplying the multiplied signal to the transmission outward path of the wave guide means. Having a storage unit for storing the nonlinear characteristic of the sounding body, a signal from the low-pass filter unit as a read address, and reading out the nonlinear characteristic from the storage unit according to an integer part of the read address. Conversion means for converting the signal from the low-pass filter means to a non-linear output signal, and an interpolation coefficient for interpolating the non-linear characteristic are stored, and the interpolation coefficient is read and read according to the decimal part of the read address. Interpolation means for interpolating and outputting the nonlinear output signal according to an interpolation coefficient,It is characterized by having.
[0007]
In the invention according to claim 2,Exciting means for generating an excitation signal based on the non-linear characteristics of the sounding body, and an exciting signal output from the exciting means for simulating the transmission characteristics of the resonator by circulating repeatedly while delaying the excitation signal for a predetermined time. A wave guide means for generating a musical tone having a pitch corresponding to the resonance operation of the musical instrument, wherein the exciting means comprises a signal corresponding to a blowing pressure supplied from the outside and a return path of the wave guide means. A first calculating means for calculating a pressure difference signal representing a difference from a signal corresponding to a reflected pressure wave fed back via the first calculating means, and a pressure difference signal supplied from the first calculating means in accordance with a blowing mode. Low-pass filter means for low-pass filtering based on the obtained resonance value and cutoff frequency, and a non-linear transformation of a signal from the low-pass filter means. A non-linear table means, and a second calculating means for multiplying a signal from the non-linear table means by the pressure difference signal and supplying the multiplied signal to the transmission outward path of the wave guide means. A memory for storing the non-linear characteristics of the sounding body, and using the signal from the low-pass filter as a read address to generate the read address integer part INT and the address decimal part FRC, while generating the read address integer part INT. A plurality of addresses INT ± n before and after and a plurality of addresses FRC ± n before and after the address fraction part FRC are generated, and the non-linear characteristic is referred to from the storage means according to the address integer part INT and the plurality of addresses INT ± n before and after the address integer part INT. Non-linear output means for reading and generating a plurality of corresponding non-linear output signals, A plurality of interpolation coefficients obtained from an interpolation function for interpolating the non-linear characteristic are stored, and a corresponding interpolation coefficient is read and read in accordance with the address decimal part FRC generated by the address generation means and a plurality of addresses FRC ± n before and after the address. Interpolation calculation means for forming and outputting a signal obtained by weighting and adding the plurality of non-linear output signals by each interpolation coefficient,
It is characterized by having.
[0009]
In the present invention,When a read address corresponding to the sound excitation operation is generated,The conversion means for storing the nonlinear characteristic of the sounding body refers to the nonlinear characteristic in accordance with the integer part of the read address and reads the nonlinear characteristic to convert the sound generation excitation operation into a nonlinear output signal. In the stored interpolation means,The interpolation coefficient is read according to the decimal part of the read address,Since the excitation signal is generated by interpolating the nonlinear output signal according to the read interpolation coefficient,The table capacity for storing the non-linear characteristics can be reduced.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
INDUSTRIAL APPLICABILITY The present invention can be applied to a tone generator that simulates a sounding mechanism of a wind instrument. Hereinafter, the principle of the invention will be described, and then a tone generator according to an embodiment of the present invention will be described as an example with reference to the drawings.
[0011]
A. Inventive principle
(1) Sound generation mechanism
Before describing the embodiment, the sounding mechanism of a wind instrument (single lead instrument) according to the present invention will be described. As is well known, a wind instrument includes a tube surrounding an air column, and a mechanism for generating vibration in the air column. The tube operates as a resonator, and a mechanism for generating air vibration operates as a sounding body. In wind instruments, in many cases, a thin piece of an elastic body called a reed is vibrated to produce sound. Here, the sounding mechanism will be described with reference to the cross-sectional structure of a single reed instrument shown in FIG.
[0012]
When the blowing pressure P is blown into the gap between the mouthpiece MP and the lead RD, a positive pressure wave generated by the blowing pressure P propagates through the bore BO (tube) to the bell BL. When this state is illustrated as an impulse response, as shown in FIG. 2, when the pipe length is L and the sound velocity is c, it can be seen that standing waves having pressure peaks occur at 2 L / c, 4 L / c, and 6 L / c.
By the way, since the bell BL is an open end, the sign of the pressure wave passing through the bore BO (tubular body) is inverted and becomes negative. This time, the pressure wave propagates in the order of the bell BL, the bore BO, and the lead RD to reduce the pressure p near the lead RD. Lower. As a result, the air velocity flowing through the gap (slit) between the mouthpiece MP and the lead RD increases, and at that time, the lead RD is closed by the Bernoulli force acting on the lead RD, and the slit width decreases.
[0013]
As described above, when the blowing pressure P is applied, the volume flow rate rises to some extent, but the slit width gradually decreases accordingly, and finally the slit closes and the volume flow rate becomes zero. When the lead RD is closed, the tube is completely closed immediately in the vicinity of the lead. In this case, the pressure wave arriving in the order of the bell BL and the bore BO is transmitted without being inverted in sign.
When the pressure p immediately near the reed rises to a certain extent, the reed RD is opened and the state returns to the state where the blowing pressure P is applied, and the same repetition is performed thereafter. Parameters relating to this repetitive behavior include the resonance frequency Fc and resonance value Q determined by the shape of the bore BO and the bell BL, the blowing pressure P, the elastic coefficient of the lead RD and the lips of the player, and the like. Propagation loss and the like.
[0014]
The relationship of the volume flow rate F with respect to the pressure difference between the blowing pressure P and the pressure p immediately near the lead can be approximately expressed by a quadratic function shown in FIG. The physical properties of the reed instrument can be considered as a feedback loop that is converted in the order of “flow rate F” → “pressure p” → “displacement x” → “flow rate F” when the displacement x of the lead RD is also considered. In other words, the nonlinear characteristic is converted in the order of “pressure p (pressure difference between the blowing pressure P and the pressure p closest to the lead)” → “displacement x of the lead RD” → “flow rate F” starting from “the blowing pressure P”. That is, the sound generation mechanism is realized by the linear characteristic converted from “flow rate F” → “pressure p (pressure difference between the blowing pressure P and the pressure p immediately near the lead)”.
[0015]
Here, starting from the "blowing pressure P", the nonlinear pressure is converted in the order of "pressure p (pressure difference between the blowing pressure P and the pressure p closest to the lead)" → "displacement x of the lead RD" → "flow rate F". Since the characteristic, that is, the non-linear operation of the lead RD is a quadratic function shown in FIG. 3 as described above, it can be approximated by a quadratic equation F = -p (p + a) shown in FIG. However, it is assumed that F = 0 when p <a. In addition, this can naturally be treated as a non-linear table, and a method of converting the (p + a) term of the above approximate expression into a table can be considered.
[0016]
(2) Signal processing algorithm
Next, a signal processing algorithm for simulating the sounding operation of the lead musical instrument based on the sounding mechanism described above will be described with reference to FIG. In this figure, reference numeral 1 denotes a multiplier, which multiplies a pressure p output from a non-linear table unit 10 described later by an output (−p + a) of an adder 8 described later to generate a blowing volume velocity (flow rate F). . M is a coefficient multiplier that multiplies the output of the multiplier 1 by a proportional constant g1 and outputs the result. Reference numeral 2 denotes a signal scattering junction for simulating the scattering (reflection and transmission) of the reflected pressure wave in the gap between the mouthpiece and the lead, and is a multiplication grating composed of adders 2a and 2c, multipliers 2e and 2b, and a delay element 2d. Is formed. Here, the delay element 2d outputs the input signal with a delay of one sample period.
