JPS6315915Y2 - - Google Patents

Description

[Detailed Description of the Invention] This invention relates to a musical tone synthesis device used in electronic musical instruments, etc., and more particularly to a device for synthesizing a musical tone of a desired timbre by digital calculation.

A method or apparatus for synthesizing a musical tone of a desired tone by frequency modulation calculation or similar calculation is disclosed in U.S. Pat.
126406) or the specification of JP-A-55-7733. However, the prior art as shown in those prior applications only shows the basic structure of musical tone synthesis using frequency modulation calculations or similar calculations; The means by which this was done had not been fully elucidated. Furthermore, in the prior art in this field, emphasis has been placed on elucidating how to give calculation parameters to obtain a desired tone or musical instrument sound, and elucidating this has been extremely difficult.

This idea was made in view of the above points,
In a device that synthesizes musical tones by frequency modulation calculation or similar calculation, a predetermined timbre,
It is an object of the present invention to provide a practical musical tone synthesis device particularly suitable for synthesizing human voice sounds, pipe organ sounds, etc. This invention achieved the above objectives with a relatively simple calculation method and configuration by using multiple operators, which are the basic elements for musical tone synthesis calculations, and by elucidating the combination of operators suitable for the above-mentioned predetermined timbre. It is something.

An embodiment of this invention will be described in detail below with reference to the accompanying drawings.

Figure 1 shows the basic configuration of the musical tone synthesizer according to this invention, in which three operators OP1, OP2,
OP3 is provided in parallel, and a cyclic frequency modulation calculation circuit 10 (hereinafter, frequency modulation calculation is performed by FM) for these operators OP1, OP2, and OP3.
) is input. The outputs of these operators OP1, OP2, and OP3 are added by an adder 11, and the added output is output as a musical tone signal. One operator OP0 is also used in the cyclic FM circuit 10. The internal configuration of each operator OP0 to OP3 is the same, for example, as shown in FIG.

In FIG. 2, operators OP0 to OP3 include a sine wave table 12, and read out the sine wave sample point amplitude value corresponding to a predetermined phase angle from the table 12 using data given from an adder 13 as an address signal. This operator OP0 to
OP3 functions as a basic calculation element for frequency modulation calculation. That is, one input of the adder 13 is given data indicating the instantaneous amplitude value of a suitable modulated wave signal (temporarily denoted by (ω _n t)) via the modulated wave signal input 14, and the other input is given Data indicating the instantaneous phase angle of the carrier signal (temporarily denoted by ω _c t) is provided via a carrier phase input 15 and a multiplier 16 . The multiplier 16 is provided with a coefficient (temporarily designated k) for controlling the carrier frequency via a carrier frequency control input 17. Therefore,
A value “kω _c t” obtained by multiplying the phase angle data ω _c t given from the input 15 by the coefficient k is input to the adder 13 . With the above configuration, the sine wave table 12
The instantaneous amplitude value sin {kω _c t+(ω _n t)} (1) of a signal obtained by frequency-modulating the carrier wave signal with the modulation signal is read from .

The output of the sine wave table 12 passes through a multiplier 18 and becomes output signals of operators OP0 to OP3. The multiplier 18 is provided with a coefficient for controlling the amplitude of the output signal of the sine wave table 12 via an amplitude control input 19. The coefficient given externally to this input 19 can be a coefficient for setting the amplitude envelope of the musical tone (temporarily denoted as E(t)) or a coefficient corresponding to the modulation index (this can be (temporarily denoted by I(t))
Either. When the output of the operator is used as a modulated wave signal, the coefficient applied to the multiplier 18 indicates the modulation index I(t), and when used as a musical tone signal, the coefficient indicates the instantaneous value E(t) of the amplitude envelope.

