JP2015111286A - Chord detection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a proper chord with respect to user's free performance such as playing with both hands.SOLUTION: This chord detection device is configured to, when performance data based on playing with both hands or the like are input, extract a candidate chord at a chord detection timing such as each of a prescribed number of bars or a prescribed number of beats in accordance with keys at that time, and to identify what type of role does each performance sound of the input performance data play with respect to each extracted candidate chord, and to acquire points indicating a degree of musical significance of the identified role from a degree name-categorized point table PT having a note point table, and to sum up points acquired for all performance sounds for each candidate chord, and to add priority points (Prior) of the candidate chord itself to the total points to obtain the sum total of points, and to define the candidate chord having the maximum sum total of points as the optimal chord.

Description

  The present invention relates to a chord detection apparatus capable of detecting chords for a free performance such as playing both hands by a user.

  Conventionally, an electronic music apparatus having a function of detecting a chord by a performance operation is known. For example, in the automatic accompaniment device disclosed in Patent Document 1, a user performance sound is searched for at the chord detection timing, matching with a pre-stored chord pattern is performed, and if there are only two keys pressed, a fractional chord is obtained. One of the designated two sounds is regarded as a root, and a fractional code is determined based on a predetermined rule, and automatic accompaniment based on the determined code is performed.

Japanese Patent No. 2526430

  However, in the conventional chord detection technique, chords are detected when chord constituent sounds are arranged by the user's performance. Therefore, the user must press the chord constituent sounds with the keyboard accurately at the detection timing. Since it is necessary to press the key so that it becomes the chord you want to detect, it is not suitable for free performances such as playing both hands. Further, since code pattern matching and route determination are performed by a predetermined calculation method, the detected code may become monotonous.

  An object of this invention is to provide the chord detection apparatus which can detect a chord suitable for free user performances, such as playing both hands, in view of such a situation.

  In order to achieve the above object, a chord detection apparatus according to the present invention instructs a chord detection at a predetermined chord detection timing, a performance data acquisition means for acquiring performance data, a key information acquisition means for acquiring key information. A code detection instruction means; a candidate code extraction means for extracting a candidate code based on the key information acquired by the key information acquisition means when code detection is instructed by the code detection instruction means; and the code detection instruction Pitch information extraction means for extracting pitch information deemed valid at the chord detection timing from the performance data acquired by the performance data acquisition means when a chord detection is instructed by the means, and a candidate code For all chords, the musical importance for each role of the chord components that make up the chord is represented by a point value. A point table by degree name having a point table and each candidate code extracted by the candidate code extracting means when the chord detection is instructed by the chord detection instructing means are extracted by the pitch information extracting means. Role identification and points for identifying the role of each pitch information in each candidate code and extracting a point indicating musical importance according to the identified role from the note point table of the degree name point table When the chord detection is instructed by the extracting means and the chord detection instructing means, for each candidate code extracted by the candidate chord extracting means, the whole pitch information extracted by the pitch information extracting means has the role Sum the points extracted by the identification and point extraction means A point total value acquisition unit that acquires a point total value of a candidate code, and when code detection is instructed by the code detection instruction unit, a point total value of each candidate code acquired by the point total value acquisition unit And a code detecting means for detecting one code from the candidate codes extracted by the candidate code extracting means, and the degree name-specific point table can be selected from a plurality of prepared degree name-specific point tables, Its main feature is that it can be created and edited.

  In the code detection device according to the present invention, a priority point indicating a priority for each of a plurality of codes is set in the degree name point table, and the code detection means is the point total value acquisition means When the point total value is obtained for each candidate code, the priority points of the codes corresponding to each candidate code are further obtained from the degree name point table, and the obtained priority points are summed for the points of each candidate code. It can be configured to add to the value and perform code detection based on the added point total value.

  In the chord detection device according to the present invention, for example, a chord detection timing, that is, chord detection timing, is set on the basis of a predetermined music timing such as every predetermined bar or every predetermined beat. Once the performance data is acquired and key information is acquired, when chord detection is instructed at the set chord detection timing, candidate codes are extracted based on the acquired key information, and chord detection is performed from the acquired performance data Extract pitch information that is considered valid for timing. Next, with respect to all chords that can be candidate chords, each of the extracted chords is provided with a point table by degree name comprising a note point table that represents the musical importance for each role of the chord constituent sounds constituting the chord with point values. For the candidate chords, the role of each pitch information is identified, and the points indicating musical importance according to the identified roles are extracted from the note point table of the degree name point table, and for each candidate chord, all pitches are extracted. Sum the points extracted for information. In other words, at the chord detection timing, the musical roles played by each performance sound, such as which constituent sound, which tension note, and the void note in the candidate chord, are identified and identified. The point value based on each musical importance is acquired. Furthermore, the point value with respect to all the performance sounds is totaled for every extracted candidate code, and the total point of each candidate code is calculated | required. Then, an optimum code is detected from the extracted candidate codes based on the point total value of each candidate code obtained in this way. Furthermore, the degree name-specific point table can be selected from a plurality of prepared degree-name point tables, and can be created and edited, and automatic performance data can be generated based on the detected chords. .

  As described above, according to the present invention, when chord detection based on performance data such as double-handed playing is instructed at chord detection timing corresponding to measure or beat timing, each performance sound candidate input at the chord detection timing is indicated. Identifies the musical role in the chord, obtains point values based on the musical importance for each identified role, sums the point values for all keystroke sounds for each candidate chord, and total points for each candidate chord The candidate code with the highest total point is selected as the optimum desired code.

