JPH11161273A - Musical sound generating device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate another impulse response signal different from an impulse response signal before synthesized, by synthesizing plural impulse response signals. SOLUTION: An impulse response signal ISj(t) of each channel is multiplied with a weighted data WT by a multiplier 89 channel by channel, and is accumulation-synthesized (s) piece by (s) piece by an adder 84 and an accumulation memory 85 to be output as one musical sound from an output memory 86. The impulse response signal ISj(t) having a characteristic provided by averaging frequency characteristics of the impulse response signals before synthesis is synthesized thereby.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a tone generator, and more particularly, to a tone generator using an impulse response signal.

[0002]

2. Description of the Related Art Hitherto, in the field of musical tone generating devices, there has been almost no device for controlling an impulse response waveform signal. However, in some cases, the impulse response waveform signal itself is sampled and stored, and this is repeated and read out, and the repetition cycle is made in accordance with the designated pitch. In this case, if the designated pitch is high, the read repetition cycle is short, and if the designated pitch is low, the read repetition cycle is long. The reading speed of the impulse response waveform signal itself was constant, and the reading speed of the impulse response waveform signal itself did not change even if the pitch was high or low.

[0003]

However, in such a device, the impulse response waveform signal is simply read out and applied to a musical tone, and the contents such as the tone color of the musical tone cannot be changed in various ways. To change the tone color or the like of a normal tone, the frequency component (characteristic) of the tone is changed using a filter or the like.

However, the frequency characteristics (spectral envelope, frequency spectrum component, formant shape) of a musical tone can be achieved only by changing the reading speed (generation speed) of the impulse response waveform signal itself. If the generation speed of the impulse response waveform signal is changed, the frequency characteristic expands or contracts on the frequency axis to change the frequency characteristic, and as a result, the timbre also changes. In this case, if the reading speed of the impulse response waveform signal itself is linked with the above-described repetition period of the reading, the tone cannot be determined independently, but the tone is determined according to the pitch, and is dependent on the pitch. You will end up with a toned sound.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and a first object of the present invention is to determine an independent timbre independent of a pitch, and a second object is to generate an impulse response waveform signal. To generate a novel tone signal using the same.

[0006]

In order to achieve the above object, according to the present invention, an impulse response signal having a predetermined length corresponding to the frequency characteristic of a tone signal to be generated is repeatedly generated,
The repetition period of the repetitively generated impulse response signal is changed according to a pitch determining factor, and the generation speed of the generated impulse response signal itself is determined according to a tone determining factor different from the pitch determining factor. Was changed independently of the repetition period.

[0007] Thus, the timbre of the tone signal generated from the impulse response signal is changed independently of the pitch, and the timbre can be freely changed without depending on the pitch. In addition, a desired tone signal is generated by repeatedly generating the impulse response signal, and a novel tone can be generated.

In the present invention, a plurality of impulse response signals different from the impulse response signal before being synthesized are generated by synthesizing a plurality of impulse response signals whose repetition period and generation speed are determined. Thus, more impulse response signals can be generated with less impulse response signals.

Further, in the present invention, the waveform shape of the repeatedly generated impulse response signal is switched according to the generated musical factor. This allows the change in the waveform shape of the impulse response signal itself to be linked to the pitch change,
It can be linked to a tone change.

In the present invention, if the repetition period or readout speed of the impulse response signal changes, the power, energy or volume of the tone signal output based on the impulse response signal also changes. In the present invention, the magnitude of the generated impulse response signal changes based on the repetition period or readout speed of the impulse response signal. As a result, the power, energy, or volume of the output tone signal is kept constant, and does not change according to the repetition period or readout speed of the impulse response signal.

Further, according to the present invention, an impulse response signal having a predetermined length corresponding to the frequency characteristic of the musical tone signal to be generated is repeatedly generated, and some of the repeatedly generated impulse response signals are inverted. As a result, the frequency characteristics of the tone signal change, and, for example, a specific overtone is eliminated / attenuated from the frequency characteristics of the tone, thereby generating a tone having only the remaining harmonic components or a tone having a relatively strong remaining harmonic component. can do.

Further, in the present invention, a fixed formant type and a moving formant type tone are selectively generated with the generation speed of the impulse response signal itself being constant with respect to a change in pitch. That is, based on the pitch determining factor, whether the repetition period of the impulse response signal is changed and whether the generation speed of the impulse response signal itself is changed are selected. This makes it possible to generate, from one impulse response signal, a fixed formant-type musical sound in which the formant of the frequency characteristic does not move on the frequency axis and a moving formant-type musical sound in which the formant shifts on the frequency axis.

Further, according to the present invention, the formant has realized a change which does not respond to the pitch change by setting the generation speed of the impulse response signal itself to a change characteristic different from the pitch change. That is, the repetition period of the impulse response signal is changed based on the pitch determining factor, and the generation speed of the impulse response signal itself is changed with characteristics different from the change in the repetition period. As a result, the change in formant movement can be changed according to the change in pitch, and the change in formant movement can be changed differently from the change in pitch of the generated musical tone. Timbre change can be realized.

In the present invention, based on musical factors,
The amount by which the end of the impulse response signal is not generated was changed.
Further, the amount by which the ends of the repeatedly generated impulse response signals overlap was detected, and the amount by which the ends of the impulse response signals were not generated was determined according to the detected amount.
As a result, the number of signals that do not overlap or overlap each other at the end of each impulse response signal is reduced, and the amount of calculation and processing of information on the impulse response signal is reduced accordingly.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION Principle of the Present Invention FIGS. 12 and 13 show the principle of the present invention. In FIG. 12, the repetition period of the impulse response signal ISj (t) is T1.
Or at T2, this impulse response signal ISj
The frequency characteristics (spectrum envelope (envelope), frequency spectrum component, formant shape) (hereinafter the same) of the musical tone synthesized and generated by the output of (t) are shown. This repetition period T determines the pitch of the generated musical tone.
Therefore, the repetition period T is determined by the pitch information, and is taken in as the period coefficient f. In the impulse response signal ISj (t), the period coefficient f is pitch information.

When the repetition period T of the impulse response signal ISj (t) is shortened and the pitch is increased, FIG.
As shown in (B1) and (B2), the frequency characteristics do not change much or at all, the density of each frequency component decreases, and the width of the formants does not change. Therefore, the repetition period T
Changes, the pitch changes, but the timbre (frequency component,
The formant characteristic itself does not change.

This impulse response signal ISj (t) is stored for each musical factor in an impulse signal storage section 51 (FIG. 2) in the impulse signal generation section 50 of FIG. This stored impulse response signal ISj
(T) is one of the two impulse response signals ISj (t) of FIG. 13 (A1) or (A1), and FIG. 13 shows the state of repeated reading.

FIG. 13 shows an impulse response signal ISj.
(T) has the same repetition period and the impulse response signal I
When the reading speed of Sj (t) is different, the frequency characteristics (spectrum envelope (envelope), frequency spectrum component, formant shape) of the musical tone synthesized and generated by the output of the impulse response signal ISj (t) are shown. The reading speed of the impulse response signal ISj (t) itself determines the timbre (frequency component, formant characteristics) of the generated musical tone. Therefore, the reading speed is determined by musical factor information that is not related to the pitch, for example, timbre information, and is taken in as the envelope coefficient r.
In the impulse response signal ISj (t), the envelope coefficient r is timbre information.

The reading speed of the impulse response signal ISj (t) is increased, and the impulse response signal ISj (t) is increased.
When the time length of (t) is shortened, FIG. 13 (B1) and (B2)
As shown in (2), the density of each frequency component does not change much or at all, the frequency characteristics (spectral envelope, frequency spectrum component, formant shape) change, and the width of the formant expands. Therefore, when the signal reading speed changes, the timbre (frequency component, formant characteristics) changes, but the pitch itself does not change.

This impulse response signal ISj (t) is
For a predetermined length of the tone signal to be generated, a linear prediction method,
It is generated and stored by applying a calculation method such as the cepstrum method. For example, in the cepstrum method, a tone signal having a predetermined length is subjected to a fast Fourier transform by a Fourier transformer and converted into a frequency power spectrum.

The converted power spectrum is logarithmically converted by a logarithmic converter, and further subjected to inverse fast Fourier transform by an inverse Fourier transformer to be converted into a cepstral in a time domain (Kefrange). This cepstrum is causally windowed from the multiplier and window function generator and converted to a complex cepstrum. This complex cepstrum is again subjected to fast Fourier transform by the Fourier transformer and returned to the frequency domain, and the spectrum envelope is obtained.

FIG. 12 shows an example of this spectrum envelope.
(B1) and (B2) and FIGS. 13 (B1) and (B2). This spectrum envelope is exponentially transformed by an exponential transformer, and is inversely fast-Fourier-transformed by the above-mentioned inverse Fourier transformer and returned to the time domain. As a result, sampling data of the impulse response signal of the minimum phase is generated, and this is stored in the impulse signal storage unit 51 as the impulse response signal ISj (t).

If the repetition period of the impulse response signal ISj (t) changes, this impulse response signal ISj
The power, energy or volume of the tone signal output based on (t) also changes. This impulse response signal I
The magnitude of the generated impulse response signal ISj (t) changes based on the repetition period of Sj (t). As a result, the power, energy or volume of the output tone signal is made constant, and the impulse response signal ISj
It does not change according to the repetition period of (t).

2. 1. Overall Circuit FIG. 1 shows an overall circuit of the musical sound generation device. The performance information generation section 10 generates performance information (tone generation information). The performance information (tone generation information) is information for generating a tone. The performance information generating unit 10 includes a sounding instruction device, an automatic performance device,
Various switches or interfaces.

The performance information (musical tone generation information) is musical factor (factor) information, pitch (tone range) information (pitch determining factor), sounding time information, performance field information, number of sounds, resonance degree. Information. The pronunciation time information indicates the elapsed time from the start of the tone generation. The performance field information indicates performance part information, musical tone part information, instrument part information, and the like, and corresponds to, for example, a melody, accompaniment, chord, bass, rhythm, or the like, or an upper keyboard, a lower keyboard, a foot keyboard, and the like.

The pitch information is taken in as key number data KN. The key number data KN includes octave data (sound range data) and note name data. The performance field information is fetched as part number data PN. The part number data PN is data for identifying each performance area, and is set according to the performance area from which the musical tone for which the sounding operation has been performed originates.

The sounding time information is the tone time data TM
And are based on time count data from key-on events or substituted in the envelope phase. This pronunciation time information is in Japanese Patent Application No. 6-2193.
This is shown in detail in the description and drawings of No. 24 as information on the elapsed time from the start of sounding.

The number-of-sounds information indicates the number of tones that are sounding simultaneously. For example, based on the number of tones whose on / off data of the assignment memory 42 is "1", this number is based on Japanese Patent Application No. Hei.
9 and FIG. 15 of Japanese Patent Application No. -242878, Japanese Patent Application No. 6-247.
8 and 18 of Japanese Patent Application No. 6855, Japanese Patent Application No. 6-276857.
9 and FIG. 20 of Japanese Patent Application No. 6-276858 of Japanese Patent Application No. 6-276858.
And the flowchart of FIG.

The resonance degree information indicates the degree of resonance between one musical tone being produced and another musical tone. The pitch frequency of this one tone and the pitch frequency of the other tone are 1: 2, 2: 3, 3:
If the ratio is a small integer multiple such as 4, 4: 5, 5: 6, the value of the resonance degree information is large, and 9: 8, 15: 8, 15:16,
If the ratio is a large integer multiple such as 45:32 or 64:45, the value of the resonance degree information becomes small. This resonance degree information is read from the resonance correlation table 53 or the resonance ratio table 54 in FIG. 7 of Japanese Patent Application No. 1-314818.

The pronunciation instruction device is a keyboard instrument, a string instrument, a wind instrument, a percussion instrument, a computer keyboard, or the like. The automatic performance device automatically reproduces stored performance information. The interface is MIDI
(Musical instrument digital interface) or the like, which receives and sends out performance information from a connected device.

