JPS6230639B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a musical tone forming method, and more particularly to an improvement of a musical tone forming method using a frequency modulation method.

Conventionally, amplitude information A, modulation index information I, and carrier frequency information (hereinafter referred to as carrier frequency information) corresponding to the pressed key pitch of the musical tone to be formed are used.
ωct and modulated wave frequency information (hereinafter referred to as modulation frequency information) ωmt are given, and based on these information, the instantaneous amplitude value e(t) of the frequency modulated signal is determined as e(t)=A sin [ωct+I sinωmt]
......calculated based on the formula shown in (1), the instantaneous amplitude value e
A frequency modulation method for forming musical tones using (t) as a musical tone signal is known.

By the way, in general natural musical instruments, the overtone structure and the like fluctuate randomly over time at the rising or sustaining parts of a musical tone, resulting in a musical tone that has a pleasant natural quality.

However, in the conventional frequency modulation tone forming method, the amplitude information A and the modulation index information I are merely changed according to a predetermined time function, so the timbre, etc. of the generated musical tone is changed. The disadvantage is that the state of change does not feel natural.

This invention was made in view of the shortcomings of the conventional musical tone forming methods mentioned above, and its purpose is to provide a musical tone forming method in which the harmonic structure fluctuates randomly over time and a rich musical tone signal with pleasant naturalness can be obtained. Our goal is to provide the following.

For this purpose, the present invention temporally randomly changes the frequency of at least one of the carrier wave and the modulated wave so that the relationship between the frequency of the carrier wave and the frequency of the modulated wave changes randomly over time in a non-integer multiple relationship. It is made to change according to a changing random function.

In addition, the relationship between the frequency of the carrier wave and the frequency of the modulated wave is made to change randomly over time in a non-integer multiple relationship at the rising edge of the musical tone signal, and the modulation index in the frequency modulation is adjusted to the sustaining area of the musical tone signal. It is made to change according to a random function that changes randomly over time for a corresponding period of time.

Hereinafter, this invention will be explained in detail using the drawings.

FIG. 1 is a block diagram showing an embodiment of an electronic musical instrument to which the musical tone forming method according to the present invention is applied, and can be roughly divided into a key switch circuit 10, a frequency information memory 20, a frequency modulation circuit 30, an amplitude control circuit 40, Consisting of a DA converter (DAC) 50 and a sound system 60, the frequency modulation circuit 30 and amplitude control circuit 40 change all of carrier frequency information ωct, modulation frequency ωmt, modulation index information I, and amplitude information A over time. and is configured to change randomly.

In FIG. 1, when a key on the keyboard section is pressed, the key switch circuit 10 outputs a key code KC corresponding to the pressed key, and also outputs a key-on signal KON indicating that any key has been pressed. do. The key code KC output from the key switch circuit 10 is supplied to the frequency information memory 20 as an address signal.

The frequency number memory 20 stores modulation angle frequency information ωm corresponding to the pitch of each key of the keyboard section at each address.
I remember. Therefore, the key switch circuit 10
When the key code KC corresponding to the pressed key is input as an address signal, the frequency information memory 20 outputs modulation angle frequency information ωm corresponding to the pitch of the pressed key. This modulation angle frequency information ωm is supplied to an adder 300 in the frequency modulation circuit 30.

The adder 300 is provided to temporally and randomly change the modulation angular frequency information ωm corresponding to the pitch of the pressed key, albeit slightly. function F ₁
(t) and the random function R ₁ (t) by an adder 301
The function obtained by adding [F ₁ (t) + R ₁
(t)] is input.

The time function F ₁ (t) has its value at the attack (rising portion) A, the first decay 1d, the sustain (sustained portion) S, and the second decay 2D of the musical tone signal, for example, as shown in the waveform diagram of FIG. 2. As shown, it is a function that changes over time in a predetermined manner, and is generated from a first time function generator (hereinafter referred to as the first FG) 302 in synchronization with the key-on signal KON.

The random function R ₁ (t) is a function whose value changes randomly over time, and is generated by a first random function generator (hereinafter referred to as first RG) 303. Therefore, the function [F ₁ (t) + R ₁ (t)] obtained from the adder 301 basically shows the change mode as shown in Fig. 2, but it changes according to the random function R ₁ (t). The value will change randomly.

Therefore, the function [F ₁
(t)+R ₁ (t)] is input to the adder 300
In, the modulation angular frequency information ωm and the function [F ₁
(t)+R ₁ (t)], modulation angular frequency information ω'm that changes temporally and randomly is obtained. Modulation angular frequency information ω'm obtained in this adder 300 is supplied to a multiplier 304 and an adder 305.

The multiplier 304 multiplies the modulation angular frequency information ω′m by a predetermined constant k (integer), and the multiplier value kω′m is applied to the first accumulator (first accumulator).
1ACC) 306 as carrier angular frequency information ωc.

