JPH02239287A - Musical sound waveform generating device - Google Patents

Abstract

PURPOSE:To obtain a desired musical sound waveform of the musical sound of a natural musical instrument or the like by controlling the mixture rate of a modulated signal to a carrier signal in the range from 0 to an arbitrary value. CONSTITUTION:Phase difference data M(t) (rad) is multiplied in a multiplier 4 by a modulation index I and is added to period data F(t) (rad) to obtain phase angle data F(t)+I.M(t) for modulation of a sin wave (sine wave). This data is inputted to a sin ROM 2, which is the ROM memory to store sin wave data, as the address signal which is used to modulate and read out the sin wave from the ROM 2. A modulated output D(t) read out from the ROM 2 is multiplied in a multiplier 5 by a normalization coefficient K (section) read out from a waveform data ROM 1 in each waveform section and is outputted as a waveform output OUT(t). In this case, the mixture rate is set to 1 just after the start of generation of the musical sound and is approximated to 0 in accordance with the elapse of time thereafter, thereby gradually controlling the frequency characteristic of the musical sound waveform so that the state of the desired musical sound waveform is changed to the state including only a single sine wave component or a single cosine wave component.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a musical sound waveform generator for an electronic musical instrument,
More specifically, it relates to a musical sound waveform generation device that performs frequency modulation to generate musical sound waveforms with various overtone characteristics. [Prior Art] With the advancement of digital signal processing technology, the first conventional example of an electronic musical instrument using the digital processing is to not only generate musical sound waveforms with simple characteristics, but also to generate musical sound waves of natural instruments, human beings, or the natural world. A PCM-based electronic musical instrument has been realized that can directly sample and store sounds, etc. (hereinafter collectively referred to as natural sounds) and reproduce them at any pitch.

On the other hand, as a second conventional example of an electronic musical instrument that can digitally generate musical sound waveforms with various types of complex characteristics,
There are electronic musical instruments based on the FM system described in Japanese Patent Application Laid-open No. 4-33525 or Japanese Patent Application Laid-open No. 126406/1982. This method is basically e=A-sin (ωet+I(t) sin ω.
t)...(A) The waveform output e obtained by the calculation formula is a musical sound waveform, and the carrier wave frequency ω. and the modulation wave frequency ω for modulating it. by selecting an appropriate ratio of Moreover, it is possible to obtain a very unique synthesized sound whose overtone characteristics can change over time.

Further, as a third conventional example of an improved FM system, there is an electronic musical instrument described in Japanese Patent Publication No. 12279/1983.

This method uses triangular wave operation instead of the sin operation in formula (A), and calculates e=A−T (cx+ I(L) T (θ)l・・−
The waveform output e obtained by the calculation formula (B) is taken as a musical sound waveform. Here, T(θ) is a triangular wave function generated by the carrier phase angle θ.

Then, by advancing the carrier wave phase angle α and the modulating wave phase angle θ at an appropriate advancing speed ratio, and setting the modulation index 1 (t) and amplitude coefficient A in the same manner as in the first conventional example, the musical sound waveform is can be synthesized.

[Problem to be Solved by the Invention] With the background of the above-mentioned conventional technology, in recent years, electronic musical instruments have been improved in their ability to dynamically produce sounds ranging from the very individual musical tones unique to electronic instruments to natural sounds. It has been demanded.

However, the first conventional example, the PCM electronic musical instrument,
Although it is very good at producing natural sounds themselves, it is not good at processing natural sounds to produce unique tones. In other words, for example, when you want to change the original sound continuously into a sine wave, etc., you use a digital filter or an analog filter to remove the harmonic components of the original sound to obtain a sine wave, but with a digital filter, the circuit size is small. becomes relatively large, and if it is attempted to change its characteristics using a time function such as an envelope, it is necessary to store filter coefficients corresponding to the filter characteristics in addition to the natural sound data. On the other hand, analog filters have problems in that it is difficult to obtain the desired characteristics, and it is not possible to perform time-sharing operations to generate multiple musical tones in parallel. Furthermore, contrary to the above, if you want to continuously change the original sound to a musical tone with a more complex harmonic structure, it is not possible to generate new harmonic components using methods such as removing the harmonic structure of the original sound with the above filter. The problem is that it is impossible.

On the other hand, the musical tones of an actual musical instrument such as a piano, for example, include not only the fundamental wave component (based on the pitch frequency), but also harmonic components of multiple frequencies that are integral multiples of the fundamental wave component, and even harmonic components of considerably higher order exist. do. Furthermore, overtone components of non-integer multiples may be included. Further, the proportion of each higher-order overtone varies depending on the type of musical instrument, and various overtone characteristics exist depending on the instrument. In this way, musical tones with rich tonal quality are generated by the presence of overtone components inherent in each instrument. However, electronic musical instruments based on the FM method, which is the second or third conventional method, are very good at manipulating the overtone structure of the musical tones produced, but as an output, they have the above-mentioned characteristics unique to each instrument. When you want to obtain a desired musical tone, it is difficult to set the parameters optimally. That is, in the second conventional example,
Since modulation is based on sine wave modulation, the frequency components of the musical tones generated by equation (A) are concentrated in low-order (low-frequency) overtone components, and the modulation is performed by setting the modulation index (rct) to a large value. Even if you apply it deeply, the high-order (high-frequency) harmonic components do not appear well. Therefore, the second conventional example has a problem in that it is not possible to generate musical tones with a rich quality similar to actual musical tones, and the quality of musical tones that can be generated is limited.

