JP2722727B2 - Electronic musical instrument - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a musical sound synthesizer and an electronic musical instrument using the request sound synthesizer, and more particularly, to a musical sound synthesizer for synthesizing musical sounds using a so-called vibration analysis technique. And an electronic musical instrument using the musical sound synthesizer.

[Prior Art] Conventionally, musical tone synthesizing methods in electronic musical instruments can be roughly classified into a harmonic synthesizing method, a waveform reading method, a Farmant method, and the like.

Here, the harmonic synthesis method is the desired tone waveform F
(T) is represented by a Fourier series. First, a sample value is calculated for each harmonic component at each scheduled time interval of a sine wave having the same frequency as the harmonic, and the amplitude value of the harmonic component is added to this value. (See, for example, Japanese Patent Application Laid-Open No. 48-90217).

In the waveform reading method, a basic waveform of a musical sound waveform to be generated is stored in a memory in advance, and the stored waveform is read at a desired reading speed (for example, US Pat. No. 3,515,792).

In the formant method, various frequency components are extracted from a rectangular wave using a filter and are synthesized.

A tone generator that outputs a new sample value from a plurality of past sample values (sample values at fixed time intervals of a tone signal) using a regression device is disclosed in Japanese Patent Publication No. 58-27.
No. 518 is disclosed.

[Problems to be Solved by the Invention] However, according to the above-mentioned prior arts, there is an inconvenience that various sounds cannot be created based on the physical structure of a natural musical instrument whose pitch, volume, and timbre change with time. was there.

The present invention has been made in view of the above-mentioned problems in the conventional type, and provides a tone synthesizer and a tone synthesizer capable of producing various sounds based on the physical structure of a natural musical instrument, which cannot be obtained by the conventional method. It is intended to provide an electronic musical instrument used.

[Means for Solving the Problems] To achieve this object, an electronic musical instrument according to claim 1 approximates a musical instrument to synthesize a musical sound with a finite set of discrete elements connected by discrete nodes. Parameter setting means for setting parameter data including data indicating the shape or material of the musical instrument corresponding to the pitch, and a predetermined equation of motion for each of the discrete elements. Initial condition setting means for giving an initial condition corresponding to the initial displacement given to the musical instrument, and time-varying musical sound waveform data is generated by repeatedly calculating the solution of the above equation of motion based on the above parameter data and the initial condition. Waveform calculating means for generating musical tone waveform data for each tone range, and a storage means for storing the generated tone waveform data for each tone range. Steps and
Instruction means for instructing generation of a musical tone by designating a pitch, and sound generating means for reading out musical tone waveform data corresponding to the pitch designated by the instruction means from the storage means and generating sound. I do.

An electronic musical instrument according to a second aspect is characterized in that, in the first aspect, data indicating the shape or material of the musical instrument is scaled by multiplying the data by a predetermined coefficient corresponding to a pitch.

An electronic musical instrument according to a third aspect is characterized in that, in the first or second aspect, the initial condition includes a condition of displacement, speed, and / or acceleration corresponding to a touch of a performance operator of the electronic musical instrument.

According to a fourth aspect of the present invention, in the electronic musical instrument according to any one of the first to third aspects, the waveform calculating means uses the parameter data and the initial condition together with external force data including aftertouch data or breath data of the electronic musical instrument. It is characterized in that the solution of the equation is calculated.

[Operation] According to the electronic musical instrument of the present invention, the musical tone is represented by parameters approximated by predetermined discrete elements of the musical instrument, and initial conditions corresponding to the initial displacement are given to the musical instrument. Can be Under these conditions, a solution of a predetermined equation of motion is repeatedly calculated by a so-called vibration analysis technique. Thus, time-varying musical sound waveform data is generated. The generated musical sound waveform data is stored in the storage means in advance. On the other hand, when there is an instruction from the instruction means for instructing generation of a musical tone, the musical tone waveform stored in the storage means is read out and sounded. Since the parameter data includes data indicating the shape or material of the musical instrument corresponding to the pitch, musical tone waveform data can be generated and stored for each tone range. The instructing means can designate the pitch of a musical tone to be generated, and musical tone waveform data corresponding to the designated pitch is read out from the storage means and emitted.