The first accumulator 306 sequentially accumulates the carrier angular frequency information ωc supplied from the multiplier 304 in accordance with the clock pulse φ of a predetermined period, and calculates the accumulated value q·ωc (q=1, 2,...) Carrier frequency information ωct of the repetition period corresponding to the pitch of the pressed key
Output as . In this case, carrier frequency information ωct
Since the modulation angular frequency information ω'm, which is the basis for creating the It changes over time and randomly.

On the other hand, the modulation angular frequency information ω′m supplied to the adder 305 is a random function F ₂ generated only in the attack part of the musical tone signal to be formed.
(t) is added. The random function R ₂ (t) is
Second random function generator (hereinafter referred to as 2RG)
It is generated from 307, but the 2nd RG/3
07, the attack mode signal AT indicating that the musical tone signal currently being formed is the attack part is the 1st FG/3
It is enabled only when it is given from 02 and outputs a random function F ₂ (t). Therefore, in the adder 305, modulation angular frequency information ω''m (=ω'm+R ₂ (t)) to which random changes are further added in the attack portion is obtained.
In this case, the base information ω′m for creating the modulation angular frequency information ω″m has already been given randomness in the adder 300, but a random function R ₂ (t) is further added to it. The reason why this is added is to obtain a nonharmonic musical tone by making the carrier frequency (ωct) and modulation frequency (ωmt) have a non-integer multiple relationship in the attack part.

The modulation angular frequency information ω″m obtained in this adder 305 is transmitted to the second accumulator (second ACC).
308. Second accumulator 308
When modulation angular frequency information ω″m is supplied, this information ω″m is accumulated sequentially according to the clock pulse φ, and the accumulated value qω″m (q=1, 2, etc.)
is output as modulation frequency information ω″mt with a repetition period corresponding to the pitch of the pressed key. This modulation frequency information ω″mt is supplied to the sine function memory 309 as an address signal.

The sine function memory 309 stores the sine amplitude value at each sample point phase of the sine waveform-period at each address.
I remember sinωt. Therefore, when the modulation frequency information ω"mt is supplied as an address signal, the instantaneous amplitude value sinω"mt of the modulation signal corresponding to the repetition frequency of the information ω"mt is output.

This modulation signal sinω"mt is supplied to a multiplier 310. The multiplier 310 multiplies the modulation signal sinω"mt by modulation index information. Modulation index information I(t)+
R ₃ (t) is provided from adder 311. That is, the second time function generator (hereinafter referred to as second FG) 312 generates modulation index information I(t) that changes over time as shown in FIG. 2 in synchronization with the rise of the key-on signal KON, and this information I (t) is supplied to the adder 311. On the other hand, the third random function generator (hereinafter referred to as the 3rd RG) 313 is enabled only when the sustain mode signal ST indicating that the musical tone signal currently being formed is a sustain portion is given from the 2nd FG 312. , generates a random function R ₃ (t) that changes randomly over time, and supplies this function R ₃ (t) to the adder 311. Then, the adder 311 adds these modulation index information I(t) and the random function R ₃ (t), and the attack, first decay, and second decay change over time as shown in FIG. 2, and the sustain changes with time. Modulation index information I(t)+R ₃ (t) to which randomness is imparted by random function R ₃ (t) is output. As a result, in the multiplier 310,
The modulation signal sinω″mt is multiplied by the modulation index information [I(t)+R ₃ (t)] given by the adder 311, and the modulation signal [I(t)+R ₃ ( t)] sinω″mt is output. This modulation signal [I(t)+R ₃ (t) sin
ω″mt is provided to an adder 314.

The adder 314 modulates the carrier frequency information ωct by a modulation signal [I(t)+R ₃ (t)] sinω″mt and outputs the modulated signal, and outputs the modulated signal ωct and the signal [I
(t) + R ₃ (t)] sinω″mt is input, these are added, and the added value is used as the modulated frequency information {ωct
+[I(t)+R ₃ (t)] sinω″mt}. This modulated frequency information {ωct+[I(t)+R ₃
(t)]sinω″mt} is supplied to the sine function memory 315 as an address signal.

Like the sine function memory 309 described above, the sine function memory 315 stores the sine amplitude value sinωt at each sample point phase of each address sine waveform-period. Therefore, the modulated frequency information {ωct[I(t)+R ₃ (t)] sin
ω″mt} is supplied as an address signal, this sine function memory 315 generates a frequency modulated signal represented by e ₀ (t)=sin{ωct+[I(t)+R ₃ (t)]sinω″mt} Output e ₀ (t).

