JP3290188B2 - Score interpreter - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a musical score interpretation device.

[0002]

2. Description of the Related Art An electronic musical instrument having a sequencer function and an inter-instrument interface function such as MIDI can perform an automatic performance according to information recorded in a sequencer and a performance according to input information from another musical instrument. The sequencer information and input information are basically performance information. For example, in a MIDI interface, a note number indicating a pitch of a note to be played as information, and a note-on / instruction for instructing sounding / muting are provided.
An off-chord, a velocity representing a sound strength, and the like are included. In other words, this kind of electronic musical instrument does not deal with the problem of music interpretation, but only accepts performance information that the user interprets and inputs.

[0003]

Therefore, there is a need for a music apparatus that automatically interprets music based on information such as music scores and obtains performance information that is specific information. A major technical problem of automatic music interpretation is to provide a music interpretation device with interpretation ability that simulates human-like music interpretation. SUMMARY OF THE INVENTION It is an object of the present invention to provide a music score interpretation apparatus that interprets music so that the musical expression of a note in a song varies in various ways depending on the presence or absence and the type of a musical score symbol acting on the note.

[0004]

To achieve Means for achieving that this object, according to this invention, to store a sequence of encoded encoded music score marks the score symbols used in music as information representing the music score Symbol string storage means, and music interpretation means for interpreting the sequence of the encoded musical score symbols to generate a performance data sequence including performance parameters for each note, wherein the music interpreting means comprises the encoded sequence score. a symbol detection means for detecting a predetermined symbol acting on notes from among the symbols of a column, the detected predetermined symbol original note multiple small nodes acting
And the number of divisions and the information of the original note
The pitch of each small note to be divided,
Determine the performance parameters of interval and duration, and
Each performance parameter independently
Score interpretation device is provided characterized by having a that allocation control means.

[0005]

[Action] According to the configuration of this, the performance parameter of the pitch and pronunciation time and the sound length of the node on capital is divided into a plurality of notes to each note of that for notes with a predetermined score symbol Assigned respectively. Therefore, it is possible to make the musical expression of a note with a predetermined musical notation richer than that of a note without a musical notation.

[0006]

Embodiments of the present invention will be described below with reference to the drawings.

[0007]overall structure  FIG. 1 is a functional block diagram of a musical score interpretation apparatus according to the present invention.
It is. The score symbol string memory 10 is included in a normal score.
Basically, one-to-one correspondence between various symbols
Memorize the sequence (score of musical notation) and express the music by this
ing. The music interpreting section 20 reads the music from the score symbol string memory 10
To receive a sequence of music notation symbols
Interpret columns. The interpretation result of the music interpretation unit 20 is played.
The actual pitch, onset time, duration and strength of each note
Information. The result of such interpretation is played as a performance symbol string.
It is stored in the symbol string memory 30. Score symbol string memory 10
Score (ML-G) file 10F
FIG. 2A shows an example of the configuration of FIG. This score code file 1
The score described by 0F is a declaration block and one or more
And each staff block is one.
It is composed of the above voice blocks. Performance symbol string memory 3
Of the performance symbol (ML-P) file 30F placed at 0
An example is shown in FIG. The performance symbol file 30F is
Consists of a word block and one or more voice blocks
Is done. Appendix-1 Example of description of ML-G file 10F
(Score symbol sequence). Syntax rules of this ML-G file
The rules (ML-G language syntax) are shown in Appendix-2. Ma
Also, the notation of the score symbol used in the ML-G file 10F
Appendix (set of terminal symbols in syntax)
3 is shown. Appendix-4 Example of description of ML-P file 30F
(Performance symbol string). Of this ML-P file 30F
The syntax rules (ML-P language syntax) are shown in Appendix-5.
You. In the description example of ML-P file (Appendix-4),
[] Indicates performance parameters of a note.
In [], the first number from the left is the actual
It is an actual pitch parameter representing the pitch. The second number is
Step tie representing the time from one note to the next
Parameter, which changes the note on time
Determined. The third number is the note actually ringing
It is a real tone length parameter representing time (gate time).
You. The fourth numerical value is a sound intensity parameter representing the sound intensity.
You. In this ML-P file description example, each performance parameter
Has numerical values conforming to the MIDI standard. Step
Notes in the musical score
FIG. 3 shows the relationship with (note) length. Below, this score
The main features of the interpretation device will be described.

[0008]Dynamics control  The first feature of this score interpreter is the dynamic symbol (Dynamic
Symbol). Functions for interpreting dynamics
A block diagram is shown in FIG. Illustrated wide dynamics interpretation unit 2
00 is one of the functions of the music interpretation unit 20 in FIG.
Interpret dynamic symbols like (forte) and p (piano)
Things. The wide range dynamic symbol interpretation unit 200
Dynamic symbol detection unit 20 for detecting a dynamic symbol from 100
Including 1. Dynamic symbol detected by dynamic symbol detection unit 201
In response to this, the reference sound strength evaluation unit 202
Evaluate the sound strength. Based on the evaluation result of this reference tone strength,
The separate sound strength determination unit 203 determines the sound strength for each dynamic symbol.
Set. The sound strength assigning unit 204 assigns the determined sound strength to a dynamic symbol.
Of each note group in the range
Assigned as the sound strength parameter of the sound
Create 00. In this way, the score interpretation device
For each weak symbol, assign a predetermined tone strength on a one-to-one basis.
Rather than hitting the
Each dynamic symbol has nothing to do with the song
Depends on the song, not as an absolute sound strength
Will be interpreted relative to each other.

