JPH04171496A - Musical sound synthesizer - Google Patents

Abstract

PURPOSE:To cope with an actual noise generating mechanism by providing a noise generating means for generating a noise signal having a predetermined spectrum distribution according to an exciting signal and a noise overlapping means interposed in a predetermined position in a loop circuit to overlap its own input signal and the noise signal. CONSTITUTION:In a musical sound synthesizing circuit 9 is given noise data N of turbulence as offset according to data (f) corresponding to flow rate of air flow passing through a gap between a mouthpiece and reed to perform a signal process adapted to a noise generating mechanism in an actual wind instrument. A musical sound signal outputted through such signal process is supplied to a sound system 40 which processes a speaker SP to sound the musical sound. Thus, in actual wind instrument blowing, noise effects of sub-tones, panting sounds or the like magnifying noises can be reproduced, so that the musical sounds adapted to an actual noise generating mechanism are composed to reproduce the noises generated in performance with high fidelity.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a device for simulating a wind instrument, and more particularly to a musical tone synthesis device capable of faithfully reproducing wind instrument sounds accompanied by noise.

"Prior Art" In recent years, various devices have been developed that operate a model simulating the sound production mechanism of a natural musical instrument and thereby synthesize the musical tones of a natural musical instrument. This type of technology is, for example,
JP 63-40199 or JP 58-58
It is disclosed in Japanese Patent No. 679.

FIG. 9 shows the configuration of a musical tone synthesizer that simulates the sound production mechanism of a wind instrument.

In this figure, 11 is ROM (read only memory)
, 12 is an adder, 13 is a subtracter, and 14 and 15 are multipliers. These components 11 to 15 simulate the operation of the mouthpiece and reed of a wind instrument such as a clarinet, and constitute the excitation circuit 10.

Reference numeral 20 denotes a bidirectional transmission circuit that simulates the transmission characteristics of a wind instrument's pipe section, that is, a resonant pipe.

This bidirectional transmission circuit 20 is a delay circuit that simulates the propagation of air pressure waves in a resonant tube.

D, . . ., junctions JU, JU, . . ., inserted between these delay circuits; low-pass filter LPF that simulates energy loss when air pressure waves are reflected at the end of the resonant tube; and a bypass filter I (PF) that blocks DC components of data propagating within the bidirectional transmission circuit 20.

These junctions JU, JU, . A case is shown in which a 4-multiplication grid consisting of adders A s and A ++ is used. Here, "1+k", '-k", "1
-k", "k", etc. are multiplication coefficients, and the numerical value k is determined so as to obtain transmission characteristics close to those of an actual resonance tube.

In such a configuration, the adder 12 and the subtracter 13
Data P corresponding to the blowing pressure is input to . Then, the output data of the adder 12 is transmitted through the bidirectional transmission circuit 20 from delay circuit D to junction JU to delay circuit D.
... It propagates as follows, and the low pass filter LPF
reach. Then, after passing through the low-pass filter LPF and the bypass filter HPF, it propagates in the opposite direction to that described above, such as from the delay circuit D to the junction JtJ-, and is output from the bidirectional transmission circuit 20 to the subtracter 13.
is input.

Then, data P is subtracted by the subtracter 13 from the output data of the bidirectional transmission circuit 20 (this data corresponds to the pressure of the air pressure wave returned from the terminal end side of the resonance tube to the gap between the mouthpiece and the reed). be done. By this subtraction, data P corresponding to the air pressure in the gap between the reed and the mouthpiece is obtained. And this data P, or ROM
ROMI 1 outputs data Y corresponding to the cross-sectional area of the gap between the reed and the mouthpiece, that is, the admittance to air flow. FIG. 10 shows an example of a nonlinear function A stored in the ROMI 1 that shows the relationship between the air pressure in the gap between the reed and the mouthpiece (input) and the cross-sectional area of the gap (output).