On the other hand, in the third conventional example based on the above formula (B), since the modulation is based on a triangular wave that originally contains many overtones, musical tones in which even high-order harmonic components clearly exist as frequency components can be used. However, if you want to obtain a desired musical tone as an output, the carrier wave phase angle α and the modulated wave phase angle θ in the above equation (B) can be changed accordingly.
It is difficult to optimally determine the traveling speed ratio, modulation index Bt), amplitude coefficient A, etc. In addition to this, the third conventional example is a method of driving a triangular wave with a triangular wave, so for example, in the process of a musical tone gradually attenuating after it begins to sound, the amplitude decreases in order from high-order harmonic components. However, there is a problem in that it is impossible to realize a process in which there is only a single sine wave component corresponding to the pitch frequency.

Furthermore, if the musical sound waveform to be obtained is a waveform that changes over time after the start of sound generation, in both the second and third conventional examples, each parameter is set so as to obtain the desired waveform over time. is even more difficult.

The object of the present invention is to be able to faithfully produce natural sounds whose characteristics change continuously with a small circuit scale, and to be able to easily and continuously control the harmonic components, and to produce musical sounds such as single sine waves. The aim is to make it easy to synthesize.

[Means to solve the problem]

The present invention first includes a carrier signal generating means for generating a carrier signal. The means receives as input a carrier wave phase angle signal in which the phase angle repeats an operation in which the phase angle increases sequentially and linearly over time, for example, and converts it according to a certain function and outputs it as a carrier signal. , a ROM, etc., which receives a carrier wave phase angle signal as an address input.
Note that the characteristics of the output carrier signal will be described later.

Next, it has a means for generating a change w418 signal. The means is a means for inputting, for example, the carrier wave phase angle signal, converting it according to a certain function and outputting it as a modulation signal, and is constituted by a ROM or the like having the carrier wave phase angle signal as an address input. The characteristics of the output modulated signal will be discussed later. Further, when the modulated signal is mixed with the carrier signal generated from the carrier signal generating means, the mixing ratio of the modulated signal to the carrier signal is controlled between 0 and an arbitrary mixing ratio, It has a mixing control means for outputting a mixed signal in which the modulated signal is mixed at the mixing rate. The means includes, for example, a multiplier that multiplies the modulation signal output from the modulation signal generation means by a modulation index whose value can vary between O and 1, and an output signal of the multiplier and the carrier. Adding the carrier signals generated from the signal generating means,
This is an adder that outputs a mixed signal. Note that the above-mentioned mixing ratio may change over time after the start of sounding the musical sound waveform. In other words, for example, with the above multiplier 1v device! f! The modulation index to be used is operated to be multiplied by the multiplier at each time that elapses after the start of the sound waveform. Furthermore, it has a waveform output means for inputting the mixed signal output from the mixing control means and outputting a modulated musical sound waveform, the input and output of which have a predetermined functional relationship. The means is, for example, a decoder that converts the mixed signal according to the predetermined functional relationship and outputs it as a musical sound waveform. Alternatively, it is a ROM etc. that uses the mixed signal as an address input.

In addition to the above configuration, the predetermined functional relationship and the carrier signal are such that when the mixing control means controls the mixing ratio of the modulation signal to O, the musical sound waveform generated from the waveform output means is The relationship is such that it becomes a sine wave or cosine wave with a single frequency. Further, the carrier signal generating means and the modulating signal generating means each control the mixing ratio of the modulating signal to a predetermined mixing ratio, for example, 1, by the mixing control means for each of a plurality of waveform sections of arbitrary periods.
A carrier signal and a modulation signal are generated such that when the waveform output means is controlled so that the waveform output means outputs a desired musical waveform of the corresponding waveform section. In addition to the above configurations, the present invention includes an amplitude envelope control means for temporally changing the amplitude envelope characteristic of the musical sound waveform output from the waveform output means. The means multiplies, for example, the musical sound waveform outputted from the waveform output means by an amplitude coefficient whose value can vary over time, for example between 0 and 1, according to a predetermined amplitude envelope function after the musical sound waveform starts producing sound. It is a multiplier. Note that the means may be a means for normalizing the amplitude of the output of the waveform output means to match the desired musical sound waveform. [Function] The function of the present invention is as follows. The musical sound waveform outputted from the waveform output means basically has characteristics obtained by converting the carrier signal outputted from the carrier signal generation means according to a predetermined functional relationship, and furthermore, the mixing control means modulates the carrier signal. By mixing the signals, a characteristic that the musical sound waveform is modulated by the modulation signal is added.