[0017]
Reference numeral 3 denotes a waveguide that simulates the transmission characteristics of the resonance tube, and includes components 4 to 6. Reference numerals 4-1 to 4-6 denote delay circuits inserted in the transmission outward path of the waveguide 3 to simulate the propagation delay of the air pressure wave in the resonance pipe. Each of the delay circuits 4 is formed of a ring buffer in which a volatile memory such as a RAM is circularly addressed, and generates a delay time corresponding to a relative difference between a write address and a read address.
As shown in FIG. 6, the delay circuit 4 can set two different read addresses a and b in the ring buffer 4a, and add interpolation coefficients r and 1−1 to delay times corresponding to the addresses a and b, respectively. It has multipliers 4b and 4c for multiplying r, and an adder 4d for adding outputs of the multipliers 4b and 4c.
[0018]
In the delay circuit 4 having the above configuration, the read addresses a and b are set alternately at each note-on timing. For example, as shown in FIG.₁, A read address a is set, and when sound is generated at a pitch corresponding to the delay time Da, "1" is set as the interpolation coefficient r in the multiplier 6b, while the read address b is used. Not done. Next, in the process of changing the pitch to a new pitch, the note-on timing t₂By setting a new read address b, the delay interpolation is performed from the delay time Da corresponding to the previous pitch to the delay time Db corresponding to the current pitch.
[0019]
Next, returning to FIG. 5, the signal processing algorithm will be described. In FIG. 5, reference numerals 5-1 to 5-5 denote signal scattering junctions for simulating scattering of an air pressure wave generated at a portion where the diameter of the resonance tube changes.
As shown in FIG. 8, the signal scattering junction 5 is a lattice structure filter including a subtractor 5a, a coefficient multiplier 5b, and adders 5c and 5d. As shown in FIG. Simulate the reflection (p1 ′, p2 ′) and transmission (p1, p2) of the pressure wave when changing from S1 to S2. The multiplication coefficient k of the coefficient multiplier 5b is a value given by the relationship of k = S1-S2 / S1 + S2 based on the values S1 and S2 of the cross-sectional area.
[0020]
FIG. 9 shows a junction equivalent to the lattice structure filter shown in FIG. 8. In this case, one multiplier can be omitted as compared with the configuration shown in FIG. One of the junction configurations shown in FIGS. 8 and 9 is selected according to the characteristics of the DSP (digital signal processor) to be used. That is, when it is desired to reduce the number of times of multiplication, it is advantageous to use a junction having the configuration shown in FIG. A junction may be used.
[0021]
Returning to the description of the signal processing algorithm with reference to FIG. 5 again. In FIG. 5, reference numerals 6-1 to 6-6 denote delay elements which are interposed in the transmission return path of the waveguide 3 and delay one sampling period, and 7 denotes a multiplier which simulates a loss at the end of the tube. The adder 8 generates a pressure difference (−p + a) between the reflected pressure wave fed back via the return path of the waveguide 3 and the blowing pressure, and outputs this to the next stage, while supplying the same to the multiplier 1 described above. I do. Reference numeral 9 denotes a low-pass filter (hereinafter, referred to as LPF) that simulates the resonance characteristics of the lead RD (see FIG. 1).
The LPF 9 is composed of, for example, an IIR filter composed of subtractors 9a and 9b, multipliers 9c, 9e and 9i, adders 9d and 9f, and delay elements 9g and 9h, as shown in FIG. The resonance value Q and the cutoff frequency Fc are given as coefficients of the multipliers 9c, 9e, 9i, and the self-excited vibration mode of the lead is determined by this.
[0022]
The non-linear table section 10 simulates the non-linear operation of the lead RD, and shows the relationship shown in FIG. 12, that is, the pressure (reflection pressure) with respect to the pressure difference (-p + a) between the blowing pressure P and the pressure immediately near the reed. p is stored in a table. As shown in FIG. 13, the non-linear table section 10 is composed of components 10a to 10f.
Here, the configuration of the nonlinear table unit 10 will be described with reference to FIG. In the figure, reference numeral 10a denotes a non-linear table memory composed of an electrically rewritable nonvolatile memory such as a flash memory or an EPROM, and as described above, the pressure between the blowing pressure P and the pressure immediately adjacent to the read. The reflection pressure p for the difference (-p + a) is stored as a data table. Reference numeral 10b denotes an interpolation coefficient memory, which stores interpolation coefficients obtained by the interpolation function Jn shown in FIG.
[0023]
To these memories 10a and 10b, an integer part INT and a decimal part FRC of an address AD generated corresponding to the blowing pressure P are supplied, respectively. The adder 10c-1 adds "1" to the decimal part FRC of the address AD to generate a decimal part FRC + 1, and the adder 10c-2 subtracts "1" from the decimal part FRC to generate a decimal part FRC-1. The subtractor 10c-3 is a subtractor that subtracts "2" from the decimal part FRC to generate a decimal part FRC-2. Thus, the interpolation coefficient IPC1 corresponding to the decimal part FRC of the address AD and the three interpolation coefficients IPC2 to IPC4 of the address immediately before and after this are stored in the interpolation coefficient memory according to the decimal parts FRC + 1, FRC-1, and FRC-2. 10b.
[0024]
On the other hand, 10d-1 is an adder which adds "2" to the integer part INT of the address AD to generate an integer part INT + 2, and 10d-2 adds "1" to the integer part INT of the address AD to add an integer part INT + 1. Adder 10d-3 is a subtractor that subtracts "2" from the integer part INT to generate the integer part INT, and corresponds to the integer part INT of the address AD as in the decimal part FRC. The reflection pressure p1 and the three reflection pressures p2 to p4 at addresses adjacent to the front and rear thereof are read from the non-linear table memory 10a in accordance with the integer parts INT-1, INT + 1, and INT + 2.
[0025]
Reference numerals 10e-1 to 10e-4 denote multipliers, each of which is a fraction of the address AD with respect to the reflection pressures p2 to p4 read from the non-linear table memory 10a in accordance with the integer parts INT, INT-1, INT + 1, and INT + 2 of the address AD. The interpolation coefficients IPC1 to IPC4 read from the interpolation coefficient memory 10b are multiplied by each other according to the sections FRC, FRC + 1, FRC-1, and FRC-2, and output. An adder 10f generates the interpolated reflected pressure p by adding the outputs of the multipliers 10e-1 to 10e-4.
[0026]
As described above, according to the present invention, the interpolation coefficient memory 10b for storing the interpolation coefficient derived by the interpolation function is provided, the read address is divided into an integer part and a decimal part, and the read address is read from the nonlinear table memory 10a according to the address integer part. Since a plurality of table values (pressure p) are interpolated based on a plurality of interpolation coefficients read from the memory 10b in accordance with the decimal part of the address, the memory capacity of the non-linear table memory 10a can be reduced. It has become.
[0027]
In the signal processing algorithm described above, when the blowing pressure P is applied to the excitation unit including the components 8, 9, 10, and 1, the adder 8 first determines the reflected pressure propagating through the return path of the waveguide 3 and the reflected pressure. Is generated and input to the LPF 9, while being supplied to the multiplier 1 as a multiplication coefficient. The LPF 9 performs low-pass filtering on the pressure difference (−p + a) based on the resonance value Q and the cutoff frequency Fc according to the blowing mode, and changes the pressure in the pipe based on the self-excited vibration mode of the reed, that is, the mouthpiece and the reed. The change in pressure in the pipe due to the air flow passing through the gap between the table and the non-linear table section 10 is given.