Frequency modulated signal e ₀ obtained as above
(t) is supplied to the amplitude control circuit 40 to control the setting of the amplitude value. That is, the frequency modulated signal e ₀ (t) is supplied to a multiplier 400, where it is multiplied by amplitude setting information A(t)+R ₄ (t) supplied from an adder 401. here,
The amplitude setting information A(t)+R ₄ (t) is the amplitude information A(t) generated as a time function from the third time function generator (hereinafter referred to as the third FG) 402, and the fourth
Random function generator (hereinafter referred to as 4th RG) 40
3 and a random function R ₄ (t) generated from R 4 (t). The 3rd FG/402 is the aforementioned 3rd FG.
Similar to the 1FG/302 and the 2nd FG/312, the amplitude changes over time in each of the attack, first decay, sustain, and second decay parts of the musical tone signal that should be formed in synchronization with the key-on signal KON, as shown in Figure 2. It outputs information A(t).
The fourth RG 403, like the third RG 313 described above, has a sustain mode signal ST indicating that the musical tone signal currently being formed is a sustain portion.
It is enabled only when given from 3FG 402, and outputs a random function R ₄ (t) that changes randomly over time. Therefore,
Amplitude setting information A obtained in adder 401
(t)+R ₄ (t) basically changes over time as shown in FIG. 2, but in the sustain portion it changes randomly according to the random function R ₄ (t). Therefore, by using such amplitude setting information A(t)+R ₄ (t), the frequency modulated signal
When the amplitude of e ₀ (t) is set, the multiplier 400 outputs e(t) = [A(t) + R ₄ (t)] sin {ωct + [I (t) + R ₃ (t)] sin ω″mt ] A frequency modulated signal e(t) is output.This frequency modulated signal e(t) is
0, it is converted into an analog signal and supplied to the sound system 60, and this sound system 6
Starting from 0, it is pronounced as a musical tone.

Here, the information ωct that is the basis for forming the frequency modulated signal e(t) is basically a time function.
Although it changes over time according to F ₁ (t), a slight randomness is imparted by the random function R ₁ (t). In addition, the information ω″mt basically changes over time according to the time function F ₁ (t), but it is given randomness such that the relationship with the information ωct becomes a non-integer multiple relationship in the attack part. Although the information I(t) basically changes over time according to a predetermined time function, it is given randomness by the random function R ₃ (t) in the sustain part. (t) basically changes over time according to a predetermined time function,
Random function R ₄ (t) in the sustain part
Randomness is imparted by .

Therefore, the frequency modulated signal e(t) formed based on such information, that is, the musical tone signal e
In (t), (a) the frequency (pitch), timbre, and amplitude change over time during the period from the rise to the fall of the musical sound signal, and randomness is added; (b) among them, the attack (c) In the sustain part, the timbre and amplitude also change randomly, causing the frequency, timbre, and amplitude of the musical tone to fluctuate over time. As a result, it is possible to obtain a natural and pleasant musical tone similar to that of a natural musical instrument.

Note that the naturalness of musical tones can be further increased by changing the information for imparting such fluctuations in accordance with the set timbre.

By the way, adders 300, 301, 305, 311, 401 for imparting such fluctuations
may be a multiplier.

Further, the part that forms carrier angular frequency information ωc and modulation angular frequency information ω″m based on the key code KC may be configured as shown in FIG. The adder ₃₀₀ generates carrier angular frequency information ωc corresponding to the high
(t)+R ₁ (t)], and the added value is output as carrier angular frequency information ωc to which fluctuation is added. Furthermore, the output information ωc of the adder 300 and the constant 1/k (k is an integer) are applied to the multiplier 30.
4', and the adder 30 uses the multiplied value as the fluctuated modulation angular frequency information ω'm.
5'. Then, the adder 305' adds the random function R ₂ (t) to the information ω'm and outputs it as modulation angle frequency information ω''m with enhanced fluctuations. In this case, the random function
The adder 305' for adding R ₂ (t) may be provided at the position indicated by the broken line in FIG. 3 to add the function R ₂ (t) to the carrier angular frequency information ωc.

Next, detailed configurations of the time function generators 302, 312, and 402 used in this embodiment will be explained.

Figure 4 shows time function generators 302, 312, 402.
This time function generator (hereinafter referred to as FG) corresponds to attack A, first decay 1D, sustain S, and second decay 2D of the musical tone signal to be formed. and four segment modes M [A], M [1D], M
[S), M [2D]. Then, in each segment mode M[A] to M[2D], the increment (or decrement) information per unit time given for each segment mode is accumulated at a predetermined speed as shown in Fig. 2. Outputs a time function that changes over time.

Here, segment mode M [A], M
[1D], M[S], M[2D], M[A]="00",
M [1D] = “01”, M [S] = “10”, M [2D] =
It is given as a 2-bit number “11”.

For example, attack segment mode M [A]
, the attack rate information AR (positive value) is accumulated sequentially from a predetermined initial level (IL), and this accumulation is calculated as the accumulated value IL + q・AR (however, q=
1, 2, 3...) until they match the attack level (AL), and the accumulated value in this accumulation process is output as a time function in attack segment mode M[A].