  Therefore, according to the present invention, an appropriate chord can be detected for a free user performance such as playing both hands, and the user can obtain an automatic accompaniment with an appropriate chord while devoting himself to the performance. Also, from the available chords, based on the user's playing sound, the musical meaning of the role such as chord constituent sounds is considered with the point value to detect the appropriate chord, so it fits musically and playing sound Code can be obtained. Further, in the prior art, when a specific tension sound is played for a certain chord (for example, D # for CMaj), the chord is sometimes controlled not to be detected. The importance of the role of harmony structure in each candidate sound at the time of chord detection is identified, and the chord is determined based on it. Therefore, it is prohibited to detect chords simply by combining performance sounds. Rather, an appropriate chord can be detected by identifying the musical role of the performance sound. In addition, the point table for each degree name can be selected according to the mode of user performance and can be created and edited, so that more appropriate chord detection can be performed.

  In the code detection device according to the present invention, the priority point indicating the priority for each of the plurality of codes is set in the point table by degree name, and at the time of code detection, the priority point corresponding to each candidate code Is added to the point total value of each candidate code, and code detection is performed based on the added point total value. That is, as described above, the highest point value is obtained as a result of adding the priority point indicating the priority of the chord to the total point value obtained in correspondence with the role of all key-pressing sounds in each candidate code. A candidate code is obtained and selected as the optimum code. Therefore, according to the present invention, it is possible to detect a chord more appropriate for the user performance.

1 shows a hardware configuration example of a code detection device according to an embodiment of the present invention. 2 shows an image of code detection timing according to an embodiment of the present invention. 4 shows an example of a point table by degree name according to an embodiment of the present invention. The 1st part of the automatic accompaniment performance processing flow by one Example of this invention is shown. The 2nd part of the automatic accompaniment combined performance performance processing flow by one Example of this invention is shown. The 3rd part of the automatic accompaniment combined performance processing flow by one Example of this invention is shown. 2 shows a note event processing flow according to one embodiment of the present invention. 2 shows a part of a code detection processing flow according to an embodiment of the present invention. The other part of the code | cord | chord detection processing flow by one Example of this invention is shown. 2 shows a part of a role extraction processing flow according to an embodiment of the present invention. The other part of the role extraction processing flow by one Example of this invention is shown. A part of an example of a point table by degree name according to an embodiment of the present invention is shown. The other part of the example according to the point table according to degree name by one Example of this invention is shown.

[Overview of system configuration]
FIG. 1 shows a hardware configuration example of a code detection apparatus according to an embodiment of the present invention. A chord detection apparatus according to an embodiment of the present invention is a kind of computer having an electronic music information processing function, and can be embodied in the form of an electronic musical instrument. As shown in FIG. 1, the chord detection device includes a central processing unit (CPU) 1, a random access memory (RAM) 2, a read only memory (ROM) 3, an external storage device 4, a performance information acquisition unit (performance information input unit). 5), other operation elements 6, a display 7, a sound source 8, an automatic accompaniment device 9, a communication interface (I / F) 10, and the like, and these elements 1 to 10 are connected via a bus 11. .

  The CPU 1 constitutes a data processing unit together with the RAM 2 and the ROM 3, and uses the clock of the timer 12 to execute various music information processing according to a predetermined control program including an automatic accompaniment performance processing program. It is used as a work area for temporarily storing various data necessary for these processes. For example, in the automatic accompaniment performance mode in which automatic accompaniment performance processing is executed, the timer 12 measures various set times set for detecting and applying chords under the control of the CPU 1. Various registers (NoteList, ChordList, Chord, sTime, eTime, Key, ResultList, etc.) necessary for code detection and application are provided. In addition, various control programs and various control data necessary for executing these processes are stored in the ROM 3 in advance.

  The external storage device 4 includes a storage medium such as a semiconductor memory such as an HD (hard disk), an FD (flexible disk), a CD (compact disk), a DVD (digital multipurpose disk), and a flash memory, and a driving device thereof. Arbitrary information such as a control program, degree name point table (PT), control data such as various rules, and various music data can be stored in an arbitrary storage medium. The storage medium may be detachable or may be built in the code detection device, and the detachable recording medium includes a USB memory.

  The performance information acquisition unit 5 includes a performance operator such as a keyboard, and a performance operation detection circuit that detects operation of the performance operator and introduces performance operation information corresponding to the detected contents to the data processing units (1 to 3). Is done. The other operation unit 6 includes other operation units (setting operation units) such as key switches, and a setting operation detection circuit that detects operations of these operation units and introduces corresponding operation information to the data processing unit. The The data processing unit (1-3) controls each part of the chord detection device based on operation information from the performance information acquisition unit 5 and the other operation unit 6, and for example, a performance corresponding to a performance operation input from the performance operation unit 5 Data can be sent to the sound source 8 or can be operated in the automatic accompaniment performance mode by operating a mode setting button in the operation unit 6.

  The display unit 7 includes a display such as an LCD for displaying various information necessary for performance and setting, a built-in lamp attached to the switches in the operation units 5 to 6, and the like. Controls with commands from, and provides display assistance for performance and setting operations.

  The sound source 8 includes a sound source unit and a DSP, and generates a musical sound signal based on performance data obtained from the performance operation information of the performance information acquisition unit 5 or based on music data read from the storage device 4. A predetermined effect is given to this musical sound signal. The sound system 13 connected to the sound source 8 includes a D / A converter, an amplifier, a speaker, and the like, and generates a musical sound based on a musical sound signal output from the sound source 8. In the automatic accompaniment performance mode, the automatic accompaniment device 9 sets the tempo of automatic accompaniment with the tempo clock from the timer 12, and according to this tempo clock, the accompaniment corresponding to the chord provided from the data processing unit (1-3) Generate automatic accompaniment sounds based on style data.

