JPS619693A - Musical sound generator - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a musical tone generating device that generates a musical waveform signal using a data memory that stores waveform data or filter parameters for waveform formation, and in particular, The present invention relates to a musical tone generating device such as the above, in which a tone change according to a tone change parameter such as a key touch or pitch is realized by a high quality musical waveform signal.

[Conventional technology]

The entire waveform from the start to the end of the sound, or the entire waveform at the rising edge and a portion of the subsequent waveform, is stored in the waveform memory, and if the former is stored, the entire waveform can be read out in order to create a high-quality musical waveform signal. JP-A No. 52-121313 discloses that when the latter is stored, a high-quality musical waveform signal can be generated by reading out the waveform at the rising edge and then repeatedly reading out part of the waveform after that. has been disclosed.

[Problem that the invention seeks to solve]

Although the above-mentioned prior art method of storing multi-cycle continuous waveforms in the waveform memory can obtain high-quality musical waveform signals, it requires a huge amount of memory capacity. It was unsuitable for realizing various tonal changes depending on the situation.

That is, in order to realize various timbre changes in response to key touches or tone pitches, it is necessary to prepare in advance a large number of different waveform memories corresponding to all i-type key touches or tone pitch timbre change parameters. This has the disadvantage that the overall memory capacity becomes enormous.

This invention was made in view of the above points, and it is possible to generate musical waveform signals capable of various timbre changes according to various timbre change parameters such as key touch or pitch, using a small memory capacity but with high quality. It is an object of the present invention to provide a musical tone generator that can generate musical tones.

[Means for solving problems]

In this invention, N different musical sound waveforms having timbres corresponding to each content of the timbre change parameter are created.
), and for each group, a reference waveform is determined, which serves as a reference for each musical sound waveform belonging to the group.・
This reference waveform may be any one tone waveform in the group or a part thereof, the average waveform of these tone waveforms, or all or any of these tone waveforms. It may be a waveform that approximates one waveform or another waveform that is appropriately set as a standard, and the number of cycles may be one cycle, seven cycles, one cycle, multiple cycles, or the entire corresponding cycle from the start to the end of the sound. etc., can be determined as appropriate in the design.

The reference waveform generation means includes a memory that stores data for generating reference waveforms for each group determined as described above, and generates the parameters according to the timbre change parameters for changing the timbre of the musical sound. The data regarding the group corresponding to the content is read from the memory, and based on this data, a waveform signal corresponding to the reference waveform of the group is generated at a frequency corresponding to a desired pitch. - By way of example, the data stored in the memory is waveform data of each sample point of the reference waveform, in which case the reference waveform signal is generated by reading the memory at a frequency corresponding to a desired pitch. As another example, the data stored in the memory may be filter parameters for obtaining the reference waveform, in which case the reference waveform signal is generated by controlling a digital filter according to the filter parameters.

Further, a residual waveform generating means including a residual waveform memory synthesizes the reference waveform signal generated by the reference waveform generating means and the residual waveform signal generated from the residual waveform generating means to generate the timbre change. A synthesizing means for obtaining a desired musical waveform signal according to the parameters is provided. The residual waveform memory stores residual waveforms corresponding to the differences between the ideal musical sound waveform corresponding to each content of the timbre change parameter and the reference waveform signal regarding the group into which each of these musical sound waveforms is classified,
The residual waveform is selected according to the contents of the timbre change parameter, and the selected residual waveform is read out.

[Effect]

When generating a musical tone, the content of the timbre change parameter is specified by a key touch or pitch, and a residual waveform signal corresponding to the specified content of the timbre change parameter is generated from the residual waveform generating means, and the timbre is A reference waveform signal of the group corresponding to the content of the change parameter is generated from the reference waveform generating means.The generated reference waveform signal and the residual waveform signal are synthesized by the synthesizing means, and the specified timbre is generated. A musical waveform signal having a timbre corresponding to the content of the change parameter is obtained.