[0028]
In the non-linear table section 10, the reflected pressure p corresponding to the change in the pipe pressure is obtained and output by the above-described interpolation calculation, and this is multiplied by the output (-p + a) of the adder 8 in the multiplier 1, and the blowing volume velocity (flow rate F ). This volume velocity is then input to the transmission path of the waveguide 3 via the signal scattering junction 2 which simulates the transmission / reflection of pressure waves in the gap between the mouthpiece and the air lead. The waveguide 3 simulates propagation delay of the resonance tube, scattering at a portion where the diameter of the tube changes, and reflection loss at the end of the tube. Thereafter, the transmission path of the waveguide 3 is traced in the order of delay element → junction → delay element..., And is fed back to the adder 8 again.
In this way, during the data circulation in the closed loop composed of the components 1 to 10, that is, in the process of performing the resonance operation, the tone signal is output from the multiplier 7 at the end of the waveguide 3.
[0029]
By the way, the excitation unit including the components 8, 9, 10, and 1 can adopt not only the configuration shown in FIG. 5 but also various modes. For example, the configuration shown in FIG. 15 simulates the non-linear characteristic in the simplest manner. In this case, the non-linear table section 10 stores the blowing volume velocity (flow rate F) corresponding to the change in the pipe pressure. In this case, the output of the LPF 9 is directly converted into the flow rate F.
Further, in the configuration shown in FIG. 16, an adder ADD and a logic gate G are provided instead of the non-linear table to simulate the quadratic equation F = -p (p + a) shown in FIG. Here, the logic gate G functions as an excitation unit that outputs zero when the input signal is equal to or less than zero, and otherwise outputs the input signal as it is.
[0030]
The configuration shown in FIG. 17 is effective for simulating the so-called nonlinear characteristics of the air lead. In this case, the reflected pressure wave fed back via the return path of the waveguide 3 is directly supplied to the LPF 9. The LPF 9 performs low-pass filtering on the reflected pressure wave based on the resonance value Q and the cut-off frequency Fc corresponding to the blowing mode to generate a flow velocity distribution change based on the self-excited vibration mode of the air lead, and generates a non-linear table section 10. Give to. In the non-linear table section 10, pressure waves generated according to the change in the flow velocity distribution acting on the reflected pressure waves traveling in the pipe are stored in a table, and the pressure waves corresponding to the output of the LPF 9 are read out. The read pressure wave is multiplied by the blowing pressure P in the multiplier 1 to obtain a blowing volume velocity (air flow velocity).
[0031]
B. Configuration of the embodiment
(1) Overall configuration
Next, with reference to FIG. 18, an overall configuration of an embodiment for generating a musical tone based on the above-described signal processing algorithm will be described. In this figure, reference numeral 20 denotes a keyboard for generating performance information such as a key-on, a key-off, a note number (pitch), and a velocity value corresponding to a key-depressing speed according to a key-depressing / releasing operation. Reference numeral 21 denotes a panel switch provided on the operation panel, and generates an operation signal according to the operation of the switch. As shown in FIG. 19, for example, as shown in FIG. 19, a numeric keypad 21a operated when inputting a numerical value, a cursor key 21b operated when moving a cursor, and the above-described nonlinear table memory 10a (see FIG. 13). ), A key EDIT operated when editing the contents, etc., a key EXIT operated when transiting the operation mode, a key Enter operated when finalizing the input, and the like.
[0032]
Reference numeral 22 denotes a display device including an LCD panel, an LCD drive circuit, and the like. Based on a display control signal supplied from a microcomputer 23 to be described later via a bus B, for example, a panel setting state corresponding to the operation signal, etc. Is displayed. The microcomputer 23 is, for example, a so-called one-chip microcomputer in which a CPU, a ROM, a RAM, and the like are formed on the same chip, and controls each unit of the apparatus by an operation described later. Reference numeral 24 denotes a digital signal processor (hereinafter referred to as a DSP), which forms a tone based on the above-described signal processing algorithm according to a microprogram incorporated therein.
[0033]
The signal processing algorithm executed in the DSP 24 is controlled based on various parameters (described later) supplied from the microcomputer 23 via the bus B. Reference numeral 25 denotes a delay RAM used as the delay circuits 4-1 to 4-5 and the delay elements 6-1 to 6-6. Reference numeral 26 denotes a ROM, which is provided with the above-described nonlinear table memory 10a and interpolation coefficient memory 10b (see FIG. 13). Reference numeral 27 denotes a D / A converter that converts a tone output output from the DSP 14 into an analog signal and outputs the analog signal.
[0034]
(2) Functional configuration of DSP 24
(1) Schematic configuration
Next, the functional configuration of the DSP 24 will be described with reference to FIGS. First, a schematic configuration of the DSP 24 will be described with reference to FIG. In this figure, reference numeral 30 denotes a microcomputer interface unit (MIF) for exchanging data with the microcomputer 23 (hereinafter abbreviated as the microcomputer 23). Reference numeral 31 denotes a microprogram memory (MPRAM), which stores a microprogram that defines the contents of the above-described signal processing algorithm.
[0035]
The microprogram stored in the MPRAM 31 is sequentially read out according to a control signal supplied from the microcomputer 23 via the MIF 30. The microprogram thus read is stored in the instruction register (IR) 32 for each processing unit, that is, for each instruction set. A program decoder (PD) 33 decodes an instruction set read from the IR 32 and generates a control signal Sc. The operation of the registers, multipliers, and logical operation units inside the DSP 24 is controlled according to the control signal Sc.
[0036]
Numeral 34 denotes an enveloper which performs envelope control so that various musical tone parameters (blowing pressure P, interpolation coefficient r, cutoff frequency F, etc.) supplied from the microcomputer 23 via the MIF 30 become a target value at a set rate. The configuration and the schematic operation will be described later. 35 is an interpolator / external RAM interface. The interpolator / external RAM interface 35 controls the read addresses for the non-linear table memory 10a and the interpolation coefficient memory 10b provided in the ROM 26 to perform interpolation at the time of non-linear table conversion, while the RAM 25 used as a delay circuit , And delay control is performed by controlling the read address, the details of which will be described later.
[0037]
36 is a parameter memory (PRAM) for storing parameters interpolated by the interpolator / external RAM interface 35, and 37 is a work memory for temporarily storing calculation results and the like performed inside the DSP 24. 38 is a bus, 39 and 40 are multiplication registers for temporarily storing multiplied data, and 41 and 42 are logical operation registers for temporarily storing data to be logically operated. 43 to 45 are selectors, 46 is a multiplier, and 47 is a multiplication result register that holds the output of the multiplier 46. 48 is a logical operation unit, and 49 is a logical operation result register holding the output of the logical operation unit 48. Reference numeral 50 denotes a shifter / clipper for shifting or clipping the output of the logical operation result register 49, and reference numeral 51 denotes a result register for storing the output result of the shifter / clipper 50. 52 to 53 are output registers, and 54 is a parallel / serial converter for converting the contents of these registers 52 and 53 into serial data and outputting the data.
[0038]
(2) Configuration of the enveloper 34
FIG. 21 is a block diagram showing the configuration of the enveloper 34. As shown in FIG. In the figure, a selector 34a selects either the output of the PRAM 36 or the output of the TRAM 1 and supplies it to the next stage. Reference numeral 34b denotes a target data register (TGR) in which a parameter serving as a set target value is set. Reference numeral 34c denotes a rate data register (RTR), which stores rate data representing a parameter change rate. Reference numeral 34d denotes a current data register (CRR), which stores a currently set parameter value.
[0039]
A selector 34e selects one of the outputs of the TGR 34b and the RTR 34c and supplies it to the next stage. 34f is an addition / subtraction comparator for adding, subtracting or comparing the output of the selector 34e and the output of the CRR 34d. The operation of the addition / subtraction comparator 34f is determined according to the control signal Sc described above. A comparison result flag register 34g stores a comparison result flag when the addition / subtraction comparator 34f performs a comparison operation. A result register 34h temporarily stores a parameter value (envelope) generated by the operation of the addition / subtraction comparator 34f. 34i is a selector for selecting and outputting any of the output of the result register 34h or the signal supplied from the MIF 30 described above.