In Figure 1, each operator OP0 to OP3
The outputs of shift circuits 20, 21, 22, and 23 are applied to the modulated wave signal input (point 14 in FIG. 2), respectively. In the cyclic FM circuit 10, the output signal of the operator OP0 is sent to the averaging circuit 2.
4 to a shift circuit 20, and returns to the modulated wave signal input 14 of the operator OP0 via this shift circuit 20. Of course, the operator
It is assumed that there is an appropriate time delay between the input and output of OP0. Operators installed in parallel
In the case of OP1, OP2, and OP3, the output signals of the cyclic FM circuit 10, that is, the operator OP0, are provided in shift circuits 21, 22, and 2 corresponding to each other.
3 to each modulated wave signal input 14. Shift circuits 20 to 23 are each operator OP
It shifts the digital value of the modulated wave signal given to OP3 from 0 to OP3 by an appropriate amount to the upper or lower digits, and the shift amount for each is determined by the data S ₀ , S ₁ , S ₂ ,
It is designed to be variable according to _S3 . By this shift, the regression rate is controlled in the cyclic FM circuit 10, and the modulation index is controlled in the operators OP1 to OP3.

The carrier wave phase input (15 in Figure 2) of each operator OP0 to OP3 is provided with instantaneous phase angle data qF (corresponding to ω _c t in Figure 2) that repeatedly changes in accordance with the frequency of the musical tone to be generated. ) are input respectively. Although a device that generates this phase angle data qF is not particularly shown, it reads out a constant F corresponding to a predetermined frequency in response to a key press on the keyboard, and this number F
A device that regularly accumulates the value "qF" (q is a variable that accompanies the progress of calculation timing) or other well-known devices can be used. Each operator OP
Carrier frequency control input from 0 to OP3 (17 in Figure 2)
Carrier frequency control coefficients k ₀ , k ₁ , k ₂ , and k ₃ are input separately to . Common phase angle data qF is input to each operator OP0 to OP3, and this is used as a coefficient
By controlling each of _k0 to _k3 separately, the carrier frequency for each operator can be controlled independently.

Operator OP0 amplitude control input (1 in Figure 2)
9) is given a coefficient I(t) corresponding to the modulation index. _{Operator OP1 ,}
The modulation index of the modulated wave signal (corresponding to (ω _n t) in FIG. 2) used in OP2 and OP3 is set. Coefficients E ₁ (t), E 2 (t), and E ₃ (t) indicating amplitude envelopes are given to the amplitude control inputs (19 in FIG. 2) of operators OP1, OP2, and _OP3 , respectively. In accordance with these coefficients E ₁ (t), E ₂ (t), and E ₃ (t), the amplitude envelope of each musical tone signal to be synthesized in parallel by each operator OP1, OP2, and OP3 is set, and these musical tone signals are The ratio at which the adder 11 adds and synthesizes the signals is controlled.

In the cyclic FM circuit 10, the frequency-modulated signal is fed back as a modulated wave signal at an arbitrary regression rate, so the harmonic components are abundant, and the lower the harmonic, the higher the level, and the higher the harmonic, the higher the level. It has the advantage that it is possible to synthesize a signal having a highly applicable spectral distribution with a monotonically decreasing tendency, and that the number of overtones can be easily controlled by controlling the regression rate (that is, the shift amount of the shift circuit 20). ing. That is, the shift circuit 20
By increasing the regression rate by controlling the shift amount of
In the signal output from the FM circuit 10 (i.e., the modulated wave signal used by operators OP1 to OP3), relatively high-order harmonic components are enhanced both in number and level, and conversely, when the regression rate is lowered, relatively high-order harmonic components are enhanced. The number and level of the next overtone component decreases. The averaging circuit 24 is provided to prevent the hunting phenomenon of signal amplitude caused by circulation, and calculates the average value of the amplitudes of adjacent sample points of the signal input from the operator OP0 to the circuit 24. It is set to output.