In this case, the relationship between the predetermined functional relationship in the waveform output means and the carrier signal from the carrier signal generating means is
The relationship is set such that when the mixing ratio of the modulation signal is controlled to be 0 by the mixing control means, the musical sound waveform generated from the waveform outputting means becomes a sine wave or a cosine wave of a single frequency.

With this, if the mixing rate of the modulation signal is set to 0 in advance by the mixing control means, it is possible to generate a musical sound waveform consisting only of a sine wave or a cosine wave of a single frequency. Furthermore, when the carrier signal generating means and the modulating signal generating means are each controlled by the mixing control means so that the mixing ratio of the modulating signal is, for example, 1 for each of a plurality of waveform sections of an arbitrary period, , the waveform output means generates a carrier signal and a modulation signal such that a desired musical waveform of the corresponding waveform section can be obtained. With this, if the mixing rate of the modulation signal is set to, for example, 1 in advance by the mixing control means, a desired musical sound waveform, such as a musical tone of a natural instrument, whose characteristics change continuously over time after the start of sound generation, can be obtained. It is possible to obtain In this case, the waveform section does not necessarily need to be divided in synchronization with the pitch period of the desired waveform, but may be divided at an appropriate period (for example, every zero cross point), so that each modified signal corresponding to the desired waveform and the carrier signal can also be easily determined. Also, during a performance, by setting the mixing ratio to 1 immediately after the start of sound generation, and then increasing the mixing ratio to 0 as time passes, it is possible to change the state of the desired musical sound waveform to a single sine wave component or a single sine wave component. It is possible to gradually control the frequency characteristics of a musical sound waveform so that it contains only one cosine wave component. Alternatively, by continuously changing the mixing ratio to, for example, 1 or more, it is possible to control a desired musical sound waveform so that a unique musical tone having a more complex overtone structure is produced.

Along with the above operation, the amplitude envelope characteristic of the musical sound wave output from the waveform output means is also controlled by the amplitude envelope control means so as to attenuate over time, so that it can be reproduced from the start of sound generation, like the musical tone of an actual musical instrument. , it is possible to realize a process in which the musical sound waveform gradually attenuates. [Example] Hereinafter, an example of the present invention will be described with reference to the drawings.

1 is a diagram showing the principle structure of one embodiment of the musical sound waveform generator according to the present invention. First, from waveform data ROM 1, periodic data F
(t), phase difference data M(t), and a normalization coefficient K (section) having a constant value for each waveform section (described later) are read out in synchronization with each other. First, phase difference data A(1)(r
ad) is multiplied by the modulation index I in the multiplier 4, and then added to the periodic data F(t) (rad) in the adder 3 to obtain phase angle data F for modulating the sine wave.
(t)+I − M(t) is obtained. The same data is s
Sin is a ROM memory that stores in wave data.
It is input to the ROM 2 as an address signal for modulating and reading a sine wave from the ROM 2. sin ROM
The modulation output D(t) read from the waveform data ROM 1 is multiplied by the normalization coefficient K (section) read from the waveform data ROM 1 for each waveform section in a multiplier 5, and then the modulation output D(t) read from the waveform output OU
It is output as T(t). Here, sin ROM2
The maximum absolute value of the amplitude of the sine wave stored in is normalized to be 1. The operation of the musical waveform generator shown in FIG. 1 based on the above-mentioned principle structure will be explained below. First, the original waveform ORG(t) of the musical sound waveform of a natural instrument, etc. is
As shown in FIG. 2(a), for example, each period section (pitch period) of the fundamental wave is divided into waveform sections A to D.

Then, within each waveform section, O (rad) or more 2πD
phase angle data that linearly increases sequentially with the passage of time as shown in FIG.
) (rad). Now, set the variable subindex I in Figure 1 to O
Then, due to the periodic data F(t) itself, sin
When the sine wave stored in ROM 2 is read out by specifying its phase angle linearly to obtain the output sin(F(t)), each waveform section A-D is A sine wave of one period corresponding to a phase angle of O to 2π (rad) is read out without modulation. In addition, each waveform section A
-D does not need to correspond exactly to the pitch period; in particular, for musical sounds with weak periodicity such as percussion sounds, one waveform section may be defined as, for example, from an appropriate zero-crossing point (the point in time when the amplitude is 0) to the zero-crossing point. .

Next, the phase difference data M(t) read from the waveform data ROM 1 in FIG. 1 is output from the adder 3 with the value of the modulation index I multiplied by the multiplier 4 as 1)
Using the phase angle data F (t) + M (t) shown in
When the in-wave is modulated and read out, the original waveform ORG for one waveform section with the maximum absolute value of the amplitude normalized to 1 is used as the waveform output OUT (t) as shown in FIG. 3(a).
(t) is the data to be read out, and in Fig. 3 (C
) is shown.

Further, the normalization coefficient K (section) for each waveform section read from the waveform data ROM 1 in FIG.
Since the maximum absolute value of the wave amplitude is normalized to 1, this is a coefficient for returning the amplitude to the final original waveform ORG(t) for each waveform section.