[0009]Voice ratio control between voices  The second feature of this score interpreter is voice in double melody music.
Increasing the volume between parts and the structure of multiple sounds (chords)
It has the function of controlling the volume ratio between synthesized sounds. FIG.
FIG. Voice identification unit (or chord structure
Each of the voice parts included in the musical score symbol string 100
(Or, when a chord is detected,
Type). In response to the type identification result from the identification unit 210,
The type-specific volume ratio determining unit 211 determines each voice (or chord configuration).
Sound) is determined. Determine volume ratio by type
Sound intensity correction unit 212 according to the volume ratio determined by unit 211
Modifies the sound intensity of the sound intensity train 300. For example, the sound between voices
In the case of quantitative ratio control, each voice starting from the voice start symbol
Modify the note strength parameter for each note in the
In the case of volume ratio control between sound components, each component of the detected chord is
Modify the note strength parameter of a synthesized note by volume ratio
You. As described above, according to the present score interpretation apparatus,
Set the preferred volume parameters according to the type of chord
It can be provided between sections and between chords.

[0010]Hierarchical control  Multiple symbols that can act on notes in music
is there. Certain symbols act on notes over a wide area,
Another kind of symbol acts locally on notes. Therefore
Note performance parameter values are determined in consideration of multiple symbols.
Need to be For this purpose, the score interpreter uses hierarchical control.
Has a function. FIG. 6A illustrates one mode of hierarchical control.
FIG. 6B shows another example of the hierarchical control. The structure shown in FIG.
The composition performs a composition type hierarchical control. Detecting section 221
A wide-area symbol (for example, a wide-area symbol such as f or p)
Weak sign). The detected global symbol is the global symbol solution
It is interpreted by the parsing unit 222. On the other hand, the detection unit 223
Local symbol 102 (for example, a sound intensity
Xent symbol). Local symbol detected is local
It is interpreted by the symbol interpretation unit 224. The synthesis unit 225
Global interpretation value and local symbol interpretation from global symbol interpretation unit 222
By synthesizing the local interpretation value from the section 224
Determine the parameters. In the case of hierarchical control of sound strength,
For example, f acts as a broad dynamic symbol on a note,
If a local accent change sign is attached
Then, the wide-area symbol interpretation unit 222 sets the sound intensity value 8 as the interpretation value of f.
0, and the local symbol interpretation value is
Then, it is assumed that a 1.1 times sound strength instruction is given. In contrast,
The synthesis unit 225 multiplies 80 by 1.1 to obtain 88, and
Set as the note strength parameter value of the note. Same f
Another note in the region where
If not, the local symbol interpreter 22
4 indicates no change in sound intensity. Synthesizing unit 22 for this
The synthesized sound strength value of 5 is 80. 6 (B) is replaced
Performs type hierarchy control. The detecting unit 231 is a part of the score symbol sequence.
From a wide area symbol 111 (for example, a key
No.) is detected. On the other hand, the wide-area symbol interpretation unit 232
Give its interpretation value. Detecting section 233 is a part of the score symbol sequence.
From the local symbol 112 (for example,
Symbol), and the interpretation for it is detected by the local symbol interpreter 2
34. The selection unit 235 can be any one according to the situation.
Is selected as the performance parameter value of the note.
Taking hierarchical pitch control as an example,
The pitch of the note with the value of E is determined as follows
Is done. Now, the key signature for this E note is G
And raises this note a semitone as an accidental
Assume that the symbol of the loop is operating. In this case, the global symbolic solution
The interpreter 232 interprets the note of E as an interpretation value for key signature G.
The pitch indicates no change, while the local symbol interpretation unit 234
Gives an instruction to raise the pitch of note E by a semitone. to this
On the other hand, the selection unit 235 performs the interpretation from the local symbol interpretation unit 234.
Select the value to change the actual pitch of note E to F, which is a semitone higher than E
Set. When accidentals do not act on note E
The selection unit 235 receives the key signature solution from the wide-area symbol interpretation unit 232
Using the pitch value, set the actual pitch of note E to E. What
6 (A) and 6 (B) show that the number of music symbol hierarchies is two.
However, the present invention is not limited to this.
Also in this case, the hierarchical control function of the musical score interpretation device can be applied.

[0011]Series note control  In addition, the musical score interpreter performs a performance
A series of note control that controls by changing the parameter over time
Has a function. This functional block diagram is shown in FIG.
The detection unit 241 converts a series of notes from the score symbol sequence.
Working musical notation symbols 121 (eg slurs, crescens
And ritardand). A series of detected notes
The action symbol 121 is passed to the time change interpretation unit 242. Time
The inter-interpretation interpreting unit 242 determines whether the detected symbol 121 operates.
For a series of notes, the performance parameters of each note
To control. For example, for the slur symbol
The time change interpreter 242 is a note sound near the center of the slur.
Time-varying sound intensity interpretation with the strongest being the strongest
I do. Series note control function is grouped into a series of notes
It is effective for giving naturalness and
It contributes to the musicality.

[0012]Simultaneous sound strength and length control  In addition, the musical score interpreter uses a predetermined musical score acting on a note.
When interpreting symbols, notes are interpreted according to the interpretation
Controlling both the sound intensity parameter and the duration parameter
It has a strong / duration simultaneous control function. This function block
The figure is shown in FIG. The detecting unit 251 detects the
Detect predetermined musical notation symbols related to notes. Sound strength / sound
The long simultaneous control unit 252 interprets the detected music symbol,
Both the strength parameter and the duration parameter of the relevant note
To control. In this way, this music score interpretation device
The relevant score symbol is simply the performance parameter of one of the notes
Rather than interpreting it as a symbol
Regarding the staff symbols, the interpretation of the tone and
Because both are controlled, it is possible to interpret musical performance more musically
Become.

[0013]System configuration  FIG. 9 is a representative diagram for realizing the above-described musical score interpretation device.
It is a hardware block diagram. Control of the entire device is C
PU1 does this. Sound to interpret music score in program ROM2
The required programs including the music interpretation program are stored
You. The score symbol RAM 3 is stored in the score symbol string memory 10 of FIG.
And input from the score input device 6.
The stored musical notation symbol sequence (a specific example is shown in Appendix-1).
The performance symbol RAM 4 corresponds to the performance symbol string memory 30 of FIG.
The performance symbol string (
An example is shown in Appendix-4). The working RAM 5
For work used by CPU1 during program execution
Memory. The CPU 1 stores in the performance symbol RAM 4
By controlling the sound source 7 based on the performance symbol sequence
Automatically play music according to the music interpretation of the score. In addition,
Instead of the key input type musical notation input device 6,
Use a score image reader to read the image of the score
May be used. In this case, the music score is stored in the program ROM2.
Recognize music symbols from image data and exemplify in Appendix-1
Music score recognition program that obtains musical notation symbol strings
There is a need to.