If the parameter data is obtained by scaling data indicating the shape or material of the musical instrument by a predetermined coefficient corresponding to the pitch, the pitch of the generated musical tone waveform data is controlled as desired. be able to.

The initial conditions to be provided include conditions of displacement, velocity, and / or acceleration corresponding to a touch of a performance operator of the electronic musical instrument. The waveform calculating means calculates the solution of the equation of motion using external force data including after-touch data or breath data (data representing the degree of blowing when a wind instrument is assumed) together with parameter data and initial conditions. By calculating, musical tone waveform data is generated in a state closer to the actual sounding state of the musical instrument.

Embodiment An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows a schematic configuration of a musical sound synthesizer according to an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes parameter setting means, 2 denotes initial condition setting means, 3 denotes waveform calculation means, and 4 denotes tone waveform data output from the waveform calculation means 3.

FIG. 2 is a schematic diagram showing a model of a musical instrument for synthesizing a musical tone, and FIG. 3 is an external view of the musical instrument. Here, as a simple example, an example imitating Zanza, a Zaire folk instrument, will be described. Sanza is a body music instrument that has the appearance shown in FIG. 3 and is widely distributed in Africa.
Sanza is a box-shaped (or plate-shaped) resonator 8 as shown in the figure.
The tongue-shaped flakes 9 are arranged on the top. This lower flake 9
Play sound with your finger.

In this embodiment, a model as shown in FIG. 2 was created by imitating a Sanza, and parameters of the material such as its shape, density and elasticity were set. In the figure, P1 to P8 indicate respective nodes of a rectangular plate-shaped cross section. Nodes P7 and P8 are assumed to be fixed. T1 to T6 indicate discrete elements (triangular elements) constituting the rectangular plate.

Using this model as an example, a tone is synthesized by the following procedure.

(1) Create a stiffness matrix K for each of the triangular elements T1 to T6 from given parameters.

(2) Create a mass matrix M for each of the triangular elements T1 to T6 from the given parameters.

(3) Create the attenuation matrix C of each of the triangular elements T1 to S6 from the given parameters.

(4) Set initial conditions (position, speed, acceleration, etc.) of each triangle element.

(5) Based on the initial conditions set in the respective matrices (4) created in the above (1) to (3) and, if necessary, parameters such as external force data taking into account arithmetic means, and the like. Synthesize the sound waveform data.

(1) First, creation of a matrix K of triangular elements will be described.

In the present invention, a so-called two-dimensional finite element method is used. First, displacement, strain, and stress are set as follows for each of the points P1 to P8 of the model in FIG.

Displacement: Where (u (x, y) is the displacement in the x direction and v (x, y) is the y
Represents displacement in direction.

distorted: Stress (average stress model): Here, the Young's modulus is e, the Poisson's ratio is ν, and d ₁ = E / 1−ν ² ) d ₂ = νE (1−ν ² ) G = E / (2 (1 + ν)).

Here, each term of the stress equation (3) is replaced with one symbol, and the equation (3) is expressed by the following equation (4).

(Σ) = (D) (ε) (4) Using the above symbols, the strain energy U is Can be expressed as Here, h is the model thickness in FIG. 2, and the integral for D has the meaning of the integral for the object.

Next, using a so-called triangular element method as an approximation method, a relation between the displacement (u) of the vertex of the triangular element and the distortion (ε) in the element is obtained.

First, pay attention to one triangle element, and number its vertex
i, j, k. The displacement in the noted triangular element is approximated by a linear expression. This assumes that the strain (ε) is constant within the element.

u = α ₁₀ + α ₁₁ x + α ₁₂ y v = α ₂₀ + α ₂₁ x + α ₂₂ y (6) For convenience, all displacements of three vertices (i-th vertex, j-th vertex, k-th vertex) are written in one line Shall be.