In addition, the segment mode M of the first decade
In [1D], the first decay rate information 1DR (negative value) is accumulated sequentially for the attack level (AL), and the accumulated value AL-q・1DR is the first decay level ( 1DL), and the accumulated value in this accumulation process is output as a time function in the segment mode M[1D] of the first decade.

Also, sustain segment mode M [S]
In this step, rate information having a value of "0" is sequentially accumulated for the first decay level (1DL). That is, in the sustain segment mode M[S], a time function that maintains the first decay level (1DL) is output.

Furthermore, the segment mode M of the second decade
In [2D], the 1st Decay level (1DL)
For, the second decay rate information 2DR (negative value)
are accumulated sequentially, and the sequential accumulation is the accumulated value 1DL−q・
The execution is performed until 2DR matches the initial level (IL), and the accumulated value in this accumulation process is output as a time function in segment mode M [2D] of the second decade.

In FIG. 4, the one-shot circuit 700 receives the key-on signal KON from the key switch circuit 10.
(“1”) is output, the key-on pulse KONP with a narrow pulse width is synchronized with the rising timing.
Output. This key-on pulse KONP is applied to the counter 701 as a reset signal.

The counter 701 outputs a 2-bit mode signal M indicating each segment mode.
When the key-on pulse KONP is applied as a reset signal, the mode signal M becomes "00". This mode signal M is supplied to selectors 702 and 703 as a select control signal.

The selector 702 attaches rate information AR (positive value) and first decay rate information 1DR to selection inputs (00, 01, 11) selected by the mode signal M.
(negative value), second decay rate information 2DR (negative value) is rate information memory 704, 705, 706
Each is given from Also, selector 7
03, attack level information AL, first decay level information 1DL, and initial level information IL are input to the selection inputs (00, 01, 11) selected by the mode signal M to the level information memories 707, 708, 70.
Each is given from 9 to 9.

In this case, the rate information memories 704 to 706 and the level information memories 707 to 709 have a plurality of memory blocks corresponding to the type of tone, and one memory block corresponding to the set tone is selected from among these memory blocks. It is configured so that information AR, 1DR, 2DR, AL, 1DL, and IL corresponding to the set tone color is output.

The rate information (AR, 1DR, or 2DR) selectively output from the selector 702 is output by the AND gate 7.
10 to an accumulator ACC consisting of an adder 711 and a register 712, and in the adder 711, the accumulated value up to the present and the level information output from the register 712 are input at a predetermined period according to the clock pulse φ. The configuration is such that the added value is held in the register 712 as a new accumulated value. This accumulator
The accumulated value output from ACC is a time function F(t)
Sent as .

On the other hand, the level information (AL, 1DL, or IL) selectively output from the selector 703 is sent to the comparator 71.
3, wherein the accumulator
It is configured to be compared with the accumulated value output from ACC.

Here, the AND gate 710 selects the selector 702 when the output signal of the NAND gate 717 is "1".
The rate information (AR or 1DR or 2DR) output from is passed through and supplied to the accumulator ACC. The NAND gate 717 sends out an output signal of "1" only when the coincidence signal EQ output from the comparator 713 is "0" or when the output signal of the inverter 715 is "0". The output of a NAND gate 714 which outputs a "1" signal when the mode signal M output from the counter 701 becomes a value other than "11" is applied to the input of the counter 701. In addition, Nand Gate 714
The output signal of is also supplied to the counter 701 as a count enable signal, and this NAND gate 71
The counter 701 is configured to be enabled only when the output signal of No. 4 is "1". In this case, the counter 701 stops at the state of the mode signal M of "11" after completing the entire operation from the attack segment mode M [A] to the second decade segment mode M [2D]. It is something that In addition, the cumulative value output from the accumulator ACC is the segment mode M [2D] of the second decade, as can be seen from the following explanation.
Since the output signal of the NAND gate 717 becomes "0" at the stage when the operation in is completed and the initial level (IL) is reached, the initial level (IL) is maintained. Along with this, comparator 7
The coincidence signal EQ output from 13 is also maintained at "1".

Now, the counter 701 is the key-on pulse KONP
When the mode signal M becomes "00" after being reset by the mode signal M, a "1" signal is output from the NAND gate 714. As a result, the counter 701 enters a count enable state, and a "1" signal is output from the NAND gate 717, allowing the AND gate 710 to operate. Furthermore, when the mode signal M becomes "00", the selector 702 selects the attack rate information AR added to the selection input (00).
(a positive value), and the selector 703 selects and outputs the attack level information AL added to the selection input (00). Therefore, the mode signal M
becomes “00” and the attack segment mode M
When [A] starts, the attack rate information AR passes through the AND gate 710 and is sent to the accumulator ACC.
At the same time, attack level information AL is also supplied to the comparator 713. Therefore, the accumulator ACC sequentially accumulates the attach rate information AR at a predetermined speed. When this cumulative value (IL+q.AR) matches the attack level information AL, the comparator 713 outputs a match signal EQ. This coincidence signal EQ is applied to the count input (CK) of counter 701 via OR gate 718. As a result, the mode signal M output from the counter 701 is updated from "00" to "01" and shifts to the first decade segment mode M[1D].