The reference waveform data to be stored in the memory in the reference waveform generating means is n? It is sufficient to prepare the number of musical waveforms corresponding to the number of rapes, which is smaller than the number N of musical waveforms of different tones that can be generated (the total number of tone change parameters).

Therefore, memory capacity can be reduced. On the other hand, the number of residual waveforms to be stored in the residual waveform memory in the residual waveform generating means is the same as the total number N of timbre change parameters, but the residual waveform itself is different from the ideal musical waveform and reference waveform signal. Since it is a difference between the two, the data can be expressed in decimal focus, and an excessive memory capacity is not required. Even though such a compressed memory configuration is sufficient, a high-quality musical waveform signal having a desired tone color can be obtained by combining the reference waveform signal and the residual waveform signal.

〔Example〕

First, N different musical sound waveforms having tones corresponding to each content of the timbre change parameter are classified into n groups,
A specific example of processing for creating a reference waveform and residual waveform based on this will be described.

Assuming that the timbre change parameter is the key touch strength, and that the key touch strength changes in 16 steps, then N-1
6, and 16 types of original musical sound waveforms CW1 to CW16 with slightly different tones corresponding to each of the 16 levels of touch intensity (each content of the timbre change parameter) are prepared in advance as listed in Fig. 2. . These 16 types of original musical sound waveforms CW1 to CW16 are extracted from, for example, actual performance sounds of natural musical instruments, and are composed of a plurality of cycles. Further, the amplitude level is standardized so that it is a constant level regardless of the amplitude envelope of the actual performance sound or the touch intensity. Note that such standardization of the amplitude level is not essential, and there is no problem even if the standardization is not performed.

Next, the original musical sound waveform CW1 prepared as described above.
-Classify CW16 into n groups (for example, n-4).

Regarding groups, waveforms that are most similar to each other are classified into the same group. The change in waveform is not so large between musical sound waveforms with adjacent touch intensity levels, so for example, if there are four successive levels of touch intensity 1 to 4,
, 5-8° 9-12. Bind the same group every 13-16, original musical sound waveform CW1-CW4 group,
It is classified into a total of four groups: a group of CW5 to CW8, a group of CW9 to CWI 2, and a group of cw13 to CW16.

Next, for each group, a waveform (reference waveform) that serves as a reference for each original musical sound waveform in that group is created. The reference waveform can be created in any manner; for example, one arbitrary period is cut out from any one original musical sound waveform in the group and this is used as the reference waveform. An example of the reference waveforms RW1 to RW4 determined for each group in this way is as shown in FIG. 3.

Next, the reference waveforms RWI to RW4 of each group are converted to the original musical sound waveforms CW1 to CW16 prepared as described above.
Create a repeated waveform of the reference waveform by repeating it for the same length of time. Then, for each group, the difference between the repetitive waveform of the reference waveform for that group (for example, RWI for the first group) and each original musical sound waveform within that group (CW1 to CW4 for the first group) is determined, and each original Residual waveforms DW1 to DWI6 (see FIG. 4 for an example) corresponding to the musical tone waveforms CW1 to CWI6 are created.

Next, an embodiment of the present invention will be described with reference to FIG.

The reference waveform memory 1 stores in advance data of reference waveforms RWI to RW4 for each group created in the manner described above. In both memories 1 and 2, such reference waveforms RWI~R, W4 and residual waveforms DWI~DWI
6 data sets corresponding to the number of selectable timbre types are stored in the timbre selection circuit 3, and in response to the timbre selection signal TC given from the circuit 3, the data set corresponding to the selected timbre type is stored. Select (make readable). It is assumed that the format of the waveform data stored in each of the memories 1 and 2 is PCM (Pulse Code Modulation) as an example, but the format is of course not limited to this.

A keyboard section 4 having a plurality of keys is used as a means for specifying the pitch of musical tones to be generated.

A touch detection device 5 is provided in association with each key of the keyboard section 4 to detect whether the pressed key is pressed or pressed. The touch detection device 5 may be either an initial touch detection type that detects the touch intensity when the key is pressed or an aftertouch detection type that detects the touch intensity while the key is continuously pressed. It includes an analog/digital conversion means and outputs touch data TD indicating the detected touch intensity in a digital quantity. - As an example, the touch data TD represents 16 levels of touch intensity using a 4-bit digital signal.