_{(U) = (u i,} v i, u j, v j, u k, v k) T ...... (7) displacement u of (6), v is the vertex i, j, the in k (7 )
To stop the value of Must be However, (X _i , y _i ) (x _j , y _j ) (x _k , y _k ) are the i-th,
These are the coordinates of the j-th and k-th points.

When formula (9) is put into a matrix form, The above is the method of the triangle element. By using this method, the displacement (u) of the vertex of the triangle (u) = (u _i , v _i ,
_{u j, v j, u k} , u k) T and strain within the element _{(ε) = (ε x,} ε y,
ε _xy ) The relationship with ^T is obtained as follows. That is, from equation (2) and equation (II), If the coefficient matrix on the right side of the above equation (12) is represented by (B), then (ε) = (B) (u) (13)

The above calculation focuses on one triangular element that is the vertices i, j, k.
If the symbol of each matrix in 3) is given a subscript _ijk , then (ε) _ijk (B _ijk ) (u) _ijk (14)

Summarizing the above relationship, the triangular element component (δ) _ijk of the stress can be obtained from the equations (4) and (14) as (σ) _ijk = (D) (ε) _ijk = (D) (B _ijk ) ( u) _ijk ... (15)

On the other hand, since the triangular element component U _ijk of the strain energy is obtained by integrating the inner product of the strain and the stress, Here, (K _ijk ) = h × (triangle element area) × (B _ijk ^T (D) (B _ijk ) (17) Next, the external force (f) applied to the triangular element (vertex i, j, k) _ijk
Is (f) _ijk ^T = (f _i , g _i , f _j , g _j , f _k , g _k ) (18) However, the external force applied to the point n is represented by f _{n in the} x direction component and f y in the y direction component. Is g _n , Expressed by

If the external force in the x direction is f, the displacement in the x direction is u, the external force in the y direction is g, and the displacement in the y direction is v, the work performed by the external force is Becomes

Since the total energy of the triangular element is the sum of the strain energy and the energy of work, When this equation (2) is partially differentiated with u ₁ and set to = 0, However, since (K _ijk ) is a symmetric matrix according to the definition equation (17), K _i 1 = k _1i .

Therefore, k ₁₁ u _i + k ₁₂ v _i + k ₁₃ u _j + k ₁₄ v _j + k ₁₅ u _k + k ₁₆ v _k = f _i (22) Other external force components g _i , f _j , g _j , f _k , g _k can be calculated in the same manner. Put together in a matrix, That is, (K _ijk ) (u) _ijk = (f) _ijk (24) From the above, the value of the element of the rigidity matrix (K _ijk ) focusing on the triangular element having the vertices of i, j, k is obtained. It will be. Note that (K _ijk ) obtained here represents only a part of the i, j, k components of the rigidity matrix K in consideration of all elements. That is, as shown in FIG. 4, only a part (hatched portion) of the rigidity matrix K considering all the elements is obtained. Therefore, in order to obtain the rigidity matrix K, first, a matrix K having a size corresponding to all elements is prepared and all elements are set to 0. And some i,
The rigidity matrix (K _ijk ) is obtained for j and k as described above, and is added to K. Thereby, the rigidity matrix K in which all the elements are considered is obtained.

As described above, the rigidity matrix K of this musical instrument model has been obtained.

In this embodiment, the equation of motion in which components such as inertia and damping are inserted at (f) of the same equation obtained by expanding the above equation (23) (or equation (24)) to all elements is solved. This generates a waveform.

(2) Next, creation of the mass matrix M will be described.

In this embodiment, for the sake of simplicity, a so-called ramped mass matrix method is used, which assumes that the mass of each triangular element is concentrated at one vertex. As a result, the entire mass is divided into a plurality of regions around each point, and the mass of each region is assigned as the mass of the point.

Since it is assumed that this Sanza models evenly divide the uniform plate, the portion corresponding to the end plate of the diagonal elements of the matrix (1/2) m _u is, m _u other diagonal A matrix in which Here, the mass matrix M is created using the ramped mass matrix method. However, the present invention is not limited to this, and CM (consistent mass) is used.
Such a method can also be used.