[0040]
The enveloper 34 having the above configuration outputs the parameter envelope control result at the timing shown in FIG. That is, the time t when the set target value D_target is set in the TGR 34b in synchronization with the basic clock φ.₀T after one cycle of the basic clock φ₁The current value D_Current is stored in the CRR 34d.
As a result, the time t₂Then, the set target value D_target is compared with the current value D_Current in the addition / subtraction comparator 34f, while the rate data is set in the RTR 34d. Then, based on the comparison result of the addition / subtraction comparator 34f, the time t₃In, the result of rate addition or rate subtraction with respect to the current value D_Current stored in the CRR d according to the rate data stored in the RTR c is set in the EVR h.
[0041]
(3) Configuration of interpolator / external RAM interface 35
Next, the configuration of the interpolator / external RAM interface 35 will be described with reference to FIG. In the figure, reference numeral 351 denotes an address memory (TRAM1) in which a real number type offset address is stored. A selector 352 selects one of the offset address supplied from the bus TR and the offset address read from the TRAM 1 and supplies it to the next stage. An address register (TXR) 353 stores a real-type address supplied via the selector 352, that is, an address integer part and an address decimal part.
[0042]
Reference numeral 354 denotes an address memory (TRAM2) for storing an integer address supplied from the MIF 30, and reference numeral 355 denotes an address register (TR) for temporarily storing an integer address read from the address memory 354. A selector 356 selects one of the address integer parts output from the TXR 353 or TR 355 and supplies it to the next stage. A selector 357 selects and outputs any one of the output of the operation result register 359 described later and the sampling cycle counter value SC.
[0043]
358 is a logical operation unit that generates three address integer parts before and after the designated address based on the output of the selector 356 and the output of the selector 357. In the operation result register 359, a designated address (integer part) generated by the logical operation unit 358 and three addresses (integer part) before and after the designated address are sequentially stored. Reference numeral 360 denotes an external memory address register which stores a designated address (integer part) output from the operation result register 359 and three addresses (integer part) before and after the designated address as an external memory address EADR.
Based on the external memory address EADR, a table value for interpolation calculation is read from the non-linear table memory 10a (see FIG. 13).
[0044]
Reference numeral 361 denotes a decimal part register (TTR) for temporarily storing the decimal part of the address output from the TXR 353. An adder / subtractor 362 generates three address decimal parts before and after the address decimal part stored in the TTR 361. Reference numeral 363 denotes an interpolation coefficient memory provided in the above-described ROM 26 (see FIG. 18), and reads an interpolation coefficient corresponding to four adjacent address decimal parts output from the adder / subtractor 362. An interpolation coefficient register 364 temporarily stores interpolation coefficients sequentially read from the interpolation coefficient memory 363. 365 is a register for temporarily storing the table value EDAT read from the non-linear table memory 10a based on the external memory address EADR described above.
[0045]
A multiplier 366 multiplies the interpolation coefficient output from the interpolation coefficient register 364 by the table value EDAT output from the register 365. An adder 367 performs an interpolation operation by adding the output of the multiplier 366 and its own output. 368 is an operation result register for temporarily storing the interpolation result of the adder 367, and 369 is a holding register for holding the interpolation result. Reference numeral 370 denotes a register for temporarily storing the velocity value p supplied via the bus TR in order to write the velocity value p into the delay RAM 25.
[0046]
As a result, the interpolator / external RAM interface 35 having the above configuration performs an interpolation operation on the table value read from the non-linear table memory 10a at the timing shown in FIG.
In other words, according to the address notation shown in FIG. 13, first, when the address integer part INT to be read is set in the TXR 353, the address is sequentially stored in the AR 359 at the next clock timing based on the processing of the logical operation unit 358. While the integer parts INT-1, INT, INT + 1, INT + 2 are set, the address decimal part FRC corresponding to the TTR 361 is stored. When the address decimal part FRC is set in the TTR 361, the adder / subtractor 362 sequentially generates the address decimal parts FRC-2, FRC-1, FRC, and FRC + 1.
[0047]
In this way, address integer parts INT-1, INT, INT + 1, INT + 2 for reading the non-linear table memory 10a and address decimal parts FRC-2, FRC-1, FRC, FRC + 1 for reading the interpolation coefficient memory 10b are generated. Table values (pressure values) p2, p1, p3, and p4 sequentially read from the memory 10a are sequentially set in the EIR 365, and the interpolation coefficients IPC2, IPC1, IPC3 and IPC4 are set.
Then, the ADR 368 outputs “p2 × IPC2”, “p2 × IPC2 + p1 × IPC1”, “p2 × IPC2 + p1 × IPC1 + p3 × IPC3”, and “p2 × “IPC2 + p1 × IPC1 + p3 × IPC3 + p4 × IPC4” are sequentially stored.
[0048]
On the other hand, when a delay occurs in the interpolator / external RAM interface 35, as shown in FIG. 25, the address integer part from the TRAM 2 and the sampling cycle counter value SC are added in the logical operation unit 358. The result is stored in AR359. The data read from the delay RAM 25 by this address is given to the EIR 365.
In the case of writing, data to be written to the EOR 370 is set in advance from the bus TR, a write address generated by a similar address operation is set to the ADR 368, and the data of the EOR 370 is written to the delay RAM 25 according to the address. become.
[0049]
C. Operation of the embodiment
Next, the operation of the embodiment having the above configuration will be described with reference to FIGS.
(1) Initial processing
First, when the power is turned on to the musical sound generating apparatus according to the present embodiment, the microcomputer 23 reads out and loads the control program stored in the internal ROM, and then executes initial processing shown in FIG. The process proceeds to SA1. In step SA1, the control flag stored in the register c_flag is set to “0”, the delay interpolation coefficient target value stored in the register r_target is set to “0”, and the register p_target storing the DC component DC corresponding to the breathing pressure is set to “0”. ”, The interrupt mask is released, and the process waits until an event interrupt occurs. Here, the event interrupt refers to an interrupt corresponding to the switch operation of the panel switch 11 described above, an interrupt corresponding to the key press / release operation, and a timer interrupt generated at regular intervals.
[0050]
(2) Edit processing
Here, for example, it is assumed that a key EDIT (see FIG. 19) arranged on the operation panel is operated to edit the contents of the musical tone parameters prior to the performance. Then, based on the key-on event generated in response to the operation of the key EDIT, the microcomputer 23 starts an edit processing routine shown in FIG. 27 and proceeds to step SB1.
In step SB1, the type of the parameter designated for editing in response to the switch operation is determined. Here, when updating the attack rate representing the rising shape of the musical tone, the process proceeds to step SB2, and the newly input attack rate is set in the register atk_rate.
[0051]
When updating the release rate, the process proceeds to step SB3, and the newly input release rate is set in the register rls_rate. Further, when updating the portamento rate for smoothly changing the pronunciation pitch, the process proceeds to step SB4, and the newly input portamento rate is set in the register plt_rate.
When the content of the non-linear table memory 10a is updated, the process proceeds to step SB5 to display all the table values forming the table 10a on a screen (LCD panel). Of all the table values displayed, the table value number No. to be updated. Is input or specified by operating the cursor key 11b. Then, in the next step SB7, the newly input value for the specified table value is stored in the register NLT [NO. ] To update the contents of the nonlinear table.