The address generation circuit 6 generates a key code KC indicating the pressed key and sample point address data ADD of the key press synchronized with the start of key pressing.

This circuit 6 temporarily sets the old address data ADD by, for example, a key-on pulse KONP, and then
Address data ADD at a rate according to key code KC
The configuration is such that the values are increased sequentially. .

The upper two bits of the touch data TD can identify four groups ("oo" means touch intensity 1).
~4 groups, "01'" for touch strength groups of 5 to 8, "10" for touch strength groups of 9 to 12, "11" for touch strength groups of 13 to 16), for each group. It is input to the reference waveform memory 1 as a waveform address signal for selecting the corresponding reference waveforms RWI to R, W4.All 4 bits of the touch data TD correspond to the contents of the individual tone change parameters, that is, the individual touches in 16 steps. The touch intensity is inputted to the residual waveform memory 2 as a waveform address signal for selecting residual waveforms DWI to DW16 corresponding to each touch intensity.

Predetermined lower n bits of the sample point address data ADD (here, 2 corresponds to the number of sample points of one reference waveform R, W1 to RW4) are given as an input example of the sample point address of the reference waveform memory 1. On the other hand, all bits of the address data ADD are applied to the sample point address input of the residual waveform memory 2.

In this way, the reference waveform (
'RWI to RW4) is repeatedly read out from the reference waveform memory 1 at a frequency corresponding to the pitch of the pressed key, while the residual waveform (DWI
~DW16) is read out from the residual waveform memory 2 at a rate corresponding to the pitch of the pressed key.

The read outputs of the memories 1 and 2 are given to an adder 7, and the repeatedly read reference waveform signal and the residual waveform signal are added and synthesized. In this way, musical sound waveform signals (original musical sound waveforms CW1 to CW
I (same as one of 6) is obtained from adder 7). The output of the adder 7 is given to a multiplier 8, where it is subjected to amplitude level control according to the amplitude envelope and key touch strength, and then converted to an analog signal by a digital/analog converter 9, and finally converted into a sound signal. System 10 is reached.

An envelope waveform signal for setting the amplitude envelope is generated by an envelope generator 11 in a well-known manner. That is, the envelope generator 11 is configured to generate an envelope waveform signal having a predetermined shape in response to the key-on signal KON applied from the keyboard section 4, and this envelope shape is determined in response to the tone selection signal TC. . The level data generating circuit 12 receives the touch data TD and generates level data corresponding to the key touch strength based on the touch data TD, and is composed of, for example, R and OM. The generated envelope waveform signal and level data are multiplied by the multiplier 13 to obtain a level-conquered envelope waveform signal corresponding to the touch intensity, which is fed to the multiplier 8 and multiplied by the musical tone waveform signal. Ru.

Note that the touch data TD may be inputted to the envelope generator 11 as shown by the dotted line, and the peak level of the envelope waveform signal generated therein may be controlled. In such a case, the level data generation circuit 12 and multiplier 13 are unnecessary.

In addition, in this embodiment, the reference waveform rtw is based on the original musical sound waveforms CW1 to CW1'6 whose amplitude levels are standardized.
1 to RW4 and residual waveforms DW1 to DW16 are created and stored in memories 1 and 2.
The waveform data itself stored in No. 2 does not include the volume change corresponding to the touch intensity. Therefore, a means is provided to control the level of the memory readout output according to the touch intensity, but the original musical sound waveforms CW1~
If the same reference waveform and residual waveform creation process as described above is performed without standardizing the amplitude level of cw16, memory 1
, 2 includes not only timbre changes depending on the touch intensity but also volume changes.
, 13, etc.) are no longer necessary.