  The communication I / F 10 includes a music wired I / F such as MIDI, a general-purpose network I / F such as USB, or a general-purpose short-range wireless I / F such as wireless Lan, and can communicate with the external electronic music apparatus MM. it can. For example, MIDI format music data can be acquired from the external electronic music device MM and stored in the storage device 4.

[Code detection timing]
In the chord detection apparatus according to one embodiment of the present invention, the chord is detected based on the performance data input from the performance information acquisition unit 5 at a predetermined bar or beat timing by the automatic accompaniment performance processing. Is automatically applied to automatic accompaniment. FIG. 2 shows an example of code detection timing according to one embodiment of the present invention.

  FIG. 2A is an example of the code detection reference position, and the horizontal axis indicates a time axis in which time advances in the right direction. The chord detection reference position T2 indicated by ◎ is a musical timing such as a predetermined bar or beat timing preset by a user operation with respect to the performance data, and instructs the user to detect or apply chords. It becomes a standard of timing. In this example, the first beat and the third beat are set as the chord detection reference position T2 for the performance data of one measure = 4 beats. On the other hand, FIG. 2B is an example of the code detection timing, and the scale of the time axis (horizontal axis) in which the time advances in the direction of the right arrow is larger than that in FIG. .

  The code detection timing starts at a code detection timing start position (time) T1 a predetermined time before the code detection reference position T2, and ends at a code detection timing end position (time) T3 after a predetermined time from the position T2. As described above, the timings T1 to T3 for performing the chord detection are obtained by setting the time widths T1 to T2 and T2 to T3 before and after the predetermined beat position T2, respectively. Then, T1 to T2 = 250 ms and T2 to T3 = 50 ms. Here, when performance data (note event) is input in response to a performance operation (key depression) such as playing both hands during the chord detection timing (time width) T1 to T3, chord detection is instructed at the chord detection timing end time T3. Then, chord detection is performed based on the input performance data (note event) and the key at that time.

  In this chord detection device, performance data sounded within the time interval T1 to T3 is a target of chord detection, and the pitch of each performance sound (note event, note data) constituting the target performance data at time T3. By referring to the point table (PT) by degree name, candidate chords that can be chords in the key set at that time are extracted, and the role of each performance sound identified for each extracted candidate chord is extracted. Find the corresponding musical point (importance). Then, the sound points of the candidate codes are summed, and the candidate code with the highest total point is detected as the optimum code. Further, at the same time as detecting the chord at time T3 in this way, the detected chord is applied to the automatic accompaniment and is sequentially added to the chord detection result list.

  In this case, if the chord detection is performed after removing the ornamental sound or the note data that is presumed to be mistouched from the performance data within the target time width T1 to T3, the accuracy is further improved. Further, the key information used at the time of detecting the chord may be specified by the user before the performance, or may be used at any time along the performance. In the latter case, the key may be detected by the conventional technique.

  During the automatic accompaniment performance process, the chord detection timing start position T1, chord detection timing end position T3, performance data (note event), key information, candidate chord, detected chord, chord detection result list are stored in the RAM 2 respectively. Registers: sTime, eTime, NoteList, Key, ChordList, Chord, and ResultList. Each register name is also used to indicate the contents of the stored information. Each time a chord is detected, the key information (Key) at that time is preferably stored in the chord detection result list (ResultList) in association with the detected chord (Chord).

  The chord detection reference position is not limited to the beat position as in this example, but can be a bar position (for example, every two bars) depending on the tempo and time signature. Further, the time widths T1 to T2 and T2 to T3 before and after the reference position are set in units of ms in this example, but may be set in terms of the note length, the number of beats, the number of clocks, or the like.

[Point table by degree name]
In this code detection apparatus, a point table for each degree name (also referred to as a code-specific point table) is used to obtain points corresponding to each candidate code at the time of code detection. FIG. 3 shows an example of a point table by degree name according to one embodiment of the present invention. The degree name-specific point table PT includes a note point table (columns 5 to 11) for obtaining points of each note (r, 3, 5,. The priority point table (fourth column) for obtaining priority points (Prior) is grouped into one table for Major and Minor keys, and information on whether code functions and diatonic scale codes ( The second and third columns) are added. The degree name point table (Major key) PT in FIG. 3 is a table relating to all codes that can be candidate codes in the Major key (major key), and a similar table is also prepared for the minor key (minor key) (not shown). )

  In FIG. 3, the first column (first column) indicates degree name information (DegreeName) (also referred to as code information, which is simply referred to as a degree name in the following description) prepared in the degree name-specific point table PT. There are many degree names other than those shown. Each degree name (DegreeName) is a frequency based on the key of the key (tonic) [the frequency of the root (root). Example: I, II, III, IV, V, ...] and code type [Example: No symbol (major), m (minor), 7, 6, Maj7 (major 7), m6 (minor 6), m7 (minor) 7), add9 (9th added major),... In the figure, “b” represents “♭” (flat), and the same applies to FIGS. 10, 12, and 13 described later.

  In the degree name point table PT, various information in the second column (second column) and below is recorded corresponding to each degree name (DegreeName) in the first column (first column). The diatonic information in the second column indicates whether the code represented by each degree name (DegreeName) is a diatonic scale code (○) or not (×), and the function information in the third column is the degree information This indicates whether the function of the name (DegreeName) is tonic (T), subdominant (S), dominant (D), or subdominant minor (SM). The priority point information (Prior) in the fourth column represents the priority of the degree name (DegreeName) itself as a point value, and is also called a code priority point or a degree name priority point. On the other hand, the fifth to eleventh columns constitute a note point table section in which note point information representing musical importance is set for each note (role: r, 3, 5,...) Characterizing the chord. ing.

