JP6658785B2 - Automatic accompaniment method and automatic accompaniment device - Google Patents

Description

The present invention relates to an automatic accompaniment method and an automatic accompaniment device .

  In a musical instrument having a keyboard such as a piano, an organ, or an electronic keyboard or an electronic piano, it is common to play a melody with the right hand and play an accompaniment with the left hand in accordance with the melody.

  However, in order to perform the performance by moving the right hand and the left hand at the same time, appropriate training is necessary. In particular, although it is relatively easy to move the right hand to play a melody, many beginners find it difficult to play an accompaniment different from the melody with the left hand at the same time. For this reason, an accompaniment corresponding to a left-handed performance is automatically created at the same time that the player plays the melody with the right hand so that even a beginner can enjoy the performance easily. Electronic musical instruments have been realized.

  For example, Patent Literature 1 discloses a technique of an electronic musical instrument that can appropriately determine a chord name based on a melody sound weight and its transition.

JP 2011-158855 A

  In the electronic musical instrument described in Patent Literature 1, although chord judgment is performed using a beat position and a beat weight of a melody sound, a melody sound and a chord name (chord name) at the front of the previous beat, etc. The melody progression and coding of classical music can be quite numerous and different, especially in the field of music, where the coding of melody progression is often determined by human perception and the accuracy of machine coding. Has been required to be further improved.

The present invention has been made in view of the above points, and an object of the present invention is to provide an automatic accompaniment method and an automatic accompaniment apparatus that can quickly perform accurate and appropriate chording based on a melody sound history. I do.

In order to achieve the above object, an automatic accompaniment method according to one embodiment of the present invention is an automatic accompaniment method used for an automatic accompaniment device that generates an accompaniment sound corresponding to a melody to be played, and for each of a plurality of music pieces, A music database corresponding to a melody to be played is searched from a music database in which a code corresponding to the melody is stored as music data, and a code included in the searched music data is determined as a code corresponding to the melody. Then, an accompaniment sound based on the determined chord is generated in correspondence with the generation of a musical tone corresponding to the melody to be played, and in the search for the music data , among the plurality of musical tones constituting the melody , 1 and sound length of the sound-th tone, music de corresponding from the music database, based on the ratio of the durations of the tones of each of the second sound and subsequent Characterized in that it searches the data.

  According to the present invention, based on the history of the melody sound, it is possible to appropriately use both the automatic coding and the coding using the music database to perform the coding. It can be done quickly.

FIG. 1 is a diagram illustrating an appearance of an electronic musical instrument according to the present embodiment. FIG. 2 is a block diagram showing a configuration of the electronic musical instrument according to the embodiment of the present invention. FIG. 3 is a flowchart illustrating an example of a main flow executed in the electronic musical instrument according to the present embodiment. FIG. 4 is a flowchart illustrating an example of keyboard processing according to the present embodiment. FIG. 5 is a flowchart illustrating an example of the song candidate search process according to the present embodiment. FIG. 6 is a flowchart illustrating an example of the F-listing process according to the present embodiment. FIG. 7 is a flowchart illustrating an example of the K list-up process according to the present embodiment. FIG. 8 is a flowchart illustrating an example of the N list-up process according to the present embodiment. FIG. 9 is a flowchart illustrating an example of the D-listing process according to the present embodiment. FIG. 10 is a flowchart illustrating an example of a process of creating a comprehensive list from the N list and the D list according to the present embodiment. FIG. 11 is a flowchart illustrating an example of the code selection process according to the present embodiment. FIG. 12 is a diagram illustrating an example of the music database according to the present embodiment. FIG. 13 is a diagram showing a musical score corresponding to the music database of FIG. FIG. 14 is a diagram showing an example of recorded data of the performance input melody history according to the present embodiment. FIG. 15 is a diagram illustrating an example of the F-list according to the present embodiment. FIG. 16A shows an example of the input melody history according to the present embodiment, and FIG. 16B shows an example of a K list corresponding to the input melody of FIG. 16A. FIG. 17A illustrates an example of the input melody history according to the present embodiment, and FIG. 17B illustrates an example of an N list corresponding to the input melody of FIG. 17A. FIG. 18A is an example of the input melody history according to the present embodiment, FIG. 18B is an example of an N list corresponding to the input melody of FIG. 18A, and FIG. It is a figure showing the example of the D list corresponding to the input melody of 18 (A). FIG. 19 is a diagram illustrating an example of the comprehensive list according to the present embodiment. FIG. 20 is a diagram illustrating an example of the diatonic register according to the present embodiment. FIG. 21 is a diagram illustrating an example of the diatonic scale table according to the present embodiment. FIG. 22 is a diagram illustrating an example of a temporary code determination map according to the present embodiment. FIG. 23 is a diagram illustrating an example of the code database according to the present embodiment. FIG. 24 is a diagram illustrating an example of the code determination table according to the present embodiment. FIG. 25 is a flowchart illustrating an example of the automatic accompaniment process according to the present embodiment. FIG. 26 is a flowchart illustrating an example of a code selection process according to the second embodiment of the present invention.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 is a diagram illustrating an appearance of an electronic musical instrument to which the electronic musical instrument according to the present embodiment is applied. As shown in FIG. 1, an electronic musical instrument 10 according to the present embodiment has a keyboard 11 as a performance operator. On the upper part of the keyboard 11, there are provided switches (see reference numerals 12 and 13) for designating a timbre, start / end of an automatic accompaniment, designation of a rhythm pattern, etc., and various information relating to a music to be played, for example, The display unit 15 displays a tone color, a rhythm pattern, a chord name, and the like. The electronic musical instrument 10 according to the present embodiment has, for example, 61 keys (C2 to C7). In addition, the electronic musical instrument 10 can perform in one of two performance modes, an automatic accompaniment mode in which automatic accompaniment is turned on, and a normal mode in which automatic accompaniment is turned off.