Now, according to the combination of operators shown in Figure 1, it is possible to synthesize musical sounds with multiple formants, such as "human voice sound" and "pipe organ sound."
It is ideal for the synthesis of

First, the center frequencies _c1 , _c2 ,
Multiple formants F1, F2, F with different _c3
Synthesis of a timbre, for example, a human voice, will be explained. In Figure 3, each formant F1, F2, F3
The spectral envelope of is illustrated. In this case, each operator OP1, OP installed in parallel
2. Individual fixed formants may be synthesized in each of OP3. For example, fixed formant F1 is synthesized with operator OP1, and F2 is synthesized with operator OP1.
2, F3 is combined with OP3, respectively. Assuming that the frequency ratio of center frequencies _c1 , _c2 , and _c3 is "4:12:18" when the fundamental frequency is 1,
The coefficient k ₁ may be set to "4", the coefficient k ₂ to "12", and the coefficient k ₃ to "18". At this time, in order to seamlessly include harmonics of integer order, set the coefficient k ₀ of operator OP0 to "1".
Set to . Then, the cyclic FM circuit 10 outputs a waveform signal that seamlessly contains integer-order harmonic components from the fundamental wave to harmonics of a predetermined order at a level determined according to I(t) and _S0 . Each operator OP1, OP as a modulated wave signal
2. Calculation is performed in OP3. As a result, musical tone signals corresponding to the respective formants F1 to F3 are output from the respective operators OP1 to OP3, and these are added and synthesized by the adder 11 to obtain a musical tone signal including a plurality of formants. The spectral envelope shape and level of each formant are determined by the shift amount and coefficient E ₁ (t) in shift circuits 21 to 23,
It can be appropriately controlled by E ₂ (t) and E ₃ (t). Furthermore, as described above, the center frequency of each formant is controlled by carrier frequency control coefficients k ₁ , k ₂ , and k ₃ . Here, the pitch of the musical tone (that is, the phase angle data)
If the coefficients k ₁ , k ₂ , k ₃ are determined without regard to the rate of change of qF, so-called moving formant timbre synthesis is achieved in which the formant center frequency also changes in accordance with changes in pitch. However, in order to synthesize human voice sounds, it is necessary to use a fixed formant, and for that purpose,
A so-called key scaling method may be used to set the coefficients k ₁ , k ₂ , and k ₃ . That is, the phase angle data
The values of k ₁ , k ₂ , k ₃ are automatically changed according to the change rate of qF (the pitch or range of the pressed key on the keyboard), and the center frequency _c1 , This is done so that _c2 and _c3 are always approximately constant.

To synthesize the unique sound of a pipe organ, which is a mixture of multiple pipes with different tones, each operator
Carrier frequency control coefficients k ₀ , k ₁ , k ₂ ,
What is necessary is to appropriately set k ₃ and the shift amount of the shift circuits 21 to 23. In the case of musical tone synthesis using frequency modulation calculations, it is known that by changing the modulation index, it is possible to widen or narrow the band of sidebands (i.e., flatten or steepen the spectral envelope). . By arbitrarily setting the shift amount of the shift circuits 21 to 23, the modulation index of each operator OP1 to OP3 can be variably controlled, and the spectrum envelope can be controlled as described above. Fourth
The figure shows an example of the spectral envelope and spectral distribution of the sound of a 16-foot flute 16' pipe, an 8-foot flute 8' pipe, and a 4-foot reed 4' pipe.
Each spectral line is distinguished by a solid line, a dashed line, and a broken line as shown. Horizontal axis scale 1
indicates the position of the fundamental frequency of the 8-foot note. Fourth
To synthesize a tone with a spectrum configuration as shown in the figure, set the carrier frequency control coefficients of each operator OP0 to OP3 to k ₀ = 2, k ₁ = 0.5, k ₂ = 1, k ₃ = 2, and Flute 16′ in OP1, OP
2 to synthesize the flute 8' tone and OP3 to synthesize the reed 4' tone, respectively. that way,
Musical tone signals in which overtone components are located at the positions shown by the spectral lines are obtained by each of the operators OP1 to OP3. Since a flat spectrum envelope like the lead 4' is achieved by increasing the modulation index, the shift amount of the shift circuit 23 is set accordingly. Also, since a relatively steep spectral envelope like the flute 16' is achieved by reducing the modulation index, the shift circuit 2
The shift amount of 1 is set as such. In the case of the flute 8', the shift amount of the shift circuit 22 is similarly set appropriately.