  First, the point information in each column of the fifth to ninth columns represents the musical importance of each chord constituent sound constituting the chord as a point value. Specifically, the route point information in the fifth column indicates a point corresponding to the root (root note) of the chord constituent sound <abbreviation “r”>, and the third-degree point information in the sixth column indicates the chord constituent The 5th point information in the seventh column indicates the point corresponding to the 5th tone <abbreviation “5”> of the chord constituent sound, The fourth voice point information in the 8th column is the fourth voice of the 6th (6th), 7th (7th), or 7th (Maj7th) long voice <abbreviation for the root The point corresponding to “4”> indicates the alternative point information in the ninth column, which is 5 degrees (♭ 5th) or 5 degrees (# 5th) less than the root (root sound) in the chord constituent sound. Points corresponding to the fifth-degree changing sound (alternate chord tone) <abbreviation "a"> are shown.

  The tension note point information in the 10th column points to the musical importance of the tension note <abbreviation “t”> that gives a sense of tension with a non-harmonic sound placed above the basic chord structure of the chord. The other point information in the 11th column (the rightmost column) is a musical importance for other notes that do not correspond to chord constituent sounds or tension notes, such as the void note that should be excluded in the structure of the chord sound. Is represented by a point value.

  In this note point table section (columns 5 to 11), performance sounds (keypress sounds, pitch information) that play the role of root among chord constituent sounds are evaluated as having high importance and higher points are set. Is done. Also, among the constituent sounds, performance sounds that play a role of 3rd or 7th (fourth voice) that are largely involved in chord type determination are also considered important. On the other hand, for the performance sound (others) that causes dissonance in the chord, the importance is lowered and the point is set to be lower.

  In this chord detection device, when chord detection is performed, chords (ChordList) to be candidates are extracted according to the key (Key) when performance data to be detected is input. For example, some combinations of types of codes that can be used in a predetermined key are stored in advance, a desired combination is selected at the time of code detection, and codes corresponding to the selected combination are extracted as candidate codes (ChordList). In addition, in the point table PT by degree name shown in FIG. 3, not only the point information (4th to 11th columns) but also the degree name (DegreeName: 1st) of each code as shown in the 2nd and 3rd columns. Since function information indicating what is a diatonic scale code or function is added to the column), refer to the additional information using the degree name point table PT. Thus, a candidate code (ChordList) can be extracted.

  For example, all the codes (DegreeName) corresponding to the diatonic scale code (second column information = “◯”) of the key at that time may be extracted as candidate codes (ChordList). Alternatively, in the case of code detection after the second time, the function (third column information) of the previously detected code (Chord) is examined. For example, if it is a dominant (D), the process proceeds to a tonic (T). Thus, a chord (DegreeName) that can be musically advanced from the previously detected chord (Chord) may be extracted as a candidate chord (ChordList). Note that such additional information may not be included in the degree name-specific point table PT, and a reference table different from the table PT may be prepared.

  Once the candidate chord (ChordList) is extracted, what role each performance sound (key press sound, Note) included in the target performance data (NoteList) plays in the extracted candidate chord (ChordList) (Which component sound, tension note, or void note) is identified, and points corresponding to the identified roles (each component sound, tension note, and void note) are acquired from the degree name point table PT. Then, the points (importance) for all performance sounds (NoteList) are summed for each candidate chord, and the total points are calculated by adding the priority points (Prior) of the candidate chords themselves. Select a chord and determine that this is the best chord for the target performance data.

  As described above, in the chord detection system according to one embodiment of the present invention, when performance data (NoteList) based on a performance operation such as double-handed playing by a user is input, chord detection timing for every predetermined bar or every predetermined beat. At T1 to T3, a candidate code (ChordList) is extracted according to the key (Key) at that time. Next, for each candidate chord (ChordList), each performance sound (Note) of the input performance data (NoteList) is what sound [root (r), 3rd (3), 5th (5), 4 (4), chord composing sounds such as alter (a), tension notes (t), and void notes). Furthermore, points indicating the musical importance of the identified roles are obtained from the degree name point table PT, and the total points obtained for all performance sounds (NoteList) are obtained for each candidate code to obtain the total points. The total points are obtained by adding the priority point (Prior) of the candidate code itself to the total points. Then, the candidate code with the highest total points obtained is set as the optimum code (Chord).

[Example of automatic accompaniment performance processing flow]
4 to 6 are flowcharts showing an example of the operation of the automatic accompaniment performance processing according to one embodiment of the present invention. In this chord detection apparatus, the automatic accompaniment combined performance mode setting button in the controller 6 is operated to operate the automatic accompaniment combined performance mode when the electronic musical instrument is operating in the normal performance mode. The process is executed and a performance using automatic accompaniment can be performed. When the automatic accompaniment performance process starts, the CPU 1 displays a performance condition setting screen on the display unit 7 in the first step P1, and according to the user operation, the performance tempo, time signature, accompaniment style selection, volume, Set performance conditions such as the performance tone before the performance starts.