[0014]Music notation interpretation details  The interpretation of each score symbol is described in detail above.
You.

[0015]Interpretation of dynamics (Figs. 10 to 14)  The dynamic symbol is dy in the ML-G language as shown in FIG.
It is represented by namics (a1). Here a1 is the strength
The issue name. The dynamic symbol interpretation program executed by the CPU 1
The program flow is shown in FIGS. 11 to 13, and FIG.
An explanatory diagram of weak symbol interpretation is shown. Interpretation of dynamics (Fig. 11)
Detects strong and weak symbols from music symbol strings representing music
Then, determine the ratio of each dynamic symbol in the song (11-
1) The dynamic symbol with the highest percentage is
Select as the dynamic symbol Z indicating the quasi-sound strength (11-
2). And the reference tone strength for the selected reference strength symbol Z
The value start is obtained (11-4 to 11-6). And
Based on this reference sound strength value start, the reference strength symbol Z
Sound intensity values for dynamic symbols other than (11-7,
11-8). In the example of FIG. 14, mf indicated by numerical value 5 is the reference strength.
Selected as the weak symbol Z, and its reference sound intensity value sta
rt is 75. Note that this device uses
Data corresponds to MIDI velocity, 0-127
It has a numerical range of. Selected reference strength
The strength values for the dynamics that are stronger than the symbols are shown in Figure 13.
Determined according to the standard (top) routine,
The sound strength value for the weak dynamic symbol is determined by the sound strength shown in FIG.
(Lower) Determined according to the routine. Nothing important
The strength value of these dynamic symbols is also the reference dynamic symbol detected from the song.
A point determined as a function of the sound intensity start for Z
It is. In the example of the routine of FIG. 12 and FIG.
For weak dynamics, W = (start) / (1-summinclear)
/ (1−increase) is defined as a decrease width (however, summinclear = in
create to the power of Z)
Is: W = (127-start) / (1-summincre
as) / (1-increase), where sumincrease = inc
release to the power of (8-Z)). Here W
Is the selected reference strength symbol Z and the reference strength symbol start
Has a value that depends on One rank lower than the reference strength mark
Is the reference sound intensity value sta
It takes a value smaller than rt by W, and two ranks
The sound intensity for the lower dynamic symbol (Z-2) is the reference sound intensity value st.
from art

(Equation 1) Similarly, the tone strength value Point (ZC) for the dynamic symbol (ZC) is

(Equation 2) Given by On the other hand, the sound intensity value Point (Z + C) for the dynamic symbol (Z + C) whose rank is higher by C than the reference is

(Equation 3) Given by As is clear from the above description, FIGS.
By executing the flow in FIG. 13, the function of interpreting the dynamic symbol as described in FIG. 4 is realized.

[0016]Interpretation of changes in dynamics (Figs. 15-17)  Next, dynamic changes such as credit and decrescendo
Explain the interpretation of the symbols. Cressi in the ML-G language
End is crescendo start symbol bCR and crescendo
Decrescen, represented by the end symbol eCR
Is decrescendo start symbol bDE and decrescendo
This is represented by the end symbol eDE (FIG. 15). I
Therefore, in the musical notation symbol sequence, the symbol bCR is converted to the symbol eC.
Up to R is the crescendo section, from the symbol bDE
Represents a decrescendo section up to the symbol eDE. Figure
In the example of 15, the 16th note of E height of the 4th octave
(Represented by the symbol G4: 16) is the decrescendo
The starting note, one of the fourth octave B height
Sixth note (symbol B4: 16) ends decrescendo
A note to do. Flow of dynamics sign interpretation routine
Is shown in FIG. Crescendo from the music symbol sequence
Symbol vs. bCR, eCR or decrescendo symbol
When paired bDE and eDE are searched and a symbol pair is found
Starts crescendo or decrescendo
Note strength and note number, crescendo or dece
Read the note strength and number of the note where the crescendo ends.
(16-1 to 16-4). Start or end here
The strength of the note is obtained by the interpretation of the dynamics
Value is used. Next, the dynamics sign interpretation routine
Crescendo (or decrescendo) start note
To a series of notes between
16-5 to 16-7 to assign the changing sound intensity
Execute. Due to this sound intensity allocation, the routine of FIG.
A technique is used in which the sound intensity is related to the pitch. That is, start
Suitable for changing pitch of note group from note to end note
Find the sound intensity change function that matches the sound intensity of each note
It has gained. For details, the sound from the start note to the end note
The waveform formed by the high row and the type
Function sound1,2,3 start sound strength so and end
Find the error from the function waveform obtained by substituting the sound strength eo,
Select the function waveform that minimizes the error as the sound intensity change curve
Then, using this sound intensity change curve, crescendo (or
Is decrescendo) from the start note to the end note
Determines the pitch value of a series of notes. Function funct
When the sound intensity change curve according to ion1 is selected, the sound intensity is
It is controlled so that the change becomes smaller over time,
In the case of function2, the sound intensity changes over time
It changes more greatly, and in the case of function 3, the sound intensity is
It changes linearly with time. Such dynamic changes
By performing the number interpretation process, the musical score interpretation device is described in FIG.
Realizes a solid notebook control function. In addition, the dynamic symbolic solution
Fig. 6A
The layer control function as described above is realized.

