JP3419261B2 - Waveform memory type tone generator - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electronic musical instrument, and more particularly to a waveform memory type electronic musical instrument having a waveform memory sound source.

[0002]

2. Description of the Related Art Natural musical instruments have their own range. Some instruments have a wide range such as a piano, and some instruments have a relatively narrow range such as a wind instrument such as a trumpet. Human voice is also one of the natural musical instruments. Human voices usually have a range of about 2 octaves, at most 3 octaves.

The waveform memory type electronic musical instrument stores waveforms of various timbres in a memory, and reads out and reproduces the waveforms to generate a musical tone signal. The stored waveform is usually formed by sampling the musical tone of a natural musical instrument.

"Voice guide" is known as a technique for electronic musical instruments. According to the music score, floor names such as "do", "re", and "mi" are generated by the human voice from the electronic musical instrument, and the performer plays the keyboard according to the guide by the human voice. The performer only needs to depress the keyboard in accordance with the floor name instruction by the human voice, and can practice efficiently.

However, as described above, the natural human voice has a range of at most about 3 octaves. When practicing a piano song, the range is insufficient at 3 octaves.

In the electronic musical instrument, there is a technique called pitch change, which makes it possible to produce sampled waveforms with different pitches. That is, if the sampled waveform is read at a speed faster than that at the time of sampling, the pitch becomes high, and if read at a low speed, the pitch becomes low.

However, if the sampled waveform of a natural musical instrument is converted into a pitch lower or higher than one octave and then read out, the generated musical sound becomes unnatural. That is, when reading is performed at a frequency lower than that at the time of sampling, the number of samples per unit time becomes low. Lowering the pitch one octave will halve the number of samples. Therefore, the sound quality deteriorates.

Further, if the data is read at a frequency higher than the number of samplings, some of the data is not read at the sampling points, so-called skipping occurs. Then, anharmonic noise due to folding occurs from the high frequency components. In the case of human voice waveform data, not only the tone color changes, but also the words themselves are likely to be unclear.

Therefore, the range in which the pitch of the musical tone generated by the pitch change can be changed is extremely limited. Then, for example, when trying to practice a piano song with a voice guide, it becomes impossible to deal with a song with a wide range.

When a musical piece having a wide timbre, such as a clarinet, is to be played, it may be necessary to produce a sound range that a natural musical instrument cannot produce. However, there is no waveform data in such a range.

A cyclic scale is known as a technique for playing a musical piece having a wide range using a musical instrument having a narrow range. In other words, it is a technique in which a musical range that cannot be pronounced by the musical instrument is moved in octave units, and is replaced by a musical range that can be pronounced. According to the cyclic scale, it is possible to play a musical composition with a wide range, but there arises a problem that the pitch of a musical tone generated at a position where the pitch increases on the score decreases.
Therefore, the performance is unnatural in terms of hearing.

[0012]

As described above,
When a musical piece having a wide range is played by using a musical tone waveform having a limited range, the performance in an area outside the range becomes unnatural.

An object of the present invention is to provide a musical tone generating apparatus which can use a musical tone waveform having a limited tone range and can play a music having a wider tone range with less unnaturalness.

[0014]

According to one aspect of the present invention, a waveform memory type musical tone generating device includes a sounding instructing unit capable of instructing to sound a series of musical tones in a first range, and a series of the musical tones. Conversion means for converting a pitch to a second pitch in one octave narrower than the first pitch range; and waveform data of a scale name sound obtained by sampling a scale name sound of each scale produced by a human voice, A waveform memory that stores waveform data of a plurality of scale names that are wider than one octave and that correspond to each scale of the third scale included in the first scale, and in accordance with the sounding instruction of the sounding instructing means, Read-out combining means for reading out waveform data of a plurality of scale names corresponding to the second pitch from the waveform memory and weighting-combining waveform data of a plurality of scale names read according to the second pitch. Based on the synthesized waveform data And a tone generating unit for generating a tone based on the waveform data of the graded note read from the waveform memory, the reading and synthesizing unit continuously changes the octave by one octave, and It is characterized in that weighting is performed by a weighting coefficient that takes a small value at both ends of the octave.