  FIG. 2 is a block diagram showing a configuration of the electronic musical instrument according to the embodiment of the present invention. As shown in FIG. 2, the electronic musical instrument 10 according to the present embodiment includes a CPU 21, a ROM 22, a RAM 23, a sound system 24, a switch group 25, a keyboard 11, and a display unit 15.

  The CPU 21 controls the entire electronic musical instrument 10, detects the depression of a key on the keyboard 11 and the operation of the switches (for example, reference numerals 12 and 13 in FIG. 1), and operates the keys and switches. Various processes such as control of the sound system 24, determination of a chord name according to the pitch of a depressed musical tone, and performance of an automatic accompaniment according to an automatic accompaniment pattern and a chord name are executed.

  The ROM 22 stores various processes to be executed by the CPU 21, for example, a process executed in response to a switch operation or a key depression of any key on the keyboard, a tone generated in response to a key depression, and a tone of a depressed tone. A program for determining a chord name according to the height, an automatic accompaniment pattern, and playing an automatic accompaniment according to the chord name is stored. The ROM 22 stores a waveform data area storing waveform data for generating various musical tones such as a piano, a guitar, a bass drum, a snare drum, and a cymbal, and data indicating various automatic accompaniment patterns (automatic accompaniment data). Is stored in the automatic accompaniment pattern area.

  The RAM 23 stores programs read from the ROM 22 and data generated during the processing. Further, the RAM 23 stores melody history data generated by the CPU 21 in response to key depression / release of keys on the keyboard 11. In the present embodiment, the automatic accompaniment pattern includes a melody automatic accompaniment pattern including a melody sound and an obligato sound, a chord automatic accompaniment pattern including a constituent sound for each chord name, and a rhythm pattern including a drum sound. For example, the record of the data of the melody automatic accompaniment pattern includes the timbre, pitch, sounding timing (sounding time), tone length, and the like of the musical tone. The record of the data of the chord automatic accompaniment pattern includes, in addition to the above information, data indicating the chord constituent sounds. The data of the rhythm pattern includes the timbre of the musical tone and the sounding timing.

  The sound system 24 has a sound source unit 26, an audio circuit 27, and a speaker 28. When the tone generator 26 receives, for example, information about a depressed key or information about an automatic accompaniment pattern from the CPU 21, the tone generator 26 reads out predetermined waveform data from the waveform data area of the ROM 22 and converts musical tone data of a predetermined pitch. Generate and output. The tone generator 26 can also output waveform data, in particular, waveform data of timbres of percussion instruments such as snare drums, bass drums, and cymbals, as tone data. The audio circuit 27 D / A converts and amplifies the musical sound data. As a result, an audio signal is output from the speaker 28.

  Further, the electronic musical instrument 10 according to the embodiment of the present invention includes a music database 30. As described later, the music database 30 is a database used as a dictionary for comparing the input melody history with each music in the database and referring to a code (chord) of a suitable music. FIG. 12 is a diagram showing an example of the actual data structure of the music database 30.

  FIG. 12 exemplifies the beginning of some of the songs starting with "CDE", that is, "Doremi" in the database of various songs. FIG. 13 shows actual music scores corresponding to the data of each music illustrated in FIG. The music database that is actually read by the CPU 21 and used for various processes is, for example, an XML (Extensible Markup Language) file format or the like, with tags and headers, etc., as necessary, configured to be recognized by the CPU 21. It shall be possible.

  As shown in FIG. 12, the song database first records a song number and a song title for each song (the song number column and the title column in FIG. 12). Also, as shown in the top row of the “Key / Beat” column of each song, how many keys the song is recorded (Key) and how many beats it is (Beat) are recorded. I have. For example, “C / 4” is recorded in “fox” of song number 1, indicating that the song is recorded in C major—4 beats. Similarly, “C / 3” is recorded in “Minato” of the music number 6, indicating that the music is recorded in C major and three beats.

