JPH04191893A - Keyboard musical instrument - Google Patents

Abstract

PURPOSE:To improve timbres and sound volume by calculating the vibration state of a tensed string and a virtual string of the same interval as this tensed string in accordance with the line spectrum indicating the vibration state of the tensed string, converting this state to a sound and producing the sound after the prescribed time set shorter than the audible identification time from the sounding time of the tensed string. CONSTITUTION:The sound corresponding to the tensed string 11 is produced if this string 11 is struck. A detecting means 12 simultaneously detects the vibration state of the tensed string 11 and a calculating means 13 calculates the vibration state of the virtual string of the same interval as the interval of this tensed string 11 in accordance with the ray spectra indicating this vibration state. The sound producing means 14 produces the sound by converting the vibration state of the virtual string to the sound. The sound producing means 14 produces the sound within the time shorter than the audible identification time after the sound producing by the striking of the tensed string 11 in such a case. The sound produced by the tensed string and the sound produced by the virtual string are eventually simultaneously audible to a listener. Both the timbres and sound volume of the sound produced by a key touch are improved.

Description

[Detailed Description of the Invention] <Industrial Application Field> The present invention relates to a single-string keyboard instrument, in which the vibration state of a virtual string having the same pitch as the stretched string is determined based on the vibration state when the attached string is struck. This method improves the timbre of a single-string keyboard instrument by calculating the sound and superimposing it on the sound of the strings being struck.

<Prior Art> Conventionally, as this type of keyboard instrument, a monochord piano (or double chord piano), in which only one (or two) strings are attached to each key, has been known. That is,
Unlike a regular three-string piano, there is one string for each key.
Only the book is mounted on a frame.

<Problems to be Solved by the Invention> However, such conventional monochord pianos have inferior tone quality compared to ordinary three-string pianos.
There was also the issue of low volume.

In other words, the rich sound of a normal piano is due to
This is because there are three (rarely four) strings that act as sound sources, and each string has slightly different pitches, line spectra, attenuation, etc.

Therefore, the present invention aims to provide a single-string keyboard instrument with excellent tone color and increased volume by correcting the partial tones that make up the amount of strings struck in a short period of time close to real time and producing additional sounds. That is the purpose.

<Means for Solving the Problems> The present invention will be described with reference to FIG. 1. The present invention will be described with reference to FIG.
2, a calculation means 13 for calculating the vibration state of a virtual string having the same pitch as the tension value 11 based on a line spectrum indicating the vibration state of the tension value 11 detected by the detection means 12; Sound generation means 14 converts the vibration state of the imaginary string calculated by the calculation means 13 into sound, and generates the sound after a predetermined time set shorter than the audible identification time from the sound generation of the tension value 11. It is a keyboard instrument.

Effect> In the keyboard instrument according to the present invention, by striking a string with a tension value of 11, a sound corresponding to the string 11 is produced. at the same time,
The detection means 12 detects the vibration state of this tension value 11,
Based on the line spectrum indicating this vibration state, the calculating means 13 calculates the vibration state of the virtual string having the same pitch as this tension value 11.

Then, the sound generating means 14 converts the vibration state of the virtual string into sound and generates sound.

In this case, the sound generating means 14, after sounding the string with the tension value 11,
Pronounce the sound within a time shorter than the audible identification time.

As a result, the sound produced by the tension value and the sound produced by the virtual strings are heard by the listener at the same time (they are heard as the same sound), and the tone and volume of the sound produced by pressing the keys are both improved. It is possible to obtain a tone similar to or better than that of a string piano.

More specifically, as shown in FIG. 1, in this keyboard instrument, when a key is pressed, a hammer 15 rotates and strikes a string 11 stretched on a frame. The detection means 12, which is composed of a piezoelectric element or the like, detects the vibration state of the string 1 as a time waveform representing string vibration, for example.

This string vibration is generated when the detected signal is Fourier-transformed by, for example, an FFT (fast Fourier transformer) 16 (see FIG. 2) and corrected by a data correction circuit 17. It is expressed as a line spectrum showing the vibration state of two virtual strings with the same pitch. That is, by detecting the vibration of one string 1, the vibration state of the other two virtual strings is calculated.

These FFT 16 and data correction circuits 17 and 18 constitute the calculation means 13.