Such data Y and data P are multiplied by a multiplier 14, and as a result, data FL corresponding to the flow velocity of air passing through the gap between the reed and the mouthpiece is obtained. This data FL is multiplied by a multiplication coefficient G by a multiplier 15. Here, the multiplication coefficient G is a constant determined according to the diameter of the pipe near the reed attachment section of the wind instrument.
It corresponds to the impedance to the air flow, that is, the impedance to the air flow. Therefore, the multiplier 15 calculates the product of the velocity of the airflow passing through the gap between the mouthpiece and the reed and the impedance of the tube to the airflow, which corresponds to the pressure change in the tube due to the airflow passing through the gap. Data P.

is obtained. And this data P, data P and force□
The signals are added by an adder 12 and input to a bidirectional transmission circuit 20.

In this way, data circulation, that is, resonance operation, is performed in the closed loop composed of the excitation circuit 1o and the bidirectional transmission circuit 20, and the data at the connection point of the low-pass filter LPF of the bidirectional transmission path 20 is extracted. , musical tones are generated based on such data.

By the way, in actual wind instruments, there is a playing style called subtone. This subtone is a playing technique that exaggerates the noise component that occurs when blowing into the gap between the reed and the mouthpiece. Conventionally, such noise synthesis has been performed by superimposing data corresponding to noise on data P corresponding to blowing pressure.

``Problems to be Solved by the Invention'' However, the above-mentioned noise component is caused by the turbulence that occurs in the gap between the mouthpiece and the reed when you breathe, and this turbulence causes air pressure waves in the tube. disturbed,
This is the noise that is generated. Therefore, the conventional noise reproduction method described above, that is, simply superimposing data corresponding to noise, does not correspond to the actual noise generation mechanism □, and the noise generated by this method lacks naturalness. There was a drawback.

This invention was made in view of the above-mentioned circumstances, and aims to provide a musical tone synthesis device that is compatible with the actual noise generation mechanism, and thereby to faithfully reproduce the noise effect when playing a wind instrument. There is.

[Means for Solving the Problems] The present invention comprises an excitation means that generates an excitation signal corresponding to performance information, and a loop circuit that delays its own input signal for at least a predetermined period of time and repeatedly circulates the excitation signal. In a musical tone synthesis device configured to use a circulating signal of the loop circuit as a musical tone signal, a noise generating means for generating a noise signal having a predetermined spectral distribution according to the excitation signal; The apparatus is characterized by comprising a noise superimposing means inserted at a position of the apparatus for superimposing the noise signal on its own input signal.

"Operation" According to the above configuration, the noise generating means generates a noise signal that approximates the turbulent flow in the pipe in response to the excitation signal, and the noise superimposing means superimposes the noise signal on the circulating signal of the loop circuit. This allows musical tone synthesis to be performed in accordance with the actual noise generation mechanism.

"Embodiments" Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the overall configuration of a musical tone synthesizer according to an embodiment of the present invention.

In this figure, reference numeral 1 denotes an operator imitating a wind instrument such as a clarinet, which generates various signals in response to operations by a wind player. Here, the configuration of the operator 1 will be explained with reference to FIGS. 2 and 3.

First, FIG. 2(A) is an external view showing an example of this operator 1. As shown in FIG. In the figure, 1a is a key switch that generates a key-on signal Kon and a key-off signal Koff. 1b is the mouthpiece, and inside this is the same figure (b).
As shown in the figure, a cantilever 1c and a pressure sensor 1d are provided. The cantilever IC detects and outputs the pressure (this pressure is called amperage) applied to the reed when the blow player holds the mouthpiece 1b in his/her mouth.

On the other hand, the pressure sensor 1d detects and outputs the breath pressure blown into the mouthpiece 1b by the blow player.

The operator 1 having such a configuration supplies each of the above-mentioned detection signals to the microcomputer 5, as shown in FIG. This microcomputer 5 is a C
PU2, ROM3, RAM4 and A not shown
It consists of a D converter. The microcomputer 5 having such a configuration converts the signals supplied from the key switch la, the cantilever IC, and the pressure sensor 1d into digital data, and then performs processing such as scaling to generate various data.