  Although FIG. 12 shows an example in which all songs are recorded in C major, even if the songs are unified in other than C major, each song has a database in a different key. Such a configuration is of course acceptable. If all songs are stored in a database with the same key, if the tonality can be determined by a coding algorithm or key determination algorithm described later, the estimated tonality of the melody played by the user is determined. Based on the pitch (the number of semitones) of the tonality (the C key in FIG. 12) used in the database, all the songs registered in the music database are uniformly assigned to the determined tonality. If transposition is performed, the processing can be performed, and the processing can be simplified. On the other hand, it is also possible to register a database in a different key for each song, for example, in the key of the original song of each song.

  Further, at the right of the Key / Beat column (hereinafter, referred to as “Note column”) for each song in FIG. 12, the melody pitch of each song, the pitch with the previous sound, and the ST (Step Time) with the start sound are shown. ) Ratio and chord (chord) for the melody are recorded.

  Referring again to the song number 1 “fox” as an example, the melody of this song is “Doremi Fasososo”, so the top row of the Note column of song number 1 in FIG. Are sequentially recorded as “CDEFGGG”.

  In the second row of the Note column of each music piece in FIG. 12, the pitch between the start sound and the immediately preceding sound is recorded. That is, for the first sound of the melody, the pitch of the start sound (all C in the example of FIG. 12) is recorded as it is, and after the second sound of the melody, the pitch (the number of semitones) with the immediately preceding sound is recorded. ing. For example, in the case of “fox” of song number 1, “C” is recorded as it is for the first sound of the melody, and “D” of the second sound of the following melody is immediately preceding the melody sound (first sound). Since the sound is two semitones higher than the sound “C”, “2” is recorded. Similarly, for the third melody note “E”, the pitch (the number of semitones) “2” with the second melody “D” is recorded, and for the fourth melody tone “F”, , The pitch (the number of semitones) “1” with the third sound “E” is recorded.

  When the melody sound is lower than the immediately preceding sound, as indicated by a part after the music number 2, the pitch with the immediately preceding sound is recorded as a negative value (for example, the music number). The fourth note of "Sparrow's Song" of number 2 and the fifth and sixth notes of "Hato" of song number 7).

  Further, "ST ratio with start sound" is recorded in the third line of the Note column of each song in FIG. Here, ST is an abbreviation of Step Time, and is a numerical value representing the length from the sounding timing of one sound to the sounding timing of the next sound. Accordingly, in the “ST ratio with start sound” of the music database, data representing the ratio of the ST of the start sound (first sound) of the melody to the ST of each note is recorded.

  Here, in FIG. 12, a note having the same length as the ST of the start sound is recorded such that the “ST ratio with respect to the start sound” is 1. Therefore, for example, in the case of the “fox” of the song number 1, all the notes from the second note to the sixth note are eighth notes (there are no rests) of the same length as the start note, so these notes Is recorded as 1 in the “ST ratio with start sound”. Also, the seventh sound (G of a quarter note) of the “fox” of the song number 1 is twice as long as the start sound (eighth note), so the “ST ratio with the start sound” is 2 It is recorded. Further, in the case of the song number 8 “Tonbi”, the start note is a dotted quarter note, whereas the second to sixth notes are eighth notes with a third of the length. The "ST ratio with the start sound" is recorded as "0.33" for these notes.

  Further, in the fourth line of the Note column of each song in FIG. 12, there is a chord column, and a chord name (chord symbol) is recorded. In this chord column, for example, a letter "C" represents a chord of "Domiso" which is a so-called "C" chord. What is expressed as "G7" represents a chord of "Socirefha" which is a so-called "G7" seventh chord. A chord column having “→” means the same chord as the previous chord, and means that the previous chord is maintained without being changed. The correspondence between the chord name and the actual sound is made by referring to a chord name correspondence table (not shown) stored in a ROM or the like, for example.

  This chord column is referred to when a chord to be added to a melody as an accompaniment is determined in a chord selection process described later. The chords that can be added to a certain melody are not always one type, and vary depending on the arrangement, the player's individuality and sensitivity, etc. From the viewpoint of performing the above, it is desirable to prepare a chord table such as "At least, if this chord is added, it is not wrong".

  The data format of the music database is not necessarily limited to the format described here. For example, based on MIDI format data, the ST ratio to the starting sound is calculated from the length of each note in the melody, Alternatively, the pitches of the notes before and after the melody sound may be compared, and the pitch with the immediately preceding note may be obtained by a calculation process. Further, the chord portion may have chord data, or the chord data may be executed in a form in which a chord symbol or a plurality of notes representing a certain chord is recorded on a separate track from the melody. .