When the mode signal M is updated to “01”, the selector 702 selects and outputs the first decay rate information 1DR (negative value), and supplies this rate information 1DR to the accumulator ACC via the AND gate 710.

On the other hand, the selector 703 selectively outputs the first decay level information 1DL and supplies this level information 1DL to the comparator 713.

Therefore, the accumulator ACC accumulates the first decay rate information 1DR at a predetermined speed with respect to the accumulated value up to now, that is, the attack level (AL). By this, the accumulated value (AL−q・
1DR) gradually changes to smaller values, and when this cumulative value q·RD[01] matches the first decay level information 1DL, a match signal EQ is output from the comparator 713. This match signal EQ is the OR gate 718
Counter 701 count input (CK) via
is applied to Therefore, the mode signal M output from the counter 701 is updated from "01" to "10" and the sustain segment mode M
Move to [S].

When the mode signal M is updated to "10", the selectors 702 and 703 do not output any information because there is no selection input corresponding to the selector control input of "10". Therefore, in the sustain segment mode M[S] in which the mode signal M indicates "10", the accumulated value of the accumulator ACC does not change and remains at the first decay level (1DL).

However, when the pressed key is released after a predetermined period of time and the key-on signal KON falls, the one-shot circuit 719 generates a narrow-width key-off pulse synchronized with the falling timing of the key-on signal KON.
KOFP is output. This key off pulse KOFP
is applied to the count input (CK) of counter 701 via OR gate 718. Therefore, the mode signal M output from the counter 701 shifts to the segment mode M [2D] of the second decade, which is updated from "10" to "11".

When the mode signal M is updated to “11”, the NAND gate 714 outputs a “0” signal and the counter 7
01 to the inoperable state, and the inverter 71
The output signal of 5 is set to "1". This inverter 7
The “1” signal output from 15 is the NAND gate 7
The match signal EQ from the comparator 713 is also supplied to the NAND gate 717. However, at this time, the comparator 713 is given the initial level information IL from the selector 703 as the mode signal M is updated to "11", and the first decay level (1DL) is supplied to one comparison input. ) is given.
Therefore, the coincidence signal output from the comparator 713
EQ indicates "0", and accordingly, the output signal of NAND gate 717 indicates "1". This causes the AND gate 710 to
2nd decay rate information selected and output from 2
Supply 2DR (negative value) to the accumulator ACC.

Therefore, the accumulator ACC is the cumulative value up to now, that is, the first decade level (1DL).
The second decay rate information 2DR is sequentially accumulated at a predetermined speed. As a result, the cumulative value (1DL−
q・2DR) gradually changes to smaller values. after that,
When the accumulated value matches the initial information IL, the comparator 713 outputs a match signal EQ of "1". Therefore, the two input signals of the NAND gate 717 are both “1”, and the AND gate 71
7 sends out an output signal of "0". As a result, the AND gate 710 is connected to the accumulator ACC.
The supply of the second decay rate information 2DR to is stopped. That is, when the accumulated value matches the initial level (IL), information RD[M] having a value of "0" is given to the accumulator ACC. Therefore, after reaching the initial level IL, the accumulated value of the accumulator ACC is maintained at the initial level (IL). As the accumulated value is maintained at the initial level (IL), the comparator 713 continues to send out the coincidence signal EQ of "1".

In this way, from attack A to the second day
Time function F(t) to reach 2D segment mode
is formed, but when the key is pressed again,
The counter 701 is reset by the key-on pulse KONP, and the time function F from the attack A to the segment mode of the second Decay 2D is similarly reset.
(t) is formed. In this case, the accumulator
The cumulative value of ACC, that is, the time function F(t), is
The initial level (IL) is maintained at the stage when the accumulation operation of segment M [2D] of the second Decay 2D is completed. Therefore, the time function F(t) due to a new key press operation changes sequentially to the attack level (AL), the first decay level (1DL), and the initial level (IL) with the initial level (IL) as the initial value. Become something to do.

The time function F(t) thus formed is supplied to a multiplier 720. The multiplier 720 multiplies the time function F(t) by a constant m corresponding to the pitch range to which the pressed key belongs to obtain a time function m·F that differs for each pitch range.
(t) is output. In this case, the constant m is read out from the key scaling memory 721, which stores a different constant m for each range of the key at each address, using the key code KC as an address signal. Therefore, the time function F
By multiplying by (t), a different time function m·F(t) is output for each range of the pressed key.

