JP2503744B2 - Music synthesizer - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a musical sound synthesizer used as a sound source of an electronic musical instrument, and in particular, a conventional musical sound synthesizer cannot synthesize music without complicating the configuration. The present invention relates to a musical tone synthesizer capable of synthesizing musical tones that have been possible or difficult with a simple configuration.

[Prior Art] FM sound sources, memory sound sources (AWM), physical model sound sources, and the like are conventionally known as sound sources for electronic musical instruments.

However, in the FM sound source, there is no example in which a musical sound having an unstable amplitude behavior such as when the clarinet is blown by excessive pressure, or a musical sound in which the amplitude fluctuates irregularly is synthesized. Further, since the memory sound source stores PCM data for a half cycle or a plurality of cycles of a musical sound, any musical sound can be synthesized as long as it is represented by a data string. However, as the number of types of musical tones increases, the memory capacity must be increased correspondingly, and the configuration becomes complicated.

On the other hand, in many physical model sound sources, musical sounds may show very complicated chaotic behavior while changing control parameters. In fact, some models, such as clarinet, have been qualitatively explained that the conditions under which the system behaves chaotically have been established (Maganza, Causse: “Bifur
cation, period doublings and chaos in clarinet like
systems ", Europhysics Letters, 1 (6), PP.295-302,
1986).

However, such chaotic behavior is not reliably reproduced to generate a desired musical sound.

[Problems to be Solved by the Invention] The present invention has been made in view of the problems in the above-described conventional example, and conventionally, synthesis is difficult, the configuration becomes complicated, or a relatively special It is an object of the present invention to provide a musical tone synthesizing device capable of synthesizing a musical tone that has not been synthesized due to a musical tone having a low possibility of being used with a simple structure and good reproducibility.

[Means for Solving the Problem] In order to achieve the above-mentioned object, the tone synthesis apparatus of the present invention sets an initial value at the start of tone generation and sets one input sample signal based on predetermined frequency information. And a signal stored in the 1-sample signal storage means, and the input signal is converted according to a predetermined non-linear function characteristic. And a non-linear conversion means for inputting the converted one sample signal to the one-sample storage means, and a series of sample signals sequentially stored and converted one by one according to the timing signal as a waveform data string. It is characterized by having a waveform generator for outputting. The predetermined frequency information, timing signal, 1
The sample signal storage means and the non-linear conversion means respectively correspond to ΔF, the chaos update pulse SP, the delay circuit 1 and the chaos arithmetic circuit 3 of the embodiment described later.

The waveform data string output from this waveform generator is as the tone waveform itself, as the envelope waveform of the tone.
Or used as a calculation parameter of FM sound source.

In particular, when the waveform data string is used as the musical tone waveform itself, the waveform generator stores and converts the sample signals one by one at a period proportional to the pitch of the musical tone to be generated, and sequentially outputs the sample signals.

[Operation] The waveform generator constitutes a dynamical system for executing the regression operation x _{n + 1} = f (x _n ) (where n = 0,1,2,3, ...), and A chaotic sequence is generated as a waveform data sequence.

The following expression x _{n + 1} = f (x _n ) (1) (n = 0,1,2,3, ...) f (x) = ax (1-x) (2) A very simple dynamical system (recurrence formula with a nonlinear term) is a very popular formula for generating chaos.

Here, a is an arbitrary constant, but is a number larger than 3 and slightly smaller than 4 for generating chaos.
Further, the closer it is to 4, the more complex the chaos characteristic is generated. Also, an initial value x ₀ must be given first, but this value is generally chosen to be a small positive value.

FIG. 1 is a graph in which the appearance of chaos is traced on a graphic. The curve in the graph shows f (x) when the value of a is about 3.5. It becomes a parabola showing the maximum value a / 4 at x = 0.5. To follow the movement of x _n on the graph, find the intersection point of f (x) with a straight line that rises from the bottom of x _n-1 in parallel with the y axis, and y coordinate of the intersection point y = f (x) As shown by a straight line drawn diagonally from the y-axis, it is sufficient to repeat the work of setting a point lying on the x-axis to a new x _n-1 . When starting from the initial value x ₀ and following the locus from the first point, it returns to the vicinity of the value twice before with the cycle of the work of finding x as a cycle, but never returns to the exact same value. That is, it looks like a periodic signal and is random, and if it is completely noise, there is something like periodicity. Such an almost periodic signal is called chaos. In the dynamical system of this example, the larger the value of a, the longer the cycle, which is generally a power of 2.