  In the normal mode, the electronic musical instrument 10 according to the present embodiment generates musical tones based on depression of keys on the keyboard 11. On the other hand, the electronic musical instrument 10 is set to the automatic accompaniment mode by operating an automatic accompaniment switch, which is one of the various switches 12 and 13 shown in FIG. 1, for example. In the automatic accompaniment mode, a tone of the pitch of the key is generated by depressing the key. Further, a chord name is determined based on the information on the key that has been depressed, and a musical tone is generated according to the automatic accompaniment pattern including the chord constituent sound of the chord name. The automatic accompaniment pattern includes a melody automatic accompaniment pattern with a change in pitch, such as a piano, guitar, and bass, a chord automatic accompaniment pattern, and a rhythm pattern without a change in pitch, such as a bass drum, a snare drum, and a cymbal. be able to. Hereinafter, a case where the electronic musical instrument 10 operates in the automatic accompaniment mode will be described.

  Hereinafter, processing executed in the electronic musical instrument 10 according to the present embodiment will be described in detail. FIG. 3 is a flowchart illustrating an example of a main flow executed in the electronic musical instrument according to the present embodiment. Although not shown, a timer increment process for incrementing the counter value of the interrupt counter is also executed at predetermined time intervals during execution of the main flow.

  As shown in FIG. 3, when the power of the electronic musical instrument 10 is turned on, the CPU 21 of the electronic musical instrument 10 executes an initial process (initialization process) including clearing and clearing of data in the RAM 23 and an image on the display unit 15. (Step S301). When the initial process is completed, the CPU 21 detects each operation of the switches constituting the switch group 25, and executes a switch process for executing a process according to the detected operation (step S302).

  For example, in the switch processing (step S302), operation of various switches such as a tone color designation switch, a designation switch of an automatic accompaniment pattern type, and a switch for designating ON / OFF of an automatic accompaniment pattern is detected. When the automatic accompaniment pattern is turned on, the CPU 21 switches the performance mode to the automatic accompaniment mode. Data indicating the performance mode is specified in a predetermined area of the RAM 23. Data indicating the type of the timbre and the automatic accompaniment pattern is also stored in a predetermined area of the RAM 23.

  Next, the CPU 21 executes a keyboard process (step S303). FIG. 4 is a flowchart illustrating an example of keyboard processing according to the present embodiment in more detail. In the keyboard processing, first, the CPU 21 scans the keys of the keyboard 11 (step S401). An event (key on or off) as a key scanning result is temporarily stored in the RAM 23 together with information on the time when the event occurs. The CPU 21 determines whether or not there is an event for a certain key with reference to the key scanning result stored in the RAM 23 (step S402). If it is determined as Yes in step S402, the CPU 11 determines whether the event is key-on (step S403).

  If Yes in step S403, that is, if it is determined that the event is a key-on, the CPU 21 executes a tone generation process for the key for which the key-on was performed (step S404). In the tone generation process, the CPU 21 reads out the tone color data for the melody key and the data indicating the pitch of the key, which are designated by the tone color designation switch and stored in the RAM 23, and temporarily stores the data in the RAM 23. . The data indicating these timbres and pitches is given to the sound source unit 26 in a sound source tone generation process (step S306 in FIG. 3) described later, and the sound source unit 26 stores the data in the ROM 22 according to the data indicating the timbres and pitches. The waveform data is read to generate tone data having a predetermined pitch. As a result, a predetermined musical sound is generated from the speaker 28.

  Thereafter, the CPU 21 stores the pitch information (for example, key number) and the key pressing timing (for example, key pressing time) of the key for which the key is turned on in the RAM 23 (step S405). The key depression timing can be calculated based on the counter value of the interrupt counter.

  If No is determined in step S403, it means that the event is key-off. Therefore, the CPU 21 executes a mute process for the key whose key has been turned off (step S406). In the silencing process, the CPU 21 generates data indicating the pitch of a musical tone to be silenced, and temporarily stores the data in the RAM 23. Also in this case, in the sound source tone generation process (step S306) described later, data indicating the tone color and pitch of the musical tone to be silenced is given to the sound source unit 26. The sound source 26 silences a predetermined musical tone based on the given data. After that, the CPU 21 stores the key pressed time (key pressed time) in the RAM 23 for the key whose key has been turned off (step S407). Further, in storing the melody history, a ratio (ST ratio to the start sound) between the sound length of the melody sound input first (ST: Step Time) and the sound length of the melody sound input this time is calculated. Then, it is added to the melody history information and stored.

  The CPU 21 determines whether the processing has been completed for all key events (step S408). If No is determined in step S408, the process returns to step S402. When the processing has been completed for all key events (step S408: Yes), the keyboard processing of FIG. 4 ends.

  When the keyboard process (Step 303 in FIG. 3) is completed, the CPU 21 executes a music candidate search process (Step S304). FIG. 5 is a flowchart illustrating an example of the song candidate search process according to the present embodiment in more detail.

  In the song candidate search process, first, the CPU 21 determines whether or not a melody has been input by the key press / key release process detected in the keyboard process described above with reference to FIG. 4 (step S501). If there is no melody input (step S501: No), the music candidate search process of FIG. 5 ends.