From the above relationship, 01JT(t) −K(interval) ・sin(F(t)+
I − M (t)) ... (1) ORG (t) = K
(Interval) ・sin(F(t)+M(t)) ・・
It can be seen that there is a relationship (2). As can be seen from these relationships, for each waveform section, when the value of the modulation index I is set to 1, the desired original waveform OR is set as the waveform output OUT(t).
If you want to obtain G(t), OR the original waveform of that waveform section.
It is necessary to obtain period data F(t), phase difference data M(t), and normalization coefficient K (interval) corresponding to G(t). How to obtain it will be explained by taking as an example the case of waveform section A of the original waveform ORG (t) in FIG. 2.

First, the method for deriving the periodic data p(t) has already been explained with reference to FIG. 2(b).

Next, in the waveform section A of FIG. 2(a), the original waveform ORG(t
) is set as the normalization coefficient K (interval A). Then, by dividing each amplitude value of the original waveform ORG (t) in the same interval by the normalization coefficient K (interval),
The amplitude is normalized to within ±1 as shown in FIG. 3(a). In FIG. 3(a), the maximum value of both the positive and negative absolute values is 1, but the maximum value of either one may be 1 and the other may be 1 or less.

Next, the normalized original waveform OR obtained in this way
Using G(t), phase difference data M(t) is obtained by the following processes (1) to (2) (see FIG. 4).

■First, waveform section A of the normalized original waveform ORG(t)
For any time t8 within, that time t. Find the amplitude 80 of the normalized original waveform ORG (t) corresponding to (
Figure 4(a)).

■On a sine wave whose maximum absolute value is 1, the amplitude A. (Fig. 4 Φ)) In this case, the position at time t8 in the original waveform ORG(t) and the position of the phase angle P8 in the sine wave should have approximately the same relationship. to ask. That is, for example, if time t8 is within about 174 from the beginning of waveform section 8, the phase angle is also determined to be around 0 to π/2, which is within about 1/4 from the beginning.

■Phase angle PM<! determined by ■ above! :, Using the periodic data F(t) of waveform section A obtained in advance, PX − F
(tx), phase difference data M(t,) corresponding to the time is determined (FIG. 4(C)).

■By changing the time tX over the entire range of waveform section A, the above ■～
Repeat the process (2) to obtain phase difference data M(L) corresponding to each time t of waveform section A. The above processing of ■ to ■ is performed for each waveform section A-D of the original waveform ORG (t) in Fig. 2(a).
Periodic data F(L
), phase difference data M(t) and normalization coefficient K (interval) are stored in the waveform data ROM 1 in FIG. By performing the modulation operation with the value of the modulation index 1 multiplied by the multiplier 4 in FIG. 1 as 1 together with each of the above data, the waveform output o'r(t) in FIG. a
) can be obtained.

Next, in FIG. 1, by changing the value of the modulation index (2) multiplied by the multiplier 4, various modulated waveform outputs OMIT (t) can be obtained.

First, if the modulation index I=0, it is possible to obtain an unmodulated sine wave within each waveform section as already shown in FIG. , 1.5, and 2.0, the modulated output D(t) from the sin ROM 2 in FIG. 1 becomes as shown in FIGS. 5(a), (b),
It is possible to obtain waveforms that are sequentially and deeply modulated as shown in (C).

As described above, by changing the value of the modulation index I, various modulated waveforms can be obtained, from a sine wave to a deeply modulated waveform, centering on the original waveform. Also, the modulation index I
By continuously changing OUT from the start of sound until the end of sound, for example, the waveform output OUT (
It is also possible to obtain t). Next, FIG. 6 is a concrete configuration diagram of one embodiment of the above-mentioned musical sound waveform generator. In the same figure, is it the origin of figure 1? Parts with the same numbers as the configuration have the same functions as in Figure 1. Note that the waveform data ROM 1 in FIG. 1 is FM in FIG.
It is composed of two ROMs: ROM1+ and K ROMI■.

In FIG. 6, periodic data F(t
．． ) and the address data ADD#1 for reading out the phase difference data M(t) are latched into the latch 6 in the accumulator consisting of the adder 7, the selector 8 and the latch 9, with the start address a as the initial value. The pitch data d is sequentially accumulated at address intervals in synchronization with the rising edge of the clock CLK#1 input to the latch 9. In this case, the start address a is output from a control section (not shown), selected by the selector 8 while the start pulse STRT is at a high level, and given to the latch 9 as the initial value of the accumulated value.

If the start pulse STRT is at low level, the output of the adder 7 is selected and the accumulation operation is executed. Also, pitch data d
is output from a control section (not shown) and latched by the latch 6 in synchronization with the rising edge of the start pulse STRT. Phase difference data M(t) output from FM ROMII
is multiplied by the modulation index 1 latched by the latch 16 in the multiplier 4, and added to the periodic data F(t) output from the PM ROMI in the adder 3. The phase angle data F(t)+I − M(
t) is input to the sin ROM 2 as a read address. Note that the modulation index I is outputted from a control section (not shown) and latched by the latch 16 in synchronization with the rising edge of the start pulse STRT.