  In the next step P2, a chord detection condition setting screen is displayed on the display, and in accordance with a user operation, chord detection timings T1 to T3 and various rules related to chord detection, for example, key acquisition method, chord candidate extraction Set method, degree name-specific point table PT, code priority point (Prior), bottom note point, etc. Here, the bottom note point in the rule is a point that is added when the performance sound with the lowest pitch is the root of the candidate chord in the performance data (NoteList) to be detected. That is, if the bottom note point is further added when the lowest note of the performance sound at the time of chord detection hits the root of the candidate chord, for example, chords of the same constituent sound such as C6 and Am7 can be distinguished. This bottom note point may be a fixed value such as 20 points, or may be set for each chord type. Editing settings may be possible by the user.

  An example of setting various rules will be described. For key acquisition, a method specified in advance by the user, a method specified by operating the key specifying operator in the operator 6 during performance, and as required by the user's performance There is a method of using the detected key (including specification of a specific key detection method). For chord candidates, the method of extracting only the key diatonic scale chord, the method of extracting all the chords that can be used in the key, and the constituent sounds of the chords that can be used in the key are included in the target performance data (NoteList) There is a method of extracting a chord having one sound or more. For the degree name-specific point table, there are setting operations such as selecting a desired type from a plurality of types of tables and editing point values in the table. If code priority points (Prior) and bottom note points are not set in the degree name-specific point table, they are controlled according to control data such as the code priority point table (also called the degree name priority point table) in which these points are set. Set.

  In subsequent step P3, the point table PT by degree name is read. Here, if the degree name point table PT is not a table integrated with the code priority point table as shown in FIG. 3, the code priority point table provided separately is also read. In the next step P4 (FIG. 5), various registers in the RAM 2 used for code detection and application are initialized. These registers include NoteList, ChordList, Chord, sTime (code detection timing start position), eTime (code detection timing end position), Key, ResultList, and the like.

  When the preparations before the performance are completed by the processes of Steps P1 to P4, the first chord detection timing start position (sTime) and the chord detection timing end are determined from the contents of the chord detection timings T1 to T3 set in Step P2 in Step P5. The position (eTime) is calculated and stored in the corresponding start and end position registers: sTime and eTime, respectively, and the process proceeds to Step P6. In the example of FIG. 2, sTime = 250 ms before the beat as the reference position and eTime = 50 ms after the beat as the reference position. However, only the first code detection timing start position (sTime) stored in the start position register: sTime is set to the first code detection reference position T2.

  In Step P6, the user is asked whether or not to return to the normal performance mode without using the automatic accompaniment, and when it is determined that the normal performance mode return button is operated to return to the normal performance mode (P6 = YES), this automatic End the accompaniment performance mode and return to the original normal performance mode. On the other hand, when it is determined that the automatic accompaniment performance mode is to be continued by operating the continue button (P6 = NO), the process proceeds to step P7, and an instruction to start automatic accompaniment by operating the automatic accompaniment start button has been accepted. If no automatic accompaniment start instruction is given (P7 = NO), the process returns to step P6. Further, when an instruction to start automatic accompaniment is received at step P6 (P7 = YES), the timer 12 is started at step P8, and the process proceeds to step P8.

  In Step P9, it is determined whether or not an automatic accompaniment stop instruction by operating the automatic accompaniment stop button has been accepted. When an automatic accompaniment stop instruction is received (P9 = YES), the timer 12 is stopped in Step P10, In step P11, the user is asked whether or not to change the setting of performance conditions such as accompaniment style and chord detection conditions such as chord detection timing. When it is determined that the setting change button is operated and the setting is changed (P11 = YES), the process returns to Step P1 (FIG. 4), and necessary settings and calculations are performed in Steps P1 to P5. Return. On the other hand, when the setting is not changed (P11 = NO), the process immediately returns to step P6.

  When the automatic accompaniment stop instruction is not accepted in step P9 (P9 = NO), the process proceeds to step P12 to determine whether or not a note event based on the performance operation has been accepted. When a note event is accepted (P12 = YES), note event processing (see FIG. 7) is performed in step P13, and then the process proceeds to step P14 (FIG. 6). In the note event processing, note events that occur between chord detection timings T1 to T3 are recorded in the performance data register: NoteList as performance data to be chord detected. If no note event is accepted (P12 = NO), the process immediately proceeds to step P14.

  In step P14, it is determined whether or not the code detection timing is reached, that is, whether or not the end time eTime has been reached. When the time eTime has been reached (P14 = YES), first, in step P15, key information is acquired and adjusted. Store in register: Key. In the next step P16, code candidates corresponding to the key information key are extracted in accordance with the extraction method set in step P2, stored in the candidate code register: ChordList, and the process proceeds to step P17 to execute code detection processing (FIGS. 8 to 8). 9). In the chord detection process, a desired chord Chord is detected from the target performance data (NoteList) and the extracted chord candidate (ChordList), and the detected chord Chord is output to the automatic accompaniment device 9 in the next step P18. In step P19, the code Chord is added to the end of the code detection result register: ResultList, and the process proceeds to step P20. That is, every time a code is detected, the key information at that time is stored in association with the detected code. If there is no chord, that is, if the desired chord cannot be detected by the chord detection process, the process proceeds from step P17 to step P20.

  In step P20, the start and end positions (sTime, eTime) of the next chord detection timing are calculated from the set contents of chord detection timing, and the respective registers: sTime, eTime are updated. In step P21, the performance data and chord are updated. Candidate registers: NoteList and ChordList are initialized. Then, when it is determined in step P14 that the end time eTime has not been reached (P14 = NO) or after the process of step P21, the process returns to step P9 (FIG. 5).