  When there is a melody input (step S501: Yes), the input melody is recorded as a melody history in a predetermined melody history storage area of the RAM 23 in addition to the already recorded melody history (step S502). ). FIG. 14 shows an example of the data structure of the melody history recorded in the RAM 23.

  In the example of FIG. 14, in accordance with the data structure of the above-described music database (FIG. 12), the input pitch, the pitch with the immediately preceding tone (start tone), and the ST with the start tone in response to the input of the melody. Calculate and store the ratio. The specific meaning and calculation method of each data are the same as those described in FIG. Of course, as described in the description of FIG. 12, it is also possible to store the data in another data structure in accordance with, for example, changing the data structure in FIG.

  At the time of recording the melody history in step S502, for example, a sound having an extremely short input sound length can be excluded from the recorded sounds of the melody history in order to remove a slight mistouch or the like. For example, of the input melody, for a sound whose key pressing time (duration) is equal to or shorter than a predetermined threshold value, the melody history is not recorded, and the processing of step S503 and the subsequent steps is not performed, and the music candidate search is performed. Processing such as returning from the processing may be performed.

  When the recording of the melody history in step S502 ends, the CPU 21 determines whether or not the input melody sound is the first melody sound (step S503). If the input melody sound is the first melody sound (step S503: Yes), an F list up process is performed (step S504). The F-listing process is shown in detail in FIG. 6, and will be described with reference to FIG.

  The F-list to be listed in FIG. 6 is a list obtained by searching the above-mentioned music database (FIG. 12) for a match of the first sound among the inputted melody sounds.

  Since the music is compared only with the first sound, it is not always possible to narrow down to one music. In consideration of the case where the user transposes and plays, it is not always possible to perform accurate narrowing down. For example, assuming that the user often plays the original melody key when playing the melody, a certain amount of music can be obtained by searching the music database recorded with the original music key using the first sound of the input melody sound. Can be narrowed down.

  In FIG. 6, first, the CPU 21 acquires the start sound, that is, the information of the first sound of the input melody as the variable FN (step S601). Next, the CPU 21 searches for a music whose melody start sound FN matches a sound registered as a start sound of the music database (a start sound of the melody of each music) (step S602).

  Then, the CPU 21 lists each music searched in step S602 as an F list (step S603). At this time, when there are a plurality of corresponding songs, all of them may be listed, or a predetermined number of randomly selected songs may be listed. If there is no corresponding music, a flag indicating that there is no corresponding music may be returned.

  FIG. 15 shows an example of the F list created in this way. Here, an example of the F-list obtained by searching the music database shown in FIGS. 12 and 13 when the first sound “C” of the melody input shown in FIG. 14 is input is shown. In this case, the first sound “C” of the melody input matches all the songs of the song numbers 1 to 9 in the song database, so all the songs of the song numbers 1 to 9 are listed in the F list.

  The rankings given to the music pieces are, by way of example only, a case where the music database is searched and the music pieces determined to be applicable are sequentially stored in a list. Here, a case is shown where the music database is stored in the list in order from 1 and the ranking is given as it is. However, besides this, it is also possible to record all of the plurality of music pieces in the same order (first place).

  When the F list is created by the processing shown in FIG. 6 through step S504 in FIG. 5 in this way, the CPU 21 returns the processing to the flowchart in FIG. 5, and in step S511, the F list is converted into a comprehensive list ( Step S511), the song candidate search process ends.

  On the other hand, returning to step S503, if the input melody sound is not the first sound (step S503: No), it is determined whether or not the Key (tonality) for the input melody has been determined. (Step S505).

  Whether or not the key (tonality) has been determined can be determined based on whether or not the key has been determined as a “determination key” in a key (tonality) determination process in step S305 in FIG. 3 described below. It is possible. The details of the key (tonality) determination processing will be described later.

  If it is determined in step S505 that the Key (tonality) has been determined (step S505: Yes), the CPU 21 proceeds to step S506 to list up a K list. The K list-up process is shown in detail in FIG. 7, and will be described with reference to FIG.

  The K list to be listed in FIG. 7 refers to each of the songs recorded in the above-described song database (FIG. 12) based on the determination result that the key (tonality) of the input melody has been determined. After transposing to the determined Key (tonality), the list is compared with the input melody history.

  By performing such processing, even when the user transposes the melody to an arbitrary key and plays it, regardless of the key (tonality) of each song registered in the song database, the song can be quickly played. To find and specify the correct code.

  In FIG. 7, first, the CPU 21 acquires information on the key of the input melody determined to be determined as the variable Key (step S701). Next, the CPU 21 compares the tune with the input melody based on the key (tonality) of each tune registered in the above-described tune database (FIG. 12) and the final tone of the melody acquired in the variable Key. (Step S702).