As a result, the modulated output o(L) is output from the sin ROM 2 to the multiplier 5.

On the other hand, the section identification data 1B output from the FM ROM1+ is input to a D flip-flop (F/F, the same applies hereinafter) 10 that operates in synchronization with the rising edge of the clock obtained by inverting the clock CLK#1 with the inverter 11. At the same time, it is input to the first input of the exclusive OR circuit (EOR, hereinafter the same) 12. Furthermore, the positive logic output Q of the F/F 1 0 is input to the second input of the EOR 12. Does the above circuit configuration change the section identification data IBO value that is sequentially output from FM ROMII? If there is, EOR1
The output of 2 becomes logic "1".

The address data ADD#2, which is the address input to the K ROM 12, is sent to the adder 13, selector 14, and latch 1.
In the accumulating unit consisting of 5, each air address is accumulated sequentially each time the output of EOR 1 2 becomes logic 1'l'' with the start address b as an initial value. Note that the start address b is output from a control section (not shown), is selected by the selector 14 while the start pulse STRT is at a high level, and is given to the latch 15 as the initial value of the accumulated value. If the start pulse STRT is at a low level, the output of the adder 13 is selected and an accumulation operation is executed.

As a result, address data ADD#2 is transferred to KRO old 2.
is given, and a normalization coefficient K (interval) is output from K ROMI to the multiplier 5. The multiplier 5 multiplies the modulation output D(t) output from the sin ROM 2 by the normalization coefficient K (interval),
This multiplication result is further multiplied by the envelope value generated from the envelope generator 17 in the multiplier 18 .

The result of this multiplication is latched by the latch 19 in synchronization with the rising edge of the clock CLK#1 inverted by the inverter 20, and is output as the waveform output OUT (t).

In the above configuration, clocks CLK#1, CLK#
2 and the start pulse STRT are output from a control section not particularly shown. Next, the FM ROMI. Figure 7 shows the structure of the data stored in the . In the figure, one set of waveform data from the start of sound generation to the end of sound is stored in order from the first address a, and one address contains one cycle data F(t), one phase difference data M(t), and One-bit section identification data IB is stored in sets. In this case, as the address advances, data at each sampling point of each waveform section A, B, C, D, . . . in FIG. 2(a) is stored. Also, as the waveform section changes from A to BSC to D, etc., the value of each address of the section identification data 1B becomes "0" in section A, "1" in section B, and "0" in section C. , "1" in section D
, . . . and so on, it changes alternately in units of sections.

This section identification data IB is data for identifying the boundary of the section and updating the address specified in the K ROM 1z as described later.Next, the structure of the data stored in the K ROM 12 in FIG. is shown in Figure 8.

In the figure, each waveform section A from the start of sound generation to the end of sound,
Corresponding to BSC, D, . . . , normalization coefficients K (sections) are stored one by one in order from the top address b.

The operation of the embodiment having the above configuration will be explained according to the operation timing chart of FIG. Hereinafter, reference will be made to FIG. 6 unless otherwise specified.

At the start of sound generation, a start pulse STR↑ is output from a control section (hereinafter simply referred to as the control section) not shown in the drawings.
However, at the timing t1 in FIG. 9, the logic becomes "1" (hereinafter,
Simply call it "1". The same goes for logical "0". ), and immediately after that clock CLκ#1 becomes "1" from the timing t2? Start the sound action. In this case, the cycle of clock CLK#1 corresponds to the sampling cycle of musical tone generation, and the clock CLK#1 generated from the control section
LK#2 is a clock that has the same cycle as clock CLK#1 and is delayed by 174 cycles from the same clock. The following operation is based on the above two clocks CLK#1 and CLK#.
2. In addition, the start pulse STRT
maintains "1" for one cycle from when clock CLK#1 falls to "0" until it falls to "r■" at the start of sound generation, and then remains "0" until the next sound generation starts.
” to maintain.

First, the start pulse STRT from the control section rises to logic "1" at a timing t1 as shown in FIG. 9(a).
Then, the pitch data d from the control section is latched into the latch 6 as shown in FIG. 6(C).

Subsequently, during timings t1 to t4 when the start pulse STRT is "1", the selector 8 selects the start address a from the control section, and this start address a is selected by the clock CLK.
#1 is latched by latch 9 at timing t2 when it rises to "1", and address data AD is output as shown in FIG. 9(d).
The initial value of D#1 is determined. As a result, after a short delay time from t2, the start address a of FM RO)(1,
The period data F(t), phase difference data M(t), and section identification data 1B (see FIG. 7) are read out as shown in FIGS. 9(e), (g), and (1).

Thereafter, the address data ADD#1 output from the latch 9 is fed back to the adder 7, and the pitch data d set in the latch 6 is successively accumulated.

In this case, the above cumulative value is
” at each timing L5 and t8 in FIG. 9(d).
, tll, tl4, tl7, etc., are sequentially launched into the latch 9 via the selector 8 and designated as new address data ADD#1, and the FM ROMI. Period data F(t) and phase difference data M( of the corresponding address above)
t) and section identification data IB (see Figure 7) are shown in Figure 9 (
e), (g) and (1). In addition,
Since the start pulse STRT falls to "0" at t4, the selector 8 selects the output of the adder 7 after t4.