[0052]
(3) Note-on processing
When the editing process is completed by updating the various parameters and the contents of the non-linear table in this way, the microcomputer 23 returns to the interrupt waiting state. In this state, it is assumed that the operator depresses any key of the keyboard 10 by performing the performance. Then, based on the key-on event generated in response to the key depression, the microcomputer 23 executes a note-on process shown in FIG. 28 and proceeds to step SC1.
In step SC1, the note number of the depressed key and the velocity value corresponding to the depressing speed are stored in the registers Note and Vel, respectively, and the process proceeds to the next step SC2.
[0053]
In step SC2, it is determined whether the control flag stored in the register c_flag is “0”, that is, whether or not to perform delay interpolation when changing from the previous pitch to the current pitch. Here, when delay interpolation is not performed, the control flag stored in the register c_flag is “0”, so the determination result is “YES”, and the process proceeds to the next Step SC3. In step SC3, "1" is set as the interpolation coefficient target value in the register r_target, and a read address a for giving a delay time corresponding to the note number stored in the register Note is set in the register D3a.
That is, the read address a of the delay circuit 4-3 in the above-described signal processing algorithm (see FIG. 5) is set in the register D3a. Next, the microcomputer 23 proceeds to step SC4, sets the control flag stored in the register c_flag to “1”, and then proceeds to step SC7 described later.
[0054]
On the other hand, in the case of sound generation with delay interpolation, that is, when interpolation is performed from the delay time corresponding to the preceding pitch to the delay time corresponding to the current pitch, the control flag stored in the register c_flag is set to “1”. Therefore, the result of the determination in step SC2 is "NO", and the flow proceeds to step SC5. In Step SC5, the delay interpolation coefficient target value of the register r_target is set to “0”, and the read address b that gives the delay time corresponding to the note number (pitch) is set in the register D3b.
Thus, the delay time in the delay circuit 4-3 is interpolated by the delay interpolation coefficient r continuously changing from "1" to "0" as shown in FIG. Next, in step SC6, since the delay interpolation has been completed, the control flag stored in the register c_flag is reset to zero.
[0055]
Then, after that, the microcomputer 23 proceeds to step SC7, sets the target cutoff frequency corresponding to the note number (pitch) in the register F_target, and sets the portamento rate stored in the register plt_rate in the register F_rate. Store. Further, the velocity value stored in the register Vel is stored in the register P_target as a DC component DC corresponding to the blowing breath pressure, and the attack rate stored in the register atk_rate is stored in the register P_rate. Further, the portamento rate stored in the register plt_rate is set in the register r_rate.
[0056]
The target cutoff frequency stored in the register F_target is tabulated in advance as shown in FIG. 29, and is read out according to the pitch (note number) to be sounded.
After all, in step SC7, the target cutoff frequency corresponding to the pitch to be generated and the change rate (portamento rate) thereof are set, and the velocity value (corresponding to the blowing pressure) corresponding to the key pressing speed and its value are set. A change rate (attack rate) is set. Further, a portamento rate is assigned to a rate at which the delay interpolation coefficient is changed from “1” to “0”.
[0057]
(4) Note-off processing
Next, when the depressed key is released, the microcomputer 23 executes a note-on process shown in FIG. 30 based on a key-off event generated in response to the key-release operation, and proceeds to step SD1. In the note-on process, the velocity value (corresponding to the blowing pressure) stored in the register P_target is reset to zero, and the release rate stored in the register rls_rate is set in the register P_rate.
[0058]
(5) Timer interrupt processing
The microcomputer 23 starts a timer interrupt process at regular intervals and executes steps SE1 to SE3 shown in FIG. That is, in step SE1, the delay interpolation coefficient r is increased or decreased according to the portamento rate so as to reach the target value. Further, in step SE2, the cutoff frequency F is increased or decreased according to the portamento rate so as to become the cutoff frequency target value. Further, in step SE3, the velocity is increased or decreased to a velocity value corresponding to the blowing pressure according to the attack rate. In other words, these steps SE1 to SE3 are all envelope processes for controlling parameters to target values in accordance with a given rate, and the details thereof will be described using the delay interpolation coefficient generation process in step SE1 as an example. I do.
[0059]
Now, when a timer interrupt event occurs and the timer interrupt process shown in FIG. 31 is started and the process proceeds to step SE1, the microcomputer 23 executes the delay interpolation coefficient generating process shown in FIG. 32 and proceeds to step SF1. .
In step SF1, it is determined whether or not the delay interpolation coefficient r has reached the target value stored in the register r_target. Here, if the target value has been reached, the determination result is "YES" and this process is completed. If not, the determination result is "NO" and the process proceeds to the next step SF2.
[0060]
In step SF2, it is determined whether or not the current delay interpolation coefficient r is smaller than the target value. Here, when the current delay interpolation coefficient r is smaller than the target value, the determination result is “YES”, and the process proceeds to Step SF3. In step SF3, the portamento rate stored in the register r_rate is added to the current delay interpolation coefficient r to update the coefficient value. That is, the rate of the delay interpolation coefficient r is added according to the pitch change.
On the other hand, when the interpolation coefficient r is larger than the target value in step SF2, the determination result is “NO”, and the process proceeds to step SF4.
[0061]
In step SF4, the portamento rate stored in the register r_rate is subtracted from the current delay interpolation coefficient r, whereby the rate of the delay interpolation coefficient r is subtracted according to the pitch change. Then, when proceeding to the next step SF5, the microcomputer 23 transfers the delay interpolation coefficient r to which the rate addition or the rate subtraction has been performed to the DSP 14 side. Thus, the DSP 24 performs delay interpolation according to the delay interpolation coefficient r supplied from the microcomputer 23.
[0062]
As described above, in the delay interpolation coefficient generating process, the delay interpolation coefficient r is updated by rate addition or rate subtraction until it reaches the target value, and the updated delay interpolation coefficient r is sequentially transferred to the DSP 14 side. Such processing is the same in steps SE2 and SE3 described above.
That is, when step SE2 is executed, a cutoff frequency generation process (not shown) is executed, and the cutoff frequency F is updated by rate addition or rate subtraction until the cutoff frequency F reaches the target value. The off-frequency is sequentially transferred to the DSP 14 side. When step SE3 is executed, a velocity generation process (not shown) is executed, and the velocity is added or decremented until the velocity value corresponding to the breathing pressure reaches the target value. The velocity value is sequentially transferred to the DSP 14 side.
[0063]
When the parameters according to the performance operation are generated in this way, the microcomputer 23 sends these parameters to the DSP 24 side. Then, in the DSP 24, based on the above-described signal processing algorithm (see FIG. 5), first, the excitation section gives the blowing pressure P, whereby the adder 8 determines the reflected pressure propagating through the return path of the waveguide 3 and the reflected pressure. Is generated and input to the LPF 9 and supplied to the multiplier 1 as a multiplication coefficient.
In the LPF 9, the pressure difference (-p + a) is subjected to low-pass filtering based on the resonance value Q and the cutoff frequency Fc according to the blowing mode, and the pressure change in the pipe based on the self-excited vibration mode of the reed, that is, the mouthpiece The pressure change in the pipe due to the air flow passing through the gap between the lead and the lead is given to the nonlinear table section 10.
[0064]
As described above, the DSP 24 generates the output of the LPF 9, that is, the read address AD corresponding to the change in the pressure in the pipe. This address AD is formed of an integer part INT and a decimal part FRC, and based on this, the interpolator / external RAM interface 35 generates the decimal parts FRC-2, FRC-1, FRC and FRC + 1, and The sections INT-1, INT, INT + 1 and INT + 2 are generated. Then, the decimal parts FRC-2, FRC-1, FRC and FRC + 1 of the address AD are given to the interpolation coefficient memory 10b to read out the interpolation coefficients IPC2 to IPC4, while the integer parts INT-1, INT, INT + 1 and INT + 2 of the address AD are read. The pressures p2 to p4 are read from the non-linear table memory 10a based on the values.