<Note event processing flow example>
FIG. 7 is a flowchart showing an operation example of the note event process performed in step P13 (FIG. 5) of the automatic accompaniment performance process. First, in step N1, the CPU 1 determines whether or not the timer 12 currently indicates a code detection timing (between sTime and eTime). If it is within the range of the code detection timing (N1 = YES), step N2 Thus, it is determined whether or not the note event received in step P12 (FIG. 5) is note-on. If the note is on (N2 = YES), in step N3, the sound source 8 is caused to perform sound generation processing in accordance with the note event, and in the next step N4, the note event information on which the note is turned on is stored in the performance data register. : Add to NoteList. On the other hand, if the note is off (N2 = NO), the musical sound is silenced by the note event in step N5, and the note event information that has been turned off is deleted from the register: NoteList in step N6.

  If it is not the chord detection timing (sTime to eTime) (N1 = NO), the process proceeds to step N7, where it is determined whether the note event received in step P11 (FIG. 4) is note-on. If there is (N7 = YES), in step N8, the sound source 8 performs sound generation processing, and if it is note-off (N7 = NO), the mute processing is performed. After the processes of steps N4, N6, N8, and N9, the note event process is terminated, and the process returns to step P14 (FIG. 6) of the automatic accompaniment performance process.

<Example of code detection processing flow>
8 to 9 are flowcharts showing an example of the chord detection process performed in step P17 (FIG. 6) of the automatic accompaniment performance process. In the first step C1, the CPU 1 determines whether or not the performance data register: NoteList is empty (no data). If not (C1 = NO), the chord candidate register: ChordList is set in step C2. It is determined whether or not it is empty. If it is not empty (C2 = NO), the process proceeds to step C3. If the performance data register: NoteList is empty (C2 = YES) or the chord candidate register: ChordList is empty (C2 = YES), the process returns to step P18 (FIG. 6) of the automatic accompaniment performance process.

  In step C3, one candidate code is converted into a degree name in order from the top of the code candidate register: ChordList (the second and subsequent codes), and this is converted to a degree name register (also referred to as a code information register): DegreeName In the next step C4, the value “0” is entered in the point value register: point. In the subsequent step C5 (FIG. 9), pitch information is extracted from the performance data register: NoteList in order from the top of the note list (the second and subsequent items are the next), and this is extracted as the performance sound register: Store in the Note and proceed to Step C6.

  In step C6, role extraction processing (see FIGS. 10 to 11) for extracting the role of the performance sound Note in the chord indicated by the degree name DegreeName is executed. In the next step C7, referring to the degree name point table PT, the point corresponding to the “role in the code of DegreeName” extracted in step C6 is acquired, added to the point value point, and the process proceeds to step C8. Note that this role can be further divided according to the type of tension note depending on the specification of the degree name point table PT (see FIGS. 12 to 13).

  In step C8, it is determined whether or not the performance sound register: Note indicates the pitch information of the last note event information in the performance data register: NoteList. If it is not the last note event pitch information (C8 = NO), returning to step C5, the processes of steps C5 to C7 are repeated for the next note event information in the performance data NoteList. And if the process of step C5-C7 is performed about the last note event pitch information in performance data NoteList (C8 = YES), it will progress to step C9.

  In step C9, referring to the degree name point table PT, a priority point (Prior) corresponding to the code indicated by the degree name DegreeName is acquired and added to the point value point. In the next step C10, it is determined whether or not the pitch information of the note event information that is the lowest note in the performance data NoteList (the performance tone of the lowest pitch) corresponds to the root of the chord indicated by the degree name DegreeName. When it corresponds to the route (C10 = YES), a bottom note point (for example, 20 points) is acquired in step C11 and added to the point value point. When it is determined in step C10 that it does not correspond to the route (C10 = NO) or after the processing of step C11, the process proceeds to step C12, and the candidate values in the chordList are all candidate codes indicated by the degree name DegreeName. Is stored in association with.

  In the next step C13, it is determined whether or not the code indicated by the degree name DegreeName is the last code in all candidate codes ChordList. If it is not the last code (C13 = NO), the process returns to step C3 to return to all candidate codes. Steps C3 to C12 are repeated for the next candidate code in the ChordList. Then, when the processes of steps C5 to C7 are performed on the last candidate code (C13 = YES), the process proceeds to step C14, and the candidate code having the highest point value point stored in association with each candidate code in all candidate codes ChordList Is stored in the chord register: Chord. When the process of step C14 is completed, the chord detection process is terminated, and the process returns to step P18 (FIG. 6) of the automatic accompaniment performance process.

  When determining one code from among the candidate codes in step C14, if there are a plurality of candidate codes having the highest point value point (when the points are arranged at the same point), one is determined by an appropriate method. To do. For example, a priority rule that defines the characteristics of a code to be selected with priority, such as a chord whose lowest tone matches the route, a code with a small number of detections, or a code with a large number of detections, is prepared and set in step P2. It is preferable to design in advance.

<Role extraction process flow>
10 to 11 are flowcharts showing an example of the code detection process performed in step C6 (FIG. 9) of the code detection process. In the first step F1, the CPU 1 starts from the main tone (tonic) of the key information key and the degree name DegreeName, the chord root, 3rd and 5th [minor 7th flat 5th [m7 (♭ 5)] and augment [aug] Is stored in the “root” register, “3rd” register, and “5th” register in the RAM 2. Here, the numbers of semitones from the main sound are determined for the root, 3rd and 5th sounds according to the chord type. For example, in the case of Key = CMajor, DegreeName = IVMaj7, “root” = F ), “3 degrees” = A (La), and “5 degrees” = C (Do).