The output signals from each data correction circuit 7.18 are converted into analog signals by digital/analog converters 19 and 20, respectively, and input to the amplifier 21 constituting the sound generation means 14.

The output from the amplifier 21 is supplied to a soundboard drive actuator 23 that is fixed to the soundboard 22 and generates sound by vibrating the soundboard 22, or to an external speaker 24.

That is, the sound generating means 14 is composed of an amplifier 21, a soundboard drive actuator 23, and an external speaker 24.

In this case, the sound generating means 14 generates the sound within a certain period of time after the sound is generated by the vibration of the string 11.

In other words, signal processing is performed at high speed using FFT16 etc.
The structure is such that the sound is generated within the audible identification time.

Note that 25 is a piece for transmitting vibrations of the strings 11 to the soundboard 22.

FIG. 2 is a graph of a line spectrum for explaining the process of calculating the vibration state of the virtual string by the calculation means 13.

For example, the signal output from the FFT 16 at time t8 causes a line spectrum to be represented by a graph of Sg. The string vibration represented by this line spectrum rises sharply with the passage of time, and then gradually decays by $5. At times t1 and t2, the line spectra of Sl and S2 are displayed, respectively.As shown in Figure 2, the vibration state of the strung string 11 changes with the passage of time. Therefore, a predetermined correction process is also required for calculation of the vibration state of the virtual string.

In addition, the vibration state of the virtual string calculated by ・corresponding to two strings with the same pitch as ``Strung Strings】1 is slightly different on the treble side and very slightly on the bass side, and the strung strings 11
The line spectra of vibrations are different from each other.

Therefore, also from this point of view, when calculating the vibration state of each virtual string based on that of the stretched string 11, a correction process is required for each virtual string.

As shown in the figure, this kind of correction is performed by frequency correction A.
1 and intensity correction A2. For example, the intensity ■o of a certain frequency in the line spectrum S2 at time t2 is calculated by IQ with different frequency and intensity by these corrections.

Then, the sound is produced based on this correction value.

For example, the tonal effect caused by the slight difference in the tuning of each string mentioned above can be achieved by adjusting the frequency of vibration caused by the electrical signal in the above correction, for example, from 0.2 cents to 0.2 cents. This can be achieved by reproducing (pronunciation) with a difference of 3 cents.

When correcting a sound that is clogged, the major cause of the blockage is that a particular dominant partial attenuates quickly, so if that particular partial attenuates faster than its neighboring partials, it is indicated by a line spectrum. Increase strength.

Deterioration in sound quality due to excessive inharmonicity is corrected by correcting frequencies in the line spectrum. For example, for the second overtone of 1,000 [H2], 2,005 [Hz: 1 to 2,002 [H
z]. Note that in the case of this excessive inharmonicity, the correction value differs for each overtone.

Furthermore, if the attenuation is fast overall, correction is made to uniformly increase the intensity of the line spectrum as time passes.

In addition, if the tension of the strings is low and the sound is harpsichord-like, the line spectrum intensity of the higher harmonics will be reduced.
Correct to increase the lower harmonics. In this case, since a fixed envelope correction is sufficient for the amount of generation, this can also be performed using an analog filter.

<Examples> Examples of the present invention will be described below with reference to the drawings.

FIG. 3 is a block diagram showing the general configuration of a keyboard instrument according to an embodiment of the present invention.

In this figure, reference numeral 31 denotes a frame in which a vertical piano is placed, and one string 32 for each key is stretched with a tuning pin, hitch pin, or the like. 33 is a sound board arranged on the back side of this frame 3], and 34 is a piece. This frame 31 has a small number of tensioned strings (one string instead of the usual three for six keys), so the overall tension that acts on it is significantly reduced, making it compact. , using lightweight ones.

The other components and mechanisms are the same as those of a conventional three-string vertical piano. For example, an action 36 is arranged for 6 keys 35, and the hammer 3 of the action 36 is
7 is a t-shaped structure that strikes the single string 32.

A piezoelectric element 38 for detecting the vibration state of the string 32 is attached to the string rest (or a bridge) in the upper part of the frame 31. The piezoelectric elements 38 are arranged corresponding to all the strings 32 (for example, 88 strings). Note that this piezoelectric element 38 may be anything that picks up the vibrations of the strings 32, and may be constructed from other mechanical-electrical conversion elements.