Among these data, key-on Kon/key-off Koff
The data is generated by determining whether the output signal of the pressure sensor 1d is greater than or equal to a predetermined value. The cent value data C is data representing pitch, and is generated based on a key code corresponding to the operation of the key switch 1a.

The breath pressure data P is generated according to the output signal of the pressure sensor 1d, and is modified by the amperage data E obtained from the cantilever IC to produce a blowing feeling.

Next, the configuration of the embodiment will be described with reference to FIG. 1 again. In the figure, 6 is a noise generator, which generates noise data N for reproducing the aforementioned subtones and breath leak sounds.
occurs. This noise data N is used to simulate the turbulence phenomenon in the gap between the mouthpiece and the reed, which will be described later. Note that the configuration of this noise generator 6 will be described later. Reference numeral 7 denotes a reed section, which simulates the operation of a portion consisting of a mouthpiece and a reed of a wind instrument. Reference numeral 8 denotes a pipe section of a wind instrument, that is, a pipe body section that simulates the transmission characteristics of a resonant tube. These components 6 to 8 constitute a musical tone synthesis circuit 9.

Next, the configuration of this musical tone synthesis circuit 9 will be explained with reference to FIG. 4. In this figure, parts corresponding to those in FIG. 9 described above are given the same reference numerals. This musical tone synthesis circuit 9 is composed of an excitation circuit 1O, a junction 22, and a tube forming circuit 20. Excitation circuit 1
0 is the subtracter 13, adder 16, 33, multiplier 31, 3
2,34, ROMI la, 1 lb and filter 3
0a and 30b, this circuit simulates the mouthpiece section of a wind instrument. Junction 22 consists of adders 22a and 22b. The tube forming circuit 20 is a circuit that simulates a resonant tube in a wind instrument, and its details will be described later.

First, at this junction 22, the output data of the multiplier 34 and the tube forming circuit 20 are added by the adder 22a and input to the tube forming circuit 20, and the output data of the tube forming circuit 20 and the adder 22a are added together. The output data of the adder 22b is further input to the subtracter 13 via the filter 30b.

In this way, the scattering of air pressure waves at the end of the resonance tube on the mouthpiece side is simulated.

Here, the reason why the filter 30b is inserted is to prevent the signal circulating between the excitation circuit 10 and the tube forming circuit 20 from becoming significantly large at a specific frequency.

The subtracter 13 is input with breath pressure data P corresponding to the blowing pressure, and feedback data via the filter 30b (this data is the air pressure reflected at the end of the resonance tube and returned to the mouthpiece side). wave) is input. Then, data corresponding to the air pressure in the gap between the mouthpiece and the reed is output from the subtractor 13, and the filter 30
input to a. This filter 30a simulates the movement of the lead, limits the band of the input signal, and outputs the signal. This is because when the pressure on the reed is changed, there is a delay in the displacement of the reed due to the inertia of the reed itself. Furthermore, if the frequency of this pressure change is high, the reed will not respond. Bandwidth limitation is performed in this filter 30a so as to simulate the followability of the lead in response to such pressure changes. In addition, this filter 30a is configured to apply an initial displacement to the leads in accordance with the above-mentioned amp joule data E.

The adder 16 adds the amplifier joule data E as an offset to the output data P of the filter 30a to obtain data P corresponding to the pressure actually applied to the lead.

Then, this data P is input to R,0M11a and is equivalent to the admittance for the air flow in the gap between the mouthpiece and the reed, according to the nonlinear function A (see Figure 10) stored in R'0M11a. Data Y
The table is converted to Subsequently, this data Y and the constant S are multiplied by a multiplier 3]. Here, since this constant is a value corresponding to the gap width between the mouthpiece and the reed, data FL corresponding to the flow velocity of the airflow passing through the gap between the mouthpiece and the reed is obtained by this multiplication.