According to another aspect of the present invention, a waveform memory type musical tone generating device includes a pronunciation instructing means capable of instructing to produce a series of musical tones in a first range and a first pitch of the series of musical tones. Conversion means for converting to a second pitch in one octave narrower than the first pitch range, and waveform data of a scale name note obtained by sampling a scale name note of each scale produced by a human voice. A plurality of data obtained by weighting and combining waveform data of a plurality of scale notes that are wider and correspond to each scale of the third scale included in the first scale, corresponding to each scale in one octave. A waveform memory for storing the synthesized floor name sound waveform data, and according to the sounding instruction of the sounding instructing means,
Reading means for selectively reading out one synthesized tone name sound waveform data corresponding to the second pitch from the waveform memory; and tone generation means for generating a tone based on the read synthesized tone name sound waveform data. In the waveform memory, the waveform data of a plurality of scale names corresponding to each scale are stored in the waveform memory.
It is characterized in that the plurality of synthesized name sound waveform data obtained by weighted synthesis by a weighting coefficient that continuously changes over an octave and has a small value at both ends of the one octave is stored. Further, according to still another aspect of the present invention, a waveform memory type musical sound generator is
Pronunciation instructing means capable of instructing to pronounce a series of musical tones in the first range, and converting means for converting the pitch of the series of musical tones into a second pitch in one octave narrower than the first range.
Waveform data of a scale name obtained by sampling a scale name of each scale produced by a human voice, which is wider than one octave and corresponds to each scale of a third scale included in the first scale. A waveform memory for storing waveform data of a scale name tone; and, in response to a sounding instruction of the sounding instructing means, reading out waveform data of one or a plurality of scale name sounds corresponding to the second pitch from the waveform memory, The reading / synthesizing means for weighting and synthesizing the waveform data of one or a plurality of scale names read according to the second pitch and the musical tone generating means for generating a musical tone based on the synthesized waveform data. In the reading and synthesizing means, the waveform data of the graded sound read out from the waveform memory is weighted by a continuously changing weighting coefficient except at both ends of one octave, and one octave. At the lower end, the waveform data of the scale name sound read from the waveform memory in the high range is output as it is, and at the upper end of one octave, the waveform data of the scale name sound read from the waveform memory in the low range is stored as it is. It is characterized by outputting.

By forming the cyclic scale by a musical tone waveform in which a plurality of scale sounds are mixed, it is possible to reduce the unnaturalness caused by using the cyclic scale.

For example, the scale notes from "do" to "shi" are cyclic scales, and each of them is formed by a plurality of waveform components separated by one octave or more by an octave unit. The constituent sound of "do" emphasizes the waveform component of the high pitch, and the constituent sound of "si" emphasizes the waveform component of the low pitch. Then, it seems as if the pitch rises even when shifting from "Si" to "Do", which is the joint of the cyclic scale.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 2 is a graph showing the concept of a conventional cyclic scale. In the present specification, when waveform data in a specific range is cyclically used in different ranges and in a fixed repeating unit, the scale is called a “cyclic scale”.

In FIG. 2, the horizontal axis represents the MIDI performance data, the note name input from an input device such as a keyboard, and the vertical axis represents the note name of the output waveform. The symbol vN surrounded by an ellipse in the graph indicates the reference volume during sounding. v10 indicates that the sound occurs at full volume.

The solid line shows the case where the input note name and the output note name match. For example, using the keyboard of an electronic musical instrument as the input device,
Such a relationship can be realized when the output musical sound is a piano sound. When a musical sound with a narrow range is used, the pitch that can be pronounced is limited. In such a case, for example, the pitch of the musical tone to be output is moved in octave units from the pitch of the input note name, as indicated by the broken line.

In the figure, a cyclic scale is formed with one octave of C3 to B3 as a repeating unit. That is, even if the pitch of the input signal changes in octave units, the pitch of the output signal does not change. By using such a cyclic scale, it is possible to realize, for example, a voice guide for a piano song.

According to this method, it is possible to accurately know which scale the pronounced sound is pronounced, but the change of the pitch of the joint of the cyclic scale, that is, the change of the pitch from B to C. In this case, there is a strong sense of discontinuity because the pitch shifts by one octave to a lower pitch. In the case of a voice guide,
The performer will not know which key to press. Examples of the present invention for reducing this discontinuity will be described below.

FIG. 1 is a graph showing a cyclic scale according to an embodiment of the present invention. The horizontal axis and the vertical axis in the figure are the same as in FIG. The fact that the output sound for one octave is used is shown in FIG.
But each output sound is composed of a plurality of waveform components having different pitches. The characteristics of this cyclic scale can be summarized in the following three points.

First, as indicated by a block X indicated by a broken line, a cyclic scale is formed with a specific range as a repeating unit. In the illustrated example, C to B (“do”)
A cyclic scale is formed with one octave up to "shi" as a repeating unit of the scale.

Secondly, in response to the input of one note name, a two-component synthesized sound having the same scale name and separated by one octave is generated.
However, since it is an octave difference, in the case of the voice guide, the sounded floor names are the same. Hereinafter, of these two components, the one with a higher pitch is called a treble component and the one with a lower pitch is called a bass component.

Thirdly, the volume ratio of the components of the synthesized voice changes depending on the floor name. Similar to FIG. 2, a symbol surrounded by an ellipse indicates the volume. In the lowest scale of the cyclic scale, the treble component is dominant, and in the highest scale, the bass component is dominant. The volume of the bass component monotonically increases as the pitch increases, and the volume of the treble component monotonically decreases as the pitch increases. It should be noted that the volume of the entire synthetic sound is made constant.

This cyclic scale utilizes human auditory characteristics. Human hearing is insensitive to changes in pitch in octave units (2nd overtone, 4th overtone, etc.).
It is sensitive to changes in pitch within the octave, in other words, changes in floor names.

For example, "C3, E3, G3 (" do ",
C major chord consisting of "mi" and "so") and its expanded chords "E3, G3, C4 (" mi "," so ",
Consider the case of listening to the C major of "do"). At this time, it is insensitive to the change in the absolute pitch of the chord, that is, the difference between the "tone of C (tone)" being C3 or C4, but at least this chord is "C major". That is, this chord is "1st degree, 3rd" with "C as the tonic".
Recognize sensitively that it is a "major" chord made with a "5th degree" pitch.