  This processing will be described more specifically. For example, consider a case where the final tone of the input melody acquired by the variable Key is F (F major). Now, suppose that a comparison is made with "Ause Sleeping" of song number 4, since the registered key of this song is "C", that is, in C major, the semitone between Key "F" and Key "C" Calculate the number. Since this can be calculated as “the number of semitones = 5” based on the well-known music theory, each tone of the melody (the top line of the Note column of the song number 4 in FIG. 12) registered with the song number 4 “Ause sleeping” is , The number of semitones +5, and transposes to the F key. In the case of the above example, the original melody “CDECCDE” is changed to “FGAFFGA” by the transposing operation of the number of semitones + 5.

  Then, each sound of the melody sound of the transposed music database is compared with each sound of the input melody history. As a result of the comparison, a music piece satisfying a predetermined condition is picked up as a search result in step S702.

  Here, various conditions can be considered as the conditions for pickup. For example, it is possible to pick up music in the music database in which all pitches from the first sound to the latest sound match the input melody history. Alternatively, it is possible to pick up a predetermined number of music pieces from the past sound of the input melody history to the current sound in order from the music piece with the largest number of matching sounds. Further, a method of comparing sounds from the first sound to the current sound and picking up a predetermined number of music pieces in descending order of the number of matching sounds (probability) is also conceivable.

  Subsequently, as a result of the search in step S702, the CPU 21 determines whether or not any music has been picked up (step S703). If no music has been picked up, the K list listing process ends (step S703: No).

  On the other hand, if the picked-up song exists (step S703: Yes), the ST ratio of each sound of the input melody history to the melody start sound (first sound) is calculated and picked up. For each song, the data is compared with the data of “ST ratio with start sound” recorded on the song database, and “correct answer rate” is calculated (step S704).

  Various methods are also possible for “calculation of the correct answer rate” in step S704.

  For example, the ST ratio between each sound of the input melody history and the melody start sound is determined, and based on the difference between the ST ratio of the input melody and the ST ratio registered on the music database, “ A method of determining the “accuracy rate” and adding (averaging) them may be considered. That is, for example, for the second note “D” of song number 4 “Ause sleeping”, the ST ratio = “1” is recorded as a length equal to the start tone on the song database, but on the contrary, In the melody actually played, if the length of the second sound is 90% or 110% of the melody start sound, the correct answer rate for the second sound is calculated. For example, "90 points".

  As described above, in step S704, “the accuracy rate of each song”, which is an index corresponding to how close (similar) each song is to the input melody history, can be calculated. Then, referring to the “correct rate of each song”, a predetermined number of songs are listed as a K list in order from the one with the highest correct rate (step S705), and the process of increasing the K list is ended.

  FIG. 16 shows an example of the K list created in this way.

  Here, FIG. 16B shows an example of a K list obtained by searching the music database shown in FIGS. 12 and 13 when a melody as shown in FIG. 16A is input. .

  As shown in FIG. 16A, four sounds “FGAB #” are input in order for the melody input. Also, it is assumed that F (F major) has been obtained as a confirmation key based on this input melody. Here, assuming that the melody input shown in FIG. 16A is a melody input by the player playing the keyboard 11, the human performance may include some fluctuation and inaccuracy. As an example in consideration of this point, in the example of FIG. 16A, the second note “G” is pressed a little shorter, and the ST ratio with the start sound is 0.9, which is slightly smaller than 1. It shows the case where it has become. The third sound “A” indicates that the key is pressed a little longer and the ST ratio with the start sound is 1.1, which is slightly larger than 1.

  When the K list-up process of FIG. 7 is performed on such an input melody, first, the key (key) determined in the variable Key in step S701, that is, “F” is recorded. Next, in FIG. For each song in the twelve song databases, the pitch connection is compared after transposing to "F" which is a determination key. In other words, of the songs in each music database, those having a pitch of “FGAB 音” transposed in F major, that is, the pitch concatenation of “CDEF” from the first to fourth notes in the C major. The music that has is searched. Therefore, in this case, the song number 1 “fox” and the song number 3 “frog song” are extracted as the corresponding songs whose pitch concatenation matches.

  Since there is a corresponding song, step S703 becomes Yes, and the process proceeds to step S704, and the correct answer ratio P of the ST ratio is calculated for each corresponding song. Here, using the ST value of the i-th note of the corresponding song in the song database: STdb (i) and the ST value of the i-th note of the input melody: STin (i), the following expression is used. The obtained example is shown.

  As shown in the equation (1), here, for the ST value of each sound, the absolute value of the difference between the value on the database and the input value is divided by the ST value of the sound on the reference database to obtain a ratio. This value is defined as a 100% value, and among the sounds input up to this point, the values from the second sound to the latest sound are averaged, and the value obtained by subtracting 100 from the average is used as the “correct answer rate”.

  That is, when the input melody of FIG. 16A is compared with the “fox” of the music number 1, the difference value of the ST value for the second sound is 0.1, so that Dividing by the ST value (= 1) results in 0.1, which is 100 fractionalized to 10. Similarly, the third note has a score of 10, and the fourth note has a score of 0. When these are added, 20 points are obtained.