Further, the performance information generating section 10 is provided with various switches. The various switches are a tone tablet, an effect switch, a rhythm switch, a pedal, a wheel, a lever, a dial, a handle, a touch switch, and the like. belongs to. These various switches generate tone control information. The tone control information is information for controlling the generated tone, which is musical factor (factor) information, tone color information (tone color determining factor), touch information (sound generation). Speed / strength of instruction operation), number of pronunciation information, resonance degree information, effect information, rhythm information, sound image (stereo)
Information, quantization information, modulation information, tempo information, volume information, envelope information, and the like.

The musical factor information is also combined with the performance information (tone information) and input from the various switches, combined with the automatic performance information, or combined with the performance information transmitted and received by the interface. . The touch switches are provided for each of the sounding instruction devices, and generate initial touch data and after touch data indicating the speed and intensity of the touch.

The timbre information includes a keyboard instrument (such as a piano),
It corresponds to the type of musical instrument (producing medium / producing means) of wind instruments (such as flutes), string instruments (such as violins), and percussion instruments (such as drums), and is taken in as tone number data. The envelope information includes an envelope level, an envelope speed, an envelope phase, and the like.

Such musical factor information is sent to the controller 20, where various signals, data, and parameters, which will be described later, are switched to determine the content of the musical sound. The performance information (tone generating information) and tone control information are processed by the controller 20, various data are sent to the impulse signal generating section 50, and an impulse response signal ISj (t) is generated. The controller 20 includes a CPU, ROM, and R
AM etc.

The program / data storage unit 30 (internal storage medium / means) comprises a storage device such as a ROM or a writable RAM, a flash memory or an EEPROM, and an information storage unit 41 such as an optical disk or a magnetic disk.
The computer program stored in the (external storage medium / means) is copied and stored (installed / transferred). The program / data storage unit 30 also stores (installs / transfers) a program transmitted from an external electronic musical instrument or a computer via the MIDI device or the transmitting / receiving device. The storage medium of the program includes a communication medium.

This installation (transfer / copy) is automatically executed when the information storage unit 41 is set in the musical sound generating device or when the power of the musical sound generating device is turned on. Installed by The above-mentioned program is a program according to a flowchart described later for the controller 20 to perform various processes.

It is to be noted that another operating system, a system program (OS), and other programs are stored in the apparatus in advance, and the above-described program may be executed together with these OSs and other programs. This program only needs to be able to execute the processes and functions described in the claims together with another program or alone when installed and executed in the present apparatus (computer body).

Further, part or all of the program is stored and executed in one or more other devices other than the present device, and data to be processed is communicated between the present device and another device via communication means. The present invention may be executed by transmitting / receiving already processed data / programs and as a whole of this apparatus and another apparatus.

The program / data storage unit 30 includes:
The above-mentioned musical factor information, the above-mentioned various data and other various data are also stored. These various data include data necessary for time division processing, data for assignment to time division channels, and the like.

In the impulse signal generating section 50, an impulse response signal ISj (t) having a predetermined length is repeatedly generated, and sound output from the sound output section 60. The repetition period of the repetitively generated impulse response signal ISj (t) is changed in accordance with the pitch information, and is changed to the above-mentioned timbre information or musical factor information irrelevant to the pitch. Accordingly, the reading speed (generation speed) of the generated impulse response signal ISj (t) itself is changed, and the repetition period and the reading speed are changed independently of each other.

The impulse response signal ISj (t) has a predetermined finite length L corresponding to the spectrum envelope of the tone signal to be generated. The impulse signal generator 50 generates a plurality of tone signals simultaneously by time-division processing and generates polyphonic sounds.

From the timing generator 30, a timing control signal for synchronizing all the circuits of the tone generator is output to each circuit. The timing control signal includes a clock signal of each cycle, a signal obtained by ANDing or ORing these clock signals, a signal having a cycle of the channel division time of the time division processing, channel number data CHNo, j, and the like.

3. Assignment Memory 42 FIG. 2 shows the assignment memory 42 of the program / data storage unit 40. In the assignment memory 42,
A plurality of (16, 32 or 64, etc.) channel memory areas are formed, and the impulse signal generating section 50 is provided.
The data relating to the musical sounds assigned to the plurality of musical sound generation channels formed in the storage area are stored.

In each of these channel memory areas, a periodic coefficient f (or key number data KN) and an envelope coefficient r (tone number data T
N), waveform start address Sa, waveform end address Ea, and cycle end value Fmax, on / off data, touch data TC, tone time data TM, part number data PN, power data PW, end cut data ΔLa.
d, window width data Wad, window selection data WS, envelope phase data EF, envelope speed data E
S, envelope level data EL and the like are stored.

The on / off data indicates whether the assigned musical tone is being keyed on or sounding ("1"), keyed off or muted ("0"). The period coefficient f (key number data KN) indicates the pitch of a musical tone assigned and produced, and is determined according to the pitch information. The upper data of the period coefficient f indicates the range or octave. The frequency number data FN is supplied from the impulse signal generator 5 to the corresponding channel timing by the controller 20.
Sent to 0.

Envelope coefficient r (tone number data TN)
Indicates the timbre of the musical tone assigned and pronounced, and is determined according to the timbre information. The envelope coefficient r is sent by the controller 20 to the impulse signal generator 50 at the corresponding channel timing. The touch data TC indicates the speed or strength of the sounding operation, and is determined according to the touch information.

The power data PW indicates the power of the energy of the waveform of the impulse response signal ISj (t), the repetition period T of each impulse response signal ISj (t) stored in the impulse signal storage unit 51, It is determined based on the reading (generation) speed or waveform shape of ISj (t) itself.

For example, if the power data PW of an impulse response signal ISj (t) having a certain pitch is set as a reference to "1", the impulse response signal ISj (t) having the same pitch one octave higher becomes "1". “１ ／” for the impulse response signal ISj (t) having the same pitch two octaves above, “「 ”for the impulse response signal ISj (t) having the same pitch one octave below. For an impulse response signal ISj (t) having the same pitch two octaves below, the result is “4”,.

If the power data PW of the impulse response signal ISj (t) at a certain read speed is set as a reference to "1", "2" is obtained for the same impulse response signal ISj (t) at twice the read speed. For the same impulse response signal ISj (t) with a reading speed three times faster, "3",
.., “「 ”For the impulse response signal ISj (t) having the same reading speed of １ ／, and“ 1 / ”for the same impulse response signal ISj (t) having the reading speed of ３.
3 ", ...

Further, different impulse response signals ISj
Even during (t), the power data PW is set. The impulse response signal ISj (t) has a waveform that changes drastically. If the energy power of the waveform is large, the power data PW is reduced, and the impulse response signal ISj (t) is reduced.
Is a waveform that changes slowly. If the energy power is small, the power data PW is increased.

Of course, if the magnitude (amplitude) of each impulse response signal ISj (t) stored in the impulse signal storage unit 51 is adjusted to have the same energy power, such different impulse response signals ISj
There is no need to set the power data PW during (t).

The power data PW indicates, for different impulse response signals ISj (t), the integrated value of the waveform of the impulse response signal ISj (t) or the sum of the absolute values of the respective maximum values and minimum values of the waveform. Determined based on

As described above, the power data PW is determined in inverse proportion to the repetition period T, that is, the pitch frequency, between the same impulse response signals ISj (t), and is further inversely proportional to the set tone color, that is, the reading speed. Is determined. However, depending on the impulse characteristics of the tone generation (sound generation) circuit and the conversion characteristics from the electric energy of the speaker to the volume energy, the following is obtained. A, B,
C and D are constants. T is the above repetition period. If it is a pitch frequency, it is reciprocal and substituted into "T". S is the reading speed.

PW = A × (T / S) (n-INV)
(N = 1, 2, 3,...) PW = B × (log−c) (T / S) PW = D × (n−RAD) (T / S) (n = 1, 2,
(3,...)) (N-INV) indicates that the preceding numerical value is raised to the “n power”, and in this case, (T / S) is raised to the “n power”. (Log-
c) shows a logarithmic formula with a base "c". (N-RAD) indicates that the "n-th root" of the following numerical value is obtained. In this case, the "n-th root" of (T / S) is calculated.

The power data PW is converted from the pitch information (key number data KN) and timbre information (tone number data TN) by a table in the program / data storage section 40. Similarly, the pitch coefficient f is also converted from the pitch information (key number data KN) by a table in the program / data storage unit 40, and the envelope coefficient r is also converted to the program / data storage unit 4.
The tone color information (tone number data TN) is converted by the table in 0.

The edge cut data ΔLad is shown in FIG.
As shown in 2), the impulse response signal ISj
(T) shows the amount of addresses that are cut off, deleted, and not read (not generated) at both ends. As a result, when the impulse response signals ISj (t) overlap, the signals do not overlap at the end of each impulse response signal ISj (t), or the number of overlapping signals decreases, and the impulse response signal ISj ( The amount of calculation and the amount of processing of the information on t) can be reduced.

The edge cut data ΔLad is detected and obtained as follows. Here, T indicates a repetition period of the impulse response signal ISj (t), f indicates a period coefficient for determining a pitch, Fmax indicates a period end value, Sa indicates a waveform head address, and the read impulse response signal ISj (T) indicates the start address, and Ea
Is the end address of the waveform, indicating the end address of the read impulse response signal ISj (t), r is the envelope coefficient that determines the tone color, and Lad is the address length of the length L of the read impulse response signal ISj (t) itself. Lt is the impulse response signal ISj to be read
(T) indicates the time length of the length L itself, ΔLt indicates the time length cut off at both ends of the impulse response signal ISj (t), and m indicates the means for alternately generating the impulse response signal ISj (t). In this embodiment, the numbers correspond to the numbers of the lower read address data R1 and R2.
"Two" as shown in FIG.

T = Fmax / f Lad = (Ea−Sa) Lt = Lad / r = (Ea−Sa) / r ΔLt = (m × Lt−T) / 2m ΔLad = Lad × ΔLt / Lt = (Ea− Sa) × Δ
Lt / Lt Here, the amount of overlap of each end of the impulse response signal ISj (t) and the cutout amount of the end of the impulse response signal ISj (t) match, whereby the respective impulse response signals ISj (t) Is connected at each end.

If ΔLad or ΔLt is made larger than the obtained value, a gap is formed between the impulse response signals ISj (t), and the impulse response signals ISj ( t)
The amount of cut off at the end of is larger. If ΔLad or ΔLt is smaller than the obtained value, the intervals between the impulse response signals ISj (t) overlap, and the amount of the impulse response signals ISj (t) overlaps with the amount of overlap of each end of the impulse response signals ISj (t). The amount of edge cut is smaller. Such a change is also possible.

M indicates the number of means for generating the impulse response signal ISj (t). If m = 1, the overlap with the end of the next impulse response signal ISj (t) will be the end cut data ΔL.
ad, and if m = 2, the overlap with the end of every other impulse response signal ISj (t) is the end cut data ΔL
ad, and if m = 3, the overlap with the end of every third impulse response signal ISj (t) is the end cut data ΔL.
ad, and if m = 4, the overlap with the end of every third impulse response signal ISj (t) is the end cut data ΔL.
ad, and so on.

The end cut data .DELTA.Lad is obtained from the periodic coefficient f, the envelope coefficient r, and the impulse response signal ISj.
It is detected and obtained from the address length Lad = (Ea-Sa) of (t). Since the period coefficient f is determined from the pitch information or other musical factor information, the edge cut data ΔLad changes according to the pitch or the like. Further, since the envelope coefficient r is determined from the timbre information or other musical factor information, the edge cut data ΔLad changes according to the timbre or the like.

The window width data Wad indicates the address amount of the window function value WF combined with the impulse response signal ISj (t) whose both ends have been cut off. This window function value WF is stored in the window function memory 71, and is arithmetically combined with the remaining both ends or the entirety of the impulse response signal ISj (t) whose both ends have been cut off.

A plurality of types of window function values WF stored in the window function memory 71 are selected by the window selection data WS. The types of this window function are shown in FIG. 26 (B1) (B2) (B
3) Shown in (B4). The window function value WF is weight data that changes from 0 times to 1 times, or changes from 1/2 times to 1 times, and changes from decreasing times to 1 times.

The envelope phase data EF indicates the attack, decay, sustain or release of the envelope, the envelope speed data ES indicates the step value of the operation per one cycle of the digital operation of the envelope, and the envelope level data EL indicates the end of each phase. Indicates a target value at which the envelope calculation value reaches.