Even if two sounds that are one octave apart are pronounced at the same time, the judgment as to whether the sound is "do" or "mi", that is, the judgment of the floor name (especially the judgment of the change of the floor name) is made. You can do it clearly. However, uncertain judgments are made regarding the pitches of different octaves. However, when the high-pitched sound component or the low-temperature component is dominant, the pitch is also judged to some extent.

For example, the sound of C is mainly formed of high-pitched sound components, and the low-pitched sound components thereof are small. The bass component gradually increases and the treble component gradually decreases as the scale increases from C to D, E, .... When, for example, a sound continuous with C and E is heard with such a mixing ratio, it is felt that the floor name is changed to C and E and the pitch is increased.

When a sound that changes from E to G is heard, the player also feels that the E and G floor names have been pronounced and that the pitch is rising.

Here, when listening to a synthesized sound that changes with C, D, ... A, the absolute pitch becomes unclear, and the treble component is suddenly felt while giving the impression that it is a continuous scale. The low tone component changes from the main sound to the main sound.

When a changing sound such as D, C or A, C is heard, the bass component changes from the main tone to the treble component, so that the user also feels that the pitch is rising. Therefore, when a musical tone that changes from C3 in FIG. 1 to the right is generated in accordance with such a cyclic scale, it is felt that the pitch gradually rises with the floor name. However, as is clear from the figure, this scale is a cyclic scale, and the overall pitch has not changed.

That is, the cyclic scale shown in FIG. 1 is a scale sound in which discontinuity peculiar to the cyclic scale is hardly felt.

For example, when such a cyclic scale is pronounced in the voice guide, the performer can clearly distinguish the scale name and understand the pitch in a natural sense. With a small change in pitch, one feels a natural change in pitch, similar to the rise and fall of the pitch on the score.

As shown by the solid line Y, in the first scale name (C = “do”) of the cyclic scale, the synthesis (C4: C3 = v8: v2 in this example) emphasizing the treble component is performed. C
Feel the pitch close to 4, and the last floor name (B =
In the case of "shi"), synthesis is performed by emphasizing the bass component (B4: B3 = v2: v8 in this example), and a pitch similar to B3 is perceived.

Therefore, the pitch change between the floor names "B⇔C", which is the switching portion of the cyclic scale, is changed to the pitch name "B3⇔C4".
It can be felt as a change similar to the change in the interval, and a natural change in the pitch can be felt even at the switching part of the cyclic scale.

FIG. 3 shows various configurations of a cyclic scale formed of synthetic sounds. In the figure, the horizontal direction indicates the note number (floor name) of the cyclic scale, and the vertical direction indicates the volume of the generated musical sound component. When there are multiple sounds in the vertical direction, the sound above is higher than the sound below.

FIGS. 3A1 and 3A2 show a configuration equivalent to the cyclic scale shown in FIG. The bass component A and the treble component B corresponding to one octave or the reference pitch are mixed while gradually changing the mixing ratio thereof to form a cyclic scale. At the juncture of the cyclic scale, as shown in FIG.
Shifts to the lowest floor name where is dominant. That is, as a change in the pitch in the sense of hearing, a natural change of one note is felt. In this orthographic scale, for all floor names,
The bass component and the treble component are mixed.

FIGS. 3B1 and 3B2 show the structure of a cyclic scale that can be composed of less waveform data. In the region where the scale name of the cyclic scale is low, only the high tone component B exists, and in the region where the scale name is high, only the low tone component A exists. In the intermediate stage, the bass component A and the treble component B are present.

From a different point of view, the bass component A does not exist in the entire octave or the entire musical scale of the reference pitch, but exists from the middle to the joint on the treble side of the circular scale. The treble component B does not exist in the entire cyclic scale, but exists from the juncture on the bass side of the cyclic scale to the intermediate scale. That is, the cyclic scale can be configured with a sampling waveform that is twice or less than the reference pitch (for example, one octave) of the cyclic scale.

FIG. 3B2 shows the structure at the joint of the cyclic scale. In the highest name of the cyclic scale,
Only the bass component A exists, and only the treble component B exists in the lowest name of the cyclic scale. Therefore, at the juncture of the cyclic scale, a pure pitch change occurs.

In this case, 2 in the middle region of the cyclic scale.
Since the components are mixed and the composition ratio thereof gradually changes, the sense of pitch gradually becomes unclear in terms of hearing, and the pitch shifts, for example, one octave below. Although it sounds natural as a scale change, a transition from a high-pitched component to a low-pitched component is performed.

FIGS. 3A1 to 3B2 show the case where the composition ratio of the bass component and the treble component changes linearly. However, the change in the composition ratio of the bass and treble components is
It is not limited to linear changes.

FIGS. 3C1 and 3C2 show the case where the composition ratio of the bass component and the treble component changes non-linearly. The high-pitched component is dominant in the low-pitched scale, and the composition ratio of the low-pitched component gradually increases, and the low-pitched component is dominant in the highest-pitched name. In this case as well, effects similar to those of FIGS. 3A1 and 3A2 are obtained.