In addition, when creating the residual waveforms DWI to DW16, instead of simply finding the difference between the repetitive waveform of the reference waveforms RWI to RW4 and the original musical sound waveform CWI - CW16 as described above, [Reference, waveform RWI to "Level control according to the envelope waveform and level control according to touch intensity are applied to the repetitive waveform of RV4" and "The amplitude level is not standardized at all (that is, level control according to the amplitude envelope and touch intensity is applied. ) Original sound waveform C
Find the difference between W1 and CWI 6 J, and use this as the residual waveform D.
WI to DWI 6 may be used. The residual waveforms DWI to DW16 created in this way are, so to speak, residual waveforms in which the width level is not standardized, and there is no need to further perform amplitude envelope control or level control according to the touch intensity. . Therefore, in that case, the arrangement of the level control circuits 8, 11, 12, and 13 in FIG. 1 is changed between the reference waveform memory 1 and the adder 7 as shown in FIG. The residual waveform memory 2 in FIG.
W16 is stored.

In the above embodiment, the timbre change parameter is the key touch intensity, but other parameters can be implemented in the same way.

A case in which the timbre change parameter is a pitch (or range) will be described below using a specific example.

Let us assume that the number of keys in the keyboard section is 61 from C2 to C7, and consider a case where this is divided into pitch blocks (ranges) of every three keys, and a different timbre change is applied to each pitch block. In this case, the number of pitch blocks is N=20, and as shown in the table below, 20
20 types of original musical sound waveforms CW1 to CW2O having slightly different tones corresponding to each pitch block (each content of the tone change parameter) are prepared in advance.

As an example in Table 1, each original musical sound waveform CW1 to CW20 has a central pitch within each pitch block (12°E2, G2
,...G6. Regarding B6, this was taken from the actual performance waveform of a natural musical instrument. As an exception, the pitch block in the highest pitch range is made up of four keys, 86 to C7.

Next, these 20 original musical sound waveforms CW1 to CW2
0 are grouped into 5 groups (n=5) for each octave.
), and create reference waveforms RWI to FLW5 for each group in the same manner as described above. Then, in the same manner as mentioned above, the reference waveforms R, Wl~ of each group (octave) are
The difference between the repeated waveform of RW5 and the original musical sound waveforms CW1 to CW20 in that group is calculated, and the residual waveform DW1 is obtained.
~ Make each DW20 good. In this case, the repetitive waveform of the reference waveforms RWI to RW5 is the original musical sound waveform C.
It is assumed that the frequency is repeated at the same pitch as WI to CW20. For example, when calculating the residual waveform DWI,
The reference waveform RW1 is the frequency of the original musical sound waveform CW1 (
That is, when repeating the pitch C2) to obtain the residual waveform DW2, the reference waveform RWI is repeated at the frequency of the original musical sound waveform CW2 (that is, the pitch E2). DW3~DW20
The same applies to The status of grouping is shown in the table below.

, In the example shown in Table 2 and FIG. 6, the reference waveforms RW1 to RW5 of each group created as described above are the reference waveform memo I.
J j A, and the residual waveforms DW1 to DW20 are stored in the residual waveform memory lJ'2A.

Similarly to the above, the memories 1A and 2A store data sets of reference waveforms RW1 to RW5 and residual waveforms DW for each tone color type.
Data sets I to DW20 are stored, respectively, and a set of reference waveforms RW1 to RW5 and residual waveforms DWI to DW20 are selected according to the timbre selection signal TC.

The key code KC of the pressed key supplied from the keyboard section 4 consists of an octave code OC and a note code NC. In this case, the octave code OC can also be used as information for identifying a group of timbre change parameters, so it is given to the reference waveform memo IJIA as a waveform address signal that specifies the reference waveform (one of aw1 to RW5) to be read. . The address generation circuit 6A generates sample point address data that changes in accordance with the pitch of the pressed key according to the key code KC, and supplies it to the reference waveform memory 1A. In this way, a designated reference waveform (one of t'tw1 to RW5) is repeatedly read out from the reference waveform memory lJ'IA at a frequency corresponding to the pitch of the musical tone to be generated.