In the above manner, the time function m·F(t) for imparting fluctuation to musical tones is formed.
2. The third FG/402 has different uses. Therefore, the rate information AR stored in the rate information memories 704 to 706 in FIG.
1DR, 2DR and level information memory 707-70
Level information AL, 1DL, and IL to be stored in 9 must be set appropriately according to each purpose.

In FIG. 4, the detection circuits 722 and 723 determine the contents of the mode signal M and select the attack and sustain segment modes M[A],
Attack mode signal indicating M[S]
This is for outputting AT and sustain mode signal ST.

Next, a random function generator (1st RG/303,
2nd RG・307, 3rd RG・313, 4th RG・40
The detailed configuration of 3) will be explained.

FIG. 5a and FIG. 5b are circuit diagrams showing detailed configuration examples of random function generators 303, 307, 313, and 403, and this random function generator (hereinafter referred to as RG) is a multi-stage shift Random function R whose value changes randomly over time using registers and exclusive or gates
(t) is generated.

In FIG. 5a, the shift register 800 has a plurality of stages, and the signals of each stage are output in parallel. Then, the final stage output signal and the output signal of the previous stage are input to the exclusive OR gate 801, and the non-inclusive OR output is input to the first stage of the shift register 800 via the OR gate 802. They are sequentially shifted toward the final stage according to the clock pulse φ. The shift register 800 is configured so that, for example, signals from the first stage to the fifth stage are output as a parallel 5-bit random function R(t). In addition, in order to prevent all the signals of each stage of the shift register 800 from becoming "0", the output signals of all stages are input to the NOR gate 803, and when the output signals of all stages become "0", This Noah Gate 8
03 outputs a “1” signal, this “1” signal is input to the first stage via the OR gate 802,
The configuration is such that a "1" signal exists in either stage.

Therefore, in such a configuration, the state of the signal at each stage of the shift register 800 differs for each predetermined shift cycle, so the output signals from the first to fifth stages also have temporally random values. It becomes a random function that shows.

FIG. 5b shows an AND gate 8 in the configuration of FIG. 5a.
04 to 808 are added, and the random function R(t) is output only when the enable signal E is given. Therefore, the 1RG 303 in FIG. 1 has the configuration shown in FIG.
2RG・307, 3rd RG・313, 4th RG・403
The configuration shown in FIG. 5b may be used. However, the
Since the 1st RG 303 to the 4th RG 403 have different uses, the number of bits constituting the random function R(t) and the number of stages of the shift register 800 may be appropriately selected according to their respective uses. Thereby, it is possible to generate a random function R(t) according to each application with an extremely simple configuration.

Figure 6 shows the function F ₁ (t) + R ₁ in Figure 1.
(t), I(t)+R ₃ (t), A(t)+R ₄ (t). Shown by chart. Note that the following explanation is based on function A
This is assumed to be the case when a time function A'(t) similar to (t)+R ₄ (t) is formed.

In FIG. 6, the control circuit 900 has a time function.
At the stage where the formation of the segment mode M [2D] of the second decade at A'(t) is completed, the mode signal MD is set to "2D" and it is in a standby state for a new key press operation. When the key-on signal KON rises to "1", the mode signal MD is updated from "2D" to "AT" as shown in step 931 of the flowchart of FIG.

When the mode signal MD is updated to "AT" and becomes the attack segment mode M[A], the selector 901 selects the attack rate information AR and supplies it to the key scaling circuit 902. This key scaling circuit 902 is for changing the added rate information according to the range of the pressed key.The attack rate information AR is changed here, and the AND gate 903 is used as the rate information AR'.
supplied to

AND gate 903 has a signal output from inverter 915 and flip-flop 905.
The gate open signal G output from the inverted set output () of
It becomes "1" when the operation of [2D] is completed.

Further, as will be clear from the following explanation, the signal becomes ="1" when the mode signal MD is updated to "AT" by a new key depression operation.

Therefore, the AND gate 903 uses the mode signal
When MD becomes “AT”, key scaling circuit 9
Attach rate information output from 02
AR' is passed through adder 906 and register 907.
It supplies to the accumulator ACC consisting of.

Then, the accumulator ACC is set to register 90.
The clock pulse φ adds the current value of the time function A′(t) output from 7 and the rate information AR′.
It is executed at predetermined intervals according to the following, and the added value is held in the register 907 as the next new current value.
Here, the attach rate information AR is set to a positive value. Therefore, the accumulator ACC is
In attack segment mode M[A], a time function A'(t) that increases sequentially according to rate information AR' is output. And the time function
When the value of A'(t) reaches the attack level (AL), level detector 908 outputs a detection signal ALD indicating this. This detection signal ALD is applied to the control circuit 900.