By the way, the flow velocity of the air flow in the gap between the mouthpiece and the reed changes depending on the air pressure, but becomes saturated when it reaches a certain velocity. FIG. 5 shows an example of the nonlinear characteristic B indicating the saturation characteristic of the flow velocity of the air flow. This saturation characteristic is stored in the ROM 11b as a table of nonlinear function B. And this RO
The output of the subtractor 13, that is, the data corresponding to the air pressure in the gap between the mouthpiece and the reed mentioned above, is input to M11b to perform table conversion and output data representing the air flow velocity in the saturated state. . This data is multiplied by the above-mentioned data FL in the multiplier 32, and as a result, data f representing the true air flow velocity in the gap is obtained. 'Next, the configuration of the noise generator 6 that generates the noise data N in accordance with the data f will be explained with reference to FIG. In this figure, 40 is a second-order low-pass filter composed of subtracters 40a, 40e, adders 40b, 40g, coefficient multipliers 40c, 40f, and delay circuits 40d, 40h. This low-pass filter 40 uses a white noise signal WN supplied from the outside to simulate the turbulence phenomenon in the gap between the mouthpiece and the reed.
is filtered and output as noise data N. This filtering approximates the spectral distribution of the white noise signal WN to the spectral distribution of turbulent flow, and the spectral distribution of turbulent flow is determined according to the Reynolds number R of data f, as will be described later. 4
Reference numeral 1 denotes a filter constant generation circuit composed of multipliers 41a and 41C and an arithmetic unit 41b, the details of which will be described later.

Here, filtering of the above-mentioned white noise signal WN will be explained. First, the Reynolds number R is a value that represents the state of flow in a fluid, and it is generally known that when this value is less than 2000 (a dimensionless quantity), the flow becomes laminar, and when it exceeds this value, the flow becomes turbulent. According to Kolmogorov's law, the larger the Reynolds number R, the more
The spectral components of turbulent flow extend to the low frequency region, and the energy of the direct current component is said to be approximately proportional to the Reynolds number R. Therefore, by filtering the white noise signal WN having a −-like spectral distribution according to the Reynolds number R, it is possible to approximate a turbulent flow having a desired spectral distribution.

Based on this principle, in this low-pass filter 40, the coefficient α of the coefficient multiplier 40f is changed according to the Reynolds number R of the data f, and its frequency characteristics are made as shown in FIG. In the figure, the horizontal axis represents the log frequency, the vertical axis represents the dB amplitude, and the coefficient α.

The frequency characteristics are shown in each case where is 01.003.001.0003 and 0.001.

By the way, looking at the correlation between this frequency characteristic and the Reynolds number R, it can be seen that the coefficient α and R-'/'' are in a proportional relationship.Therefore, in the filter constant generation circuit 41 described above, first, the multiplication In the device 41a, the above-mentioned data f is multiplied by the constant C1 to obtain the Reynolds number R. Here, the constant C7 is a value given by a/ν, and a represents the thickness, width, etc. of the flow. The value ν is the kinematic viscosity coefficient.

The Reynolds number R obtained in this way is calculated by the computing unit 41b.
The calculation of R-"" is performed, and this is multiplied by a constant C in the multiplier 41c. The constant Cm is a value for making both the coefficients α and R-"' correspond to each other. The coefficient data Dα2 output from the multiplier 41c in this way is transmitted to the coefficient multiplier 40f in the low-pass filter 40 described above.
Set the multiplication coefficient α. As a result, the cutoff frequency of the low-pass filter 40 is controlled in accordance with the Reynolds number R of the data f which changes from moment to moment, and noise data N simulating turbulent flow is output.

Next, this noise data N is sent to an adder 33 (see FIG. 4).
As a result, data FLN to which data corresponding to turbulent flow is added as an offset is output. Then, this data FLN is processed by multiplier 3.
The constant Z is multiplied by 4. This constant Z is a value determined according to the diameter of the pipe near the reed attachment part of the wind instrument, and corresponds to the difficulty of passing airflow, that is, the impedance to the airflow. Through these multiplications, data corresponding to the air pressure inside the pipe is obtained,
This data is input to the tube forming circuit 20 via the adder 22a of the junction 22. Then, the output data of the tube forming circuit 20 is output to the junction 22,
The signal is inputted to the subtracter 13 via the filter 30b, and the same signal processing as described above is performed again.