[0065]
When the pressures p2 to p4 of the four samples including the designated address AD and the corresponding interpolation coefficients IPC2 to IPC4 are read out, the DSP 24 multiplies the pressures p2 to p4 and the interpolation coefficients IPC2 to IPC4 by each other. An interpolating calculation is performed to generate an interpolated reflected pressure p. Then, the reflected pressure p generated by the non-linear table section 10 through the interpolation operation is multiplied by the output (−p + a) of the adder 8 in the multiplier 1 and converted into a blowing volume velocity (flow rate F). The signal is input to the transmission path of the waveguide 3 via the signal scattering junction 2 simulating the transmission / reflection of the pressure wave in the gap between the mouthpiece and the air lead.
[0066]
The waveguide 3 simulates propagation delay of the resonance tube, scattering at a portion where the diameter of the tube changes, and reflection loss at the end of the tube. Thereafter, the transmission path of the waveguide 3 is traced in the order of delay element → junction → delay element..., And is fed back to the adder 8 again. Thereafter, the data circulating is repeated by repeating the above-described process in the closed loop composed of the components 1 to 10 to perform a resonance operation. At this time, the multiplier 7 provided at the end of the waveguide 3 resonates at the designated pitch. A tone signal is extracted.
[0067]
As described above, in the present invention, the interpolation coefficient memory 10b for storing the interpolation coefficient derived by the interpolation function is provided, and a plurality of table values (pressure p) read from the nonlinear table memory 10a according to the integer part of the read address are provided. ) Is calculated based on a plurality of interpolation coefficients read from the memory 10b in accordance with the decimal part of the address. Therefore, unlike the related art, it is not necessary to provide a huge table capacity, and the memory capacity of the nonlinear table memory 10a can be reduced. It can be planned.
[0068]
D. Modified example
(1) Configuration
In the above-described embodiment, the interpolation technique for forming a musical tone by a so-called self-excited vibration type has been described. However, in the following, the interpolation means used for the self-excited vibration type sound source is a waveform readout type. A modified example that can be shared with a sound source will be described.
That is, in the above-described embodiment, the non-linear table memory 10a for simulating the non-linear characteristics of the read was interpolated and read. In a modified example, a waveform memory was provided instead of the non-linear table memory 10a. Is read out by interpolation to form a musical tone. FIG. 33 shows a signal processing algorithm in this case.
[0069]
In FIG. 33, reference numeral 60 denotes a temporary register (TEMPR) for temporarily storing a read address integer part and a decimal part of the waveform memory unit 63 described later. Reference numeral 61 denotes an increment register (INCR) for storing rate data for incrementing an address stored in the TEMPR 60 in accordance with a pitch to be sounded. Reference numeral 62 denotes an adder, which rate-adds the value of the INCR 61 to the output of the TEMPR 60 and feeds it back to the TEMPR 60 again.
The waveform memory unit 63 has a configuration equivalent to that of the nonlinear table unit 10 shown in FIG. 13, and is obtained by replacing the nonlinear table memory 10a with a waveform memory (not shown).
[0070]
That is, while having the interpolation coefficient memory 10b storing the interpolation coefficient obtained by the interpolation function Jn shown in FIG. 14, the decimal point FRC + 1 is generated by adding "1" to the decimal part FRC of the address supplied from the TEMPR60. Adder 10c-1 subtracts "1" from decimal part FRC to generate decimal part FRC-1, and subtractor 10c-2 subtracts "2" from decimal part FRC to generate decimal part FRC-2. A subtractor 10c-3 is provided, and the interpolation coefficient IPC1 corresponding to the decimal part FRC of the address and the three interpolation coefficients IPC2 to IPC4 of the address immediately before and after this are converted to the decimal parts FRC + 1, FRC-1, and FRC-2. Read from the interpolation coefficient memory 10b accordingly.
[0071]
Also, for the waveform memory, an adder 10d-1 that adds "2" to the integer part INT of the address to generate an integer part INT + 2, and adds "1" to the integer part INT of the address AD to add an integer part An adder 10d-2 for generating INT + 1 and a subtractor 10d-3 for subtracting "2" from the integer part INT to generate an integer part INT are provided, and the integer part INT of the address is provided in the same manner as the decimal part FRC. The corresponding waveform data WD1 and three waveform data WD2 to WD4 of addresses adjacent before, after, before and after are read from the waveform memory in accordance with the integer parts INT-1, INT + 1, INT + 2.
[0072]
Further, for the waveform data WD2 to WD4 read from the waveform memory in accordance with the integer parts INT, INT-1, INT + 1, INT + 2 of the address, the decimal part FRC, FRC + 1, FRC-1, FRC-2 of the address is used. Multipliers 10e-1 to 10e-4 for multiplying and outputting the interpolation coefficients IPC1 to IPC4 read from the interpolation coefficient memory 10b, respectively, and an adder 10f for adding these multiplication results and outputting an interpolation waveform. I have.
The LPF 9 is the same as that applied to the above-described signal processing algorithm of the self-excited oscillation type, is constituted by a second-order IIR filter, and has a low-pass filtering coefficient based on a resonance value Q and a cutoff frequency Fc corresponding to the blowing mode. The output of the LPF 9 is input to the coefficient multiplier 1 and multiplied by the enveloping coefficient ENV to form a musical tone.
[0073]
As described above, since the interpolation technique in the above embodiment can be shared with the sound source of the conventional waveform readout method, it should be generated by including the self-excited vibration type sound source and the waveform readout sound source of the embodiment. In the case of realizing a musical tone generator that switches between these sound sources according to the type of musical tone and operates the same, the interpolation means can be shared, so that efficient design is possible and this can be a factor that leads to a reduction in product price. . In the case of generating polyphonic sound, the above-described interpolation means may be used in a time-division manner.
[0074]
(2) Operation
Next, an operation of a modification based on the above signal processing algorithm will be described. Since the above-described signal processing algorithm has the same configuration as that of FIG. 18, the detailed configuration is omitted here.
(1) Initial processing
First, when the power is turned on to the musical sound generating device according to the modification, the microcomputer 23 reads out and loads the control program stored in the internal ROM, executes the initial processing shown in FIG. 34, and proceeds to step SG1. Proceed with the process. In step SA1, after performing the initialization to set the cutoff frequency stored in the register F to "0" and the envelope coefficient value stored in the register ENV to "0", cancel the interrupt mask and generate an event interrupt. Wait until. Note that the event interrupt mentioned here includes an interrupt corresponding to a key press / release key operation and a timer interrupt which occurs at regular intervals.
[0075]
(2) Note-on processing
After the initialization is performed, it is assumed that the operator presses any key on the keyboard 10 to perform a performance operation. Then, based on the key-on event generated in response to the key depression, the microcomputer 23 executes a note-on process shown in FIG. 35 and advances the process to step SH1.
In step SH1, the note number of the depressed key and the velocity value corresponding to the key depressing speed are stored in the registers Note and Vel, respectively, and the process proceeds to the next step SH2.
[0076]
In step SH2, the waveform reading start address SAD is set in the above-described temporary register (TEMPR) 60. In step SH3, the frequency number corresponding to the note number is set in the increment register (INCR) 61. As a result, the read address AD is sequentially accumulated from the start address SAD according to the pitch to be sounded.
Then, after that, the microcomputer 23 proceeds to step SH4, stores the velocity value stored in the register Vel in the register ENV_target, sets the target envelope value, and then stores the velocity value in the register Vel in step SH5. The stored velocity value is stored in the register F_target and set as the target cutoff frequency.