(3) Next, the creation of the attenuation matrix C will be described.

The damping matrix C is a matrix that multiplies the damping matrix C by the velocity () of the displacement vector to give a damping force in the form of damping force = (C) (). By introducing this component, damped oscillation can be generated.

The damping matrix C is, as an approximation, multiplied by a constant to the stiffness matrix K to form a damping matrix C.
Alternatively, it can be created by a method such as multiplying a mass matrix M by a constant to obtain an attenuation matrix C.

(4) Next, a method of giving initial values and parameters to the tone synthesizer of this embodiment will be described.

One of the initial values given at the beginning is the initial displacement of the musical instrument model, and by changing this, it is possible to change the assumed playing strength of the musical instrument.

In the Sanza model shown in FIG. 2, there are eight nodes P1 to P8. Assuming that the x coordinate of the node P _n is represented by u _n and the y coordinate is represented by v _n , the initial displacement u (0) is u (0) = (u ₁ , v ₁ , u ₂ , v ₂ , u ₃ , v ₃ , u ₄ , v ₄ , u ₅ , v ₅ , u ₆ , v ₆ ,
written as _{u 7, v 7, u 8} , v 8) T. The equilibrium state u _b is all 0, and u _b = (0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0) ^T . Now, for simplicity, if only the nodes P1 and P2 are pulled by 5 mm in the y direction, the initial displacement u (0)
Is u (0) = (0,5,0,5,0,0,0,0,0,0,0,0,0,0,0,0) ^T. This is the value input as the initial displacement. Assuming that the nodes P1 and P2 are pulled more strongly by 5 mm in the y-direction, the initial displacement u (0) becomes u (0) = (0,10,0,10,0,5,0,5, 0,0,0,0,0,0,0,0)
It becomes ^T. Thus, the initial touch is input to the system as an initial condition.

Next, the case where the parameters for the system are changed will be described. Here, in the example of Sanza taken up, the vibration system is constant in time. Now, for example, if the length of the plate of the model in FIG. 2 is reduced, the cycle of vibration is shortened and the pitch is increased.

In this case, since the node coordinates change, the above-described stiffness matrix K, mass matrix M and damping matrix C
Changes. Therefore, the output pitch used for the waveform generation calculation described later changes, and the pitch (key) changes.

(5) Next, the mass matrix M, the rigidity matrix K, the damping matrix C, and the external force F prepared as described above
(T), and initial conditions u (0), (0), (0)
A description will be given of a waveform generation operation for creating a motion equation and obtaining a waveform sample value as a solution.

 First, the equation of motion is generally as follows.

M (t) + C (t) + ku (t) -f (t) = 0 (30) Note that u (t), (t), and (t) in equation (30) are
This is a vector in which data on the x direction and data on the y direction for all nodes are arranged. Also, f
(T) is also a vector in which the external force in the x direction and the external force in the y direction that act on all the nodes at regular intervals are arranged. Therefore, in practice, equation (30) can be said to be a collection of equations of motion in the x and y directions of each node.

In this equation of motion (30), the time at which Δt has elapsed from the current data u (t), (t), (t) at time t is assumed on the assumption that the change in acceleration is uniform for each section. The next sample point data u (t at t + Δt
+ Δt), (t + Δt), and (t + Δt). By repeating this, time-varying waveform data is generated.

In the example of Sanza used here, since the attenuation is a simple type, the attenuation is given later as an envelope, and the attenuation matrix C is not included in the calculation. Also, in Sanza's performance, no external force is applied except when playing for the first time. That is, the term f (t) in the equation of motion (30) is also zero, and the external force acts only on the external force that gives the initial displacement (initial displacement).

 Therefore, the equation of motion (30) is as follows.