Each impulse response signal ISj (t) generated by the tone data of each channel of the assignment memory 42 is synthesized and output as one tone for each s number, or output as one tone for each channel. Or be done. For such s impulse response signals ISj (t) synthesized as one musical tone,
The musical tone data is allocated at the same timing for every s consecutive channels, the generation timing is the same, and the phase is the same.

The combination of the s number of impulse response signals ISj (t) to be synthesized is determined by the musical factor information, for example, timbre information, touch information, pitch information, sounding time information, sounding number information, and resonance degree information. Selection is decided. A decoder or table for this is provided in the program / data storage unit 40.

The weighting data WT is obtained by synthesizing each impulse response signal ISj which is synthesized as s pieces as one musical tone.
The mixing ratio of (t) is determined and stored for each channel. The weighting data WT takes a value between “0” and “1”, and the total value of the s pieces of weighting data WT is “1”, but may be any other value. The weighting data WT is selected and determined based on the above-described musical factor information such as pitch information, timbre information, touch information, sounding time information, sounding number information, and resonance degree information. A decoder or table for this is provided in the program / data storage unit 40.

When the number of composites s = 2, the channel number for each s number is from 0 to 1, 2 to 3, 4 to 5, 6 to 7,.
.., when the number of composites s = 3, 0 to 2, 3 to 5, 6 to 8,
9 to 11,..., When the composite number s = 4, 0 to 3, 4 to
7, 8 to 11, 12 to 15,..., When the composite number s = 5, 0 to 4, 5 to 9, 10 to 14, 15 to 19,.
Becomes The synthesis number s is input by the operator, automatically set by the initial setting of the apparatus, or determined by the musical factor (tone color information), and stored in the register of the program / data storage unit 40. .

The s number of impulse response signals ISj
The waveform head address Sa, the upper read address data, or the envelope coefficient r in (t) are different from each other, have different waveform shapes or timbres, and have different frequency characteristics (spectral envelope, frequency spectrum component, formant shape). . Of course, it may be the same.

This s number of impulse response signals ISj
The period coefficient f of (t) is the same or an integer multiple, and the phases of the s impulse response signals ISj (t) are the same, and the pitch is the same or an integer multiple. Of course, a non-integer multiple may be used.

The s-number of impulse response signals ISj
(T) tone time data TM, touch data TC,
Power data PW, part number data PN, waveform end address Ea, cycle end value Fmax, thinning flag, thinning order n, edge cut data ΔLad, window selection data W
S, envelope speed ES, envelope level E
L and the envelope phase EF are the same, and may be stored collectively on behalf of each of the s channels. Of course, these s pieces of data may be different from each other.

If the envelope speed ES, the envelope level EL, and the envelope phase EF are shared for the s synthesized impulse response signals ISj (t), the number of time division channels formed by the envelope generator 56 is , And the synthesized accumulated impulse response signal AISj (t) is multiplied and synthesized by one envelope data EN.

For each of the s synthesized impulse response signals ISj (t), the envelope speed ES, the envelope level EL, and the envelope phase EF are different. These data ES, EL are different for each impulse response signal ISj (t). , EF are individually stored, the composition ratio or the mixture ratio of the respective impulse response signals ISj (t) to be composed, that is, the weighting ratio, changes with the passage of the sounding time. In this case, each envelope level data EL plays the role of the combined weight data WT.

The data in each of these channel memory areas is written at the tone generation start timing, rewritten or read out at each channel timing, and sent to the impulse signal generator 50. The assignment memory 42 may be provided not in the program / data storage unit 40 but in the impulse signal generation unit 50 or the controller 20.

A method of allocating or truncating each tone to a plurality of tone generating systems for generating a plurality of tone in parallel, ie, a channel formed by the time division processing, is described in, for example, Japanese Patent Application No. Hei. ,
Japanese Patent Application No. 1-305818, Japanese Patent Application No. 1-312175
No., Japanese Patent Application No. 2-2098978, Japanese Patent Application No. 2-4095
No. 77 and Japanese Patent Application No. 2-409578 are used.

4. FIG. 3 shows the impulse signal generating section 50 described above. Various impulse response signals ISj (t) are stored in the impulse signal storage unit 51, as shown in FIG. The waveform shapes of these impulse response signals ISj (t) are different, and the timbre when synthesized and output as a tone signal is also different. These impulse response signals ISj (t) are
It corresponds to each of the above musical factor information,
Touch, pitch (sound range) or sounding time, performance field, envelope phase, sound count information, resonance degree information, and the like are multiplexed and stored.

For example, the impulse response signal ISj (t)
Is stored differently for each timbre, of which the impulse response signal ISj (t) of one type of timbre is stored differently for each touch, and the impulse response signal ISj (t) of one type of touch is Stored differently for each sounding time,... Stored differently for each pitch,... Stored differently for each performance field,... Stored differently for each envelope phase.

The musical factor information is stored in the controller 2
0, such various impulse response signals ISj (t)
, And the read address data is stored in the impulse signal selection section 52 and supplied to the impulse signal storage section 51 to select the impulse response signal ISj (t). If the musical factor information changes for each musical tone or during tone generation, the upper read address data is switched, and the read impulse response signal ISj
(T) is also switched.

If the musical factor information is related to the pitch, the change in the waveform shape of the impulse response signal itself is linked to the change in the pitch, and the tone color of the synthesized tone signal is also changed in the pitch. Can be linked. If the musical factor information is related to information other than the pitch, the change in the waveform shape of the impulse response signal itself is linked to the change in the timbre, touch, sounding time, performance field, or synthesized output. The tone of the tone signal
It can be linked to changes in touch, pronunciation time, and performance field.

The impulse signal selecting section 52 has a memory area corresponding to the number of time-division channels, and upper read address data corresponding to a musical factor of a musical tone assigned to each channel is stored in each memory area. , Are read out by the channel number data j from the timing generator and supplied to the impulse signal storage 51.

Each of the impulse response signals ISj (t) in the impulse signal storage unit 51 is read by the lower read address data R1 and R2 from the impulse signal read unit 53. This lower read address data R
1, R2 is incremented at a speed corresponding to the musical factor information other than the pitch information. Therefore, the reading speed of the impulse response signal ISj (t) is not determined by the pitch information. The impulse response signal ISj (t) is repeatedly read out, and the cycle of the repetition is changed according to the pitch information, and the repetition cycle and the readout speed are determined independently of each other.

Each impulse response signal ISj (t) read from the impulse signal storage section 51 is individually accumulated and synthesized for each channel by an impulse accumulation section 54, and a multiplier 55 outputs a signal from an envelope generator 56. The envelope signal is multiplied and synthesized for each channel, and the tone accumulating section 57 accumulates and synthesizes the tone signals of all channels, and outputs the sound from the sound output section 60.

The envelope generator 56 has the above-mentioned musical factor information (envelope information), that is, envelope speed data ES and envelope level data E
L, envelope phase data EF
0, stored for each channel, and based on this musical factor information (envelope information),
The speed and level of each phase of the envelope of each channel are set, and the shape of each envelope is determined. This envelope signal is generated in a time-division manner for each channel and sent to the multiplier 55.

The power data PW of each channel is
The controller 20 controls the waveform power control RAM 58
Is stored in the corresponding channel area. Each of the power data PW is sent to the multiplier 55, and the tone signal (the impulse response signal ISj) from the impulse accumulating section 54
(T)) or the envelope data from the envelope generator 56 is multiplied. The waveform power control RAM 58 has a memory area corresponding to the time-division channel, and each power data PW is read in a time-division manner. This switching is performed by the timing generator 30.
Based on channel count data from.

Thus, the impulse response signal ISj
Even if the repetition period T of (t) changes according to the pitch, and the power of the waveform, that is, the volume, changes according to the pitch, this can be adjusted and eliminated. In addition, since the waveform shape of the impulse response signal ISj (t) is different, even if the power of the waveform, that is, the volume changes according to the pitch, it can be adjusted and eliminated.

FIG. 5 shows this impulse response signal ISj.
The relationship between the pitch at (t), that is, the repetition period T and the waveform power is shown. When the impulse response signal ISj (t) is read at a short cycle as shown in FIG. 5A, the waveform energy per unit time increases and the sound volume also increases. On the other hand, when the impulse response signal ISj (t) is read at a long cycle as shown in FIG. 5A, the waveform energy per unit time is reduced and the sound volume is also reduced.

Similarly, although not shown, when the impulse response signal ISj (t) is read out slowly, the waveform energy per unit time increases and the sound volume also increases.
On the other hand, when the impulse response signal ISj (t) is read out quickly, the waveform energy per unit time decreases and the sound volume also decreases.

On the other hand, by multiplying the power data PW, the output impulse response signal I
By controlling the level of Sj (t) (tone signal), it is possible to suppress an extra volume change (power) change due to a repetition period T, ie, a change in pitch or a readout speed, ie, a change in timbre.

The power data PW may be multiplied by the envelope level data EL stored in the assignment memory 42 and sent to the envelope generator 56. Further, the touch data TC stored in the assignment 41 is multiplied by the power data PW to obtain the waveform power control RAM.
58 may be stored. Further, the power data PW may be obtained by calculating the reciprocal of the periodic coefficient f and the envelope coefficient r.

5. FIG. 6 shows the impulse signal reading unit 53 in the impulse signal generating unit 50. The above controller (CP
U) The period coefficient f, envelope coefficient r, waveform start address Sa and waveform end address Ea from 20 are stored in the parameter RAM 501 for each time-division channel. In some cases, the cycle end value Fmax is also stored in the parameter RAM 501 for each time division channel. A memory area corresponding to the number of time-division channels is formed in the parameter RAM 501, and the coefficients f, r, addresses Sa, and Ea corresponding to musical sounds assigned to each channel are stored in the corresponding memory area. If the data f, r, Sa, and Ea are always sent from the assignment 42 in a time-sharing manner, the parameter RAM 501 may have one memory area.

The cycle coefficient f and cycle end value Fmax
Determines the length of the repetition period T of the impulse response signal ISj (t), determines the pitch of the generated musical tone,
The cycle coefficient f is sequentially and repeatedly accumulated, and the cycle end value Fma
Each time x is reached, the impulse response signal ISj (t) is repeatedly read. The accumulated value of the periodic coefficient f becomes the periodic count value F. This cycle coefficient f and cycle end value F
max is determined by the pitch information (key number data) and is converted from the key number data. The length of the repetition period T determines the pitch of the tone signal generated.

Either the cycle coefficient f or the cycle end value Fmax may be fixed. In the embodiment of FIG. 3, the cycle end value Fmax is the maximum value that the cycle count value F can take (“1111... 11” or “111... 1100... 0”).
, And the cycle end value Fmax is not stored.

The envelope coefficient r determines the reading (generation) speed of the impulse response signal ISj (t), and determines the frequency characteristics (spectrum envelope, frequency spectrum component, formant shape) of the impulse response signal ISj (t). The tone coefficient of the generated tone is determined, and the envelope coefficient r is repeatedly accumulated from the waveform start address Sa to the waveform end address Ea, and is supplied to the impulse signal storage unit 51 as lower read address data R1 and R2. You. The lower read address data R1,
When R2 reaches the waveform end address Ea, until the cycle count value F reaches the cycle end value Fmax next time,
The reading of the impulse response signal ISj (t) is waited.

The envelope coefficient r is determined by the musical factor information that is not related to the pitch. For example, the timbre information (tone number data), touch information (touch data), sounding time information (tone time data), performance It is converted from field information (part number data), number of pronunciation information, resonance degree information, and the like.

The envelope coefficient r or the cycle count f from the parameter RAM 501 is supplied to the FLX via the B register.
With or without passing through 503, the data is accumulated in the lower-order read address data R 1, R 2 or the cycle count value F by the adder 505 through the selector 504, through the accumulation register 506, and through the selector 507. It is stored in the arithmetic RAM 508. Memory areas corresponding to the number of time-division channels are formed in the arithmetic RAM 508, and the data R1, R2, and F corresponding to musical tones assigned to each channel are stored in the corresponding memory areas. The FLX 503 converts floating-point data to fixed-point data.