[0077]
(3) Note-off processing
Next, when the depressed key is released, the microcomputer 23 executes a note-on process shown in FIG. 36 based on a key-off event generated in response to the key-release operation, and proceeds to step SJ1. In the note-on processing, the velocity values stored in the registers ENV_target and F_target are reset to zero.
[0078]
(4) Timer interrupt processing
By the way, the microcomputer 23 starts the timer interrupt process at regular intervals and executes steps SK1 to SK2 shown in FIG. First, in step SK1, the cutoff frequency F is increased or decreased according to a predetermined rate so as to be the target cutoff frequency. Further, in step SK2, the envelope coefficient ENV is increased or decreased in accordance with a predetermined rate so as to become a target value. As described above, the steps SK1 to SK2 are all envelope processes for controlling the parameters to be the target values in accordance with the given rates, and the details thereof are substantially the same as those of the delay interpolation coefficient generation process described above (see FIG. 32). Since they are equivalent, their description is omitted.
[0079]
When each parameter according to the performance operation is generated based on the series of processes described above, the microcomputer 23 sends these parameters to the DSP 14 side. Then, based on the signal processing algorithm shown in FIG.
This address AD is formed of an integer part INT and a decimal part FRC, and based on this, the interpolator / external RAM interface 35 generates the decimal parts FRC-2, FRC-1, FRC and FRC + 1, and The sections INT-1, INT, INT + 1 and INT + 2 are generated. Then, the decimal parts FRC-2, FRC-1, FRC and FRC + 1 of the address AD are given to the interpolation coefficient memory 10b to read out the interpolation coefficients IPC2 to IPC4, while the integer parts INT-1, INT, INT + 1 and INT + 2 of the address AD are read. The waveform data WD2 to WD4 are read from a waveform memory (not shown) based on the waveform data.
[0080]
When the waveform data WD2 to WD4 of four samples before and after including the designated address AD and the corresponding interpolation coefficients IPC2 to IPC4 are read out, the DSP 24 compares the waveform data WD2 to WD4 and the interpolation coefficients IPC2 to IPC4 with each other. An interpolation operation of multiplying and adding is performed to generate interpolated waveform data WD. Then, the waveform data WD is subjected to low-pass filtering based on the cutoff frequency F subjected to the envelope processing in the LPF 9, and the filtered output is envelope-multiplied to form a musical tone.
[0081]
As described above, the modified example has the interpolation coefficient memory 10b that stores the interpolation coefficient derived by the interpolation function, and stores a plurality of waveform data read from the waveform memory according to the integer part of the read address in the memory according to the address decimal part. Since the interpolation calculation is performed based on the plurality of interpolation coefficients read from 10b, the self-excited vibration type sound source and the interpolation means can be shared while reducing the waveform memory capacity.
[0082]
E. FIG. Other examples
Next, another example based on the modified example of the delay circuit 4 which is inserted in the transmission outward path of the waveguide 3 and simulates the propagation delay of the air pressure wave in the resonance pipe in the above-described embodiment will be described.
As shown in FIG. 6, the delay circuit 4 of the above-described embodiment enables two different read addresses a and b to be set in the ring buffer 4a having the circular addressing, and the delay corresponding to the addresses a and b. It has multipliers 4b and 4c for multiplying the times Da and Db by interpolation coefficients r and 1-r, respectively, and an adder 4d for adding the outputs of the multipliers 4b and 4c. The interpolation is performed, but instead, a delay circuit 100 having an aspect shown in FIG. 38 is configured.
[0083]
In FIG. 38, reference numeral 101 denotes a ring buffer 4a to be circularly addressed, which is constituted by a RAM or the like, and can set four read addresses a to d. A subtractor 102 subtracts the delay time Db corresponding to the read address b from the delay time Da corresponding to the read address a to generate (Da−Db). A multiplier 103 multiplies the output of the subtracter 102 by a delay interpolation coefficient r to generate r · (Da−Db). Reference numeral 104 denotes an adder that adds the delay time Db to the output of the multiplier 103 to generate a delay interpolation output out1 that is r · Da + (1−r) · Db. As described above, the components 101 to 104 perform the interpolation from the delay time Da to the delay time Db according to the delay interpolation coefficient r, similarly to the delay circuit 4 described above.
[0084]
Next, a subtractor 105 subtracts the delay time Dd corresponding to the read address d from the delay time Dc corresponding to the read address c to generate (Dc-Dd). A coefficient multiplier 106 multiplies the output of the subtracter 102 by an address decimal part FRC. An adder 107 adds a delay time Dd to the output of the multiplier 106 to generate a delay interpolation output out2. An adder 108 adds data Dt representing the center delay time to the write address in_address. An adder 109 adds an LFO (low frequency oscillator) output to the output of the adder 108. The decimal part FRC of the address signal output from the adder 109 is supplied as a multiplication coefficient of the coefficient multiplier 106, while the integer part INT is supplied to the ring buffer 101 as a read address c. Further, “1” is added to the integer part INT by the adder 110, and the result is given to the ring buffer 101 as a read address d.
[0085]
According to the above configuration, the data Dt representing the center delay time is added to the write address in_address, and the read address c obtained by adding the value of the LFO and the read address d obtained by adding “1” to the address c Are generated. Then, the delay time Dd corresponding to the read address d is subtracted from the delay time Dc corresponding to the read address c to generate (Dc-Dd), and this is multiplied by the address decimal part FRC. By adding the time Dd, a delay interpolation output out2 that is linearly interpolated is output.
[0086]
When such a delay circuit 100 is applied to the above-described signal processing algorithm (see FIG. 5), the configuration is as shown in FIG. In this figure, the same elements as those shown in FIG. 5 are denoted by the same reference numerals, and description thereof will be omitted.
In this case, the delay interpolation outputs out1 and out2 output from the delay circuits 100-1 to 100-6 are multiplied by predetermined coefficients in the multipliers 120 and 123, respectively, and then output for each of the output out1 system and the output out2 system. The signals added by the adders 121 and 124 and further extracted from the terminal point of the waveguide 3 are added by the adders 122 and 125 for each system.
[0087]
In this way, since a plurality of musical tones having different delay times are output from the outputs out1 and out2 of the adders 122 and 125, a well-known chorus effect can be provided.
Further, an ensemble effect can be obtained by further increasing the number of taps. Furthermore, if only the outputs out1 and out2 of the delay circuits 100-1 to 100-6 are extracted as musical sounds, it is possible to obtain a vibrato effect. Further, since the LFO output is superimposed, a modulation effect can be provided.
[0088]
As described above, according to the present invention, the interpolation coefficient memory 10b for storing the interpolation coefficient derived by the interpolation function is provided, and a plurality of table values (pressures) read from the non-linear table memory 10a according to the integer part of the read address are provided. Since p) is interpolated based on a plurality of interpolation coefficients read from the memory 10b in accordance with the decimal part of the address, it is not necessary to provide an enormous table capacity as in the related art, and the memory capacity of the nonlinear table memory 10a is reduced. Can be achieved. If a waveform memory is provided instead of the non-linear table memory 10a, it can be applied to the waveform memory reading method, and the self-excited oscillation type sound source and the interpolation means described above can be used while reducing the waveform memory capacity. Can be shared.
[0089]
In the above-described embodiment, the microcomputer 23 generates various parameters according to the performance operation, and the DSP 24 operates the signal processing algorithm based on the various parameters to form musical sounds. May be left to the microcomputer 23.
Further, in the self-excited vibration type sound source according to the present embodiment, an initial excitation signal corresponding to a pressure change may be input to the nonlinear table unit 10 only at the beginning of generation of a musical sound. By doing so, a tone can be formed before the signal makes a round of the closed loop, so that sounding delay can be eliminated.