As in FIGS. 3A1 and 3A2,
Although the case where two components exist in all the names of the cyclic scale is shown, the two components may simultaneously exist in only a part of the cyclic scale, as in FIGS. 3B1 and 3B2.

FIGS. 3D1 and 3D2 show an example of the configuration when the sampled waveform data covers three octaves of the cyclic scale. The lowest octave (bass component) is designated by A, the middle octave (mid component) is designated by B, and the highest octave (treble component) is designated by C. In the low-pitched scale name, the middle-tone component B and the high-pitched component C exist, and as the floor name increases, the high-pitched component disappears and only the middle-tone component B exists. Further, when moving to a higher floor name, there are a bass component A and a middle tone component B. In the low-pitched area of the cyclic scale, the high-pitched component C is about 1/2.
The composition ratio is gradually decreased from 0, and eventually becomes 0. In the stage name having a high cyclic scale, the bass component A gradually increases its composition, and eventually becomes about half the composition.

FIG. 3D2 shows the composition of the components of the joint of the cyclic scale. In the transition from the highest name of the cyclic scale to the lowest name, the bass component A changes to the middle tone component B, and the middle tone component B changes to the high tone component C. Therefore, in terms of hearing, 1
You can feel the natural pitch change of the notes.

The case where the sampled waveform corresponds to a pitch of 2 octaves or 2 octaves or less has been described above. However, there may be more sampled waveforms. In this case, it is possible to take another configuration.

FIGS. 3 (E1) and 3 (E2) show an example of the configuration in the case where the sampling waveform exists for 3 octaves. That is, the sampled waveform has three octaves of a low octave, an intermediate octave, and a high octave. In this case, the middle octave B is used to construct a pure octave scale composed of a single component.
A low octave waveform A and an intermediate octave B are used to construct a cyclic scale for the bass. Similarly, the middle octave B and the treble octave C are used to construct a treble scale.

FIG. 3 (E2) shows the structure at the joint of the cyclic scale. In the treble scale on the treble side, as shown in FIG.
The same connection as the cyclic scale in 2) is made. Similarly, the same connection is made in the cyclic scale on the bass side.

At the juncture of the intermediate scale B and the high-pitched cyclic scale, a connection is made from the pure B component, which is the highest scale name of the intermediate scale, to the high-pitched cyclic scale in which the C component is dominant. You can feel the natural change of one note. Similarly, also in the connection between the cyclic scale on the bass side and the intermediate scale B, the transition from the dominant bass component A to the middle tone component B is performed,
You can feel the natural pitch change.

As described above, it is possible to obtain a voice guide in which the pitch discontinuity is small while the cyclic scale is used and the pitch is not changed, so that the sound quality is not deteriorated or the timbre is not changed.

FIG. 4 is an overall view of an electronic musical instrument using a cyclic scale according to an embodiment of the present invention. 1 bus on bus 10
1, CPU12, ROM13, RAM14 are connected,
The CPU 12 follows the performance operation input from the keyboard 11,
In accordance with the program in the ROM 13, the RAM 14 is used as a calculation resist or the like to generate a tone parameter.

A display unit 15 and a panel switch 16 are also connected to the bus 10. The player can confirm the display unit 15 and perform operations such as mode selection and tone color selection with the panel switch 16.

The CPU 12 can access the waveform RAM 21 via the access controller 22 and read out a desired tone waveform. The sound source 23 uses the waveform read from the waveform RAM 21 to generate a musical tone signal according to a musical tone parameter supplied from the CPU 12, and causes the sound system 24 to generate a musical tone. With the configuration described above, the functions of a normal waveform memory type electronic musical instrument can be realized.

In this electronic musical instrument, the waveform input section 25 is further connected to the bus 10. The waveform input section 25 is
A desired musical tone waveform can be sampled using the microphone 26 or the like. For example, it is possible to sample the waveform of the pronunciation of the floor name by a human voice and store it in the waveform RAM 21.

The CPU 12 can form a musical tone of a cyclic scale, which will be described in detail below, based on the further sampled musical tone waveform. For example, it is possible to realize a voice guide using the sampled cyclic scale based on the pronunciation of the human scale based on the MIDI format performance data stored in the RAM 14.

The voice guide is to control the sound source so that the automatic performance is performed based on the automatic performance data of the music to be practiced, and the scale name pronunciation of the corresponding scale is generated according to the MIDI event occurring at that time. . The practitioner can practice a piece of music by operating a performance operator such as a keyboard, depending on the pronounced floor name. By the performance operation on the keyboard 11,
The input performance data is the sound source 2 under the control of the CPU.
3 through the sound system 24 to generate a musical tone having a desired tone color (for example, a piano tone color). When the musical sound to be generated by the playing operation is a musical sound with a narrow range, the reproduced sound can also be a cyclic scale.

FIG. 5 shows an embodiment of a voice guide using a cyclic scale according to the present invention. FIG. 5A is a list of scale vocalization waveforms stored in the waveform RAM.