Further, a reflective optical sensor 40 is disposed opposite the end surface of the catcher 39, and this optical sensor 40 detects the speed and timing of string striking by the hammer 37.

Furthermore, a key sensor 42 is mounted on the shelf board 41 below the six keys 35.
is installed. These key sensors 42 are configured as reflected light type optical sensors.

Further, the sound board 33 is electromagnetically driven on the back surface of the sound board 33.
An actuator 43 that vibrates is provided.

And, 44 is an external speaker built into the main body of the piano, which generates sound by a drive circuit to be described later.

Here, in this vertical piano, the vibration state of the string 32 is detected, the vibration state of the virtual string is calculated based on the line spectrum indicating this vibration state, and the predetermined vibration state is calculated from the sound produced by the hammer 37 when the string is struck. It has a configuration that allows sound to be generated via the actuator 43 or the external speaker 44 within a time period (approximately 5 m5 ec).

That is, the vibration of the string 32 struck by the hammer 37 in response to the key depression is detected by the piezoelectric element 38, and this string vibration signal is filtered by the high-cut filter 45, for example,
High frequency components of 0 [Hz] or higher are gradually attenuated and removed as noise.

Further, the string vibration signal after noise removal is sampled at predetermined time intervals by an A/D converter 46 and converted into a digital signal.

This digital vibration signal is then input to an FFT (fast Fourier transformer) 47.

FFT47 performs Fourier transform on this digital signal.
A line spectrum is calculated and sent to two DSPs (digital signal processors) 48.49.

These two DSPs 48°49, one of which is selected according to instructions from the MPU (main control unit) 50, are as follows:
The following arithmetic processing is performed on the input line spectrum.

That is, it is an anharmonicity correction calculation, a virtual string vibration data calculation process, a partial tone level and attenuation correction calculation, and the first and second virtual string data are inversely Fourier transformed, respectively. This is to calculate a virtual string vibration signal.

This arithmetic processing is repeatedly executed, for example, every 5 [m5ecl] from when the string is struck until the sound is stopped or muted.

The human temporal discrimination limit for two different sounds is approximately 5 m5 ec, and this time is used as the calculation time.

The above-mentioned 11th. The second virtual string vibration signal is D/
The signals are converted into analog signals by an A converter 51 or 52, and outputted to a drive circuit 53 or 54 as first and second virtual string vibration signals, respectively.

The drive circuit 53 drives the actuator 43 for driving the soundboard. Further, the drive circuit 54 drives the external speaker 44.

Therefore, the MP according to the setting of the output selector switch 55
The drive circuit 53 or 54 on the enabled side operates according to the command from U50, and sound can be generated from the sound board 33 or external speaker 44.

The M P tJ 50 is composed of a well-known microcomputer, etc., and includes a CPU 56, a ROM 5, etc.
7, RAM 5 B and input/output section (Ilo) 59 are connected by a lotus. Therefore, the pressed key number is determined by the key sensor 42, the hammer speed is determined by the optical sensor 40, and the timing of the sound is determined by the piezoelectric element 38 (or the optical sensor 40).
0) are respectively input to the input/output section 59. The MPU 50 sends these data to each sensor 38.
Detected by 40.42 and added to MIDIID
The configuration is such that it can also be output to the DI terminal 60.

In addition, each of the above DSP48.49 has a ROM61.6
2. It is configured with RAM 63.64. R
Arithmetic processing is performed on the above line spectrum based on a program stored in OM61.62. RAM
63 and 64 function as temporary storage devices in this arithmetic processing. These DSP48.49 are
The frequency and intensity of a predetermined overtone at a predetermined time are calculated to desired values.

Here, FIG. 4 shows a flowchart executed by the MPU 50.

This program is started when power is turned on to the MPU (main control unit) 50, and is repeatedly executed at predetermined time intervals.

First, a changeover switch signal indicating the setting of the changeover switch 55 is inputted (step Sl), and an activation signal for the drive circuit 53 or 54 is outputted in accordance with the setting of the changeover switch 55 (S2). This is for setting which of the actuator 43 or the external speaker 44 the output is to be output to.

Next, detection signals from the key sensor 42 and the optical sensor 40 for hammer speed detection are input (S3). Then, based on the pressed key number and the hammer speed, 5P48 or 49 is selected as required for signal processing (S4
). This DSP selection processing subroutine establishes a map that defines the relationship between the key number and the corresponding DSP. The second map is stored in the ROM 57 in advance.