Then, the control circuit 900 performs the steps 932 and 933 in the flowchart of FIG.
Update mode signal MD to “1D”.

When the mode signal MD is updated to "1D" and shifts to the first decay segment mode M[1D], the selector 901 selects and outputs the first decay rate information 1DR instead of the attach rate information AR. This first decay rate information 1DR is also changed by the key scaling circuit 902 into rate information 1DR' corresponding to the range of the pressed key, and then supplied to the accumulator ACC via the AND gate 903.

Then, the accumulator ACC comes to accumulate the time function A'(t) indicating the attack level (AL) and the rate information 1DR' sequentially at a predetermined speed. In this case, the first decay rate information 1DR is set to a negative value, so the cumulative
The value of the time function A'(t) output from the ACC gradually decreases from the attack level (AL).

On the other hand, since the mode signal MD has been updated to "1D", the selector 909 selects the first decay level information 1DL, and the key scaling circuit 910
supply to. The key scaling circuit 910 has the same function as the above-described key scaling circuit 902, and changes the added level information 1DL to level information 1DL' according to the range of the pressed key and outputs it. This level information 1DL' is then supplied to the A side comparison input of a comparator 912 via an adder 911. Here, the random function R ₄ (t) is input to the adder 911 via the AND gate 913.
Since flip-flop 905 is in the reset state, it does not send out the random function R ₄ (t). Therefore, in the segment mode of the first decade, the adder 911 is connected to the key scaling circuit 910.
The level information 1DL' output from the comparator 912 is applied as is to the A side comparison input of the comparator 912. Here, the B side comparison input of this comparator 912 is an accumulator.
A time function A'(t) output from ACC is given.

Therefore, the segment mode M of the first decade
In [1D], the value of the time function A'(t) output from the accumulator ACC gradually decreases, and when it matches the first decay level information 1DL', the value of the time function A'(t) output from the accumulator ACC decreases.
12 outputs a coincidence signal AEB indicating this fact. This coincidence signal AEB is supplied to control circuit 900.

Then, the control circuit 900 next controls the time function in the sustain segment mode M[S].
In order to form A'(t), the mode signal MD is updated to "S" as shown in step 935 of the flowchart of FIG. 7, and the sustain mode signal
Outputs ST (“1”). Then, selector 90
Since selectors 1 and 909 do not have selection inputs corresponding to the "S" mode signal MD, the outputs of these selectors 901 and 909 are both "0".
becomes. Along with this, the addition input given from the AND gate 903 to the adder 906 of the accumulator ACC also becomes "0", and the addition input given from the key scaling circuit 910 to the adder 911 also becomes "0".

However, since the sustain mode signal ST (“1”) is output from the control circuit 900 and this signal ST is applied to the set input (S) of the flip-flop 905, the flip-flop 905 is set and the random function
R ₄ (t) is now provided to adder 911. Therefore, the adder 911 uses the random function R ₄
(t) is applied to the A side comparison input of the comparator 912. Here, the random function R ₄ (t) is generated from a random function generator (hereinafter referred to as RG) 914, and this RG 914 is generated by the adjustment section.
ADJ, and is configured such that the function can be adjusted and output using a switch SW.

When the random function R ₄ (t) is applied to the A-side comparison input of the comparator 912 in this way, the control circuit 900 performs step 936 in the flowchart of FIG.
As shown in the sections 938 to 938, the comparison result signal AEB (A=B) and
Based on AGB (A>B), accumulator ACC
"+1" or "-1" for the adder 906 of
The time function A'(t) output from the register 907 is changed to follow the random function R ₄ (t).

That is, in the case of A'(t)=R ₄ (t) (A=B), "+1" and "-" are sent to the adder 906.
1" command signal is not given at all, but A'(t)>R ₄
(t) If (A<B), a command signal of "-1" is given to adder 906 as shown in step 938. Thereby, the function A'(t) is sequentially decreased to match the random function R ₄ (t). vice versa,
If A'(t)<R ₄ (t) (A>B), a command signal of "+1" is given to adder 906 as shown in step 937. By this, the function A′(t)
increases sequentially to match the random function R ₄ (t).

In this way, in the sustain segment mode M[S], the accumulator ACC is a time function that follows changes in the random function R ₄ (t).
Output A'(t).

After that, when the key-on signal KON falls to "0", the control circuit 900 changes the mode signal MD to "2D" as shown in step 940 of the flowchart of FIG.
Update to segment mode M of the second day K.
Shift to [2D]. On the other hand, since the key-on signal KON is inverted by the inverter 904 and applied to the reset input (R) of the flip-flop 905, the flip-flop 905 receives the key-on signal.
It is reset as KON falls.