The data R1, R2, F are stored in the arithmetic RAM 5.
08 to the adder 505 via the selector 510 via the A register 509. The data R
1 and R2 are alternately selected by a selector 518 via an R1 register 511 and an R2 register 512 and sent to the impulse signal storage unit 51. Lower decimal data Fr or data “0” of the cycle count value F from the accumulation register 506 is supplied to the adder 505 via the selector 510. The selector 518 is switched by the clock signal φR1. One cycle of the clock signal φR1 is equal to a division time for one channel as shown in FIG.

The initial value of the lower read address data R1 is "0", the initial value of the lower read address data R2 is the waveform end address Ea, and the initial value of the cycle count value F is "0". The operation RAM 50 is provided by the controller 20 via the selector 507.
8 is stored.

The waveform end address value Ea (cycle end value Fmax) from the parameter RAM 501 is supplied to a comparator 516 via an Ea register and an AND gate group 514. The comparator 516 has lower read address data R1 from the adder 505,
R2 or the cycle count value F is also supplied, and when the lower read address data R1, R2 reaches the waveform end address value Ea, or when the cycle count value F reaches the maximum value “111”.
.. 11 or the maximum approximate value "111... 1100... 0" (cycle end value Fmax), the detection signal is set in the flip-flop 517, and the controller (CPU)
Sent to 20.

The carry-out signal Cout from the adder 505 is supplied as an upper bit group of the comparator 516, and the gate signal of the AND gate group 514 is inverted by an inverter 515 and supplied as an upper bit group of the comparator 516. And the number of bits is matched.

The waveform start address Sa and the waveform end address Ea are also determined by the musical factor information. For example, the pitch information (key number data), timbre information (tone number data), touch information (touch data), It is converted from sounding time information (tone time data), performance field information (part number data), sounding number information, resonance degree information, and the like, and stored in the corresponding channel memory area of the parameter RAM 501. The waveform start address Sa and the waveform end address Ea select one of the various impulse response signals ISj (t). In this case, the impulse signal selection section 52 can be omitted.

6. Overall Process FIG. 7 is a flowchart of the overall process executed by the controller (CPU) 20. The whole process is started by turning on the power of the musical tone generating apparatus, and is repeatedly executed until the power is turned off.

First, various initialization processings such as initialization of the program / data storage section 40 are performed (step 01). Based on a manual performance or an automatic performance by the sounding instruction device or the automatic performance device of the performance information generating section 10, A sound generation process is performed (step 02).

In this tone generation process, a tone related to the key-on event is assigned to the searched empty channel. The contents of the musical tones include the performance information (tone generation information) from the performance information generating unit 10, the musical factor information of the musical tone control information, and the program / data storage unit 4 at this time.
0 is determined by the musical factor information already stored.

Next, a mute (attenuation) process based on a manual performance or an automatic performance in the sounding instruction device or the automatic performance device of the performance information generating section 10 is performed (step 0).
3). In this silencing (attenuation) processing, a channel to which a tone related to the key-off event is assigned is searched, and the tone is attenuated and muted. In this case, the envelope phase of the musical tone related to the key-off event is released, and the envelope level gradually becomes “0”.

Further, if various switches of the performance information generating section 10 are operated, musical factor information corresponding to the switches is fetched, stored in the program / data storage section 40, and the musical factor information is changed. (Step 04). Thereafter, other processes are executed (Step 05), and the processes from Step 02 to Step 05 are repeated.

7. FIG. 8 shows a flowchart of the sound generation process of step 02 performed by the controller (CPU) 20.
In this process, an impulse response signal ISj (t) is generated. The flowchart of FIG. 5 is performed for all time division channels.

First, when a key-on event (sound generation event) based on a manual performance or an automatic performance is sent from the sound generation instruction device or the automatic performance device of the performance information generation unit 10 to the controller 20 (step 11), an empty channel is searched. In the area of the assignment memory 42 of the searched empty channel, ON / OFF data of "1", a period coefficient f corresponding to a pitch, an envelope coefficient r corresponding to a timbre, touch data TC, part number data PN,
The tone time data TM of "0", the power data PW corresponding to the pitch, the envelope speed ES, the envelope level EL, the envelope phase EF of "1", and other flags and data Sa, Ea, Fmax to be described later,
n, ΔLad, and WS are written. The period coefficient f and the envelope coefficient r will be described below.

Further, the lower-order read address data R1 of "0", the lower-order read address data R2 set to the waveform end address Ea, and the cycle count value F of "0" are stored in the corresponding channel memory area of the arithmetic RAM 508. The corresponding channel area of the superimposed channel counter (program / data storage unit 40) is reset to "0" (step 12).

Next, the cycle coefficient f and the cycle end value Fmax
Is converted from the key number data (pitch information) KN and stored in the corresponding channel memory area of the parameter RAM 501, and the envelope coefficient r is the tone number data (tone color information) TN, touch data (touch information) TC, tone time data (Sounding time information) TM or part number data (performance field information) PN is converted and stored in the corresponding channel memory area of the parameter RAM 501 (step 13), and other processing is executed (step 14).

The waveform start address Sa and the waveform end address Ea include key number data (pitch information), tone number (tone color information), touch data (touch information),
The parameter R is converted from tone time data (toning time information) or part number data (performance field information).
The data is stored in the corresponding channel memory area of the AM 501 (step 13). The waveform start address Sa and the waveform end address Ea select one of the various impulse response signals ISj (t). In this case, the impulse signal selection section 52 can be omitted.

Further, if there is a musical tone at the start of key-on (start of sound generation) or the key-on (during sound generation) (step 1)
5) The cycle coefficient f is added (accumulated) to the cycle count value F (step 16), and the added value (accumulated value) is the cycle end value Fmax (“1111... 11” or “111.
1100... 0 ") or more (step 17), the cycle end value Fmax is subtracted from the cycle count value F, and the fraction of the cycle count value F is corrected (step 18).

The superimposed channel is switched (steps 18, 21, 22), the lower decimal data Fr of the period count value F is added to the switched waveform start address Sa (step 23), and the lower read is performed. The initial value of the address data R1 or R2 is corrected.

In this superimposed channel, two impulse response signals ISj (t) are alternately read out in a time-division manner and output as one musical tone. As described above, a plurality of musical tones are made polyphonic through additional time-division channels. Is output. The value of this superimposed channel (c
h = 0, 1), the two impulse response signals IS
Reading j (t) is distinguished.

The time length L for reading one impulse response signal ISj (t) depends on the impulse response signal Ij (t).
If it is longer than the time length T of the repetition period of Sj (t),
The previous impulse response signal ISj (t) and the next impulse response signal ISj (t) overlap. Therefore, if the two impulse response signals ISj (t) are individually read out by channel division, the two impulse response signals ISj (t) are read out and superimposed in parallel. Of course, this number of channels may exceed two.

Next, the lower read address data R1
Is added to the waveform end address Ea (steps 24, 25, 26), and is added to the lower read address data R2 until the envelope coefficient r becomes the waveform end address Ea (steps 27, 28, 29). ). As described above, the two lower read address data R1 and R2 allow the two impulse response signals ISj
(T) is an impulse response signal ISj (t) signal storage unit 5
1 and are superimposed in the impulse accumulator 54 in parallel. The above processing from step 15 to step 29 is repeated over all time division channels (step 30), and other processing is executed (step 3).
1).

8. FIG. 9 shows a time chart of the operation of each unit of the impulse signal reading unit 53. As described above, writing / reading to / from the parameter RAM 501 and the operation RAM 508, switching of the selectors 504, 507, 510, and registers 502, 5096, 509, 511, 5
12, store to 513, strike to flip-flop 517, and open / close of AND gate group 514 are controlled. As these switching control signals, various timing control signals from the above-described timing generator are used.

The above cycle end value Fmax is the parameter RA
When stored in M501, the cycle end value Fma
Writing / reading of x is also performed. Of the waveforms in this time chart, those whose high level / low level are indicated by dotted lines are set to a high level or a low level by detection / non-detection of the comparator 516 when write data is present / absent.

9. Reading State FIG. 10 shows a reading state of the impulse response signal ISj (t) from the impulse signal storage unit 51. When the sound starts (key-on), the first impulse response signal I
Reading of Sj (t) is started by the lower read address data R1 (step 24). This lower-order read address data R1 starts incrementing at the speed of the envelope coefficient r.

At the same time, the cycle count value F starts to be incremented at the speed of the cycle coefficient f (steps 15, 1).
6). When the cycle count value F reaches the cycle end value Fmax (step 17), even if the first impulse response signal ISj (t) is still being read, the second impulse response signal ISj (t) becomes lower. Reading is started by the read address data R2 (step 27). The lower read address data R2 also starts incrementing at the speed of the envelope coefficient r.

When the second period T has elapsed from the start of sound generation (step 17), even if the second impulse response signal ISj (t) is still being read out,
The reading of the third impulse response signal ISj (t) is started by the lower read address data R1 (step 24).

Further, when the third period T elapses from the start of sound generation (step 17), even if the third impulse response signal ISj (t) is still being read out,
Reading of the fourth impulse response signal ISj (t) is started by the lower read address data R2 (step 27).

As described above, two identical impulse response signals ISj (t) are alternately read by two lower read address data R1 and R2. Therefore, the length L of the impulse response signal ISj (t) itself
Is longer than the cycle T of repetition, the generation of the previous impulse response signal ISj (t) is continued at the end of each cycle T, and the next impulse response signal ISj (t)
Can be generated repeatedly. These impulse response signals ISj (t) are synthesized and output as one tone signal.

The number of impulse response signals ISj (t) read out in a time-division manner as one tone signal is "2".
May be exceeded. Accordingly, the number of lower-order read address data also increases to R1, R2, R3, R4,..., And the number of steps 24-26, 27-29 also increases.

The number of impulse response signals ISj (t) read out in a time-division manner as one tone signal is "1".
May be. In this case, the impulse response signal ISj (t)
The length L itself becomes shorter than the repetition period T. Therefore, it is determined whether or not the length L of the impulse response signal ISj (t) itself is shorter than the repetition period T. If it is shorter, the processing of steps 27 to 29 is omitted, and the time-division reading system of the impulse response signal ISj (t) becomes one. The length L of the impulse response signal ISj (t) itself is obtained by dividing the difference between the waveform end address Ea and the waveform start address Sa by an envelope coefficient r. Similarly, the repetition cycle T is obtained by dividing the cycle end value Fmax by the cycle count f.

10. Processing of Tone Time Data TM FIG. 11 is a flowchart of the interrupt processing executed by the controller 20 at regular intervals. In this process, the tone time data TM is incremented.

In this processing, each channel area of the assignment memory 42 (steps 41 and 4)
4, 45), for tone data whose on / off data is "1" and a tone is being produced (step 42), the tone time data TM is incremented by "+1" (step 43), and other periodic processing is performed. (Step 46). In this way, the elapsed sounding time of the musical tone of each channel is counted and stored, and is used as described above.

(11) FIG. 14 shows a second embodiment of the impulse signal generator 50. In this embodiment, an alternate inversion circuit 66 as shown in FIG. 14 is inserted between the impulse signal storage unit 51 and the impulse accumulator 54. In this embodiment, two channels are assigned to one tone in step 12 or 13 described above. The periodic coefficient f of one of the first channels is in accordance with the specified pitch, and the periodic coefficient of the other second channel is not in accordance with the specified pitch, but is the periodic coefficient of the specified pitch. It is set to a period coefficient nf (2f) that is n times (2 times) f.

The allocation timings of these two channels are the same, the generation timings of the impulse response signals ISj (t) of the two channels are also the same, and the phases are the same and the same. The power data PW is further set to (n-1) / n, (1/2) from the set value, or the envelope level data EL is further set to (n-1) / n, (1/2) from the set value. Value. Others are the same as the above embodiment.