M (t) + Ku (t) = 0 (31) Here, it is assumed that the acceleration of each point of each element changes linearly and can be determined at certain time intervals. That is, if the acceleration becomes α at one time and becomes β at the next time, the acceleration linearly changes from α to β between the times. Such, the assumption of linear acceleration, u (t + Δt) = u (t) + Δt (t) + {Δt 2/3) * (t) + {Δt 2/6} * (t + Δ
t) (32) (t + Δt) = (t) + {Δt / 2} * ((t) + (t + Δt)) (33) Expression (32) is replaced by t in Expression (31). substituting into the equation below was t + Δt (34) M ( t + Δt) + Ku (t + Δt) = 0 ...... (34), M (t + Δt) + K (u (t) + Δt (t) + {Δt 2/3} * ( ^{t) + {Δt 2/6} } * (t + Δ
t)) = 0 (35) ^{Therefore, (M + {Δt 2/} 6} * K) * (t + Δt) = -K (u (t) + Δt (t) + {Δt 2/3} *
(T)) ...... (36) Therefore, (t + Δt) = - (M + {Δt 2/6} * K) -1 * K (u (t) + Δt (t) + {Δt 2/3} *
(T)) (37) Thus, the current data u (t),
(T + Δt) of the next sample point at time t + Δt at which Δt has elapsed from (t) and (t) is obtained.

Further, this (t + Δt) is substituted into the above two equations (32) and (33) to obtain u (t + Δt) and (t + Δt).
t) is obtained.

FIG. 5 is a flowchart showing a process of synthesizing a musical tone waveform in the Sanza model.

In the figure, first, in step S1, node coordinates, strain ε, and stress σ are specified based on the input physical parameters of the system, for example, the shape and material of the lowered musical instrument, and the Young's modulus E
And the stiffness matrix K and mass matrix M are calculated using the Poisson's ratio ν. Then, calculate the ^{INV = (M + {Δt 2} /6} * K) -1 at step S2.

On the other hand, in step S3, an initial condition is set. Here, the initial speed (0) is set to zero, and the initial displacement u (0) is set to the distance pulled by a finger. Next, in step S4, (t + Δt) is calculated based on the above equation (37). In addition,
In step S5, (t + Δt) is calculated based on the above equation (33),
In step S6, u (t + Δt) is calculated based on the above equation (32).
Calculate each.

Then, in step S7, an appropriate point displacement is extracted and used as a sample value of the waveform data. In this example, u (t + Δt)
Among them, the point closest to the fixed point, that is, the displacement in the y direction of the node P5 in FIG. 2 is extracted, and this is used as the sample value of the waveform data. Note that the displacement of any node may be used as a sample value, and may be selected later.

In the second and subsequent iterations, the calculated u (t + Δt)
(T + Δt) and (t + Δt) are replaced by the following u (t),
The calculation after step S4 is performed using (t) and (t). The calculation is repeated in this manner to obtain sample values of the waveform output. The waveform output is followed by an envelope to generate a musical tone.

Here, Δt is a simulation step (sequential calculation interval) and is not the reciprocal of the sampling rate. Next, an example in which the above-described musical tone synthesizer is applied to an electronic musical instrument will be described.

FIG. 6 is a schematic block diagram of an electronic musical instrument to which the tone synthesizer of the above embodiment is applied. In the figure, reference numeral 11 denotes a parameter setting and initial condition setting unit of the tone synthesizer,
Reference numeral 12 denotes a waveform calculation unit based on the finite element method, 13 denotes a write processing unit for writing the musical tone waveform synthesized by the waveform arithmetic unit 12 to a memory, and 14 denotes a waveform bank memory for storing the synthesized musical tone waveform data.

Reference numeral 15 denotes a keyboard, 16 denotes a key press detection unit that detects a key press of the keyboard 15, 17 denotes a touch detection unit that detects touch data when the key of the keyboard 15 is pressed, and 18 denotes an operation for selecting a voice number or the like. The child 19 is a voice selection unit. Reference numeral 20 denotes voice data such as a voice number output from the voice selection unit 19, touch data output from the touch detection unit 17, and a key press detection unit.
A section for selecting and reading out the waveform data in the waveform bank 14 based on the key press key code output from the key 16. Reference numeral 21 denotes data such as a voice number from the voice selection unit 19, touch data output from the touch detection unit 17, and a key press detection unit.
An envelope generation unit that inputs data such as a key press key code output from 16 and receives a key-on signal from the key press detection unit 16 to generate an envelope. 22 is an integrator for adding the envelope from the envelope generator 21 to the waveform data read from the waveform bank 14,
Reference numeral 23 denotes a D / A converter for converting a digital tone signal output from the integrator 22 into an analog signal, and reference numeral 24 denotes a sound system for generating a tone based on the tone signal output from the D / A converter 23.