[0017]Slur interpretation (Figures 18-20)  In the ML-G language, slurs are slurs with the slur start symbol bSL.
And an error end symbol eSL (FIG. 18).
In the example of FIG. 18, the quarter note of C4 is the start note of the slur.
The G4 quarter note is the end note of the slur. Slur
The symbol interpretation routine is shown in FIG. This slur sign interpretation
The routine is characterized by a series of notes with slurs (no
G) feels like a single phrase
In addition, the sound intensity of a series of notes is changed over time
is there. This enables the series of note control functions described in FIG.
expressing. Furthermore, in addition to the series control of note strength,
The note length of each note with an error
The length of the time is longer than the step time
You. In this sense, the device for simultaneous control of the sound intensity and the sound length described in FIG.
Noh is realized. Performed by the slur symbol interpretation routine
In the sound intensity control, as shown in FIG.
-The first after the center near the center Z of the slur section to the end
Include notes in slur section with maximum note strength
The note strength curve of a series of notes

(Equation 4) Seeking according to. Here, b represents the position of the note at which the sound intensity is maximized, c is the sound intensity maximum value, and (onkyo
+5). Here, onkyo is a sound intensity value that has already been obtained by the above-described interpretation of the dynamics. 20 in FIG.
From 0-1 to 20-8, the number S of the start note of the slur, the number E of the end note, and the coefficients a, b of the sound intensity change curve y,
c. In steps 20-9 to 20-14, the sound intensity and the gatetime (actual sound length) of each note from the start note to the end note (S to E) are determined.
In particular, as indicated by reference numeral 20-11, the sound intensity of the note of interest is determined according to the sound intensity change curve y, and the notetime of the note is set to be 10% longer than the steptime. Here, the initial value of steptime is
It has a value obtained by converting the notation length of a note (for example, 4 in the case of a quarter note) shown in a musical notation symbol string described in the ML-G language according to a conversion table shown in FIG. It should be noted that in the sound intensity control of the slur interpretation, the interpretation result of the dynamic symbol, which is a wide-area symbol, is used as the reference sound intensity onkyo.
The hierarchical control described in FIG. 6A is also realized. 2
As shown in 0-10, when the slur section overlaps with the section of the dynamic change symbol such as crescendo or decrescendo, 20-11 is skipped to give priority to the sound intensity change processing by the interpretation of the dynamic change symbol.

[0018]Interpretation of local sound intensity change symbols (FIGS. 21 and 2
2)  Local intensity variation to vary the intensity of a single note
Accent marks, sforzand, pho
Luzand, Lynforzand and others. like this
Interpretation of various local sound intensity change symbols
Affects a group of notes, including notes with a partial tone change symbol
Interpretation value of dynamic symbol (result of dynamic symbol interpretation routine)
And so on, based on the reference sound intensity (the dominant sound intensity)
To determine the sound intensity of the note with the local sound intensity change symbol.
In this way, the note strength is determined hierarchically.
By realizing the hierarchical control function as described in FIG.
I have. The interpretation of the local sound intensity change symbol is shown in FIGS.
It is performed according to 1st kind of local sound intensity change symbol (example
(For example, accent) is detected at 21-1 in the flow of FIG.
A second type of local intensity change symbol (eg, Sforza)
22) is detected at 22-1 in the flow of FIG. First
When a local sound intensity change symbol is detected, the change symbol
The reference sound intensity, which is the initial value of the sound intensity of the attached note, that is,
Dominates the note obtained in the dynamics interpretation routine
Strength value of the dynamic symbol or the note
When it is within the range of a crescendo or crescendo,
To the sound strength value that is the result of executing the dynamics symbol interpretation routine,
Sound intensity change data value prepared for each detected sound intensity change symbol
To determine the tone strength value of the note (21-2).
If a second type of local sound intensity change symbol is detected, the change
1 rank higher than the dynamic symbol controlling the note with the number
The strength value obtained for the upper dynamic symbol is the sound of the note
It is determined as a strong value (22-2).

[0019]Chord interpretation (Figs. 23 and 24)  For expressing multiple notes (chords) that are pronounced simultaneously
Therefore, in the ML-G language, the pitch symbols of the notes are shown in FIG.
Connected with chords of mountain signs as shown. Part of the interpretation of chords
FIG. 24 shows a chord volume control routine that performs the following. Chord volume
The control routine searches for chord symbols from the score symbol sequence.
When a chord symbol is detected (24-1),
The highest constituent sound is found (24-2). In the example of FIG.
A5 is the constituent sound of the highest sound. And the highest chord
The sound intensity of the constituent sound is reduced by 3 (24-3). to this
The highest note in the chord is louder than other constituent sounds
Will be played, and the desired chord sound will be obtained.
It is. In this way, the chord volume control routine
By giving a difference in volume between
A volume ratio control function is realized.

[0020]Volume ratio control between voices  As described with reference to FIGS. 2A and 2B,
The interpreter can handle songs with multiple voices (parts).
Can be. To the score symbol string stored in score symbol RAM3
The data for each voice follows the voice start symbol.
You. This musical score interpretation device
It has a function of controlling the volume ratio between the two (FIG. 5). FIG.
In the example of 5, the voices of soprano, alto, tenor and bass
Use 75:37, 5:37, 5:45 as the volume ratio of
are doing. FIG. 26 shows such volume ratio control between voices.
7 is a flowchart of a voice-to-voice volume ratio control routine to be implemented.
This routine completes other interpretations for each voice.
It is performed at the stage when it is completed. In 26-1, the CPU 1 displays a performance symbol.
Voice start symbol% Pa from the performance symbol string in RAM4
rt (), and finds the voice start symbol (
 Identify the type of voice shown in parentheses. Identified voice
If the type is soprano, the voice following the voice start symbol (immediately
The data of the soprano voice section is not changed. Identified voice
If the type is Alto or Tenor, the following voice data
Multiply the tone strength parameter of each note in the block by 、,
(26-2, 26-3), the type of the identified voice part is a bus
If present, the sound intensity of each note in the following voice data block
Multiply the parameter by 45/75. As a result, as shown in FIG.
The volume ratio of soprano, alto, tenor, bass
It will be attached between each voice. As a result,
The soprano part is best heard and the bus
Hear, the inner voices tenor and alto
Working to support the sound organization and overall
Sound is obtained.