The first impulse response signal ISj (t) read in the first channel is as shown in FIG. 16A1, and the first frequency characteristics (spectral envelope, frequency spectrum component, formant shape) are shown in FIG.
If 1), n times (2
The second impulse response signal ISj (t) having the frequency of (times) is as shown in FIG. 16A2, and the second frequency characteristic (spectral envelope, frequency spectrum component, formant shape) is as shown in FIG. 16B2. become. Here, “n = 2”. The interval between each frequency component of the second frequency characteristic of n = 2 times the frequency of FIG.
6 (B1) n of the interval of each frequency component of the first frequency characteristic
= 2 times, and the second frequency characteristic (see FIG. 16 (B
2)) corresponds only to the even-order harmonic of the first frequency characteristic (FIG. 16 (B1)).

Here, the first impulse response signal ISj
(T) From the waveform of (A1), the second impulse response signal IS
If 1 / n (1/2) of j (t) (FIG. 16A2) is subtracted and difference synthesis is performed, only even-order harmonics are subtracted, and a combined impulse response signal ISj (t) (FIG. 16A3) is generated. The synthesized frequency characteristic (FIG. 16B3) is only odd harmonics. The waveform of FIG. 16 (A3) shows that the impulse response signal ISj (t) is (n-1) out of n (2).
(1) The tone is output with its polarity inverted (every other alternate) and output according to the designated pitch.

Also, the decimation order n = 3, 4, 5, 6, 7,
.., Which is shown in FIG.
As shown in 4), (n-1) of the n signals are inverted and output, and the level of the non-inverted impulse response signal ISj (t) is set to (n-1) / n. The level of the impulse response signal ISj (t) is 1 / n
It is said.

The frequency characteristic of this tone signal is nth, 2n
, 3n, 4n, 5n, 6n, 7n
.., Only the frequency components of... Are eliminated / attenuated, and a musical tone having no specific overtone component can be generated. For example, if n = 2, only the second harmonic, the fourth harmonic, the sixth harmonic, the eighth harmonic,... Can be eliminated / attenuated, and this musical tone is suitable for the sound of a closed instrument. Further, for example, if n = 3, only the third harmonic, the sixth harmonic, the ninth harmonic, the twelfth harmonic,... Can be erased / attenuated. For example, if n = 7, the seventh harmonic, the 14th harmonic, the 21st harmonic, the 28th harmonic,...
Can be eliminated / attenuated, and this musical tone is suitable for piano sound.

The two impulse response signals ISj (t) successively read from the impulse signal storage section 51
Is sent directly to the impulse accumulator 54, multiplier 55, and tone accumulator 57 via the selector 63, the selector 59 and the OR gate group 61, and the other
3. Inverters are inverted by the inverter group 65 via the selector 59 and are multiplied by 1 / (n-1) and (1/2) by the multiplier 62, and are stored in the impulse accumulator 54, the multiplier 55 and the musical tone accumulator. It is sent to 57 and accumulated (added) and synthesized (difference synthesized).

The selector 59 is switched by the lower bit of the channel count data CHNo.
The first impulse response signal ISj of the first channel
(T) is output as it is, and the second channel of the second channel is output.
The impulse response signal ISj (t) is inverted in the positive / negative direction and 1 /
It is output as (n-1) times and (1/2 times).
A tone signal as shown in (B3) is generated and output. The frequency characteristic (spectrum envelope, frequency spectrum component, formant shape) of this tone signal consists of harmonics in which only integer harmonics of n have been eliminated / attenuated, and only odd odd harmonics if the decimated degree n = 2. The channel count data CHNo is sent from the timing generator 30.

The data of 1 / (n-1) and (1/2) are calculated by the controller 20, stored in the register 82, and supplied to the multiplier 62. This 1 / (n
The thinning order data n for determining the value of -1) is stored in the assignment memory 42 for each channel, and the thinning harmonic order in the frequency characteristics is selectively switched. The decimation order data n and the timbre information (tone number data T
N), and is determined by tone color information from the performance information generator 10.

A thinning flag is sent to the selector 63 as a select signal, and a normal impulse response signal ISj (t) that does not perform the odd harmonic control is sent to the OR gate group 61 via the selector 63. Cumulative (addition) synthesis is performed as shown in FIG. This thinning flag is stored in the assignment memory 42 for each channel.
It is selectively switched whether to generate a musical tone of a specific overtone that has been erased / attenuated or an odd harmonic, or a musical tone that does not have this erase / attenuate or that contains an even harmonic. The thinning flag forms a part of the tone color information (tone number data TN), and is determined by the tone color information from the performance information generating unit 10.

The thinning flag of each channel is written to the thinning RAM 64 by the controller 20 and supplied to the selector 63 in a time-division manner. This thinning R
The channel number data CHNo is supplied to the AM 64 as access address data.

As described above, the above-described impulse response signal ISj used to generate such odd-numbered overtone musical tones.
If (t) is normally read as shown in FIG. 10, a musical tone without erasure / attenuation of specific harmonics or including even harmonics is generated. Therefore, one impulse response signal IS
From j (t), it is possible to selectively generate a musical tone without elimination / attenuation of specific harmonics or including even-order harmonics and a musical tone with elimination / attenuation or no specific harmonics.

12. FIG. 15 shows a third embodiment of the impulse signal generator 50. In this embodiment, an alternate inversion circuit 66 as shown in FIG. 15 is inserted between the impulse signal storage unit 51 and the impulse accumulating unit 54. In this embodiment, the periodic coefficient f stored in the memory area of the assigned channel of the assignment 42 does not correspond to the designated pitch, but is n times (2 times) the periodic coefficient f of the designated pitch.
X) period coefficient nf (2f). The power data PW is (n-1) / n, (1/2) further from the set value.
Or the envelope level data EL is further set to (n-1) / n, (1/2) from the set value. Others are the same as the above embodiment.

When the period coefficient is n times (twice) “nf (2
f)), the impulse response signal ISj (t) read out per unit time becomes n times (2 times). As shown in FIG. 15 (A3), this is output with (n−1) (1) out of n (2) inverted (positively and alternately) inverted and output to the designated pitch. It becomes the musical tone according to. In this case, since the level of the impulse response signal ISj (t) must be set to (n-1) / n, (1/2), the power data PW is further changed from the set value by (n-1) as described above. ) / N, (１ ／) or the envelope level data EL is further set to (n−1) / n, (１ ／) from the set value.

The above impulse signal storage unit 51 stores
The impulse response signals ISj (t) alternately read by the two lower read address data R1 and R2 are:
One is the selector 63, the selector 59 and the OR gate group 6
1 is sent to the impulse accumulator 54 as it is, and the other is inverted by the inverter group 65 via the selector 63 and the selector 59 and sent to the impulse accumulator 54 to be accumulated (added) and synthesized.

The selector 59 is switched according to the 1 / n frequency-divided data from the programmable 1 / n frequency-divided counter 81, and n of the impulse response signals ISj (t) read out by the lower read address data R1 and R2. Is output as it is, and the other (n-1)
The impulse response signals ISj (t) are inverted and output, and a tone signal as shown in FIG. 16 (A3) is generated and output. Frequency characteristics of this tone signal (spectral envelope, frequency spectrum component, formant shape)
As shown in FIG. 16 (B3), only integer harmonics of n are composed of erased / attenuated harmonics. If n = 2, only harmonics of odd-order harmonics are formed. The selector 63, the thinning RAM 64, and the thinning flag are the same as those in the second embodiment.

The above programmable 1 / n frequency dividing counter 8
1, the controller 20 sets the above-described thinning-order data "n" and increments it by a clock signal 2φR1, and this clock signal 2φR1 is 1 / n
Divided. When n clock signals 2φR1 are input, the 1 / n frequency-divided data goes to a high level, and subsequently when a clock signal 2φR1 is input, it goes to a low level. Thereafter, n clock signals 2φR1 are input. Then, the 1 / n frequency-divided data becomes high level again.

This clock signal 2φR1 is a clock signal having a frequency twice that of the clock signal φR1. The programmable 1 / n frequency dividing counter 81 outputs one of the n non-inverted impulse response signals ISj (t).
And (n-1) impulse response signals IS to be inverted
Each occurrence of j (t) is detected. The programmable 1 / n frequency dividing counter 81 is reset by a reset trigger signal every time all channel times elapse or a reset trigger signal every one channel time elapses.

The odd-numbered harmonic musical tone generated by eliminating / attenuating only integer harmonics of n thus generated is optimal as a musical tone signal of a wind (wind) musical instrument shown in FIG. One end of this tube is closed and the other is open, odd harmonics are much easier to resonate than even harmonics, and each frequency component of the frequency characteristic is almost an odd multiple (1) of the fundamental wave.
Times, 3 times, 5 times, 7 times ...).

13. Variations of Odd-Order Overtone Configuration FIG. 16 shows variations of the odd-order overtone configuration in which only integer harmonics of n have been eliminated / attenuated. The level of the first impulse response signal ISj (t) (FIG. 16A1) is (n−
When the level of the second impulse response signal ISj (t) of (1) / n, (1/2) level becomes relatively large with respect to the level of the second impulse response signal ISj (t) (FIG. 16A2), a differential composite waveform as shown in FIG. The frequency characteristic is n as shown in FIG.
The characteristic has a characteristic that even-numbered overtones, which are integer overtones, are included.

Also, the first impulse response signal ISj
(T) The second impulse response signal ISj (t) having the 1 / n (1/2) level (FIG. 16A1)
When it becomes relatively small with respect to the level of 2), FIG.
A difference synthesized waveform as shown in (A2) is obtained, and the frequency characteristic becomes a characteristic in which even harmonics, which are integer harmonics of n, are negative values as shown in FIG. 18 (B1). It is possible to obtain such a musical tone waveform in which a part of frequency components which cannot be found in the natural world becomes negative.

In this case, a multiplier or a shifter is provided at the two output terminals of the selector 59, and the level of the other is relatively increased or decreased relative to one. The level of the impulse response signal ISj (t) is set to (n-1) /
Alternatively, the envelope level data may be changed to (n-1) / n, (1/2).

14. FIG. 19 shows an envelope cycle table group 43 in the program / data storage unit 40. The envelope cycle table group 43 includes pitch information (key number data KN, frequency number data F
N), the corresponding envelope coefficient r and the periodic coefficient f are read out. This envelope coefficient r is, as described above, the impulse response signal ISj of the inverter signal storage unit 51.
(T) Determine the reading speed S itself, and determine the period coefficient f
Determines the repetition period T of the impulse response signal ISj (t).

The periodic coefficient f is directly proportional to the pitch information. This pitch information indicates a value corresponding to the frequency value of each pitch. This direct proportion may be a strict direct proportion, or may be slightly deviated from the direct proportion due to “S-shaped tuning” or the like.

The envelope coefficient r is directly or not directly proportional to the pitch information, and has the same or different change characteristics as the periodic coefficient f. The characteristic r1 is constant without changing for pitch information. The characteristic r2 is directly proportional to the pitch information. The characteristic r3 is also directly proportional to the pitch information, but is more gradual than the change in the periodic coefficient f. The characteristic r4 has an inverse change characteristic such as inverse proportion to the pitch information. Characteristic r5
Has a portion that changes stepwise with respect to pitch information and changes rapidly and a portion that changes gradually. The characteristic r6 changes exponentially with the pitch information. The characteristic r7 changes exponentially with the pitch information.

The characteristics of the period coefficient f and the characteristics r1 to r7 of the envelope coefficient r for the pitch information constitute one table each, and each table constitutes the envelope period table group 43. Note that the table of the constant value characteristic r1 is omitted, and in this case, the pitch information is calculated and the envelope coefficient r is calculated.
May be required.

The cycle coefficient f read from the table in this way is written in the corresponding channel memory area of the assignment memory 42 in the step 13. The average value of the envelope coefficient r that has been read out and another envelope coefficient r that is obtained from the timbre information (tone number data TN) or the like is calculated or arithmetically synthesized as described above. At the corresponding channel memory area of the assignment memory 42.

15. rf selection table 44 FIG. 20 shows the rf selection table 44 in the program / data storage unit 40. According to the rf selection table 44, the characteristics r1 to r7 corresponding to the musical factor information are obtained.
Is selected. The musical factor information includes pitch information, pitch range (octave) information, pitch name information,
Tone information (tone determination factor), touch information (speed / strength of sounding instruction operation), sounding time information, number of sounds, resonance degree information, effect information, rhythm information, sound image (stereo) information, quantization information, modulation Information, tempo information, volume information, envelope information, and the like.
Generated from 0, etc., the corresponding table is selected.