In this case, the instruction to start producing a musical tone is given by the performer pressing any key on the keyboard (not shown), for example, and if the pressed key at that time is a high-pitched key, , pitch data d having a large value is latched into the latch 6 from the control section. As a result, the address width skipped on the FM ROl 41 increases, and a waveform output OUT (t) with a high pitch period is obtained. Conversely, for example, when the lowest pitch key is pressed, the value 1 is latched as the pitch data d, and each data is read out one address at a time on the FM ROMII, and the waveform output of the lowest pitch cycle is output ( t.) is obtained.

Among the data read out from the FM ROM1+ as described above, the phase difference data M(L) is input to the multiplier 4. Now, at t1 when the start pulse STRT rises to "L", the modulation index I is set in the latch 16 by the control section. Therefore, in the multiplier 4, the phase difference data M(t) is multiplied by the modulation index ■. This output is input to the adder 3, where it is added to the period data F(t) output from PMROM II, and the phase angle data F
(t)+I − M(t) is obtained.

From the above phase angle data F(t)+I − M(t), s
in ROM2 is accessed and the periodic data F(t
) and the output of phase difference data S(t) (Figure 9 (g), (
1) each timing t2, t5, t8, tll, t
l4, immediately after LIT, etc.) after a short delay, sin
The modulated output D is obtained from ROM 2 as shown in Fig. 9 (n).
(t) is output. On the other hand, between the timings t1 and L4 when the start pulse STRT is "1", the selector 14 selects the start address b from the control section, and this start address b is selected at the timing t3 when the clock CLK#2 rises to "1". It is latched by the latch 15, and the initial value of address data ADD#2 is determined as shown in FIG. 9(i).

This causes K RO
The normalization coefficient K (section A) of the first address b of M1z is read out as shown in FIG. 9(j). Thereafter, the address data ADD#2 output from the latch l5 is fed back to the adder 13, and the logical output value "O" or rl of the EOR12 is
, are accumulated sequentially. In this case, the above cumulative value is calculated at each timing t6, t9, tlo, t12, tl5, tl in FIG. 9(i) when the clock CLK#2 rises to "1".
8, etc., are sequentially latched into the latch 15 via the selector 14, designated as new address data ADD#2, and the normalization coefficient K (interval) of the corresponding address on the K ROM 12 becomes as shown in FIG. 9(j). It is read out as follows.

Note that the start pulse STRT is "O" at t4.
Therefore, the selector 14 selects the output of the adder 3 after t4. In parallel with the above operation, the clock CLK#1 rises to "1" at each timing t2, t5, t8 in FIG. 9(b).
, after a slight delay from tll, L14, tl7, etc.
As shown in FIG. 9(e), FM ROMI. The section identification data IB are sequentially output from. This data is clock C
Each timing t4, t7, t10, tl3, tl6, L19 in FIG. 9(b) when LK#1 falls to "0"
etc., F/F 1 is set to 0 and its positive logic output Q is determined sequentially as shown in FIG. 9(f). Then, the exclusive OR of this positive logic output Q and the section identification data IB from the FM ROMI is calculated by EOR 1 2.

Here, the section identification data output from FM ROMII
The timer IB is set to 0 for each waveform section as shown in
” or r1, are stored alternately, so each time the waveform section changes from A to B, B to C, C to D, etc., “0” is stored.
It changes from ``1'' to ``0'', and from ``0'' to ``1''. Therefore, the output of EOR 1 2 becomes "1" only when the waveform section changes, and becomes "0" at other times. In the example of FIG. 9, the waveform section A
The clock C of tl6 immediately after tl4 changes from to B.
EOR12 output is “l” only while LK#1 is “1”
becomes. From this, only within the above timing, the adder 13
1 is accumulated in the value b of the address data ADD#2 from the latch 15, and this accumulated value b+i is the clock CLK#2.
Timing tl5 in Fig. 9 (c) when ``stands up to lJ''
, new address data ADD#2 is stored in latch 15.
It is latched as Therefore, clock CLK#2 is “
After the timing tl5 when it rises to "1", <K R
The normalization coefficient K (section B) on OMlz is read out as shown in FIG. 9(j).

On the other hand, at other timings when the output of EOR12 is "0", the adder l3 does not perform the accumulation operation, so the latch 15 contains the same address data AD as one timing ago.
D#2 is latched. Therefore, at timings t6, t9, tl2, tlB, etc. in FIG.
2 as shown in FIG. 9(j).

In this way, at the timing when the modulation output D(t) based on the period data F(t) and phase difference data M(t) of the waveform section A from the FM ROM1+ is read from the sin ROM 'l, the K ROMh The normalization coefficient K (section A) corresponding to the waveform section A is read from the waveform section A, and similarly, the normalization coefficient K (section B) corresponding to the waveform section B is read from the KROM1z in the waveform section B, and so on.
Corresponding to the output of the modulated output D(t) of each waveform section, the normalization coefficient K (section) of that waveform section is set to K ROM
It is read from 1z.