[0021]Interpretation of tempo change symbols (Figs. 27 to 31)  Accelerand, ritardand, stringend
Etc. is a tempo change note to change the tempo in the middle of the song
No. As shown in FIG. 27, in the ML-G language, AL
Accelerando, RI ritardand, SG strike
Represents Lindend. FIG. 28 shows Accelerand,
Lines for ritardand and stringend respectively
This shows the function of the tempo change control performed. Ma
Fig. 2 shows the tempo interpretation routine for each symbol.
9, FIG. 30, and FIG. Accelerando Tempo Solution
In the explanation (FIG. 29), the tempo of each note is calculated by y = -log (-x + bb) + a + log (bb). Where bb is (Accelerand
(A ending position + 100) represents the position of the accumulated note length,
This is the tempo before the start of Tcceleland
Tempo (for example, score notation)
Of the song indicated in the declaration section of the issue file (see Appendix 1)
Is the value obtained by decrypting the port). x is the start of Acceleland
The variable of the cumulative duration of the note from the point. According to this interpretation
A series of notes with acceleland
Po gradually becomes faster. In the interpretation of Rittand (Fig. 30)
The tempo of each note with a ritardand symbol is interpreted according to y = log (−x + bb) + a−log (bb). The meaning of each factor in this equation is
The same as for Leland. This equation is based on the previous equation.
This is the result of folding the polygon y = a. But
The tempo of each note with ritardand is gradually slower
It becomes. In the stringend interpretation (Figure 31),
Calculate the tempo of each note marked with the nendo symbol SG according to y = (ab) exp (-x) + b. Here, string a begins
B is the extreme value of the tempo, and (a + 10)
Given. As a result, a series with stringends
The tempo of the notes gradually increases. However, Acche
Unlike Leland, the tempo changes to converge to an extreme value.
Become In this way, the tempo change symbol interpretation routine (Fig.
29 to 31), a sequence of tempo change symbols is applied.
Notes are controlled so that their tempo changes over time.
Control, a series of note control as described in FIG.
Implement the function. In addition, a wide area (top
Based on the interpreted value of the tempo symbol
6 (A) by changing the tempo of the
It has a solid hierarchical control function. FIG. 29 to FIG.
The result of routine 1 is an array of tempos for each note
Temporarily stored. Performance symbol string in ML-P language
Adding tempo data to each note in terms of storage capacity
Is disadvantageous, and when sending performance information by MIDI etc.
Changing the tempo frequently is inconvenient for control. There
In this music score interpretation device, ML-
The pitch parameter g of each note in the performance symbol string in P language
The reference tempo of the song is set to "atitem" and "steptime".
Multiplied by the ratio of the tempo at the note
By correcting the data, the duration parameter (time parameter)
Meter) incorporates tempo changes into the
The data is only the tempo of the song.

[0022]Interpretation of pitch change symbols (Figs. 32 to 36)  Breath, Fermata, Stackatesimo, Stacker
G, tenuto, etc. are music score symbols that basically change the pitch
It is. FIG. 32 shows a normal musical score of this kind of pitch change symbol.
And an expression encoded in the ML-G language. Also,
FIG. 33 to FIG. 36 show various pitch change symbol interpretation routines.
Show. The features of these routines are breath breath
(Except for the interpretation routine), simply change the note duration
Not only that, but also the sound intensity is changed. This
Therefore, the simultaneous control function of the sound intensity and the sound length described in FIG. 8 is realized.
You. To be specific, staccato, staccatesimo solution
In the parsing routine (Fig. 33), the stacker
When the symbol ST or the stacker symbol SM is detected
(33-1), the note length parameter of the note with the symbol
The tagtime is set to 10 (33-2). Change
If the symbol is the staccatesimo symbol SM,
+3 is added to the tone strength parameter ONKYO. here
And the value of the tone strength parameter ONKYO before adding 3
The upper routines (strong and weak symbol interpretation routine,
It is given from the dynamic change symbol interpretation routine). this
In a sense, the hierarchical control function described in FIG. 6A is realized.
(The same applies to FIGS. 34 and 36). This allows the stacker
Note and note with staccatesimo
Notebook with distinction and staccatesimo
Between the note group before and after it,
A distinction can also be made in terms of sound strength. Tenuto interpretation routine
(Fig. 34) finds the tenuto symbol TE from the score symbol sequence
(34-1), the symbol is attached (on the performance symbol string
) Of the note's length parameter
Multiply (34-2). Here is the initial value of the gatetime
Is equal to the interval between notes steptime, so ga
By multiplying ttime by 1.1, one of tenuto
Adjacent sounds that overlap by about 10%
Become. In addition, note strength parameters for notes with a tenuto symbol
Data ONKYO is increased by 3 (34-3). this is,
Only 3 sounds with tenuto than the sound around tenuto
It means to make it bigger and play. Breath interpretation root
(Fig. 35) finds a breath symbol BR from a score symbol sequence
Then (35-1) the note length in front of the breath symbol
The parameter "gatetime" is set to minus one. to this
The sound is cut more and the phrase before the breath
Can be clarified. Fermata interpretation routine
In FIG. 36, the (36-1) file detected from the musical score symbol string is used.
Whether the Hermata symbol FE is in front of the line
(36-2) Length of note with fermata
The way of changing the meter gatetime is different (36-
3, 36-4). Fermata symbol FE before the line
If present, the note length parameter of the note with fermata
Double the meter gatetime, otherwise
Multiply by 1.5. In addition, no with a fermata symbol
The sound strength parameter ONKYO is increased by 3 (36-
5). This clarifies the beginning and end of the song.