In particular, tone color information (tone number data T
N) are divided into two groups. For the timbre information of the sound that is not related to the pitch, the table of the characteristic r1 having the constant value is selected, and a tone signal of a fixed formant in which the formant does not move due to the pitch change is generated. The timbre information of the sound that is not related to the pitch is, for example, a voice, a musical instrument sound close to a voice, a sound of a container part of a musical instrument, a percussion instrument sound not having a pitch change property, or a percussion instrument sound close to a voice. The container portion of the musical instrument is, for example, a soundboard or piece of a piano (keyboard instrument), a body or piece of a stringed instrument, a tube itself of a wind instrument or a lead, a body of a percussion instrument, a membrane, a gong or a stick.

Further, in the timbre information of the sound related to the pitch,
A table of a changing characteristic other than the above-described characteristic r1 is selected, and a tone signal of a moving formant in which the formant moves due to a pitch change is generated. The timbre information of the sound related to the pitch includes, for example, the sound of a struck string of a keyboard instrument, the sound of a bowed string of a stringed instrument, the resonance length (tube length) of a wind (tube) instrument, a sound of a pipe diameter or a mouth shape, or Such as percussion sounds having a characteristic of pitch change.

17. Fixed Formant and Moving Formant FIG. 21 shows an example of a fixed formant, and FIG. 22 shows an example of a moving formant. In the fixed formant of FIG. 21, the envelope coefficient r is constant and the impulse response signal ISj
The reading speed S of (t) itself does not change, the period coefficient f changes according to the pitch, and the impulse response signal ISj
The repetition period T of (t) changes.

Then, the formant shape of the frequency characteristic (spectrum envelope, frequency spectrum component) of the generated tone signal does not change and becomes a fixed formant, and the interval between the respective frequency components changes to correspond to the pitch. Such control is executed by the above-mentioned human voice, animal cry, musical instrument sound close to voice / crowning, or percussion instrument sound close to voice / crowning with no pitch change.

In the movement formant of FIG. 22, the envelope coefficient r also changes, the reading speed S of the impulse response signal ISj (t) itself also changes, the period coefficient f changes according to the pitch, and the impulse response signal ISj ( The repetition period T of t) changes.

Then, the frequency characteristics (spectrum envelope, frequency spectrum component) of the generated tone signal are shifted on the frequency axis, the formant shape is changed, and the formant is shifted, and the interval between the respective frequency components is also changed to change the pitch. It also depends. Such control is executed by the above-mentioned keyboard instrument, string instrument, wind (wind) instrument, or percussion instrument sound having a pitch change.

18. End Cutting Table Group 45 FIG. 23 shows the end cutting table group 45 in the program / data storage unit 40. The end cutout table group 45 includes the above-mentioned musical factor information, for example, pitch information (key number data KN, frequency number data FN, range information, tone name information), timbre information, touch information, number of sounds, resonance degree information. The corresponding edge cut data ΔLad is read out based on the sounding time information, the power data PW, and the like. The end cut data ΔLad indicates the cut amounts of both ends of the impulse response signal ISj (t) as described above.

The edge cut data ΔLad is directly or not directly proportional to the musical factor information. The characteristic L1 does not change with respect to the musical factor information and is constant. The characteristic L2 is directly proportional to the musical factor information. The characteristic L3 is also directly proportional to the musical factor information, but is more gradual than the change in the characteristic L2. Characteristic L4
Has inverse change characteristics such as inverse proportion to musical factor information. The characteristic L5 has a portion that changes stepwise with respect to the musical factor information and changes rapidly and a portion that changes gradually. The characteristic L6 changes exponentially with respect to the musical factor information. The characteristic r7 changes exponentially with respect to the musical factor information.

The step-like characteristic L5 is used when the musical factor information is in the envelope phase EF. For example, the first attack, the second attack, the decay, the sustain, and the release are in ascending order of the edge cut data ΔLad.

FIG. 24 shows another edge cut-off table 46 in the program / data storage section 40. The corresponding end cut data ΔLad is read from the end cut table 46 based on the tone color information. In this end cutting table 46, piano, violin, drum, flute,
The value of the edge cut data ΔLad increases in the order of other instruments, voice 1, voice 2,....

As described above, the edge cut data ΔLad becomes smaller as the musical instrument sound changes more rapidly, and becomes smaller as the attack portion changes more rapidly or the sounding time is shorter in one musical tone. When a plurality of timbres are produced simultaneously, the edge cutout data ΔLad is set to be smaller for important timbres that are better heard. Therefore, the edge cut data ΔLad of the edge cut table 46 may be determined based on such importance.

Each of the characteristics L1 to L7 of the edge cut data ΔLad corresponding to the musical factor information constitutes one table, and each table constitutes the edge cut table group 45. Note that the table of the characteristic L1 having a constant value is omitted, and in this case, the value of the musical factor information may be calculated to obtain the edge cut data ΔLad.

The end cut data ΔLad of the end cut table group 45 and the end cut table 46 is obtained and stored by the above-described arithmetic expression group described in the assignment memory 42. Therefore, the musical factor information for reading out the end cut data ΔLad is the periodic coefficient f and the envelope coefficient r or the address length La.
d = (Ea-Sa) can be substituted.

Note that the end cut data ΔLad obtained by the above equation and the end cut data ΔLad read from the end cut table group 45 or the end cut table 46.
May be combined with the average value or the arithmetic combination. The edge cut data ΔLad read from this table in this manner is written to the corresponding channel memory area of the assignment memory 42.

19. FIG. 25 shows a fourth embodiment of the impulse signal generator 50. In this embodiment, an end cutout circuit 67 as shown in FIG. 25 is inserted between the impulse signal storage unit 51 and the selector 63.

The above-mentioned waveform start address data Sa and waveform end address data Ea of each channel are written in the parameter RAM 501 of the impulse signal reading section 53. At this time, the following correction is executed, and The flowchart of the tone generation process is executed.

Sa ← Sa + ΔLad Ea ← Ea−ΔLad As a result, the length of the end cut data ΔLad is cut from both ends of the impulse response signal ISj (t) read from the impulse signal storage unit 51. From the corrected waveform start address data Sa to the waveform end address data Ea, the lower read address data R1 and R
2 is incremented.

The lower read address data R1 or R2 is output from the impulse signal read unit 53 in a time-division manner. This read address data R1 (R2)
Is subtracted by the adder 68 from the correction value (Sa + ΔLad) and supplied to the window function memory 71 via the selector 69 to read the window function value WF. The window function value WF is sent to the multiplier 72 via the selector 76 and is multiplied and synthesized with the impulse response signal ISj (t), so that the front end of each impulse response signal ISj (t) is smoothed.

Since the adder 68 subtracts the correction value (Sa + ΔLad) from the read address data R 1 (R 2), the read operation is started from the top address of the window function memory 71. When the address data R1 (R2)-(Sa + .DELTA.Lad) from the adder 68 exceeds the window width data Wad, a detection signal is sent from the comparator 77 to the selector 76, and data "1" is sent to the multiplier 72. .

The read address data R1 (R
2) is also sent to the comparator 73. The correction value (Ea−ΔLa) is supplied from the adder 74 to the comparator 73.
A value obtained by further subtracting the window width data Wad from d) is sent. The read address data R1 (R2) has a value (Ea
−ΔLad−Wad), the detection signal is sent from the comparator 73 to the selector 69.

Then, the address data R1 (R2)-(Sa + ΔLad) from the adder 68 is inverted by the inverter group 70 and sent to the window function memory 71.
As a result, the window function value WF is obtained before the window width data Wad from the end of the reading of the impulse response signal ISj (t).
Are read in the reverse direction, and each impulse response signal ISj
The cut back end of (t) becomes smooth. In this way, the ends before and after each of the impulse response signals ISj (t) are smoothly connected.

A plurality of types of window function values WF are stored in the window function memory 71, and one of them is selected by the window selection data WS. The type of this window function is shown in FIG.
(B1), (B2), (B3) and (B4). This window function value WF changes from 0 times to 1 time, or 1/2.
This is weight data that changes from double to one-fold and from reduced to one-fold.

In each channel area of the end RAM 75, window selection data WS, waveform start address data Sa, and waveform end address data E of each channel area of the assignment memory 42 are stored by the controller.
a, the edge cut data ΔLad and the window width data Wad are written and copied.

The waveform head address data Sa and the edge cut data ΔLad are inverted by the inverters 78 and 79 in positive and negative directions and sent to the adder 68. Edge cut data ΔLa
The d and window width data Wad are inverted by the inverters 79 and 80 and sent to the adder 74, and the waveform end address data Ea is sent to the adder 74 as it is. The window width data Wad is sent to the comparator 77 as it is. The window selection data WS is sent to the window function memory 71 as it is. Channel number data CHNo from the timing generator 30 is supplied to the end RAM 75 as read / write address data.

20. FIG. 26 shows an end cutout state of the impulse response signal ISj (t). Reading of both ends of the impulse response signal ISj (t) (A1) stored in the impulse signal storage unit 51 is not executed by the end cut data ΔLad. Then, the impulse response signal I as shown in FIG.
Sj (t) is actually read.

This read impulse response signal IS
Both ends of j (t) are discontinuous. FIG. 26 (B1) (B
2) The window function value WF shown in (B3) and (B4) is selected,
The operation is synthesized at both ends. Thereby, each impulse response signal ISj (t) is smoothly connected at each end.

The window function value WF in FIG. 26 (B1) changes from “0” to “1”, and the window function value WF in FIG. 26 (B2) changes from “1/2” to “1”. The window function value WF of 26 (B3) changes linearly from “0” to “1”, and the window function value WF of FIG. 26 (B4) changes from “0” to “1” immediately.

The cut-off impulse response signal ISj
If the front end of (t) rises from “0” toward the local maximum value or the local minimum value, the window function value W in FIG.
F is selected. Further, the leading end of the cut-out impulse response signal ISj (t) is "0" from the local maximum value or the local minimum value.
26, the window function value WF of FIG. 26 (B2) is selected. The “rise” and “fall” are reversed at the rear end.

The detection of the "rise" and "fall" is executed as follows. Address (Sa + ΔLad) or (Ea−ΔLa) at the end of the impulse response signal ISj (t)
The level value of d) is the next address (Sa + ΔLad + 1)
Or, if it is smaller than the level value of (Ea-ΔLad-1), it is “up”, and if it is larger, it is “down”. This detection is executed by the controller 20, and the window selection data WS is determined.

Thus, the impulse response signal ISj
When the end of (t) is “rising” as shown in FIG. 26 (C1), this end is corrected to “0” as shown in FIG. 26 (D1), and the discontinuity disappears. Impulse response signal ISj
When the end of (t) is “falling” as shown in FIG. 26 (C2), the slope of this end is corrected to almost “0” as shown in FIG. Connect.

The value "1" at the center of the window function value WF is long. When the value portion of “1” becomes short, the frequency characteristics (spectrum envelope, frequency spectrum component, formant shape) of the generated tone signal change. Therefore, it is better that the value of "1" is long.

The window function value WF in FIG. 26 (B4) is a square window function without weighting, and the edge of the cut-out impulse response signal ISj (t) is not weighted and corrected. Therefore, noise may occur. However, if this end is at the maximum value or the minimum value of the level “0” or the slope “0”, no noise is generated. In this case, the end cutout circuit 67 in FIG. 25 is omitted. These FIG. 26 (B
1) Not limited to (B2), (B3), and (B4). For example,
Window function exceeding the value of "1", such as changing from "2" to "1" or "1.5" to "1", as well as a triangular window function, trapezoidal window function, Hamming window function, Hanning window function, etc. May be.

FIG. 27 shows an impulse response signal ISj (t).
Shows the connected state of. Impulse response signal ISj
If you read repeatedly without cutting both ends of (t),
When the pitch is low and the repetition period T is long, FIG.
As shown in 2), the ends of the impulse response signals ISj (t) do not overlap. However, when the pitch is high and the repetition period T is short, the ends of the impulse response signals ISj (t) overlap, and the amount of calculation and the amount of processing in the overlapping portion increase.