FIG. 7 shows a memory map of the waveform bank 14 of FIG. In the waveform bank 14, as shown by reference numeral 31, waveform data calculated by a method of vibration analysis is edited and stored for each voice number. The waveform data corresponding to one voice number is composed of waveform data for each sound range as indicated by reference numeral 32. Further, as shown by reference numeral 33, the waveform data of one range includes waveform data when the keyboard touch is weak and waveform data when the touch is strong. Each waveform data has an attack part and a loop part.

The electronic musical instrument of the embodiment shown in FIGS. 6 and 7 is basically a waveform memory type electronic musical instrument, and the waveform data calculated as described above is stored in advance in the waveform bank memory 14 prior to performance. That is, prior to the performance, first, the shape of the musical instrument and the initial displacement amount to be assumed are set by the parameters and the initial condition setting 11. Then, the waveform calculator 12 calculates the waveform data using the finite element method, and the writing unit 13 writes the waveform data to the waveform bank 14. At this time, various musical sound data are obtained by changing parameters as appropriate assuming some different types of musical instruments, and further changing the initial displacement amount, and storing the data for each voice number. As described above, the waveform bank 14
Stores waveform data as shown in FIG.

Next, assuming that the player selects the voice number with the voice selection operator 18 and presses the keyboard 15, the key press key code, touch data, and data such as the voice number are transmitted to the key press detection unit 16, the touch detection unit 17 and the like. And input from the voice selecting unit 19 to the reading unit 20. The reading unit 20 reads predetermined waveform data from the waveform bank 14 based on these data, and inputs the read data to the integrator 22. This waveform data is
At 22, the envelope is added to the envelope data output from the envelope generator 21 to give an envelope, and D /
A tone is generated based on the waveform data via the A converter 23 and the sound system 24.

The waveform data stored in the waveform bank 14 may be interpolated between different sound ranges or different touch waveforms.

FIG. 8 shows the configuration of an electronic musical instrument that synthesizes waveform data in real time using the musical sound synthesizer according to the present invention to generate musical sounds. In the figure, 41 is an operator for setting parameters and initial conditions and the like representing an assumed instrument model, 42 is an arithmetic unit that creates a predetermined matrix and initial values from data such as the set parameters,
Reference numeral 43 denotes a control unit for writing matrix data and the like created by the operation unit 42 into the parameter bank memory 44, and reference numeral 44 denotes a parameter bank memory for storing the matrix and initial base data. 46 is a keyboard of an electronic musical instrument, 47 is a key press detection unit that detects a key press of the keyboard 46, 48 is a touch detection unit that detects touch data when a key of the keyboard 46 is pressed, and 49 is a key for selecting a voice number. An operation unit 50 is a voice selection control unit.

45 selects a parameter stored in the parameter bank 44 based on the key code from the key press detection unit 47, the touch data from the touch detection unit 48, and the voice data from the voice selection control unit 50, and further scales the parameter. Is a parameter selection and scaling unit that performs 51
Is a waveform calculation unit that receives the key-on signal from the key press detection unit 47 and calculates waveform data by the finite element method, 52 is an envelope generator, 53 is the waveform data from the waveform calculation unit 51 and the envelope from the envelope generator 52 An integrator for integrating the data with the digital tone signal, a D / A converter for converting the digital tone signal into an analog tone signal, and a sound system 55. From the waveform calculator 51, to the parameter selection and scaling unit 45, timing information indicating that calculation is possible is passed.