[0023]Tuplet interpretation (Figs. 37-42)  A score symbol as shown in FIG. 37 is called a tuplet. like this
In the ML-G language, the tuplet symbol has the format shown in FIG.
Is represented by The example of FIG. 38 is a triplet of G4, A4 and B4.
In the ML-G language, <3 G4: 16-A4: 1
6-B4: 16>. Here 16 is a tuplet configuration
The length of the "notation" of each note G4, A4, B4 that is a sound
Is a sixteenth note. In fact, three
Because the entire length of the note is an eighth note, the tuplet configuration
The length of the sound is 1/3 of it, and the length of the notation (16th
）). In general, the duration of a tuplet
It is the length of the total length equally divided by the number of tuplets. Tuplet length is given
If it is, the actual length of the tuplet
In order to obtain the length, the coefficient multiplied by the length of the notation
As shown in FIG. FIGS. 40 to 42 show each tuplet interpretation route.
This shows the flow of the operation. The features of these routines are
The note length of the tuplet note written in the sequence is shown in FIG.
And change the first note of the tuplet to another tuplet
The point is that the sound intensity has been changed to be stronger than the sound intensity. Therefore
Side view of the note control function (Fig. 7) and simultaneous control of sound intensity and length
Control function (FIG. 8). Also at the beginning of the tuplet
Of the note at the upper level of the note
The reference sound intensity is increased by a predetermined amount using the result of the interpretation as the reference sound intensity.
The hierarchy control function.
(FIG. 6A). For each tuplet interpretation routine
The tone strength setting of the first note is 40-7, 41-
4, 42-6, as shown in FIG.
This is done by adding 2 to the sound intensity data that is the interpretation value.
You. 5 and 7 tuplet interpretation for tuplet note length correction
In the routine (Figure 40), the total tuplet length is
Notes other than tuplet notes (in measures containing tuplets)
The total length of this tuplet is calculated by subtracting the sum of the
Divide by the number of notes to get the duration of the tuplet note
Calculate (40-1 to 40-4), 3, 4, 9-tuple interpretation
In the routine (Fig. 41), the tuplet no
Note data of the note of the performance symbol string obtained from the note length of the note
In steptime, {(number of tuplets-1) {number of tuplets},
That is, multiplying by the coefficient shown in the table of FIG.
Has been obtained (41-1, 41-
2) In the two- and eight-tuplet interpretation routine (FIG. 42),
For the pitch data steptim representing the notation pitch
Multiply e by a factor of 3/4 to get the correct duration of each note in the tuplet
Data steptime is obtained (42-1, 42-2),
For 8-tuplet, note data step representing notation note length
Time is multiplied by a coefficient 6/8 and correct pitch data ste
ptime.

[0024]Interpretation of pre-hit sound (Figs. 43 to 49)  FIG. 43 shows an example of a pre-hit sign in the ML-G language.
Indicates a short-precedence sound symbol and a long-precedence sound symbol. Generally due to the previous hit
The sound that is decorated with is called the main note (main note). Figure
In the example of the score of 43, the sound of D5 represented by a half note is mainly used.
It is a key sound. However, the main note is actually performed in half notes.
So that it is not played and the length of the pre-hit is a half note
Will be played. In the ML-G language, the length of the note written on the score
As you can see, the pitch of each note is expressed,
When interpreting the performance, it is necessary to change the pitch of the notation. This
For each of the pre-hitting sound interpretation routines (Fig.
44 to 47) show the sound of the previous striking sound read from the score symbol string.
Correct the pitch and the length of the main note to perform the correct pre-hitting performance
I am trying to be. In addition, slightly increase the main tone
By doing so, the main soundness can be expressed
(FIG. 47). This controls both the sound intensity and the duration.
The simultaneous control of sound intensity and length as described in FIG.
Will be fulfilled. The strength of the main sound controls the main sound, etc.
Local sound intensity change in main sound to interpretation value of wide range dynamics
It is obtained by adding minute +1 (47-1 to 47-
4) This enables the hierarchical control function of the sound intensity (FIG. 6A).
Play. Explaining the interpretation of the pre-hit length by type
In the percussion length interpretation routine (FIG. 44), a long percussion
When a phonetic symbol is detected (44-1), this long hammer symbol
, Ie, the long strumming st (on the performance symbol string)
eptime and gatetime data in musical notation string
Ornament sounds obtained from the long note data shown in
Length (Step length type according to Fig. 3
(Value converted to data of (4), (44-2, 44-
3), the next note on the performance symbol array, ie, the st
eptime and gatetime are the main sounds of musical notation
Of the above grace tone from the decorated pitch obtained from the notation length of
The value is set to a value obtained by subtracting the length (44-4, 44-5).
As a result, the duration interpretation as shown in the upper part of FIG. 48 is performed.
Will be. In the double-precision sound length interpretation routine (FIG. 45)
Is detected from the musical notation symbol sequence (45-
1) The performance of each double-preceding sound, the number of double-preceding sounds (decorative sounds)
The duration parameters steptime and gateti
me is set to the notation length of each double-precision striking sound indicated in the score symbol string.
It is set to the corresponding value (45-2 to 45-8). More complex
The total Z of the percussion sounds is a value representing the notation length of the main sound
(The duration of the grace sound)
Performance length parameters steptime, gatetim
e is obtained (45-9, 45-10). Therefore, the figure
As shown in the lower half of 48, when the number of double hits is 2,
Will be interpreted as shown in the playing style. Short
In the pre-hit sound length interpretation routine (FIG. 46),
Following the detection (46-1), the tempo of the music is checked (46
-2). When the tempo is fast, the pitch length
Data of steptime and gatetime are 24
(Corresponding to a sixteenth note) (46-3, 46-
8) If the tempo is moderate, steptime and ga
set ttime to 12 (corresponding to the 32nd note)
(46-4, 46-8), step at slow tempo
time and gatetime are 6 (corresponding to the 64th note)
(46-7, 46-8). Performance of main sounds
Steptime and gateti which are pitch parameters
me is a short strike from the notation length value of the main sound (decorated length)
Subtract the sound duration value (current steptime) of the sound
(46-9, 46-10). Therefore, FIG.
As shown in Fig. 9, a slow text such as Largo
In the case of an impo, and fast like Presto
In the case of tempo, the sound of the short front hit and the main sound of the same expression
A different duration interpretation will be performed for the punctuation.
This allows the main sound to be decorated with tempo.
Can be.