Similarly, if both ends of the impulse response signal ISj (t) are repeatedly read out without being cut off, when the pitch is low and the repetition period T is long, each impulse response signal ISj (t) as shown in FIG. ) Ends do not overlap. When both ends of the impulse response signal ISj (t) are cut out and repeatedly read out, when the pitch is high and the repetition period T is short, each impulse response signal ISj (t) is
The end of (t) can be prevented from overlapping, and the amount of calculation and the amount of processing in this overlapping portion can be reduced.

Thus, the impulse response signal ISj
When the length of (t) is reduced, the frequency characteristics (spectrum envelope, frequency spectrum component, formant shape) of the generated tone signal become smooth. This allows
It is possible to smoothly change the tone of the musical tone.

21. FIG. 28 shows the impulse synthesis circuit 83. The impulse synthesizing circuit 83 is provided between the impulse accumulator 54 and the multiplier 55. The impulse accumulator 54
Response signal I of each channel sent from
Sj (t) is multiplied by the weighting data WT for each channel in a multiplier 89, added to an accumulated impulse response signal AISj (t) from an accumulation memory 85 by an adder 84, and written to the accumulation memory 85. And the impulse response signal I
Sj (t) is sequentially accumulated. The clock signal φR1 is supplied as a write signal to the accumulation memory 85, and the impulse response signal ISj (t) is accumulated for each channel timing.

Accumulated impulse response signal AISj
(T) is written to the output memory 86 and sent to the multiplier 55 to be multiplied by the envelope data EN and the power data PW. This impulse response signal ISj
The accumulation of (t) is not performed for every channel but for each impulse response signal ISj (t) to be synthesized.

The weighting data WT of each channel is written in the weighting RAM 90 by the controller 20. The channel number data CHNo is supplied to the weighting RAM 90 as access address data. The weighting data WT of each channel is sent to the multiplier 89.

The combined channel number data s is written into the latch 87 by the controller 20. The synthesized channel number data s indicates the number of impulse response signals ISj (t) synthesized as one tone, that is, the number of channels assigned to one tone. The value of the composite channel number data s is determined by the musical factor information.

The synthesized channel number data s is sent to a programmable 1 / s frequency dividing counter 88, and the clock signal φR1 is frequency-divided by 1 / s. The 1 / s frequency-divided data is written to the output memory 86 as a write signal. The signal is sent to the accumulation memory 85 with a slight delay and sent as a clear signal.
Sj (t) is accumulated and output. This programmable 1 / s frequency dividing counter 88 is reset by a reset trigger signal every time all channel times elapse.

S synthesized as one such musical tone
As for the impulse response signals ISj (t), the tone data is allocated to the s channels at the same timing in the step 12 or 13. Therefore, the s impulse response signals ISj (t) have the same generation timing and the same phase. Since the periodic coefficient f is also uniform as described above, the phases of the impulse response signals ISj (t) are also uniform.

In the multiplier 89, each impulse response signal I
Since Sj (t) is multiplied by the weighting data WT,
Each impulse response signal IS synthesized as one musical tone
The combination ratio (mixing ratio) of j (t) is changed by the weighting data WT. By changing the weighting data WT, the frequency characteristic of the tone signal generated by synthesis changes even if the combination of synthesis of the impulse response signals ISj (t) is the same.

22. FIG. 29 shows a combined state of a plurality of impulse response signals ISj (t). The frequency characteristic (spectral envelope, frequency spectrum component, formant shape) of the impulse response signal ISj (t) as shown in (A1) of FIG.
The frequency characteristics (spectral envelope, frequency spectrum component, formant shape) of the impulse response signal ISj (t) as in (A2) have less harmonic components as in (B2) as in (1). .

Both impulse response signals ISj (t)
When (A1) and (A2) are accumulated (added) and synthesized, the synthesized accumulated impulse response signal AISj (t) becomes (A
3), the composite shape has a waveform intermediate between the two signals, and an impulse response signal ISj (t) having a completely different shape.
Is generated, and its frequency characteristics (spectral envelope, frequency spectrum component, formant shape) are (B
As shown in 3), an intermediate state is obtained in which harmonic components are averaged.

By such a combination, a combined impulse response signal ISj having a frequency characteristic obtained by averaging each frequency characteristic of the impulse response signal ISj (t) before the combination.
(T) is generated. In the conventional addition and synthesis of a normal tone waveform, the synthesis is cumulative. In the addition synthesis of the impulse response signals, intermediate averaging is synthesized.

The average weight, that is, the state of the average is changed by the weight data WT. The shape of the synthesized impulse response signal ISj (t) is different from the shape of the impulse response signal ISj (t) before the synthesis, and the frequency characteristics (spectral envelope, The frequency spectrum components and formant shapes are the frequency characteristics (spectrum envelope, frequency spectrum components, and formant shapes) of the impulse response signal ISj (t) before synthesis.
Is also different.

By such a combination, an impulse response signal ISj (t) having an intermediate frequency characteristic can be synthesized only by storing the representative impulse response signal ISj (t). For example, for each note name, for each range,
For each tone group (keyboard system, string system, bowed system, woodwind system, brass system, percussion system, audio system, etc.), for each touch group, for each sounding time group, for each number of sounds, for each resonance degree group, for each performance field A typical impulse response signal ISj (t) is stored for each (memory, chord, bass, accompaniment, rhythm) and for each envelope (sustained type, attenuated type). By simply weighting and combining the representative plurality of impulse response signals ISj (t), the impulse response signals Ij in other musical factors are obtained.
Sj (t) is obtained.

For example, an impulse response signal IS in a low frequency range
If j (t) and the high-frequency range impulse response signal ISj (t) are weighted and synthesized, an intermediate-range impulse response signal ISj (t) is obtained. Here, if the weight of the synthesis is changed, the impulse response signal I of a finer sound range can be obtained.
Sj (t) is obtained.

Further, for example, a brass impulse response signal ISj (t) and a wood impulse response signal ISj
If (t) is weighted and synthesized, an intermediate impulse response signal ISj (t) is obtained. Here, if the weight of the synthesis is changed, an impulse response signal ISj (t) of a finer tubing can be obtained.

By such a combination, an impulse response signal ISj (t) having an arbitrary frequency characteristic can be combined only by storing a small number of impulse response signals ISj (t). For example, a frequency characteristic having a convex portion or a concave portion only in a specific frequency region, a frequency characteristic having a plurality of convex portions or concave portions, a frequency characteristic having a plurality of steps, and the like are possible.

Also, the same impulse response signal ISj
Even in the case of (t), the frequency characteristic having a plurality of bumps or steps can be obtained by changing the periodic coefficient f. For example, when the frequency characteristics of (B2) of FIG. 29 are synthesized by sequentially changing the period coefficient f, a frequency characteristic having irregularities such as a continuous mountain can be obtained.

The weighting data WT may be obtained from the sounding time information (tone time data TM) or may be converted from this information (TM). In this case, the tone time data TM obtained in steps 41 to 45 is arithmetically processed and converted, and is sequentially written into the weighting RAM 90 for each channel.

The weighting data WT may be computed (multiplied, etc.) with the power data PW, envelope level data EL, touch data TC, etc., and stored in the assignment memory 42.

Further, the impulse synthesizing circuit 83 may be omitted, and the function of the impulse synthesizing circuit 83 may be executed by the musical tone accumulating section 57. Further, the s number of the impulse signal generation units 50 are provided, and the s number of the impulse response signals ISj from each impulse signal generation unit 50 are provided.
The weighting data WT may be multiplied and synthesized by s multipliers in (t), and may be added and synthesized by an adder and output.

In this case, s pieces of weighting data WT are R
AM outputs the data in parallel to the s number of adders. The controller 20 copies the s weighting data WT of the assignment memory 42 to the s number of adders, and copies the channel number data CHN to the RAM.
The upper data of o is supplied as an access address.

The present invention is not limited to the above-described embodiment, and can be variously modified without departing from the gist of the present invention. For example, the setting of the cycle end value Fmax may be omitted, and the cycle count value F may be fixed to the maximum possible value. In this case, the pitch of the tone signal is determined only by the period coefficient f.

The one impulse response signal ISj
Although the first half and the second half of (t) have the same shape, they may have different shapes. Further, the impulse response signal ISj
Only the first half of (t) may be stored, and this first half may be read back and reversed to generate the second half signal. The impulse response signal ISj (t) may be a signal obtained by sampling and storing an external sound and converting it by the cepstrum method, the linear prediction method, or the like, or may be a signal artificially created by a user.

Further, the present invention can be implemented in an electronic musical instrument, a computer, or the like. The functions of the circuits in the above-described drawings may be implemented by software (flowcharts), and the functions of the flowcharts in the above-described drawings may be implemented by hardware (circuits).

In the impulse response signal ISj (t), the former half waveform and the latter half waveform are often symmetric. In this case, only the first half waveform or the second half waveform is stored in the impulse signal storage unit 51, and after reading this half waveform, the same half waveform is read in reverse and the impulse response signal ISj (T) The whole may be generated. Thereby, the storage amount of the impulse signal storage unit 51 decreases.

Further, the impulse response signal ISj (t)
May be sent from the outside or generated by calculation. In this case, the impulse response signal ISj
Each instantaneous value of the waveform is calculated and output from the accumulated value of the envelope coefficient r based on the arithmetic expression representing (t). This calculation is repeatedly executed every time the repetition period T elapses.

The control of the volume power by the power data PW is not limited to the case where the reading (generation) speed S and the repetition period T of the impulse response signal ISj (t) are controlled independently, and one is linked to the other or It is also executed when it depends. In this case, the period coefficient f and the envelope coefficient r are determined according to the same musical factor, for example, pitch, timbre, touch, and sounding time, or one of the data is obtained.
An operation is performed on one of the data to obtain the other.

The readout at both ends of the impulse response signal ISj (t) is not cut, and the impulse response signal ISj (t) is not cut.
(T) The whole may be read. In this case, the window function value W
At the end of F, a portion of a value of "0" is stored and read for a predetermined address. This also allows each impulse response signal IS
It is possible to prevent each end of j (t) from overlapping.

(1) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is changed according to the pitch determining factor. Let
A computer for generating a musical tone for causing a computer to change at least a generation speed of the generated impulse response signal itself independently of the repetition period in accordance with a tone determinant different from the pitch determinant. Medium storing program / communication method of computer program / musical sound generation device (method).

(2) Generate an impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated, and control the generated impulse response signal to be repeatedly generated at a predetermined cycle. Even if the length of the signal itself is longer than the cycle of the repetitive generation, generation of the impulse response signal is continued at the end of the cycle, and the computer generates the next impulse response signal by superimposing the impulse response signal. Medium storing computer program for communication / communication method of computer program / musical sound generation device (method).

(3) The pitch determining factor is a factor for determining the pitch of the tone signal, the timbre determining factor is a factor for determining the tone of the tone signal, and the impulse response signal is stored in the storage means. 3. The musical tone generation device according to claim 1, wherein a reading speed of the impulse response signal is determined according to the timbre determining factor, and a repetition period of repetitive reading of the impulse response signal is determined according to the pitch determining factor. Medium storing computer program for communication / communication method of computer program / musical sound generation device (method).

(4) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is changed according to the pitch determining factor. The generation speed of the generated impulse response signal itself according to a tone determinant different from the pitch determinant,
A medium storing a computer program for generating a musical tone for causing a computer to change the waveform shape of the repetitively generated impulse response signal in accordance with the generated musical factor while changing the repetition period independently of the repetition period. / Communication method of computer program / Musical sound generation device (method).

(5) An impulse response signal having a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is generated, and the generated impulse response signal is controlled to be repeatedly generated at a predetermined cycle. Even if the length of the signal itself is longer than the period of the repetitive generation, the generation of the impulse response signal is continued at the end of the period, and the next impulse response signal is superimposed by a plurality of generating means (reading means). A medium storing a computer program for generating a musical tone for causing a computer to switch the waveform shape of the repeatedly generated impulse response signal in accordance with the generated musical factor / communication method of computer program / Musical sound generator (method).