On the other hand, a key code KC is given to the waveform address input of the residual waveform memo 1J2A, which specifies the residual waveform (one of DWI to DW20) to be read out according to the pitch block described above. The address generation circuit 6B generates sample point address data for reading out a designated residual waveform, and generates sample point address data at a rate corresponding to the contents of the No 1 code NC of the pressed key. . That is, the central pitch (42) of each pitch flock that is the source of each original musical sound waveform CW1 to CW20.

F2. ... G6. Regarding F6, sample point address data is generated at a predetermined standard rate, but pitches lower than that are C2, D4I-2, . . . F6. For A6, sample point address data is generated at a rate 100 cents lower than the standard rate; however, for A6, the sample point address data is 20 cents lower than the standard rate.
0 sen!・lower), higher pitch D2, F
2,... G6. Regarding C7, sample point address data is generated at a rate 100 cents higher than the reference rate. In this way, within each pitch block, the residual waveform DW
Even if DW1 to DW20 are shared, their readout frequencies can be made to correspond to individual pitches.

Each memo! Reference waveform RW1~ read from JIA, 2A
Repeated readout signal of RW5 and residual waveforms DWl to DW20
are added and synthesized by the adder 7 in the same way as described above, and as a result, the musical sound waveform signals (original musical sound waveforms CW1 to C
W20) is output from the adder 7. The subsequent processing is similar to the embodiment shown in FIG.

However, the key code KC is input to the envelope waveformer 11, and the key scaling open side of the envelope waveform] (
It is used for controlling the shape of the envelope waveform or the peak level, etc. according to the pitch or range of the sound.

The same changes as in FIG. 5 can be made in the embodiment shown in FIG. 6 as well. Further, the grouping of each content (pitch or range) of the timbre change parameter is not limited to every octave, but may be set arbitrarily. In addition, in the above embodiment, a pitch block of every three keys is used as one unit of the timbre change parameter, but the timbre change is not limited to this, but it is possible to make a timbre change every one key (one pitch) or every other appropriate plural keys. You can also do this. l! The best way to change the timbre for each pitch (1 pitch) is to use the residual waveform memo IJ 2.
In case A, the residual waveform is stored for each key (each pitch), and the address generation circuit 6B may be configured to always generate the address data at a constant pitch.

In each embodiment, the format of the waveform data stored in the memories 1J1, 2.1A, 2A is limited to PCM, and any waveform such as DPCM (differential PCM), DM (telta modulation), A, DM (adaptive DM), etc. Of course, an encoding method may be adopted, and in that case, it goes without saying that appropriate peripheral circuits such as a demodulation circuit are added to the memory depending on the encoding method.

Further, each reference waveform stored in the reference waveform memorandums IJI and IA is not limited to one period waveform, but may be a plurality of periods, a ■ period, one period, etc.

In addition, in the residual waveform memo lJ2', 2A, it is not necessarily necessary to store the residual waveform regarding the entire content of the timbre change parameter (for example, all 16 levels of touch intensity), but it is necessary to store the residual waveform regarding the entire content of the timbre change parameter (for example, all 16 levels of touch intensity). The shape may be read out and interpolated into the changing parameter contents (for example, parameters for every other step such as touch intensity 1; 3° 5.7...).

Further, instead of storing the reference waveforms for each group independently, one reference waveform is stored as a reference for each reference waveform, and the residual waveform of the reference waveform for each group with respect to this is stored, respectively, The reference waveforms for each group may be generated by the combination of these.

Note that as a reference waveform generating means, instead of using the waveform memory as described above, a digital filter 14 as shown in FIG.
may also be used. In this case, filter parameters for obtaining a reference waveform signal for each group are stored in the parameter memory 15, and if the timbre change parameter is the key touch intensity, the filter parameters for the corresponding group are stored in accordance with the upper two bits of the touch data TD. Parameters are read from the memory 15. On the other hand, when the tone color change parameter is pitch, the filter parameters of the corresponding group are read from the memory 15 in accordance with the octave code OC. A sound source waveform generation circuit (for example, a waveform memory) 16 generates a digital sound source waveform signal 1 corresponding to the pitch of the pressed key in accordance with the key code KC, and a digital filter 14 generates a digital sound source waveform signal 1 corresponding to the pitch of the pressed key in accordance with the filter parameter given from the parameter memory 15. The digital sound source waveform signal is controlled by the filter characteristics determined by the filter, and as a result, a reference waveform signal is generated.