The human voice band is typically E2 in the case of bass.
~ D4, for soprano typically C4 to C6.
That is, the range of the same person is at most 2-3 octaves. On the other hand, the piano sound range is, for example, A.
It is 0 to C8 (in case of 88 keys) and reaches 7 octaves or more. In MIDI, sounding is instructed in a range of C2 to G8 corresponding to note numbers 0 to 127.

Therefore, the human voice of the bass is sampled,
Two octave scales are sampled and the waveform data is recorded in the waveform memory. Two waveforms, one octave apart, are recorded for one scale pronunciation, for example, the pronunciation of "do". Similarly, two waveforms are recorded for each scale.

The scale pronunciation consists of a combination of consonants and vowels, and has an attack part and a loop part on the waveform data. There is no problem when recording for a sufficiently long time, but when limiting the recording time, it is preferable to adjust the phase in the loop section.

After recording the waveform for 2 octaves, for example, seven mixed waveforms as shown in FIG. 5B are created. Here, the original waveforms stored in the waveform memory are mixed to create a mixed waveform, but as illustrated later,
The mixed waveform may be created when a musical tone is generated in response to a pronunciation instruction, or a mixed waveform is created and stored in a waveform memory prior to playing, and the mixed waveform stored in response to the pronunciation instruction is stored. May be read. In addition,
The mixed waveform shown in FIG. 5B is only the pronunciation corresponding to the white key on the keyboard. A similar mixed waveform can be created for a black key sound.

The cyclic note number NX shown at the left end in the figure
Is a note number representing the name of the cyclic scale. The floor name shown on the right side indicates the pronunciation of the floor name, and consists of a combination of bass and treble components. In addition, vN in the parentheses of each component represents the volume level, and for example, v2 represents a sound with a volume level of 20%. Similarly, v8 indicates a sound with a volume of 80%. The low-pitched sound component gradually increases from the low-ranked area to the high-ranked area, and the high-pitched sound component decreases gradually from the low-ranked area to the high-ranked area. As described above, the bass component and the treble component are cross-faded in the cyclic scale.

At the juncture of the cyclic scale, the note number NX = 2 is changed to the note number NX = 4. At this time, the low-pitched sound component D3 is 8 in the pronunciation of NX = 2.
The treble component D4 is 20%. Note number N
On the contrary, in the pronunciation of X = 4, the treble component E3 is 80% and the bass component E2 is 20%.

Aurally, D3 is dominant in the pronunciation of NX = 2, and E3 is dominant in the pronunciation of NX = 4. Therefore, when the note number NX = 2 changes to NX = 4,
It is felt that the pitch is shifting from D3 to E3.

However, if the pitch is sequentially increased by one octave after that, the bass component and the treble component are more strongly mixed in the middle of the scale, and the bass component becomes dominant before the note number NX = 2 is reached. This change means that there is a pitch change of one octave during a pitch change of one octave, but it does not seem so unnatural to the sense of hearing. At the juncture of the cyclic scale, there is a change close to the natural change in pitch, so it is felt that a very natural change in pitch is occurring.

The cyclic scale of this embodiment is shown in FIG.
It will be apparent to those skilled in the art that the configuration of 1) or the configuration of FIGS. 3C1 and 3D1 can be adopted. In addition, when the sampling waveform is more than 2 octaves or narrow, the waveforms shown in (B1), (D1), and (E) of FIGS.
It will be apparent to those skilled in the art that the configuration like 1) may be adopted.

FIG. 6 shows the CPU 12 according to the embodiment of the present invention.
6 is a flowchart of an electronic musical instrument control process executed by the.
Here, the waveform memory 21 stores the above-mentioned waveform data for two octaves, for example, the 14 waveforms E2, F2, and G in FIG.
2 ..., D4 are stored.

FIG. 6A shows a flowchart of the main flow. When the process starts, first step SA
Initialization is performed at 1. Next, step SA
At 2, voice guidance processing is performed. The voice guide process is a process of reproducing an automatic performance based on performance data stored in advance by using a sound source capable of producing a floor name by a human voice. The performer presses the corresponding key on the keyboard in accordance with the pronunciation of the floor name.

Following the voice guide processing, step SA3
At, keyboard processing is performed. This keyboard process is a sounding process based on a normal key depression operation, and generates a musical tone of a pitch and a volume corresponding to the key depressed by the player.

Next, in step SA4, switch processing is performed by operating various switches on the operation panel. The switch has a selection switch that selects performance data to be played back by voice guidance from multiple performance data, a switch that instructs start / stop of voice guidance processing, and a tone that selects the tone color of the musical sound that is produced in response to keyboard operation. There is a selection switch. In the switch processing, the operated switch is detected, and event processing is performed according to the function assigned to the switch. Then, it returns to step SA2 and repeats the same processing.

FIG. 6B shows the voice guidance process.
This is a note-on event process that is executed when a note-on event occurs in the automatic performance of performance data. When the processing is started, in step SB1, a note number NN for voice guidance processing is obtained according to the note number of the note-on event.

In the next step SB2, the obtained note number NN is converted into a note number NX for the cyclic scale.