FIG. 9 shows a memory map of the parameter bank 4 of FIG. As indicated by reference numeral 61, the parameter bank 44 has parameter data for each voice number arranged in order. The parameter data corresponding to one voice number is divided into several ranges as shown in reference numeral 62. Parameters corresponding to one range is configured as shown in numbering 63, i.e., first matrix ^{(M + {Δt 2/6} } * K) -1 and component data, the parameter K for scaling the components DOO matrix obtained by integrating ^{(M + (Δt 2/6} ) * K) stores component data of ^-1 * K. Next, a conversion table for converting touch data into initial value data numbers (1 to S) and initial value data 1 to S are stored.

The electronic musical instrument of the embodiment shown in FIGS. 8 and 9 is a type of electronic musical instrument that synthesizes waveform data in real time in parallel with performance and outputs musical tones. Since the calculation in the waveform calculation unit 51 needs to be performed at high speed, the calculations in S4, S5, and S6 are performed by a high-speed parallel calculator.

In the electronic musical instrument of this embodiment, parameters such as the shape and material of the musical instrument to be assumed are first set by the parameter setting operator 41 prior to the performance. Here, settings can be made for a plurality of musical instruments. Then, by selecting one voice number with the voice selection operator 49, a musical instrument (a musical instrument having a shape or material assumed in advance) corresponding to the voice number to generate a musical sound can be selected. The calculation unit 42 performs the above-described calculation based on the set parameters, and as a result, data as shown in FIG. 9 is stored in the parameter bank 44.

Next, it is assumed that the operator selects the voice number with the voice selection operator 49 and presses the keyboard 46. At this time, data such as a key press key code, touch data, and voice number are transmitted to the key press detection section 47, the touch detection section 48, and the voice tip
From 50, it is input to the parameter selection and scaling unit 45. The parameter selection and scaling unit 45 selects predetermined parameter data from the parameter bank 44 based on these data, and performs scaling processing so as to generate a sound of a pitch corresponding to a depressed key. The waveform calculator 51 calculates and outputs waveform data using the finite element method using these parameter data. This waveform data is stored in an integrator 53 by an envelope generator 52.
The envelope is added to the envelope data output from the PDP to give an envelope, and a tone is generated based on the waveform data via the D / A converter 54 and the sound system 55.

According to this embodiment, the selection is performed based on the selection key code (pitch data) of the parameter set in the parameter bank 45. On the other hand, if the size of the assumed musical instrument is increased or the material is set to be heavy, the pitch is lowered, and if the material is set reversely, the pitch is increased. Therefore, if appropriate parameters are set based on the pitch data, the pitch can be adjusted. Further, similarly, when the input touch data indicates a strong touch, the touch can be adjusted by setting a larger initial displacement.

The envelope waveform generator may be omitted by introducing an appropriate attenuation term. Further, a percussion sound may be generated by using a drum pad instead of the keyboard. Further, the operation of an external force may be added to the calculation by adding aftertouch detection.

FIG. 10 is a flowchart showing a procedure of tone synthesis taking into account parameters of external force and attenuation. Steps
That the damping matrix C is calculated in S11, and that a step S18 for giving an external force is newly added,
The processing procedure is the same as that of the flowchart of FIG. 4 except that the calculation formulas of steps S12 and S14 are different.

11 and 12 show examples of generation of a waveform calculated using the Sanza model. Since the parameters are different between FIG. 11 and FIG. 12, the waveform of FIG. 12 is a waveform including a harmonic component.

In the above-described embodiment, an example in which the waveform data is calculated and calculated by the vibration analysis using the linear acceleration method has been described. However, the present invention is not limited thereto, and the calculation unit may perform the calculation based on each of the following vibration analysis methods. Can also. In this case, the calculation unit of the above embodiment is set to perform calculation based on the following equation.

Here, β is a parameter for obtaining a stable solution, and 0 ≦
Select from the range of β ≦ 1/2.

In the above-described embodiment, the two-dimensional finite element method is used. However, the present invention is not limited to this. For example, a three-dimensional or one-dimensional (for example, a material having a different material for each part) may be used.