  Therefore, their average value is (10 + 10 + 0) ÷ 3 = approximately 6.7 points, which is subtracted from 100 points to obtain the correct answer rate = 93.3 points.

  In the equation (1), the loop starts from i = 2, and the reason why the ST value ratio of the first note is not included in the calculation is that the ST ratio is determined based on the first note. Therefore, the ST ratio of the first sound is always 100% and is excluded.

  In this manner, when the correct answer rate of each music piece whose pitch concatenation coincides is obtained, in step S705, the K list is listed in descending order of the score. In the example of FIG. 16, there is no difference in the score of the correct answer between the song number 1 “fox” and the song number 3 “frog song”. With respect to such ranking of songs having the same score, the ranking can be determined by the same method as described in the F list.

  When the K list is created by the processing shown in FIG. 7 through step S506 in FIG. 5 in this way, the CPU 21 returns the processing to the flowchart in FIG. 5, and in step S510, the K list is converted into a comprehensive list ( (Step S510), the song candidate search process ends.

  On the other hand, if it is determined in step S505 that the Key (tonality) has not been determined (step S505: No), the CPU 21 proceeds to step S507 to list up the N list. The details of the N list-up process are shown in FIG. 8, and will be described with reference to FIG.

  The N-list listed in FIG. 8 is a comparison between the pitch of the input melody history and the melody pitch data (the top line data of the Note column) of each song recorded in the above-mentioned song database (FIG. 12). Go and list up.

  In FIG. 8, first, the CPU 21 acquires input pitch data as pitch linked data from the input melody history data (FIG. 14) (step S801). Considering the example of the input as shown in FIG. 14, assuming that the input has been made up to the third tone of the melody, in this step S801, data "CDE" is acquired as the pitch concatenation of the input melody. You.

  Subsequently, the CPU 21 compares the acquired linked pitch data of the input melody with the melody pitch data of each song in the song database (FIG. 12) (the top row data in the Note column), and the pitches match. The music is searched (step S802).

  For example, when the input melody is “CDE” as described above, when the music database of the example of FIG. 12 is searched, all nine tunes from tune number 1 to tune number 9 shown in FIG. Is "CDE" in order from the start sound, so in this case, in step S802, all nine songs in FIG. 12 are picked up as search results. Here, in addition to picking up music pieces whose pitches match from the first sound to the latest sound, search can be performed by various methods as in the music pickup method in step S702. is there.

  Subsequently, as a result of the search in step S802, the CPU 21 determines whether or not there is any music that has been picked up (step S803). If no music has been picked up, the N list listing process ends (step S803: No).

  On the other hand, if the picked-up song exists (step S803: Yes), the ST ratio of each sound of the input melody history to the melody start sound (first sound) is calculated and picked up. For each song, the data is compared with the data of “ST ratio with start sound” recorded in the song database, and “correct answer rate” is calculated (step S804). The specific method of calculating the correct answer rate is the same as that described in step S704.

  Then, referring to the "correct rate of each song" calculated in step S804, a predetermined number of songs are listed as N lists in order from the one with the highest correct rate (step S805), and the N list-up process is performed. To end.

  FIG. 17 shows an example of the N list created in this way.

  Here, FIG. 17B shows an example of an N list obtained by searching the music database shown in FIGS. 12 and 13 when a melody as shown in FIG. 17A is input. .

  As shown in FIG. 17A, three tones “CDE” are input in order for the melody input. In the example of the melody input shown in FIG. 17A, it is assumed that all the CDEs have the same length (ST value) and the ST ratio to the first sound is 1.

  When such an input melody is subjected to the N-listing process shown in FIG. 8, first, in step S801, the pitch concatenation “CDE” is extracted from the melody history data shown in FIG. 17A. Next, in step S802, pitch concatenation is compared for each song in the song database of FIG. Here, unlike the case of listing the K list, the key is not determined in the listing of the N list in this figure, so that the comparison is performed using the pitch of the music database as it is. Therefore, in this case, since all the songs from song number 1 to song number 9 have the pitch of “CDE” from the start to the third tone, all nine songs are extracted as candidates in step S802.

  Since there is a corresponding song, step S803 becomes Yes, and the process proceeds to step S804 to calculate a correct answer ratio P of the ST ratio for each corresponding song. Here, the calculation of the correct answer rate is performed using Expression (1), similarly to step S704 in the case of the K list.

  As an example, a case will be described in which a correct answer rate is obtained by comparing with "Tonbi" of music number 8. In this case, the difference value of the ST value of the second note “D” is 0.67, so that when divided by the ST value (= 0.33) of the sound in the database, it is about 2. (For convenience, 0.33 = 1/3 and 0.67 = 2/3). Therefore, this value is converted into 100 to be 200. Similarly, for the third sound “E”, the difference value ratio is calculated as 200.