Figure 2 shows x _n when the initial value is 0.01 and a = 3.6.
Is plotted with n as the horizontal axis. From this, it can be seen that chaos can be used as the tone signal itself.

The inventors of the present invention positively incorporate a chaos-generating system into a sound source to generate a chaotic musical tone, for example, a musical tone having an unstable amplitude behavior when the clarinet is blown by excessive pressure. Or a musical sound whose amplitude fluctuates irregularly,
We have found that they can be synthesized with good reproducibility with a simple structure.

As described above, in many physical model sound sources, musical sounds may exhibit chaotic behavior while the control parameters are being changed variously. Further, according to the knowledge of the present inventor, the oscillation of the feedback FM sound source is also chaotic.

However, the chaos of these physical model sound sources and FM sound sources is not always reproducible with the desired chaotic musical sound, such as the occurrence being accidental, the generation conditions of the chaos and the nature of the generated chaos being limited. It was not something that could be obtained.

[Effect] According to the present invention, since the system for generating chaos is positively incorporated in the sound source, chaos having various properties can be generated with good reproducibility, and various chaos can be dealt with. Musical sounds with various chaotic behavior can be synthesized with good reproducibility and a simple structure.

[Embodiment] An embodiment of the present invention will be described below with reference to the drawings.

FIG. 3A shows the structure of a physical model sound source according to an embodiment of the present invention. The sound source in the figure is a clarinet physical model sound source, and uses the chaotic operation circuit, which is a feature of the present invention, as a non-linear function generating circuit in a conventional physical model sound source of the same kind.

In the figure, the delay circuit 1 delays the input signal x by the period of the drive clock SP and outputs it, simulating a clarinet tube, and its delay time corresponds to the length of the tube. The multiplier 2 simulates the tube end, and the multiplication coefficient -1 corresponds to the reflection of sound waves. The non-linear function generating circuit 3 simulates a lead movement, but here, a chaotic operation circuit is used.

The chaos arithmetic circuit 3 is an arithmetic unit that calculates a predetermined arithmetic expression (dynamic system). Here, the quadratic function expressed by the above equations (1) and (2) is used as the arithmetic expression, and the negative half is positive. Half of is used as a folded even function. By using an even function in this way, the sequence of chaos shown in FIG. 2 appears positive and negative for each half cycle. As a result, a relatively simple circuit can obtain a waveform data string of a musical tone showing an unstable amplitude behavior such as when the clarinet is blown by excessive pressure.

In the sound source of FIG. 3A, the coefficient a given to the chaos calculation circuit 3 is treated like an index in the FM sound source, and an envelope generator (EG) or low frequency oscillator (LFO) is used.
Instability that fluctuates with time can be generated by vibrating the same. Also, this chaotic operation circuit 3
As described above, when a is reduced, the output loses instability as chaos, and behaves as a normal rectangular wave oscillator.

The phase generator 4 accumulates the frequency designation information ΔF corresponding to the pitch of the musical tone to be generated in a cycle of a predetermined relock φ, and generates a carrier output generated each time the accumulated value overflows as a chaotic update pulse SP. To do. Such a phase generator 4 can be composed of, for example, an adder 31 and a delay circuit 32 as shown in FIG. 3B. Delay circuit
32 is an adder that delays the addition output of the adder 31 by a period φ.
It returns to one input terminal of 31. The adder 31 is supplied with the frequency specifying information ΔF at the other input terminal, and adds the frequency specifying information ΔF with its own addition output in a cycle φ.
That is, each time the added value exceeds the specified value, the carrier output
Generates C ₀ . This carrier output C ₀ is output as the chaos update pulse SP.

The delay circuit 1 is supplied with this chaotic update pulse SP as a drive clock. That is, the input signal f (x) of the delay circuit 1 is delayed by one cycle of the pulse SP and then output.

The signal obtained by delaying the input signal f (x) of the delay circuit 1 is inverted by the multiplier 2 and input to the chaos calculation circuit 3 as a new variable x. As a result of the above loop operation being executed each time the pulse SP is generated, one chaos number is output each time the pulse SP is generated. The initial value x _{0 of the} calculation is given to the delay circuit 1 here.

In the sound source of FIG. 3A, the number of stages of the delay may be increased or a filter may be inserted in the operation loop for modification. As such a filter, for example, a low pass filter for simulating the loss at the tube end may be used.