(6) The pitch determining factor is a factor for determining the pitch of the tone signal, the timbre determining factor is a factor for determining the tone of the tone signal, and the impulse response signal is stored in the storage means. The reading speed of the impulse response signal is determined in accordance with the timbre determining factor, the repetition period of the repetitive reading of the impulse response signal is determined in accordance with the pitch determining factor, and the musical factor is the musical tone signal. 6. A medium storing a computer program for generating a musical tone according to claim 4 or 5, wherein said factor is a factor determining a pitch or a tone determinant different from said pitch determining factor.
Musical sound generator (method).

(7) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is changed. Changing the generation speed of the signal itself, and changing the magnitude of the generated impulse response signal based on the repetition period of the changed impulse response signal or based on the generation speed of the changed impulse response signal itself A medium storing a computer program for generating a musical sound to cause a computer to execute the program / communication method of the computer program / musical sound generating apparatus (method).

(8) The magnitude of the impulse response signal is also changed in accordance with the power of the waveform of the impulse response signal or the speed or strength of the sounding operation. The generation speed of the impulse response signal itself is changed according to the timbre determinant, and is changed independently of the repetition period. The pitch determinant is a sound of the tone signal. A factor that determines the tone, the tone color determining factor is a factor that determines the tone color of the tone signal, and the impulse response signal is stored in storage means,
8. The tone generating apparatus according to claim 7, wherein a reading speed of the impulse response signal is determined according to the tone determining factor, and a repetition period of the repetitive reading of the impulse response signal is determined according to the pitch determining factor. Medium storing computer program / communication method of computer program / musical sound generation device (method).

(9) An impulse response signal having a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and some of the repetitively generated impulse response signals are inverted between positive and negative. A medium storing a computer program for generating a musical sound for causing a computer to change the frequency characteristic of the medium / a communication method of the computer program / a musical sound generating apparatus (method).

(10) The above-mentioned impulse response signals are generated repeatedly n times with the same waveform, one repetition cycle is 1 / n times the other repetition cycle, and one level is lower than the other level or one level. Is 1 / n of the other level, and these two impulse response signals are differentially synthesized and output. As a result, (n-1) impulse response signals of the n impulse response signals are inverted and generated. Or, even if the length of the impulse response signal itself is longer than the cycle of the repetitive generation, the generation of the impulse response signal is continued at the end of the cycle, and the next impulse response signal is generated by a plurality of generation means. The impulse response signals are generated repeatedly, the occurrence of each impulse response signal is detected, and the impulse repeatedly generated based on the detection result Among the response signal of n (n-1)
And the level of the non-inverted impulse response signal is set to (n-1) / n, and the level of the inverted impulse response signal is set to 1 / n. The impulse response signal is stored in storage means, 7. The computer program according to claim 6, wherein a reading speed of the impulse response signal is determined according to a timbre determining factor, and a repetition period of repeated reading of the impulse response signal is determined according to a pitch determining factor. Medium that stores
Computer program communication method / tone generator (method).

(11) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is determined. Determine the generation speed of the signal itself, and change the repetition period of the impulse response signal based on a pitch determining factor, and cause a computer to select whether or not to change the generation speed of the impulse response signal itself. Medium storing computer program for generating musical sound / communication method of computer program /
Musical sound generator (method).

(12) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is determined. The generation speed of the signal itself is determined, and the repetition period of the impulse response signal is changed based on the pitch determining factor, and the generation speed of the impulse response signal itself is changed with a characteristic different from the change of the repetition period. Medium storing a computer program for generating a musical sound to be executed by a computer to be executed / communication method of computer program / musical sound generating apparatus (method).

(13) Whether or not to change the generation speed of the impulse response signal itself is not selected when selecting a fixed formant tone signal, but is selected when selecting a moving formant tone signal. The generation speed of the generated impulse response signal itself is changed according to a timbre determining factor, the repetition period of the generated impulse response signal is changed according to a pitch determining factor, and the impulse response signal is The read speed of the impulse response signal is stored in the storage means, and the readout speed of the impulse response signal is determined in accordance with the tone determinant. A factor for determining a pitch of the tone signal, wherein the tone color determining factor is a factor for determining a tone color of the tone signal. Communication method / apparatus for generating musical tones medium / computer program stored a computer program for claim 11 tone according generation is (how).

(14) The repetition period of the impulse response signal changes almost in proportion to the pitch determining factor, and the generation speed of the impulse response signal itself changes in proportion to the pitch determining factor. The generated speed of the impulse response signal itself is changed in accordance with the timbre determining factor, the repetition period of the generated impulse response signal is changed in accordance with the pitch determining factor, and the impulse response signal is stored in the storage means. The reading speed of the impulse response signal is determined according to the tone determinant, and the repetition period of the repeated reading of the impulse response signal is determined according to the pitch determinant. 13. The timbre determining factor is a factor for determining a timbre of the musical tone signal. Communication method / apparatus for generating musical tones medium / computer program stored a computer program for the tone generating (method).

(15) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is determined. Determining the generation speed of the signal itself, preventing the end of the impulse response signal from being generated, and causing the computer to change the amount by which the end of the impulse response signal is not generated based on a musical factor. For storing a computer program for communication / communication method of computer program / musical sound generation device (method).

(16) An impulse response signal of a predetermined length corresponding to the frequency characteristic of the tone signal to be generated is repeatedly generated, and the repetition period of the repeatedly generated impulse response signal is determined. The generation speed of the signal itself is determined, the end of the impulse response signal is prevented from being generated, and the amount of overlap of the ends of the repeatedly generated impulse response signal is detected. According to the detected amount, the impulse response signal is determined. A medium storing a computer program for generating a musical tone for determining an amount that does not generate an edge of the music / a communication method of the computer program / a musical sound generating device (method).

(17) The amount at which the ends of the impulse response signals overlap is equal to the amount at which the ends of the impulse response signals are not generated, whereby the impulse response signals are connected at the ends, or The amount by which the end of the impulse response signal is not generated is greater than the amount by which the respective ends of the signal overlap, and a window function is arithmetically synthesized with the impulse response signal, or the weight at which the end changes from reduced to equal. The data is synthesized,
As a result, the end of the impulse response signal is not generated, and each impulse response signal is smoothly connected at the end. The amount of the end of the impulse response signal that is not generated is changed by a pitch determining factor. The impulse response signals generated by the plurality of generating means are alternately generated by the generating means, and the ends of the impulse response signals generated by one generating means do not overlap. The generation speed of the impulse response signal itself is changed in accordance with the tone determining factor, the repetition period of the generated impulse response signal is changed in accordance with the pitch determining factor, and the impulse response signal is stored in storage means. The read speed of the impulse response signal is determined according to the tone determinant, and The repetition period of the repeated reading of the impulse response signal is determined and the pitch determining factor is a factor that determines the pitch of the musical tone signal,
17. The medium for storing a computer program for generating a tone according to claim 15 or 16, wherein the tone determining factor is a factor for determining the tone of the tone signal. ).

[0234]

As described above in detail, according to the present invention, by combining a plurality of impulse response signals whose repetition period and generation speed are determined, another impulse response signal different from the impulse response signal before the combination is obtained. Generate an impulse response signal. Therefore, there is an effect that more impulse response signals can be generated with less impulse response signals.

[Brief description of the drawings]

FIG. 1 shows an overall circuit of a musical sound generation device.

FIG. 2 shows an assignment memory 42.

FIG. 3 shows an impulse signal generator 50.

FIG. 4 shows various impulse response signals ISj (t) stored in an impulse signal storage unit 51.

FIG. 5 shows a difference in waveform power (volume) due to a change in the repetition period T of the impulse response signal ISj (t).

FIG. 6 shows an impulse signal reading section 53 in the impulse signal generating section 50.

FIG. 7 shows a flowchart of the entire processing.

FIG. 8 shows a flowchart of a sound generation process in step 02.

FIG. 9 shows a time chart of the operation of each unit of the impulse signal reading unit 53.

FIG. 10 shows a state in which an impulse response signal ISj (t) is read from an impulse signal storage unit 51.

FIG. 11 shows a flowchart of an interrupt process.

FIG. 12 illustrates the principle of the present invention.

FIG. 13 illustrates the principle of the present invention.

FIG. 14 shows an alternate inversion circuit 66 (50).

FIG. 15 shows another embodiment of the alternate inversion circuit 66 (50).

FIG. 16 shows a waveform obtained by inverting every other impulse response signal by positive and negative and a frequency characteristic.

FIG. 17 shows the resonance principle of a wind (wind) instrument.

FIG. 18 shows waveforms and frequency characteristics obtained by inverting the polarity of every other impulse response signal and changing the level.

FIG. 19 shows an envelope cycle table group 43 in the program / data storage unit 40.

20 shows an rf selection table 44 in the program / data storage unit 40. FIG.

FIG. 21: Impulse response signal IS of a fixed formant
7 shows a change in j (t) and a change in frequency characteristics.

FIG. 22: Impulse response signal IS of a moving formant
7 shows a change in j (t) and a change in frequency characteristics.

FIG. 23 shows an edge cutout table group 45 in the program / data storage unit 40.

FIG. 24 shows an edge cutout table 46 in the program / data storage unit 40.

FIG. 25 shows an edge cutout circuit 67 (50).

FIG. 26 shows a state where an end of the impulse response signal ISj (t) is cut off.

FIG. 27 shows a connection state of an impulse response signal ISj (t).

FIG. 28 shows an impulse synthesis circuit 83.

FIG. 29 shows a composite state of a plurality of impulse response signals ISj (t).

[Explanation of symbols]

10: Performance information generator, 20: Controller (CP
U), 30: Time generation unit, 40: Program / data storage unit, 41: Information storage unit, 42: Assignment memory, 43: Envelope cycle table, 44: rf selection table, 45: Edge cutout table group, 46 ... End cutout table, 50 ... Impulse signal generator, 60 ... Sound output unit,
51: impulse signal storage unit, 52: impulse signal selection unit, 53: impulse signal read unit, 54: impulse accumulation unit, 58: waveform power control RAM, 59, 63
... selector, 61 ... OR gate group, 62 ... multiplier, 64
... Latch, 65 ... Inverter group, 66 ... Alternate inversion circuit,
67: end cutout circuit, 71: window function memory, 75:
End RAM, 81: programmable 1 / n frequency dividing counter, 83: impulse synthesis circuit, 85: accumulation memory, 8
6 ... Output memory, 88 ... Programmable 1 / s frequency dividing counter, 90 ... Weighting RAM, 501 ... Parameter RA
M, 508: arithmetic RAM, 505: adder.

Claims (3)

[Claims] An impulse response signal having a predetermined length corresponding to a frequency characteristic of a tone signal to be generated is generated a plurality of times, and a repetition period of each of the repeatedly generated impulse response signals is determined. By determining the generation speed of the impulse response signal itself and synthesizing a plurality of impulse response signals whose repetition period and generation speed have been determined, another impulse response signal different from the impulse response signal before being synthesized is obtained. A tone generator which generates a tone. 2. A method for generating a plurality of impulse response signals having a predetermined length corresponding to the frequency characteristics of a tone signal to be generated, determining a cycle of repetition of each of the repeatedly generated impulse response signals, By determining the generation speed of the impulse response signal itself and synthesizing a plurality of impulse response signals for which the repetition period and the generation speed have been determined, another impulse response signal different from the impulse response signal before the synthesis is obtained. A tone generation method characterized by generating a tone. 3. The combined impulse response signals have the same generation start timing and the same phase, and the combined impulse response signals have different or identical frequency characteristics from each other. The respective impulse response signals to be synthesized have different or the same repetition periods from each other, and the respective impulse response signals to be synthesized have different or the same signal generation speed from each other, The combination of each impulse response signal
Each impulse response signal determined by a musical factor is synthesized by being weighted and synthesized. Each weight is determined by a musical factor. The repetition period of the generated impulse response signal is
The impulse response signal is changed in accordance with a pitch determining factor, the impulse response signal is stored in a storage means, a reading speed of the impulse response signal is determined in accordance with the timbre determining factor, and the impulse response signal is read out in accordance with the pitch determining factor. The repetition period of repetitive reading is determined, the pitch determining factor is a factor for determining the pitch of the tone signal, and the timbre determining factor is a factor for determining the tone of the tone signal. A musical sound generation device according to claim 1.