  In the next step F2, it is determined whether or not the pitch name of the performance sound information Note is equal to the value of the “root” register. If it is the value of the “root” register (F2 = YES), the performance sound is determined in step F3. The role of the information note is determined to be “root”, otherwise (F2 = NO), the process proceeds to step F4. In step F4, it is determined whether or not the pitch name of the performance sound information Note is equal to the value of the “three times” register. If it is the value of the “three times” register (F4 = YES), the performance sound is determined in step F5. The role of the information note is determined to be “three times”, otherwise (F4 = NO), the process proceeds to step F6. In step F6, it is determined whether or not the note name of the performance sound information Note is equal to the value of the “5th” register. If it is the value of the “5th” register (F6 = YES), the performance sound is determined in step F7. The role of the information Note is determined to be “5 degrees”, otherwise (F6 = NO), the process proceeds to Step F8. When the processes of steps F3, F5, and F7 are completed, the role extraction process is terminated and the process returns to step C7 (FIG. 9) of the code detection process.

  In step F8, it is determined whether or not the code type of the degree name DegreeName is “m7 (♭ 5)” (minor 7 flat 5). If the chord type is “m7 (♭ 5)” (F8 = YES), the process proceeds to step F9, where the note name of the performance note information Note is reduced by 5 degrees (♭) with respect to the note name in the “root” register. 5th), and if it is a decrease of 5 degrees (F9 = YES), the role of the performance sound information Note is determined to be “alternate” in step F10. On the other hand, if the code type is not “m7 (♭ 5)” (F8 = NO), the process proceeds to step F11 to determine whether the code type of the degree name DegreeName is “aug” (an augment). If the chord type is “aug” (F11 = YES), the process proceeds to step F12, and whether the note name of the performance sound information Note is increased by 5 degrees (# 5th) with respect to the note name of the “root” register. If NO in step F10 (F12 = YES), the role of the performance sound information Note is determined to be "alternate" in step F10. When the process of step F10 is completed, the role extraction process is terminated and the process returns to step C7 (FIG. 9) of the code detection process.

  When it is determined in step F11 that the code type is not “aug” (F11 = NO), the process proceeds to step F13 (FIG. 13), and the code type of the degree name DegreeName is “6”, “6sus4” (6 suspended). 4) or “m6” (minor 6). Here, if the chord type is any of “6”, “6sus4”, and “m6” (F13 = YES), the process proceeds to step F14, and the note name of the performance sound information Note is the sound of the “root” register. It is determined whether or not the name is 6 degrees (6th), and if it is 6 degrees (F14 = YES), the role of the performance sound information Note is determined to be “fourth voice” in step F15. On the other hand, when it is determined in step F13 that the code type is neither “6”, “6sus4”, or “m6” (F13 = NO), the process proceeds to step F16, and the code type of the degree name DegreeName is “ It is determined whether or not it is a Maj7 (major 7) system such as “IMaj7” (1 degree / major 7) and “IVmMaj7” (4 degrees / minor major 7). Here, if the chord type is Maj7 type (F16 = YES), the process proceeds to step F17, and whether or not the pitch name of the performance sound information Note is 7 degrees (Maj7th) long with respect to the pitch name of the “root” register. If it is 7 degrees long (F17 = YES), the role of the performance sound information Note is determined to be “fourth voice” in step F15. If it is determined in step F16 that the code type is not Maj7 (F16 = NO), the process proceeds to step F18, and the code type of the degree name DegreeName is “IIm7” (twice / minor 7), “V7” ( It is determined whether it is an m7 (minor 7) system to a 7 (seven) system such as “5 degrees / Seven)” and “VII7sus4” (7 degrees / Seven Suspended 4). Here, if the chord type is the m7 system to the 7th system (F18 = YES), the process proceeds to step F19, and the pitch name of the performance sound information Note is 7 degrees (7th) shorter than the pitch name of the “root” register. In step F15, the role of the performance sound information Note is determined to be “fourth voice” if it is short 7 degrees (F19 = YES). When the process of step F10 is completed, the role extraction process is terminated and the process returns to step C7 (FIG. 9) of the code detection process.

  When it is determined in step F9 that it is not a decrease of 5 degrees (F9 = NO), when it is determined that it is not an increase of 5 degrees in step F12 (F12 = NO), and when it is determined that the length is not 6 degrees in step F14 (F14 = NO) When it is determined at step F17 that it is not 7 degrees long (F17 = NO), when it is determined at step F18 that the code type is neither m7 system nor 7 system (F18 = NO), or it is not short 7 degrees at step F19. (F19 = NO), the process proceeds to step F20, where it is determined whether or not the pitch name of the performance sound information Note is a tension note with respect to the chord of the degree name DegreeName in the current key. Here, the tension note has a semitone distance number from the root sound depending on the chord type (for the root, any of ♭ 9th, 9th, # 9th, 11th, # 11th, ♭ 13th, 13th) 1-3), for example, when Key = CMajor and DegreeName = IVMaj7, the sound is G (So), B (Sh), and D (Le). If it is a tension note (F20 = YES), the role of the performance sound information Note is determined as “tension note” in step F21. Otherwise (F20 = NO), the performance sound information Note is determined in step F22. The role of “Other” is determined. When the processes of steps F21 and F22 are completed, the role extraction process is terminated and the process returns to step C7 (FIG. 9) of the code detection process.