FIG. 4 shows another embodiment of the present invention. In this embodiment, a chaotic sequence is used to AM-modulate a waveform generated in another system such as an FM sound source or an AWM sound source. The update time of the chaotic sequence (dynamic system) and the basic pitch of the original waveform may be selected from among matching, an integral multiple relationship, and a completely unrelated ratio. When a chaotic sequence is used for modulation, a signal consisting of a chaotic sequence is angular and is noisy, and thus a low-pass filter 5 is preferable to obtain a good result.

In FIG. 4, the delay circuit 1 and the chaos arithmetic circuit 3 are
By the loop operation, a chaotic sequence as shown by a in FIG. 5 is generated. The low-pass filter 5 is used to smooth the waveform of the chaotic sequence to eliminate noisiness. The multiplier 6 multiplies the chaotic waveform output of the low-pass filter 5 by the envelope waveform generated by the known envelope generator 7 to give fluctuations to the envelope. The adder 8 adds a constant to the envelope output of the multiplier 6 to prevent the envelope from taking a negative value. FIG. 5b shows the envelope-added waveform output from the adder 8.

The original waveform generator 9 is for generating an original waveform (shown by c in FIG. 5; here, it is a simple rectangular wave for simplification), and uses an FM sound source or a known sound source of AWM sound source. be able to. The multiplier 10 multiplies the envelope imparting waveform shown by b in FIG. 5 and the original waveform shown by c in FIG.
Add an envelope to the original waveform. As a result, a musical tone waveform output having a fluctuation effect on the amplitude is obtained as shown by d in FIG.

FIG. 6 shows an example in which a chaotic sequence is used for the FM / PM sound source. Also in this example, unsteady fluctuation due to chaos can be added to the pitch and the waveform.

In the figure, the chaos generating circuits 61 and 62 are assumed to be composed of a low-pass filter and a dynamic system composed of the delay circuit 1 and the chaos arithmetic circuit 3 as shown in FIG. The operators 63 and 64 forming the FM sound source include a sine wave generator and a multiplier. Further, the update period of the dynamical system is preferably related to the basic pitch by the phase generators (PG) 65 and 66 as in the sound source of FIG. The sampling pulse generator 67 is for generating the clock φ.

[Modifications of the Embodiment] The present invention is not limited to the above-described embodiments, and can be implemented by being modified appropriately.

For example, there are various classes of systems that generate chaos in addition to those described by differential equations such as equations (1) and (2). Is applicable to. In this case, it is simulated by an analog circuit, which can be applied to this system by using various standard methods for solving the differential equation digitally (software).
The simplest method is to replace the derivative with Euler difference.

[Brief description of drawings]

Fig. 1 is a graph that traces the state of chaos on a figure, and Fig. 2 is a graph in which the initial value is 0.01 and a = 3.6, and the chaos sequence x _n is plotted with n as the horizontal axis. FIG. 3A is a configuration diagram of a physical model sound source according to an embodiment of the present invention, FIG. 3B is a block diagram showing a specific example of the phase generator in FIG. 3A, and FIG. FIG. 5 is a block diagram of a musical tone synthesizing apparatus according to another embodiment of the present invention, FIG. 5 is a structural diagram of a musical tone synthesizing apparatus according to another embodiment of the present invention, FIG. FIG. 8 is a configuration diagram of an FM / PM sound source according to still another embodiment of the present invention. 1: Delay circuit 2,6,10: Multiplier 3: Chaos operation circuit 4,65,66: Phase generator 7: Envelope generator 9: Original waveform generator 61,62: Chaos generation circuit 63,64: FM operator.

Claims (4)

(57) [Claims] 1. An initial value is set at the start of generation of a musical sound,
One sample signal storage means for storing one input sample signal according to a timing signal generated based on predetermined frequency information; and a signal stored in the one sample signal storage means as an input, And a non-linear conversion means for converting the converted signal according to a predetermined non-linear function characteristic and inputting the converted one sample signal to the one-sample storage means, and sequentially stores one by one according to the timing signal. A musical tone synthesizing apparatus comprising: a waveform generator that outputs a sequence of converted and converted sample signals as a waveform data sequence. 2. The musical tone synthesizer according to claim 1, wherein a musical tone is modulated by a waveform data string generated by the waveform generator. 3. The musical tone synthesizing apparatus according to claim 1, wherein said waveform generator stores and converts said sample signal at a period proportional to the pitch of a musical tone to be generated and outputs it as the musical tone waveform data string itself. 4. A musical tone synthesizer according to claim 1, wherein said sample signal output from said waveform generator comprises an FM sound source given as a calculation parameter.