Then, M P L' 50 is the selected DSP4.
A command that only 8 or 49 operates effectively is sent to that DSP4.
8 again! , t49 (S5).

As a result, the DSP 48 or 49 performs calculations such as correction processing based on the above-mentioned line spectrum, and outputs a correction value to the drive circuit 53 or 54.

Next, a MIDI signal corresponding to the string striking sound is generated based on the key number and hammer speed according to a map stored in advance in the ROM 57 (S6). This MIDI signal is output to the MIDI terminal 60 (S7).

Hereinafter, the arithmetic processing in the DSP 48 or 49 will be explained with reference to FIGS. 5(A) to 5(E).

FIG. 5(A) is a graph of a line spectrum for explaining the anharmonicity correction of the vibration data of the tensioned string 32 (actual value).

This correction is to correct the frequency of the high-order harmonic component shown in the line spectrum, for example, the frequency of the n-th harmonic to a frequency that causes appropriate anharmonicity. In the graph,
○ indicates actual vibration data, and ○ indicates vibration data after anharmonicity correction.

At the time of pronunciation (if the fundamental tone of ti is 1 day, the nth overtone nfs
is nfi after correction! ' The frequency is expressed as '.

In this case, the strength is the same. Also, t after 5m5ec
The correction at 1 and the correction of the higher harmonics at t2 5m5ec later are corrections that lower the frequency.

For example, as a method for quantitatively determining appropriate inharmonicity, an arithmetic expression (formula ■) that defines the correlation between the key number and the parameter b value or an imaginary map may be stored in advance in the ROM61 or 62.

Then, in accordance with the key number of the actual vibration data read this time, a parameter b value indicating appropriate anharmonicity is determined according to the arithmetic expression or map held in the R0M61 (or 62).

Based on the parameter b value obtained in this way, a correction is made to reduce the frequency of the actual vibration data, especially the higher-order harmonics, so that there is an appropriate pitch (pitch) difference from the n-th harmonic. What to do From here, obtain the anharmonicity corrected data. This data becomes data after time t1.

1200・10g2(fo/n-fe)=b-n2...
・・・・・・・・・・・・■ Here, f, [Hz] is the frequency of the harmonic overtone, and nfs [Hz] is the frequency of the n-th overtone. Note that n is an integer, and the unit on the right side is cents.

As described above, the frequency of the n-th overtone is corrected to decrease to f.

FIG. 5(B) is a graph for explaining calculation of the first virtual string vibration data and the second virtual string vibration data. In the figure,
The mouth indicates first virtual string vibration data (right string), Δ indicates second virtual string vibration data (left string), and O indicates data after anharmonicity correction, respectively.

This correction is based on the fundamental tone f8 and overtone 2 of the data after anharmonicity correction.
For example, the fundamental tone f1 of the first virtual string vibration data is 0.2 cents lower.
Overtones 2f1~nf1'~ and, for example, fundamental tone f2 of the second virtual string vibration data that is 0.3 cent higher, overtones 2f2~
This is to calculate nf2'~.

First, in order to create the sound of a piano with three strings for one key, anharmonicity correction is applied to the actual string vibration data obtained by hitting the first string (center string). The first virtual string vibration data (
right string) and second virtual string vibration data (left string).

Here, the pitch of the right string is 0.2 cents lower than the pitch of the center string, while the pitch of the left string is 0.2 cents lower than that of the center string.
．． An example of setting the price 3 cents higher is shown below. Note that the pitch difference between the left and right strings with respect to the center string does not have to be uniform.

It is assumed that a pitch difference parameter that determines the pitch is stored in advance in the ROM 61 (62) in correspondence with each key number.

Then, the data after anharmonicity correction (fe, 2fe.

...nfs...) to 0. 2 cents lower
1 virtual string vibration data (f+, '2f+, ...
nfl...) is calculated. The frequency of the fundamental tone f1 is calculated from the relational expression (-0,2=I20QIog2(f+/fe)). The frequencies of the overtones 2f+... are calculated in the same manner.

Furthermore, the anharmonicity corrected data (fs, 2fi!