Amplitude control coefficient E ₁ (t) of operators OP1 to OP3
By setting ~E ₃ (t) to 0, it is also possible to control the number of operators used in parallel. For example, if the coefficients E ₂ (t) and E ₃ (t) are fixed to 0,
Only the combination of the cyclic FM circuit 10 and one operator OP1 can be used.

Note that the number of operators OP1 to OP3 and the number of shift circuits 21 to 23 provided in parallel are not limited to three, but may be any number as long as they are plural. Alternatively, instead of providing the operators OP0 to OP3 individually, one operator may be used in a time-sharing manner to perform the functions of the operators OP0 to OP3. Moreover, this invention can be implemented not only by hardware logic but also by software processing of a microcomputer.

Note that the cyclic FM circuit 10 may use a plurality of operators to form a cyclic loop. Further, the sine wave table 12 included in the operator is not limited to sine waves, and may store other predetermined waveforms.

As explained above, according to this invention, a plurality of frequency modulation calculation operators are provided in parallel for the output of the cyclic frequency modulation calculation circuit. This has the excellent effect of making it possible to easily synthesize organ sounds and the like.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an example of this invention, Fig. 2 is a block diagram showing an example of the operator used in Fig. 1, and Fig. 3 is a block diagram showing an example of the operator used in Fig. 1. A graph showing an example of the spectral envelope of a musical tone containing,
FIG. 4 is a graph showing an example of the spectral envelope and distribution of a pipe organ sound that is a mixture of a plurality of pipe tones that can be easily synthesized in the same embodiment. OP0, OP1, OP2, OP3...Operator,
10...Cyclic frequency modulation calculation circuit, 11, 13
... Adder, 12 ... Sine wave table, 14 ...
Modulated wave signal input, 15... Carrier wave phase input, 1
6, 18... Multiplier, 17... Carrier frequency control input, 19... Amplitude control input, 20, 21, 22,
23...Shift circuit, (ω _n t)...Modulated wave signal, ω _c t, qF...Phase angle data, k, k ₀ , k ₁ ,
k ₂ , k ₃ ... carrier frequency control coefficient, I (t) ... modulation index, E (t), E ₁ (t), E ₂ (t), E ₃ (t) ...
Amplitude coefficient.

Claims (1)

[Claims for Utility Model Registration] 1. A cyclic frequency modulation calculation means that returns its own output signal as a modulated wave signal and performs a frequency modulation calculation using this modulated wave signal and carrier wave information corresponding to a musical tone frequency; They are provided in parallel and can independently control calculation parameters, and the output signal of the cyclic frequency modulation calculation means is added as a modulation wave signal to each of them, and this modulation wave signal and carrier wave information corresponding to the musical tone frequency are combined. 1. A musical tone synthesis device comprising: a plurality of calculation means each performing a frequency modulation calculation according to the method of the present invention, wherein a musical tone signal is obtained by adding and synthesizing the outputs of the respective calculation means provided in parallel. 2. Each of the cyclic frequency modulation calculation means and the plurality of calculation means includes a multiplication means for multiplying phase angle data that repeatedly changes in accordance with musical tone frequency by an arbitrary coefficient, and outputting the result as the carrier wave information; an adding means for adding the output of the means and a modulated wave signal; a waveform storage means for storing a predetermined waveform in advance and reading out the waveform using the output of the addition means as an address signal; and an amplitude of the output signal of the waveform storage means. and second multiplication means for variably controlling the tone synthesis apparatus according to claim 1.