In this embodiment, the note number NX for the cyclic scale is obtained according to the formula NX = mod (NN / 12).
Here, mod (NN / 12) indicates a modulo operation in which modulo is 12, the note number NN is divided by 12, and the remainder is taken. By this processing, the note number NN that can take values from 0 to 127 is converted into any note number NX from 0 to 11. Regarding the conversion method, modulo is shown here, but in general, a pitch value that can take a wide range (first range) is folded back to a numerical value in a narrower range (second range). Any conversion method may be used.

In the next step SB3, two tone generation channels are assigned to the floor name of each cyclic scale. A composition consisting of a combination of a bass component and a treble component is specified for each name of the cyclic scale.

At step SB4, the tone generation of the pitch component corresponding to the note number NX is set in one of the two tone generation channels to which the sound source is assigned. In particular,
For example, when the note number NX = 5, the pitch component F3 of FIG.
The selection information of the waveform of F3 and the volume information of v7 stored in the waveform memory 21 and other envelope control information are set so that the pronunciation of (v7) is performed in the one sounding channel.

At step SB5, the tone generation setting of the low tone component according to the note number NX is performed on the other of the two tone generation channels. Specifically, for example, when the note number NX = 5, the selection information of the waveform of F2 stored in the waveform memory 21 and the volume of v3 are stored so that the bass component F2 (v3) is sounded in the other sounding channel. Information, and other envelope control information and the like are set.

In step SB6, note-on is sent to both channels of the sound source to generate the musical tone of each sounding component.

In this way, a musical tone of the cyclic scale as shown in FIG. 5 is generated. It should be noted that although the case of the voice guidance by the human voice has been described, the sounds of other musical instruments can be generated by the same procedure.

In the above embodiment, the sampled waveform data for two octaves is stored in the waveform memory as it is, and each tone component is generated by the cyclic scale using a plurality of channels. With multiple channels,
Instead of playing the sampled waveform, create and store synthetic waveform data that forms a cyclic scale in advance.
At the time of reproduction, this composite waveform data can be reproduced from one channel.

FIG. 7 shows a note-on event process executed in a voice guide process using a cyclic scale according to another embodiment of the present invention. It is assumed that the synthetic waveform of the cyclic scale is already stored. The waveform memory 21 stores cyclic waveform composite waveform data, for example, the seven composite waveforms (E2 in FIG. 5).
(V2) + E3 (v8)), (F2 (v3) + F3 (v
7)) ..., (D3 (v8) + D4 (v2)) are stored. When the process starts, step S
At C1, a note number NN is obtained from the read note-on event generated by the automatic performance. Step SC
In No. 2, the note number NN is converted into the cyclic note number NX by the modulo operation. These processes are equivalent to the processes shown in FIG.

At step SC3, one tone generation channel is assigned. In step SC4, the sound generation channel to which the sound source is assigned according to the cyclic note number NX is set to generate a mixed waveform according to the note number NX. Specifically, for example, note number NX =
5, the mixed waveform (F2 (v3) + F3 (v
The selection information of the mixed waveform, the volume information, and other envelope control information stored in the waveform memory 21 are set so that 7)) is sounded in the sound generation channel. In step SC5, a note sound is sent to the set channel to reproduce the waveform data.

At step SC3, the sound is assigned to a channel, and at the next step SC4, the sound is set to that channel according to the note number NX for the cyclic scale. The waveform data read here is already for the cyclic scale. And stored in the waveform memory 21. In the next step SC5, note-on is sent to the set channel. In response to the note-on, the sound source 23 outputs the selected mixed waveform to the waveform memory 2
Generate a read tone from 1.

In the above embodiments, each scale of the cyclic scale is composed of two components. In order to reduce discontinuity, it is effective to make the connection between the lowest and highest notes of the cyclic scale smooth.

FIG. 8 shows a modification of the above embodiment. Figure 8
(A) shows an example of a recording waveform. The recording waveform may be 2 octaves or more, or 1 octave or more and less than 2 octaves. As an example, A less than 2 octaves
2 to D4, and as another example, E of 2 octaves or more
2 to G4.

FIG. 8B shows an example of creating a cyclic scale when the recorded waveform is less than 2 octaves. Each of the intermediate scales E, F, and G of the cyclic scale has only a single recording waveform. However, the two scales C and D on the bass side of the cyclic scale and the two scales A and B on the treble side are each 2
It has two recorded waveforms and forms a cyclic scale waveform by mixing low-pitched and high-pitched components. B, which is the highest name of the cyclic scale, is a mixed waveform of B2 and B3, and C, which is the lowest name, is C.
This is a mixed waveform of 3 and C4. The transition from B to C gives the impression that the pitch naturally rises. When transitioning from a floor name having a mixed waveform to a floor name having a single waveform, the mixing ratio changes so that a natural pitch change is felt. The mixed waveform of this cyclic scale is shown in FIG.
It is equivalent to that shown in 2).

FIG. 8C shows an example of creating a mixed waveform when the recorded waveform exceeds 2 octaves. In the middle E, F, and G of the cyclic scale, the recorded waveform is 3
There is an octave. In the upper floors and the lower floors, the recorded waveform exists for two octaves. A mixed waveform is formed by two kinds of waveforms in a low cyclic scale, a mixed waveform is formed by three kinds of waveforms in an intermediate scale, and a mixed waveform is formed by two kinds of waveforms in a high scale. Forming a corrugation.