As described above, the clock CLK# of FIG. 9 (G)) is
Each timing L2, t5, t8 when 1 rises to “1”
, tll, L14, tl7 etc. after some delay, s
Modulation output D from tn ROM 2 as shown in Figure 9(n)
(t) is output, and in parallel with this, the clock CLK#2 of FIG. 9(5) rises to "1" at each timing t.
After a slight delay from 3, t6, t9, tl2, tl5, tl8, etc., the normalization coefficient K (interval) is output from the K ROM 12 as shown in FIG. 9U). The modulation output D(t) and the normalization coefficient K (interval) at each of these timings are multiplied by the multiplier 5, and then further multiplied by the envelope data from the envelope generator 17 by the multiplier 18. Kronk CLK #1 in Figure 9 [with]
0” at each timing t4, t7, t10, t
At l3, tL6, tl9, etc., it is latched by the latch 19 as shown in FIG. 9(0), and the waveform output OUT(t) for each timing is determined. Here, although the envelope generator 17 and the multiplier 18 are not shown in FIG. 1, they may be operated when it is desired to add an envelope other than the original waveform to the waveform output OUT (t), and are not necessarily necessary. .

Through the operations described above, in the embodiment shown in FIG. 6, it is possible to realize operations similar to those of the musical sound waveform generator shown in FIG. 1 shown in FIGS. 2 to 5.

In Figure 6 above, the waveform data ROM 1 in Figure 1 is configured with two ROMs, FM ROM1+ and K ROMI2, but it can be configured with one ROM if you can accept that the address specification becomes somewhat complicated. You may also do this. Also, in FM ROMI, periodic data F(t)
As shown in Fig. 7, the values for each sampling point are stored, but within one waveform section, the periodic data F
(t) has a linear characteristic in which the phase angle changes from O to 2π (rad) in each waveform interval width as shown in Figure 2 (b), so if there is enough calculation time, By calculating and outputting the data F(t), storage capacity can be saved.

In addition, in Fig. 6, the waveform output OUT (
t), but it is also possible to output a plurality of tone waveforms in parallel by operating each section in a time-division manner.

In the principle configuration of one embodiment of the musical sound waveform generator shown in Fig. 1,
In the case of modulation index 1-0, from equation (1) above, OUT(
t) =K (interval) ・sin(F(t))
...(3), and the periodic data F(t) is shown in Figure 2)
As shown in , the data increases linearly within each waveform interval, and the normalization coefficient K (interval) is constant within each waveform interval, so a sine wave was obtained as the waveform output OtlT(t). .

In addition, in the case of a modulation index of 1-1, as in equation (2) above,
ORG (t) = K (interval) ・sin (F (t) + M
(t))...The original waveform ORG(t) was obtained as (2). By changing the value of the modulation index 1 between 0 and 1, the waveform can be continuously changed from the sine wave to the original waveform. Further, if the value of the modulation index I is set to 1 or more, it is possible to continuously change the waveform from the original waveform to a further modulated waveform.

In this way, in this embodiment, at least s
It is characterized by being able to output in-waves and original waveforms. Therefore, if a sine wave can be output when ■ - 0, and an original waveform can be output when ■ = 1, there is no need to adhere to equations (1) and (2) above, and OUT (t) - K ( interval)・f (g
(F(t))+I-M(t)) (4) Any embodiment may be used as long as it realizes the calculation having the following relationship.

For example, in equation (4) above, let f be the input α, f(α)
=(2/π)α... (0≦α≦π/2) f(α)=-1+ (2/π)(3π/2-α)...
(π/2≦α≦3π/2) f(α)=-1+(2/π)(α-3π/2)(3π/
2≦α≦2π) ... (5) Defined as a triangular wave function that satisfies the following, and with input β and g, g (β) = (π/2) sinβ ... (0≦β≦π/2) g(β)=π−(π/2) sinβ・・(π/
2≦β≦3π/2) g (β) = 2π + (π/2) sin β ・・ (3π/2≦β≦2π) ・ ・ ・ (6) If defined as a function that satisfies the following, the value of modulation index I is O, that is, there is no modulation, then the above equations (5) and (6) can be transformed into the above (
4) By substituting into the formula, f(g(β))=K(interval) ・f((π/2)
sin β) - K (interval) (2/π) (g/2)
sin β=K (interval)・si口β・(0≦β≦π/2) f(g(β))=K(interval)・f((π−(π/2)・
sin β)) =K (interval)・(-1+(2/π)(3π/2-π+
(π/2) sin β)) = K (interval)・sin β・(π/2≦β≦3π/2) f(g(β))=K(interval)・f(2π+(π/2)・
sin β) −K (interval)・(−1+(2/π)(2π++(π/2
)sin β-3π/2))-K(interval)・sin
β ・ ・ (3π/2≦β≦2π) ・ ・ ・(7) That is, a single sine wave is output when no modulation is performed.