[0090]
In addition, the gist of the present invention can be applied not only to a device that generates wind instrument sounds, but also to a device that simulates a sounding mechanism of a plucked instrument such as a guitar or a bowed instrument such as a violin. Needless to say, there is. The point is that the gist of the present invention can be applied to any device using a signal processing algorithm for reading out the non-linear characteristics in a table.
[0091]
【The invention's effect】
According to the present invention,When a read address corresponding to the sound excitation operation is generated,The conversion means for storing the nonlinear characteristic of the sounding body refers to the nonlinear characteristic in accordance with the integer part of the read address and reads the nonlinear characteristic to convert the sound generation excitation operation into a nonlinear output signal. In the stored interpolation means, the interpolation coefficient is read according to the decimal part of the read address generated by the address generation means, and the excitation signal is generated by interpolating the non-linear output signal according to the read interpolation coefficient.The table capacity for storing the non-linear characteristics can be reduced.
[Brief description of the drawings]
FIG. 1 is a diagram for explaining a sound generation mechanism according to the present invention.
FIG. 2 is a diagram for explaining a sound generation mechanism according to the present invention.
FIG. 3 is a diagram showing a relationship between a volume difference F and a pressure difference between a blowing pressure P and a pressure p immediately near a lead.
FIG. 4 is a diagram illustrating a nonlinear characteristic of a lead RD.
FIG. 5 is a block diagram illustrating a signal processing algorithm according to the present invention.
FIG. 6 is a block diagram illustrating a configuration of a delay circuit 4 in a signal processing algorithm.
FIG. 7 is a diagram for explaining delay interpolation in the delay circuit 4;
FIG. 8 is a block diagram showing a configuration of a signal scattering junction 5 in a signal processing algorithm.
FIG. 9 is a block diagram showing another configuration example of the signal scattering junction 5.
FIG. 10 is a diagram for explaining a multiplication coefficient k at a signal scattering junction 5;
FIG. 11 is a block diagram illustrating a configuration of an LPF 9 in a signal processing algorithm.
FIG. 12 is a diagram showing nonlinear characteristics stored in a nonlinear table memory 10a.
FIG. 13 is a block diagram showing a configuration of the non-linear table unit 10.
FIG. 14 is a diagram for explaining an interpolation function.
FIG. 15 is a block diagram illustrating a configuration of an excitation unit according to another embodiment.
FIG. 16 is a block diagram illustrating a configuration of an excitation unit according to another embodiment.
FIG. 17 is a block diagram illustrating a configuration of an excitation unit according to another embodiment.
FIG. 18 is a block diagram showing a configuration of an embodiment according to the present invention.
FIG. 19 is a diagram showing the appearance of the operation panel in the embodiment.
20 is a block diagram illustrating a schematic configuration of a DSP 24. FIG.
FIG. 21 is a block diagram illustrating a configuration of an enveloper 34 in the DSP 24.
FIG. 22 is a timing chart showing the operation timing of the enveloper 34;
FIG. 23 is a block diagram showing a configuration of an interpolator / external RAM interface 35 in the DSP 24.
24 is a timing chart showing the interpolation operation timing of the interpolator / external RAM interface 35. FIG.
FIG. 25 is a timing chart showing a delay generation operation timing of the interpolator / external RAM interface 35;
FIG. 26 is a flowchart showing an operation of an initial process in the embodiment.
FIG. 27 is a flowchart showing an operation of an edit process in the embodiment.
FIG. 28 is a flowchart showing an operation of a note-on process in the embodiment.
FIG. 29 is a diagram showing an example of a parameter table in the embodiment.
FIG. 30 is a flowchart showing an operation of a note-off process in the embodiment.
FIG. 31 is a flowchart showing an operation of a timer interrupt process in the embodiment.
FIG. 32 is a flowchart illustrating the operation of a delay interpolation coefficient generation process in the embodiment.
FIG. 33 is a block diagram illustrating a signal processing algorithm according to a modified example.
FIG. 34 is a flowchart showing the operation of the modification.
FIG. 35 is a flowchart showing the operation of the modification.
FIG. 36 is a flowchart showing the operation of the modification.
FIG. 37 is a flowchart showing the operation of the modification.
FIG. 38 is a block diagram showing a configuration of a delay circuit 100 according to another example.
FIG. 39 is a block diagram showing a configuration of a signal processing algorithm to which the delay circuit 100 is applied.
[Explanation of symbols]
23 Microcomputer (address generation means, conversion means)
24 DSP (conversion means, interpolation means)
26 ROM (conversion means, interpolation means)

Claims (2)

Exciting means for generating an excitation signal based on the non-linear characteristics of the sounding body, and an exciting signal output from the exciting means for simulating the transmission characteristics of the resonator by circulating repeatedly while delaying the excitation signal for a predetermined time. A wave guide means for generating a musical tone having a pitch corresponding to the resonance operation of
The excitation means,
First calculating means for calculating a pressure difference signal representing a difference between a signal corresponding to the blowing pressure supplied from the outside and a signal corresponding to the reflected pressure wave fed back via the transmission return path of the wave guide means;
Low-pass filter means for low-pass filtering the pressure difference signal supplied from the first calculating means based on a resonance value and a cut-off frequency according to a blowing mode;
Non-linear table means for non-linearly converting the signal from the low-pass filter means;
Second calculating means for multiplying the signal from the non-linear table means by the pressure difference signal and supplying the multiplied signal to the transmission outward path of the wave guide means;
And the non-linear table means further comprises:
A memory for storing the non-linear characteristics of the sounding body, a signal from the low-pass filter as a read address, and reading and reading the non-linear characteristics from the storage in accordance with an integer part of the read address; Conversion means for converting a signal from the low-pass filter means to a non-linear output signal;
Interpolation means for storing an interpolation coefficient for interpolating the nonlinear characteristic, reading an interpolation coefficient according to a decimal part of the read address, and interpolating and outputting the nonlinear output signal according to the read interpolation coefficient;
A musical sound generating device comprising: Exciting means for generating an excitation signal based on the non-linear characteristics of the sounding body, and an exciting signal output from the exciting means for simulating the transmission characteristics of the resonator by circulating repeatedly while delaying the excitation signal for a predetermined time. A wave guide means for generating a musical tone having a pitch corresponding to the resonance operation of
The excitation means,
First calculating means for calculating a pressure difference signal representing a difference between a signal corresponding to the blowing pressure supplied from the outside and a signal corresponding to the reflected pressure wave fed back via the transmission return path of the wave guide means;
Low-pass filter means for low-pass filtering the pressure difference signal supplied from the first calculating means based on a resonance value and a cut-off frequency according to a blowing mode;
Non-linear table means for non-linearly converting the signal from the low-pass filter means;
Second calculating means for multiplying the signal from the non-linear table means by the pressure difference signal and supplying the multiplied signal to the transmission outward path of the wave guide means;
And the non-linear table means further comprises:
A memory for storing the non-linear characteristic of the sounding body; generating a read address integer part INT and an address decimal part FRC using a signal from the low-pass filter means as a read address; A plurality of addresses INT ± n and a plurality of addresses FRC ± n before and after the address decimal part FRC are generated, and the non-linear characteristic is referenced and read from the storage means in accordance with the address integer part INT and the plurality of addresses INT ± n before and after the address INT. Non-linear output means for generating a plurality of corresponding non-linear output signals,
A plurality of interpolation coefficients obtained from an interpolation function for interpolating the non-linear characteristic are stored, and a corresponding interpolation coefficient is read and read in accordance with the address decimal part FRC generated by the address generation means and a plurality of addresses FRC ± n before and after the address. Interpolation calculation means for forming and outputting a signal obtained by weighting and adding the plurality of non-linear output signals by each interpolation coefficient,
A musical sound generating device comprising:

Images