[Point table by degree name (2)]
The degree name point table PT described in FIG. 3 is a table in which note points indicating the importance of the role of notes such as constituent notes, tension notes, and void notes based on the corresponding degree names are set for each possible chord. However, note points can be set in more detail. 12 to 13 show other examples of degree name point tables according to an embodiment of the present invention. In this degree name-specific point table, the same information as the table of FIG. 3 is set in the first to fifth columns, but the sixth to twenty-second columns are shown for the Root in the fifth column. Points (importance) corresponding to the role of the sound indicating the pitch of the street (m3rd, 3rd, 4th, ..., 13th) are set, and each sound including the root in the 5th to 22nd columns is characterized by a chord Role information representing a note (role) is added and displayed in the note point display frame as follows in FIGS.
The symbol “r” at the upper left of the display frame represents the root of the chord component sound, the symbol “3” represents the third tone of the chord component tone, the symbol “5” represents the fifth tone of the chord component sound, The symbol “4” represents the fourth voice of the chord constituent sound [6th, 7th, Maj7th], and the symbol “a” represents the altered chord tone [♭ 5th, # 5th (aug5th)] in the chord constituent sound. The symbol “t” represents a tension note. In addition, if the entire display frame is shaded, it indicates that it is not used due to a different name and the same name (same on the keyboard but different on the keyboard). Etc.).

  By using such a degree name point table PT, it is possible to simultaneously obtain the role and point in the chord for each performance sound (pitch information). Also, not only chord constituent sounds but also tension notes can be set for each type, and chord detection can be performed with high accuracy.

[Various Embodiments]
The preferred embodiment of the present invention has been described above with reference to the drawings. However, this is merely an example, and the present invention can be variously modified without departing from the spirit of the invention. For example, the chord detection timing may be determined by the tempo and the accompaniment style to be used, or may be set by the user. Further, it may be a timing when a specific operation element such as a switch is operated, or may be a predetermined beat timing + an operation element operation timing. For example, when a change in tempo is detected, it may be changed from every 2 beats to every 1 beat. Further, as in the embodiment, a time width (T1 to T3) for receiving input of target performance data may be provided.

In the embodiment, a point table for each degree name (PT) in which note points are set for every chord that can be a candidate is used. In addition, a desired type of point table may be selected by a user operation. Moreover, editing and creation may be possible by the user.
In addition to the note point table, a code priority point table (degree name priority point table) in which a code priority point (degree name priority point) indicating the priority of the degree name itself corresponding to the candidate code is set can be provided. In this case, similarly, a plurality of types of code priority point tables may be prepared and selectable, or the table may be edited or created by the user.
Further, a table for each degree name of the chord as in the embodiment may be used, and the note point table may be a table for each chord type.

  In the embodiment, the role of the key-pressing sound in the candidate chord is finely discriminated such as chord constituent sound, tension note, and void note. For example, two roles of chord constituent sound or non-constituent sound are identified. It may be.

In the embodiment, the automatic accompaniment is generated based on the detected chord according to the performance data, but a harmony sound may be generated based on the detected chord. Or the utilization method of displaying the information of the detected code | cord | chord on a display screen sequentially is also considered.
When storing the detected chords, all the chords detected after the start of the performance may be stored in a list format or the like, or the latest chord or an arbitrary number may be stored and updated for each detection. May be. The detected code may store a code name, or a degree name associated with the key information at that time. Both the code name and the degree name may be stored.

In the embodiment, the performance data to be input is assumed to be real-time performance data by the user. However, for example, a method of detecting chords while sequentially reproducing recorded performance data without prefetching is also conceivable. Alternatively, performance data may be input from a device such as a musical instrument connected to the outside.
The performance data input means (performance operator) is not limited to a keyboard, and may be a stringed instrument type such as a guitar. When the user's performance input sound is an audio signal, at least at each chord detection timing, it is converted into note information having at least sounding timing information and pitch information before referencing performance data to select one chord from candidate chords It only has to be done.

T1-T3 code detection timing [T1 = start (sTime), T3 = end (eTime)],
T2 code detection reference position,
Point table by PT degree name.

Claims (2)

Performance data acquisition means for acquiring performance data;
Key information acquisition means for acquiring key information;
Code detection instruction means for instructing code detection at a predetermined code detection timing;
Candidate code extraction means for extracting a candidate code based on the key information acquired by the key information acquisition means when code detection is instructed by the code detection instruction means;
Pitch information extraction means for extracting pitch information considered valid at the chord detection timing from the performance data acquired by the performance data acquisition means when the chord detection is instructed by the chord detection instruction means;
For all chords that can be candidate chords, a point table for each degree name comprising a note point table that represents the musical importance for each role of chord constituent sounds constituting the chord with a point value;
The role of each pitch information extracted by the pitch information extracting means in each candidate code for each candidate code extracted by the candidate code extracting means when code detection is instructed by the chord detection instructing means Role identification and point extraction means for extracting a point indicating musical importance according to the identified role from the note point table of the degree name point table;
When chord detection is instructed by the chord detection instructing means, the role identification and point extraction are performed for all pitch information extracted by the pitch information extracting means for each candidate code extracted by the candidate chord extracting means. Point total value acquisition means for acquiring the point total value of each candidate code by summing the points extracted by the means;
When code detection is instructed by the code detection instruction means, one of the candidate codes extracted by the candidate code extraction means based on the point total value of each candidate code acquired by the point total value acquisition means Code detecting means for detecting a code,
The degree detection point table can be selected from a plurality of degree name point tables, and can be created and edited. further,
A priority point indicating a priority for each of a plurality of codes is set in the degree name point table,
When the code detection unit acquires the point total value for each candidate code by the point total value acquisition unit, the code detection unit further acquires a priority point of a code corresponding to each candidate code from the degree name point table, The code detection apparatus according to claim 1, wherein the acquired priority point is added to the point total value of each candidate code, and code detection is performed based on the added point total value.

Images