[0025]Interpretation of decorative symbols (Figs. 50 to 53)  Mordent, pral trailer, toril, turn, turn
For turns etc., adjacent sounds (sounds of adjacent pitch) are used as decorative sounds
The name of the musical notation used to decorate the main sound using
In the ML-G language, as shown in FIG.
It is encoded with O, PR, TR, TU, and IT. On music
Is the grace note itself (Mordent note, Praltrila
-Notes, trill notes, etc.) are not specified, and the main notes
(Quarter note of G4 in the example of FIG. 50) or note and sound
Each decorative symbol between marks (Mordent symbol in the example of FIG. 50)
Just put on. Therefore, it is written in ML-G language.
The musical score symbol string does not include the symbol of the grace note itself.
For this reason, the musical score interpretation device interprets and interprets musical notation symbol strings.
When interpreting decoration symbols when converting
Performs decoration symbol interpretation processing for attaching decorations. Each decoration symbol solution
In the parsing process, the performance of each note of the decorative sound type (decorative pattern)
Determine the pitch and pitch parameters, and configure the grace tone pattern.
In order to add a change in sound intensity to a series of notes,
Some accents on important notes (eg first note)
Attach This interpretation of decorative symbols allows a series of nodes
G control function (Fig. 7), simultaneous sound strength / length control function (Fig. 7)
8) Hierarchical control function of sound intensity (FIG. 6 (A)) is performed
You. As a specific example, the routine for interpreting the Trill symbol is shown in FIGS.
As shown in FIG. In this routine, we use a trill
Consisting of four equal-length notes,
Pattern that changes to the main sound, the lower adjacent sound, and the main sound
Is adopted. Specifically, the pitch of the turn pitch is determined.
In Rule 1 (Fig. 51), a musical score symbol string is converted to a note (note).
When the turn symbol is detected (51-1), the turn record is made.
Trill played by reducing the length of notes with numbers to 1/4
Each note length (stepti) of four notes in the grace tone pattern
me, gatetime) (51-2, 51-
4). The second and fourth notes are at the pitch of the main note.
The first note is the adjacent pitch above the main note, the third note
To the pitch of the lower adjacent note.
(51-5 to 51-13). Turn pitch determination
In the constant 2 (FIG. 52), a note and a note
If the turn sign in between is detected (52-1), the tar
Each note length (steptime, g) of four notes
atetime) to 12 which is equivalent to a 32nd note
(52-6), Put the full length of the turn sound type before the turn sign.
Deducted from the note length value of the note (preceding sound)
Correct duration of sound (steptime, gatetim)
e) is obtained (52-2). Of the four notes of the turn sound type
The pitch is a turn sound with the sound preceding the turn sound type as the main sound
It is determined in the same manner as the pitch / length determination 1 (52-5 to 5-5).
2-12). Turn sign in turn sound strength control (Fig. 53)
Is attached to a note (main note) (53-
1), plus 1 for the tone of the first note of the turn ornament type
(53-2). The sound strength value before adding 1 here is above
This is the interpretation value of the dynamic symbol at the hierarchical level. Turn sign
If is between notes (53-3), the main sound
+1 is added to the sound intensity data of the preceding note (5
3-4).

[0026]Pitch interpretation (Figs. 54 and 55)  FIG. 54 shows a pitch interpretation routine. At first it is easy with 54-1
Accidental note on note (current note) of interest on staff symbol string
, Or the current note within the bar of the current note
Note staff position (Alpha and octa
Accidentals preceding the number).
Find out if When there is no accidental sign (54-
2) Check the key signature and check that the key signature is at the staff position of the current note.
Check whether to attach a chirp or flat (54
-3). If nothing can be added, the pitch indicated on the current note
Is determined as the actual pitch of the current note (54-4).
To add a semitone (54-5)
When adding notes, lower the notation pitch by a semitone (54-
6), determine the actual pitch of the current note. In this way, the station
The accidental sign that is the local pitch indicator is not attached to the note.
The pitch of the note according to the interpretation of the key signature
To determine. On the other hand, when the accidental sign is attached
(54-2), the type of the accidental symbol is identified (54-
7). Current note table when accidental type is natural
The pitch is determined as the actual pitch of the current note (54-8).
Lower the pitch by one semitone when the hour symbol type is flat
(54-9), when sharp, raise notation pitch by a semitone
(54-10), Double notation pitch in double sharp
Raised (54-11), notation pitch when double flat
(54-12) to determine the actual pitch of the current note
You. When accidentals are acting on notes,
The pitch of the note is determined according to the interpretation of the hour symbol. this
In this way, the pitch interpretation routine is performed as described in FIG.
Layer control function is realized for pitch. This result
As a result, correct pitch interpretation as illustrated in FIG.
You.

[0027]

[Table 1]

[Table 2]

[Table 3]

[Table 4]

[Table 5]

[Table 6]

[Table 7]

[Table 8]

[Table 9]

[Table 10]

[0028]Modified example  This concludes the description of the embodiments.
Various modifications and changes are possible. For example, the encoded word of the score
Any suitable score other than the ML-G language mentioned above as words
An encoding language can be used. Similarly, the encoding language of the performance
Any suitable performance encoding other than the ML-P language described above
Language available.

[0029]

As described in the foregoing, according to the musical score interpretation apparatus of the present invention, as a music interpretation of a given symbol in the score symbol string obtained by encoding, <br/> plurality of note that the symbol acts , And the parameters of the pitch, the sounding time, and the pitch can be assigned to each note, thereby enriching the performance expression of the note.

[Brief description of the drawings]

FIG. 1 is an overall functional block diagram of a musical score interpretation apparatus according to the present invention.

FIG. 2A shows M stored in the score symbol string memory of FIG.
FIG. 2B is a configuration diagram of an LG file, and FIG. 2B is a configuration diagram of an ML-P file stored in the performance symbol string memory of FIG.

FIG. 3 is a diagram showing a relationship between a pitch expression in a musical notation symbol string and a pitch expression (step time) in a performance symbol string.

FIG. 4 is a block diagram of a wide area dynamic symbol interpretation function included in the musical score interpretation apparatus.