Although all of the above embodiments have been described with respect to a single-tone electronic musical instrument, it goes without saying that they can be similarly implemented in a multi-tone electronic musical instrument using a sound generation assignment method or the like. Furthermore, the present invention is not limited to the generation of musical sound waveform signals corresponding to scale tones in electronic musical instruments, but also applies to percussion instrument music corresponding to the touch state of the operating section as shown in, for example, Japanese Utility Model Application Publication No. 58-86697. This method can be similarly implemented when generating a sound wave signal.

〔Effect of the invention〕

As described above, according to the present invention, each timbre change parameter can be handled by synthesizing the reference waveform for one n glue, which is smaller than the number N of timbre change parameters, and the residual waveform corresponding to each timbre change parameter. Since musical waveform signals with unique timbres are obtained, it is possible to reduce the memory capacity for storing takes of the reference waveform and residual waveform, and to improve the quality of the final musical waveform signals. can maintain high quality.

[Brief explanation of the drawing]

FIG. 1 is an electrical block diagram showing an embodiment of the present invention;
Figure 2 is a waveform diagram showing an example of an original musical sound waveform with different tones corresponding to each stage of key touch intensity, Figure 3 is a diagram showing an example of reference waveforms determined for each group, and Figure 4 is a diagram showing an example of a reference waveform determined for each group. The figure is a diagram illustrating a residual waveform that is the difference between the repetitive waveform of the reference waveform and the original musical sound waveform, FIG. 5 is a block diagram showing a modified example of FIG. 1, and FIG. FIG. 7 is a block diagram showing another embodiment, but does not show an example of reference waveform generating means using a digital filter. 1.1A...Reference waveform memory, 2,2A...Residual waveform memory, 4...Keyboard section, 5...Touch detection device,
6゜5 A, , 13 B...address generation circuit,
7... Adder, 8° 13... Multiplier, 11... Envelope generator, 14... Digital filter, 1
5...Parameter memory.

Claims (1)

[Scope of Claims] 1. A timbre change parameter generating means for generating a timbre change parameter for changing the timbre, and N types of different musical sound waveforms ( However, the timbre change parameter generated by the timbre change parameter generation means is divided into groups (N>n), and each group includes a first memory storing data for generating a reference waveform regarding the group. reference waveform generating means for reading the data regarding the group corresponding to the content from the memory and generating a waveform signal corresponding to the reference waveform of the group based on the data; A residual waveform corresponding to a difference between a waveform and a waveform signal of the reference waveform regarding the group into which the waveform is classified can be generated based on data stored in the second memory, and the timbre change parameter generation means a residual waveform generating means that selects and generates the residual waveform according to the content of the timbre change parameter generated from the reference waveform generating means; and a waveform signal of the reference waveform generated by the reference waveform generating means and the residual waveform generating means. A musical tone generating device comprising: synthesis means for synthesizing the residual waveform generated by the above to obtain a musical waveform signal having a timbre corresponding to the contents of the timbre change parameter. 2. The timbre change parameter generating means generates, as the timbre change parameter, data indicating the intensity of a touch applied to a key or an operating section operated to generate a musical tone. The musical tone generator described above. 3. The timbre change parameter generating means generates, as the timbre change parameter, data indicating the pitch or the range thereof specified by the pitch specifying means for specifying the pitch of the musical tone to be generated. A musical tone generator according to claim 1. 4. The musical tone generating device according to claim 1, wherein the first memory in the reference waveform generating means stores each sample point data of the reference waveform for each group. 5. The reference waveform generation means includes a sound source waveform generation circuit and a digital filter that filters the sound source waveform signal generated by the circuit, and the first memory stores the filter parameters for the digital filter. 2. The musical tone generating device according to claim 1, wherein the musical tone generating device stores the reference waveform corresponding to each group.