Here, the configuration of the tube forming circuit 20 will be explained with reference to FIG. In this figure, 20a.

..., 20a is a delay circuit composed of a shift register, which simulates the propagation delay of air pressure waves in the resonance tube. 20b, . . . , 20b are junctions inserted between these delay circuits 20a. 20
c is an inverter that simulates the reflection of air pressure waves at the end of this resonant tube. 20d is a low pass filter (LPF), 20e is a bypass filter (H
PF). According to this configuration, the signal output from the final stage delay circuit 20a is divided into one that passes through the LPF 20d and is reflected at the tube end, and one that passes through the HPF 20e and is output as musical tone output. Here, HPF2
The reason why 0e is inserted is that the acoustic radiation impedance characteristic of a wind instrument is a high-pass characteristic. Note that the delay amount d shown in the figure.

~dn, 1~kn-s for the junction coefficient and filter coefficients FCL, FCH are the aforementioned key food,
It is given as a result of calculation based on the amperage data E1 and the breath pressure data P, and this calculation is performed by the CPU2.
Executed by

In the musical tone synthesis circuit 9 configured in this manner, noise data N of turbulent flow corresponding to data f corresponding to the flow velocity of the air flow passing through the gap between the mouthpiece and the reed is added as an offset, so that the actual Signal processing is performed in accordance with the noise generation mechanism in wind instruments. The musical tone signal outputted after such signal processing is supplied to the sound system 40 shown in FIG. 1, and this sound system 40 performs processing for producing musical tone from the speaker SP.

This makes it possible to faithfully reproduce the noise effects that occur when playing a wind instrument, such as the subtones that become louder when the airflow velocity in the gap between the mouthpiece and reed becomes saturated when you blow into the wind instrument strongly, and the sound of breath leakage. become.

In the above embodiment, the noise data N is added to the data f in the adder 33 as shown in FIG. As shown in (b), an adder 35 and a multiplier 36 are provided, and after adding 1 to the noise data N, the data f
You may also do this by multiplying by In this way, data f is modulated with noise data N.

"Effects of the Invention" As explained above, according to the present invention, the noise generating means generates a noise signal that approximates the turbulent flow in the pipe according to the excitation signal, and the noise superimposing means adds the noise to the circulating signal of the loop circuit. Since the signals are superimposed, musical tones are synthesized in accordance with the actual noise generation mechanism, and the effect is that the noise generated during performance is faithfully reproduced.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the overall configuration of a musical tone synthesizer according to an embodiment of the present invention, and FIG. 2 is an external view showing an example of an operator l in the same embodiment. FIG. 3 is a block diagram showing the configuration of the information conversion circuit in the same embodiment, FIG. 4 is a block diagram showing the configuration of the musical tone synthesis circuit 9, and FIG. 5 is a diagram explaining the nonlinear function B stored in ROMI lb. , FIG. 6 is a block diagram showing the configuration of the noise generator 6, FIG. 7 is a diagram showing the frequency characteristics of the low-pass filter 40, FIG. 8 is a block diagram showing the configuration of the tube forming circuit 20, and FIG. 9 is a block diagram showing the configuration of the tube forming circuit 20. A block diagram showing the configuration of a conventional musical tone synthesis device, FIG. 10 is a diagram explaining the nonlinear function A in FIGS. 4 and 9, and FIG. 11 is a block diagram showing an example of adding noise data N to data f. It is. 5...Microcomputer, 6...
Noise generator, 7...Lead section, 8...
・Pipe body part, 40...Low pass filter, 41・
...Filter constant generation circuit.

Claims (1)

[Claims] Excitation means for generating an excitation signal corresponding to performance information;
delaying its own input signal for at least a predetermined time;
A musical tone synthesizer comprising a loop circuit that repeatedly circulates the excitation signal, and uses the circulating signal of the loop circuit as a musical tone signal, wherein a noise signal having a predetermined spectral distribution is generated in accordance with the excitation signal. A musical tone synthesis device comprising a noise generating means and a noise superimposing means inserted at a predetermined position in the loop circuit to superimpose the noise signal on its own input signal.