FIG. 5 is a block diagram of a volume ratio control function between voices or between chord constituent sounds included in the musical score interpretation apparatus.

FIG. 6A is a block diagram of a composite hierarchical control function included in the musical score interpretation device, and FIG. 6B is a block diagram of a selective hierarchical control function included in the musical score interpretation device.

FIG. 7 is a block diagram of a series of note control functions included in the musical score interpretation device.

FIG. 8 is a block diagram of a simultaneous sound intensity / length control function included in the musical score interpretation device.

FIG. 9 is a block diagram showing a typical system configuration for realizing a musical score interpretation device.

FIG. 10 is a diagram showing encoding of dynamic symbols in the ML-G language.

11 is a flowchart of a dynamic symbol interpretation routine executed by the CPU of FIG. 9;

FIG. 12 is a flowchart of a sound intensity determination (lower) block 11-7 in FIG. 11;

FIG. 13 is a flowchart of a sound intensity determination (lower) block 11-8 in FIG. 11;

FIG. 14 is a diagram expressing a mode of interpretation of a dynamic symbol in a graph.

FIG. 15 is a diagram showing dynamic change symbols in the ML-G language.

FIG. 16 is a flowchart of a dynamics change symbol interpretation routine.

FIG. 17 is a graph showing a sound intensity time change function used in interpretation of a dynamic change symbol.

FIG. 18 is a diagram illustrating a slur symbol in the ML-G language.

FIG. 19 is a graph showing a function of changing a sound intensity with respect to a series of notes with a slur over time.

FIG. 20 is a flowchart of a slur interpretation routine.

FIG. 21 is a flowchart of a local sound intensity change symbol interpretation 1.

FIG. 22 is a flowchart of a local sound intensity change symbol interpretation 2.

FIG. 23 is a diagram showing chord symbols in the ML-G language.

FIG. 24 is a flowchart of a chord volume control routine for controlling the volume ratio between chord constituent sounds.

FIG. 25 is a diagram showing a volume ratio between voice parts.

FIG. 26 is a flowchart of an inter-voice volume ratio control routine.

FIG. 27 is a diagram showing tempo change symbols in the ML-G language.

FIG. 28 is a graph showing functions for tempo change control for a series of notes performed by interpreting each tempo change symbol.

FIG. 29 is a flowchart of an interpretation routine of an accelerando symbol, which is a kind of tempo change symbol.

FIG. 30 is a flowchart of a routine for interpreting a ritard symbol, which is a kind of tempo change symbol.

FIG. 31 is a flowchart of a routine for interpreting a stringend symbol, which is a kind of tempo change symbol.

FIG. 32 is a diagram showing pitch change symbols in the ML-G language.

FIG. 33 is a flowchart of an interpretation routine for staccato and staccatesimo, which are a kind of pitch change symbol.

FIG. 34 is a flowchart of a routine for interpreting tenuto, which is a type of pitch change symbol.

FIG. 35 is a flowchart of a routine for interpreting breath, which is a type of pitch change symbol.

FIG. 36 is a flowchart of an interpretation routine of fermata, which is a type of pitch change symbol.

FIG. 37 is a diagram showing tuplet symbols in a musical score.

FIG. 38 is a diagram showing tuplet symbols in the ML-G language.

FIG. 39 is a diagram showing coefficients for converting the note length indicated by a tuplet symbol to an actual note length.

FIG. 40 is a flowchart of a 5- and 7-tuplet interpretation routine.

FIG. 41 is a flowchart of a 3, 4, 9-tuple interpretation routine;

FIG. 42 is a flowchart of a 2,8-tuple interpretation routine;

FIG. 43 is a diagram showing a pre-hitting sound symbol in the ML-G language.

FIG. 44 is a flowchart of interpreting the length of a long hammer sign.

FIG. 45 is a flowchart of interpreting the duration of a double-preceding hammer symbol.

FIG. 46 is a flowchart of interpreting the duration of a short-preceding hit symbol.

FIG. 47 is a flowchart of sound strength interpretation of a preceding hit symbol.

FIG. 48 is a diagram illustrating an example of interpretation of a long-precision sound and a double-precision sound according to a pre-hit sound interpretation routine.

FIG. 49 is a diagram illustrating an example of interpretation of a short-preceding striking sound and a double-preceding striking sound by a preceding striking sound interpretation routine.

FIG. 50 is a diagram showing decorative symbols in the ML-G language.

FIG. 51 is a flowchart of turn pitch determination 1 for a turn symbol that is a type of decorative symbol.

FIG. 52 is a flowchart of turn pitch determination 2.

FIG. 53 is a flowchart of turn sound intensity control.

FIG. 54 is a flowchart of a pitch interpretation routine for interpreting the pitch of each note in a musical score symbol string.

FIG. 55 is a diagram illustrating an example of pitch interpretation performed by a pitch interpretation routine.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Musical-symbol string memory 20 Music interpreter 30 Performance-symbol string memory 131 Predetermined symbol 251 Detection 252 Simultaneous control of sound strength and length

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-1-304498 (JP, A) JP-A-61-278993 (JP, A) JP-A-64-51992 (JP, U) 3 ", Heibonsha, April 30, 1982, p. 1444 (58) Fields investigated (Int. Cl. ⁷ , DB name) G10H 1/00 G10H 1/057

Claims (1)

(57) [Claims] A musical notation symbol sequence storing means for storing a sequence of encoded musical notation symbols obtained by encoding musical notation symbols used in a musical score as information representing a musical composition; Music interpretation means for generating a performance data sequence including performance parameters of each note, wherein the music interpretation means detects a predetermined symbol acting on a note from the sequence of the coded sequence score symbols. Detecting means and an original note on which the detected predetermined symbol acts
Along with the number of divisions and the information of the original note.
Pitch and pronunciation of each of the divided small notes according to the information
Determine the performance parameters for time and duration, and for each small note
Each of the above performance parameters independently
Score interpretation apparatus characterized by having a allocation control unit Teru.