As a result, the AND gate 913 becomes inactive and the adder 911 receives the random function R ₄ (t).
will no longer be given. Therefore, only the initial level information IL selectively output from the selector 909 is sent to the key scaling circuit 910 and the adder 91.
1 to the A side comparison input of the comparator 912.

Further, the selector 901 selectively outputs the second decay rate information 2DR, and this rate information 2DR is sent to the key scaling circuit 902 and the AND gate 90.
3 to the accumulator ACC. In this case, the AND gate 903 can pass the rate information 2DR' because the gate open signal G output from the inverted set output () of the flip-flop 905 becomes "1" in response to the key release operation.

Therefore, when the accumulator ACC is supplied with the second decay rate information 2DR' (negative value), it sequentially accumulates this rate information 2DR' with respect to the current value of the time function A'(t). Then, the function A′(t)
will gradually decrease. and the function
When A'(t) matches the initial level information IL' applied to the A-side comparison input of the comparator 912, the comparator 912 outputs a match signal AEB (“1”). This match signal AEB is sent to inverter 9
15 and supplied to AND gate 903 as a gate close signal. As a result, the rate information 2DR' is no longer supplied to the accumulator ACC, the function A'(t) maintains the initial level (IL'), and the function A'(t) in response to a new key press operation is
A standby state is entered for the formation of A'(t).

Time function formed as above
A'(t) is supplied to the adder 916, where it is added to the constant m output from the key scaling memory 917, changed to a function "A'(t)+m" according to the range of the pressed key, and output. be done.

Therefore, even in this configuration, a time function A'(t) having randomness is generated only in the sustain segment mode M[S], similar to the function A(t)+R ₄ (t) in FIG. can be done.

As is clear from the above description, the present invention uses a frequency modulation musical tone forming method that uses a random function that changes randomly over time to create a frequency relationship between a carrier wave and a modulated wave (the relationship between ωct and ωmt in the embodiment). By making it change randomly over time in a non-integer multiple relationship, it is possible to easily create a musical tone signal that has an aharmonic overtone structure and whose anharmonicity fluctuates over time. You can obtain natural, high-quality musical tones.

Further, in the present invention, the frequency relationship between the carrier wave and the modulating wave is temporally randomly changed in a non-integer multiple relationship at the rising edge of a musical tone, and
In the sustained part of the musical tone, the modulation index of the frequency modulation is changed randomly over time, so that in the rising part of the musical tone, the anharmonicity mainly fluctuates over time, and in the sustained part, the timbre fluctuates over time. A signal can be formed. Here, in natural musical instrument sounds that have an aharmonic overtone structure, anharmonicity occurs in the rising part, and after the rising part, there is a transition from an aharmonic to a harmonic overtone structure, and the harmonic overtone structure changes. There are some instruments whose timbre changes depending on their structure, but according to the present invention, it is possible to create musical tones that are almost similar to the sounds of natural musical instruments.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of an electronic musical instrument to which the musical tone forming method according to the present invention is applied, FIG. 2 is a diagram showing an example of a time function used in the embodiment of FIG. 1, and FIG. A block diagram showing another embodiment of the part forming carrier angular frequency information and modulation angular frequency information in the embodiment of FIG. 1, and FIG. 4 shows details of the time function generator used in the embodiment of FIG. 1. A circuit diagram showing an example of the configuration, FIGS. 5a and 5b are diagrams showing an example of the detailed configuration of the random function generator used in the embodiment of FIG.
The figure shows another embodiment of the part that generates the time function with randomness used in the embodiment of FIG. 1, and FIG. 7 shows a schematic operation of the function generator shown in FIG. 6. This is a flowchart showing the following. 10... Key switch circuit, 20... Frequency information memory, 30... Frequency modulation circuit, 40... Amplitude control circuit, 302, 312, 402... Time function generator, 303, 307, 313, 403...
Random function generator.

Claims (1)

[Scope of Claims] 1. A musical tone forming method in which a musical tone signal is formed by frequency modulating the frequency of a carrier wave using a modulating wave, wherein the relationship between the frequency of the carrier wave and the frequency of the modulating wave is a non-integer multiple relationship. 1. A musical tone forming method characterized in that the frequency of at least one of the carrier wave and the modulated wave is changed according to a random function that changes randomly over time so that the frequency changes randomly over time. 2. In a musical tone forming method in which a musical tone signal is formed by frequency modulating the frequency of a carrier wave with a modulating wave, the relationship between the frequency of the carrier wave and the frequency of the modulating wave is a non-integer multiple relationship at the rising edge of the musical tone signal. The frequency of at least one of the carrier wave and the modulated wave is changed according to a random function that changes randomly over time during a time corresponding to the rising portion, so that the frequency changes randomly over time, and A musical tone forming method characterized in that a modulation index in frequency modulation is changed according to a random function that changes randomly over time for a time corresponding to the duration of a musical tone signal.