The mixed waveforms of FIGS. 8 (B) and 8 (C) can also obtain a change in the rank name waveform with reduced discontinuity. What is important in these examples is that the component ratios are adjusted so that a natural change can be felt at the juncture of the cyclic scale.

In the above embodiments, the case where the cyclic scale is formed for one octave has been described. In the voice guide, if a 1-octave cyclic scale is created, it is possible to carry out the pronunciation of the scale name by a human voice.

However, the application of the cyclic scale as described above is not limited to the voice guide. For example, a similar technique can be used when playing a musical composition having a wide range by using a waveform of a tone color having a narrow range. Furthermore, in such applications, the cyclic scale need not be one octave.

FIG. 9 shows an example of creating a cyclic scale of 1 octave + 5 degrees. The sound of 1 octave + perfect 5th is an example of a combination of pitches that can be harmoniously heard.

As shown in FIG. 9A, for example, E2-
Record the F3 3 octave + 1 minute waveform.

As shown in FIG. 9B, NX = 0 to NX
Create a cyclic scale with 19 scale names of = 18. 1 octave + perfect 5 in each name of this cyclic scale
Two tones with different pitches are combined. For example,
The lowest rank name NX = 0 is formed by the combination of E2 and B3, and the highest rank name N is formed by the combination of A # 3 and F5.
X = 18 is formed.

Similar to the embodiment shown in FIG. 6, the musical tones having these mixed waveforms are read out by the waveforms of the respective constituents after the floor name is designated and sounded in another channel to generate the mixed waveform sound. Alternatively, similarly to the embodiment shown in FIG. 7, the mixed waveform may be stored and, when the floor name is designated, the mixed waveform may be read and sounded. Note that step S in FIG.
The note number NX of the cyclic scale in B2 or step SC2 of FIG. 7 is given by NX = mod [(NN-4) / 19].

The pitch difference is not limited to this example, but may be octave ±± 4 degrees, octave ±± 6 degrees, or the like.

Use natural musical sounds in a wide range as much as possible,
The cyclic scale can be used only when it is unavoidable. For example, record a human voice for 3 octaves, and record 1
The octave waveform data is used as it is as a single waveform. Octaves above this octave are cyclic scales, and using the waveform data of the middle octave and the upper octave, a cyclic scale similar to the above-described embodiment is created. Similarly, in the lower octave, as in the above-described embodiment, the cyclic scale is created using the waveforms of the low octave and the middle octave. In this way, a cyclic scale in which a simple scale and a cyclic scale are mixed as shown in FIGS. 3 (E1) and (E2) is created. In this case, the note numbers NX for the cyclic scale are NX = 0 to 11 as the lower octave, NX = 12 to 23 as the middle octave, and NX = 2 as the upper octave.
4 to 35.

The conversion from the note number NN to the cyclic scale NX is performed, for example, in the case of NN = 0 to 59, NX = mod (NN / 12), in the case of NN = 60 to 72, NX = mod (NN / 12) +12, When NN = 73 to 127, NX = mod (NN / 12) +24 may be set. It should be noted that, as in the above-described embodiment, the constituent sounds may be recorded and mixed at the time of sounding, or the sampled waveforms may be mixed and recorded and the mixed waveform may be read. Although the case where there are recorded waveforms for 3 octaves has been described, it is obvious that the same configuration can be adopted when there are recorded waveforms for 4 octaves or more.

The pitch change is not preferable for realizing a large pitch change, but can be carried out in a narrow pitch range with almost no deterioration in sound quality.

For example, in the case of a recorded waveform other than a human voice or a human voice having a single sound such as during chorus, the number of recorded waveforms to be sampled can be reduced by using the pitch change.

For example, as a recording waveform, WAVE DATE = [E2-G # 2, A2-C # 3,
D3 to F # 3, G3 to C4, C # 4 to F4], and one waveform is sampled for each unit. Pitch change this sampled waveform,
Generate the waveform of the scale in each subregion. Subsequent mixing may be performed for each scale in the same manner as in the above embodiment. If the sampled waveform is recorded and pitch change and mixing are performed at the time of sounding, the capacity of the waveform data stored in the waveform memory can be further saved.

In the case of pronunciation of a floor name, this technique can be used for floor names represented by the same sound (for example, "do" and "do #").

As the recorded waveform, in addition to the waveform sampled by the sampler or the like, arbitrary waveform data already recorded in the waveform ROM of the electronic musical instrument may be used.

The contents of the above-mentioned cyclic scale may be made editable. In this case, the user using the musical instrument can edit by operating the panel operation element on the panel. For example,
(1). Selection of waveform data to be used cyclically, (2). Setting of the range to be used for cyclic use in the selected waveform data, (3). Setting the pitch difference of multiple waveforms to be mixed,
(4). Setting of the weighting coefficient of the mixing (the minimum necessary mixing is only at the joint of the cyclic scale. The change curve of the weighting coefficient in the pitch direction may be arbitrary), (5). Allocation of waveform data to the musical tone range, designation of the region to be used for the cyclic use, and the like may be set using the panel operator.