Also, by substituting the above equations (5) and (6) into the above equation (4),
In order to obtain the original waveform ORG(t) when the modulation index is 1-1, the sine wave in FIG. 4(b) is replaced with the triangular wave defined by the equation (5), and ,
Replace the periodic data F(t) in the figure (C) with the function g(t) defined by the equation (6) above, and determine the phase difference data M(t) as the difference data from the g(t). do it.

In this case, as means for generating the triangular wave corresponding to the sin ROM 2 of FIG. 1, it is also possible to generate the triangular wave by a decoder circuit or the like in addition to the ROM.

In addition to the above embodiment, various combinations of functions can be defined as the functions f and g that satisfy the above equation (4).

[Effects of the Invention] According to the present invention, if the mixing rate of the modulation signal is set to 0 in advance by the mixing control means, a musical sound waveform consisting only of a sine wave or a cosine wave of a single frequency can be generated, and the mixing control If the mixing ratio of the modulation signal is set in advance to a predetermined mixing ratio (for example, 1) by the means, it is possible to obtain a desired musical sound waveform such as the musical tone of a natural musical instrument. Therefore, during a performance, by setting the mixing ratio to 1, for example, immediately after the start of a musical tone, and then increasing the mixing ratio to 0 as time passes, the desired musical sound waveform can be changed to a single sine wave component or a single sine wave component. The frequency characteristics of the musical sound waveform can be gradually controlled so that it contains only one cosine wave component. Alternatively, by continuously changing the mixing ratio to, for example, 1 or more, it is possible to control a desired musical sound waveform so that a unique musical tone having a more complex overtone structure is produced.

In this case, in particular, the waveform section does not necessarily need to be divided in synchronization with the pitch period of the desired waveform, but may be divided at an appropriate period (for example, every zero cross point), so that each modulated signal corresponding to the desired waveform and carrier signals can also be easily determined.

Along with the above operation, the amplitude envelope characteristic of the musical sound wave output from the waveform output means is also controlled by the amplitude envelope control means so as to be attenuated over time, so that the sound waveform can be reproduced after the start of sound generation, like the musical tone of an actual musical instrument. It is possible to realize a process in which a musical sound waveform gradually attenuates.

As described above, the present invention can easily generate both a state in which natural musical tones are produced and a state in which only a single sine wave component or a single cosine wave component is produced, and also a state in which various overtones are produced. properties can be generated.

Moreover, as a configuration to realize this, a normal RO
Since it can be realized only by combining M, decoders, adders, multipliers, etc., it is possible to realize complex musical sound waveforms with a simple circuit configuration, and as a result, high-quality electronic musical instruments can be provided at low cost. becomes possible.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the principle configuration of one embodiment of the musical sound waveform generator according to the present invention, FIG. 2 is a diagram showing the relationship between ORG(t.) and F(t), and FIG. 3 is a diagram showing the relationship between ORG (t. t), F(t), and M(t.). Figure 4 is a diagram showing how to obtain M(t). Figure 5 is
FIG. 6 is a diagram showing the relationship between D(t), F(t), and M(t) when changing . FIG. 6 is a specific configuration diagram of one embodiment of the musical sound waveform generator. FIG. , FM ROM data configuration diagram, FIG.
FIG. 9, which is a data configuration diagram of the ROM, is an operation timing chart of a specific configuration of one embodiment of the musical waveform generator. ■...Waveform data ROM, 2・-・sin ROM, 3...Adder, 4, 5...Multiplier, Add(L)・F(t)・・M(t)・・Eng.・ ・ F(t)+I・K (interval) D(L)... 0[JT(t)...Address data, ・Period data, ・Phase difference data, Modulation index, M(t)...Phase Angular data, ・・・Normalization coefficient, ・Modulation output, ・Waveform output.

Claims (1)

[Claims] 1) A carrier signal generating means for generating a carrier signal, a modulating signal generating means for generating a modulated signal, and a method for mixing the modulated signal with the carrier signal generated from the carrier signal generating means. Mixing control means that controls a mixing ratio of the modulated signal to the carrier signal between 0 and an arbitrary mixing ratio, and outputs a mixed signal in which the carrier signal and the modulated signal are mixed at the mixing ratio; waveform output means for outputting a modulated musical waveform by inputting the mixed signal output from the mixing control means, the input and output of which have a predetermined functional relationship, the predetermined functional relationship and the carrier signal; is such that when the mixing rate of the modulation signal is controlled to be 0 by the mixing control means, the musical sound waveform generated from the waveform outputting means becomes a sine wave or a cosine wave of a single frequency. The carrier signal generating means and the modulating signal generating means each control the mixing ratio of the modulating signal to a predetermined mixing ratio by the mixing control means for each of a plurality of waveform sections of arbitrary periods. A musical sound waveform generating device, characterized in that, when controlled, a carrier signal and a modulation signal are generated such that a desired musical sound waveform of the corresponding waveform section is output from the waveform output means. 2) The musical sound waveform generating device according to claim 1, further comprising an amplitude envelope control means for temporally changing the amplitude envelope characteristic of the musical sound wave outputted from the waveform output means.