...nfe...) to 0. 3 cents more and 2nd
Virtual string vibration data (f2. 2f2. ... n
f2...) is calculated. The frequency of the fundamental tone f2 is determined by the relational expression (0,3=1200 l o g2 (f2/ f
e) Calculated from ). The frequencies of the overtones 2f2... are calculated in the same manner.

FIG. 5(C) is a graph of a line spectrum for explaining intensity attenuation correction over time.

In the figure, the opening indicates first virtual string vibration data, ■ indicates first virtual string data after damping correction, Δ indicates second virtual string vibration data, and M indicates second virtual string data after damping correction.

The second correction is to make sure that the time elapsed time of the first virtual string vibration data and the second virtual string vibration data: the elapsed time after pronunciation (↑e) and the elapsed time (
t+, t2), the intensity is corrected to increase or decrease, and the first and second virtual string data are calculated.

First, the damping characteristics of the actual string vibration data included in the first virtual string vibration data and the second virtual string vibration data are corrected. Generally, the intensity I (t) over time is smaller than a suitable target value due to energy loss in the process of energy transfer from the strings to the soundboard via the bridge. On the other hand, only a specific overtone component may decay slowly and lose its balance with other overtones.

In this case, the intensity correction factor Δ for each elapsed time for the fundamental tone and each overtone
I(t,) is stored in advance in the ROM 61 (62).

After the string is struck, the intensity correction value DI (
t) is calculated as shown in the following formula based on the intensity at the time of pronunciation (0).

DI (t)=I (o)XΔI (t) After pronouncing,
The intensity correction for the fundamental tone and each harmonic at time t is performed using the following equation.

■ Gut)=r (t) 10DI (t) I on the left side
(t) indicates the intensity after attenuation correction, 1(t) on the right side indicates the intensity before correction, and DI(t) indicates the intensity correction value at time t.

Then, two correction calculations are performed for each fundamental tone, each overtone, and each time.

FIGS. 5(D) and 5(E) respectively show the first virtual string data and the second virtual string data calculated in the above manner as line spectra.

Note that the present invention can also be applied to a multi-string piano.

Furthermore, by disabling the electrical system, it can be easily used as a single-string piano, and can be used as a low-volume practice piano.

Furthermore, MIDI signals can be output via the MIDI terminal, allowing unsampled performances with other electronic musical instruments, MIDI devices (music computers, etc.), and the like. Also, delay processing [500m5ec
], it is possible to play an ensemble on a player piano.

<Effects of the Invention> As explained above, the keyboard music according to the present invention is simple. -F Juxtaposition 11- k'-marugeto ← The tone has improved to the bare stage. The volume has also been improved. Furthermore, the volume can be adjusted by adjusting the amplification factor and the like. Since the volume increases, it becomes possible to perform at a lower tone.

The timbre and volume of these can be increased to a level equal to or higher than that of a three-string piano.

Furthermore, compared to a three-string piano, a single-string piano has a smaller string force acting on the frame, so the frame can be made lighter and smaller.

[Brief explanation of the drawing]

1 is a block diagram showing a schematic configuration of a keyboard instrument according to the present invention; FIG. 2 is a graph showing a line spectrum for explaining the calculation process of the vibration state of a virtual string in a keyboard instrument according to the present invention; FIG. 4 is a block diagram showing a keyboard instrument according to an embodiment of the present invention; FIG. 4 is a flowchart showing a procedure for sound generation processing according to an embodiment; FIGS. It is a graph showing a line spectrum for explaining the calculation process of the vibration state of a virtual string. 11... String, 12... Piezoelectric element (detection means), 13.
...... Calculation means, 14... Pronunciation means. Patent Applicant Yamaha Corporation Representative Patent Attorney Kiyoshi Kuwai Figure 1 Block diagram of the GT Iwa instrument of the present invention 1n4ensijy JI2 Figure 5 A graph explaining the calculation process of the present invention Kazumi! ! i! Example 2° rough

Claims (1)

[Scope of Claims] Detection means for detecting the vibration state of a tensioned string that vibrates in response to a key press; a calculation means for calculating the vibration state of a virtual string having the same pitch as the set string; and a calculation means for converting the vibration state of the virtual string calculated by the calculation means into sound, and from the time when the set string is sounded, the vibration state is shorter than the audible identification time. A keyboard instrument characterized by comprising: a sounding means that produces a sound after a set predetermined time;