The tone generator circuit in the above-mentioned embodiment includes the waveform reading circuit (address generator, interpolation circuit) for reading the waveform data from the waveform memory, the envelope applying circuit,
It may include a channel accumulator, effector, DAC, etc.

The present invention has been described above with reference to the embodiments.
The present invention is not limited to these. For example,
It will be apparent to those skilled in the art that various modifications, improvements, combinations and the like can be made.

[0108]

As described above, according to the present invention,
The cyclic scale makes it possible to expand the musical range that can be played,
It is possible to reduce the discontinuity at the joint of the cyclic scale.

[Brief description of drawings]

FIG. 1 is a schematic graph for explaining a cyclic scale according to an embodiment of the present invention.

FIG. 2 is a schematic graph for explaining a cyclic scale according to the related art.

FIG. 3 is a conceptual diagram showing an aspect of mixing a cyclic scale according to an embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of an electronic musical instrument according to an embodiment of the present invention.

FIG. 5 is a conceptual diagram showing an example of a recording waveform and a mixed waveform according to an embodiment of the present invention.

FIG. 6 is a flowchart showing a processing program of the electronic musical instrument according to the embodiment of the present invention.

FIG. 7 is a flowchart showing a voice guide process according to another embodiment of the present invention.

FIG. 8 is a conceptual diagram showing a modification of the above-described embodiment.

FIG. 9 is a conceptual diagram showing an example of a recording waveform and a mixed waveform according to another embodiment of the present invention.

[Explanation of symbols]

10 buses, 11 keyboards, 12 CPUs,
13 ROM, 14 RAM, 15 indicator,
16 panel switch, 21 waveform RAM,
22 access control unit, 23 sound source, 24
Sound system, 25 waveform input section, 26
Sampler (microphone)

Claims (3)

(57) [Claims] 1. A pronunciation instructing means capable of instructing to pronounce a series of musical tones in a first range, and a pitch of the series of musical tones narrower than the first range.
Conversion means for converting into a second pitch in the octave; waveform data of a scale name tone obtained by sampling the scale name tone of each scale produced by a human voice, the waveform data being wider than one octave and having the first range. A waveform memory for storing waveform data of a plurality of scale names corresponding to each scale of the third tone range included in, and corresponding to the second pitch from the waveform memory in response to a sounding instruction of the sounding instructing means. Read-out synthesis means for reading out waveform data of a plurality of scale names and weighting-synthesizing the waveform data of the plurality of scale names read out in accordance with the second pitch, and a musical tone based on the synthesized waveform data. In the read / synthesize means, the read / synthesize means continuously changes over one octave with respect to the waveform data of the graded note read from the waveform memory, and both of the one octave. A waveform memory type musical tone generating apparatus characterized in that weighting is performed by a weighting coefficient that takes a small value at an end. 2. A pronunciation instructing means capable of instructing pronunciation of a series of musical tones in a first range, and a first pitch of the series of musical tones, a second pitch in an octave narrower than the first range. A conversion means for converting into a scale name and waveform data of a scale name sampled from a scale name of each scale produced by a human voice, the waveform data being wider than one octave and of a third range included in the first range. A waveform memory for storing a plurality of synthesized name sound waveform data obtained by weighting and synthesizing waveform data of a plurality of scale names corresponding to each scale in correspondence with each scale in one octave; On the basis of the reading means for selectively reading out one synthesized floor name sound waveform data corresponding to the second pitch from the waveform memory in response to the sounding instruction of the means, based on the read synthesized floor sound waveform data. Musical sound generating means for generating musical sounds The waveform memory is weighted so that waveform data of a plurality of scale notes corresponding to each of the scales is continuously changed over one octave and takes a small value at both ends of the one octave. A waveform memory type musical tone generating device characterized in that the plural synthetic name sound waveform data obtained by weighted synthesis by coefficients are stored. 3. A pronunciation instructing means capable of instructing to pronounce a series of musical tones in a first range, and a pitch of the series of musical tones narrower than the first range.
Conversion means for converting into a second pitch in the octave; waveform data of a scale name tone obtained by sampling the scale name tone of each scale produced by a human voice, the waveform data being wider than one octave and having the first range. A waveform memory for storing waveform data of a plurality of scale names corresponding to each scale of the third tone range included in, and corresponding to the second pitch from the waveform memory in response to a sounding instruction of the sounding instructing means. Read-out / combining means for reading out waveform data of one or a plurality of scale names and for weighting-combining the waveform data of the one or a plurality of scale names read according to the second pitch, and the synthesized waveform data. And a musical tone generating means for generating a musical tone based on the above, wherein the reading and synthesizing means continuously changes the waveform data of the scale name sound read from the waveform memory, except at both ends of one octave. Heavy Weighting is performed using a weighting coefficient, and at the lower end of one octave, the waveform data of the graded note read from the waveform memory in the high range is output as it is, and at the upper end of one octave, the waveform data in the low range is output. A waveform memory type musical tone generator characterized in that it outputs the waveform data of the read name note as it is.