[Effects of the Invention] As described above, according to the present invention, waveform data is calculated and calculated by a vibration analysis method based on the physical structure of a natural musical instrument. A natural electronic musical instrument capable of making various sounds based on a physical structure and having a transient response is provided. In addition, a new timbre can be synthesized by using an unrealistic parameter.

Also, since it takes time to generate musical tone waveform data,
Prior to an instruction to generate a musical tone by the instruction means, musical tone waveform data is generated in advance and stored in the storage means, and when an instruction is issued by the instruction means, the musical tone waveform data stored in the storage means is read out and sounded. Thus, a musical sound can be generated immediately in response to an instruction from the instruction means. Furthermore, data indicating the shape or material of the musical instrument corresponding to the pitch is included as parameter data, and tone waveform data for each musical range is generated and stored using the parameter data corresponding to the pitch. When a musical tone generation instruction whose pitch is designated by the instruction means is issued, the tone waveform data corresponding to the pitch is read out and pronounced. Musical sound waveform data can be generated and sounded by faithfully imitating the phenomenon of "different shapes and materials".

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a musical sound synthesizer according to an embodiment of the present invention, FIG. 2 is a schematic diagram showing a model of a musical instrument for synthesizing musical sounds, FIG. FIG. 4 is a diagram showing the correspondence between the rigidity matrix K and the calculated K _ijk , FIG. 5 is a flowchart showing the process of synthesizing the sound waveform in the Sanza model, and FIG. FIG. 7 is a block diagram of an electronic musical instrument to which the musical tone synthesizer according to the embodiment is applied, FIG. 7 is a memory map of a waveform bank of the electronic musical instrument of FIG. 6, and FIG. 8 is musical tone synthesis according to an embodiment of the present invention. FIG. 9 is a block diagram of an electronic musical instrument in which waveform data is calculated in real time using the apparatus. FIG. 9 is a memory map of a parameter bank of the electronic musical instrument in FIG. 8, and FIG. 10 takes into account external force parameters. The processing procedure of the embodiment The flowcharts shown in FIGS. 11 and 12 are examples of synthesized musical sound waveform data. 1; parameter setting means, 2; initial condition setting means, 3; waveform calculating means, 4; musical sound waveform data, 11; parameter and initial condition setting section, 12; waveform calculating section, 13; composite waveform writing section, 14; waveform Bank, 15; Keyboard, 16; Key press detection section, 17; Touch detection section, 18; Voice selection operator, 19; Voice selection section, 20; Memory readout section, 21; Envelope generation section, 22; Integration section, 23 D / A converter, 24; sound system.

Claims (4)

(57) [Claims] 1. Parameter data obtained by approximating a musical instrument for which a musical tone is to be synthesized by a finite set of discrete elements connected at discrete nodes, and indicating the shape or material of the musical instrument corresponding to the pitch. Parameter setting means for setting parameter data including data; initial condition setting means for providing an initial condition corresponding to an assumed initial displacement given to the musical instrument with respect to a predetermined equation of motion for each of the discrete elements; Waveform calculation means for generating time-varying musical waveform data by repeatedly calculating the solution of the above-mentioned equation of motion based on data and initial conditions, and generating musical waveform data for each tone range; Storage means for storing the generated musical tone waveform data for each tone range; instruction means for designating a pitch and instructing generation of a musical tone; Electronic musical instrument tone waveform data corresponding to the high-specified sound, characterized in that a Could sound generating means reads from said storage means by. 2. The electronic musical instrument according to claim 1, wherein the data indicating the shape or material of the musical instrument is scaled by multiplying the data by a predetermined coefficient corresponding to a pitch. 3. The electronic musical instrument according to claim 1, wherein the initial condition includes a condition of displacement, velocity, and / or acceleration corresponding to a touch of a performance operator of the electronic musical instrument. 4. A method according to claim 1, wherein said waveform calculating means calculates a solution of said equation of motion using external force data including aftertouch data or breath data of said electronic musical instrument together with said parameter data and initial conditions. Electronic musical instrument as described.