  Therefore, the average value is (200 + 200) ÷ 2 = 200 points, which is subtracted from 100 points to obtain the correct answer rate = −100 points (− is indicated by ▲ in the figure).

  When the correct answer rate of each music is obtained in this way, in step S805, the N lists are listed in descending order of the score. FIG. 17B shows an example of the N list created in this way. The correct answer rate is the same for some songs, and this ranking shows an example in which the songs are ranked in ascending order of song numbers, as in the F list and K list. The ranking can be done by the method.

  When the N list is created by the processing shown in FIG. 8 through step S507 of FIG. 5 in this way, the CPU 21 returns the processing to the flowchart of FIG. 5, advances the processing to step S508, and executes the processing of the D list. Make a list. This D-listing process is shown in detail in FIG. 9 and will be described with reference to FIG.

  The D-list listed in FIG. 9 uses the data of the pitch relation of the input melody history and the data of the pitch relation of the melody of each music (the second line of the Note column) recorded in the above-mentioned music database (FIG. 12). The comparison is performed.

  In the N list of FIG. 8, the input melody is compared with the pitch itself of the music database, whereas in the D list of FIG. 9, the comparison is made using a pitch relationship. Thus, even when the input melody is transposed and played in a different key from the melody registered in the music database, the comparison can be performed using the relative pitch relationship registered in the music database. Thus, music can be compared regardless of the tonality of the input melody.

  In FIG. 9, the CPU 21 first compares the data of the “starting sound / preceding sound” obtained by comparing the pitches of two adjacent sounds from the input melody history data (FIG. 14). It is acquired as interval data (step S901).

  For example, if the input melody is now “FGA”, the pitch relation is “2”, which is the pitch between the first sound “F” and the second sound “G”, and “2”, which is the second sound “G”. "2", which is the pitch relation of the third sound "A", is obtained.

  Subsequently, the CPU 21 compares the acquired pitch interval data of the input melody with the “pitch between the start tone and the immediately preceding tone” of each song in the song database (FIG. 12) (data in the second line of the Note column). Then, music pieces having the same interval are searched for (step S902).

  Referring again to the example in which the above-mentioned “FGA” is input, the pitch relation of the input melody is “2, 2” in order from the beginning, whereas 9 is registered in the music database of the example of FIG. Since all of the songs have a pitch relationship of "2, 2" in order from the beginning, all the nine songs in FIG. 12 are picked up as a search result in step S902.

  In this case, in view of the point of the above-mentioned D-list that "the music can be searched even when the melody is input by transposing to a different key from the music database", the absolute values of the start melody of the input melody and the music are absolute. The different pitches are not compared. Explaining the above example in which "FGA" is input, the pitch "F" of the start sound of the input melody is different from the pitch "C" of the start sound of each song, but this is compared. Instead, try to pick up different songs.

  As a result, when the above “FGA” and the melody are input, in the search processing of the N list in FIG. 8, since the pitch is completely different from the pitch of each music, it is not picked up. In the search processing of, music pieces starting with “CDE” are picked up even if the pitches themselves are different.

  Subsequently, as a result of the search in step S902, the CPU 21 determines whether or not there is any music that has been picked up (step S903). If no music has been picked up, the D list listing process ends (step S903: No).

  On the other hand, if the picked-up music exists (step S903: Yes), the ST ratio of each sound of the input melody history to the melody start sound (the first sound) is calculated and picked up. For each song, the data is compared with the data of “ST ratio with start sound” recorded on the song database, and “correct answer rate” is calculated (step S904). The specific method of calculating the correct answer rate is the same as that described in step S704.

  Then, with reference to the “accuracy rate of each music piece” calculated in step S904, a predetermined number of music pieces are listed as a D list in descending order of accuracy rate (step S905). To end.

  FIG. 18 shows an example of the D list created in this way.

  Here, when a melody as shown in FIG. 18A is input, an example of an N list obtained by searching the music database shown in FIGS. 12 and 13 is shown in FIG. 18B. 18C is shown in FIG.

  As shown in FIG. 18A, three sounds “FGA” are input in order for the melody input. In the example of the melody input shown in FIG. 18A, it is assumed that all the FGAs have the same length (ST value) and the ST ratio with the first sound is 1.

  As described above, comparing the melody input of FIG. 18A with the pitch concatenation of each song in the song database of FIG. 12, there is no song that starts with “FGA”. As shown in FIG. 12), none of the tunes 1 to 9 in FIG. 12 are listed. If another song is registered in the song database, and there is a song starting with “FGA”, the song is picked up in the N list of FIG. 18B.

  When the D-listing process in FIG. 9 is performed on such an input melody, first, in step S901, the first sound (starting sound) to the interval interval data from the melody history data in FIG. The pitch “2” of the second sound and the pitch “2” of the second to third sounds are extracted. Next, in step S902, the pitch intervals of the respective songs in the song database of FIG. 12 are compared with the extracted pitch intervals.

  Here, in the listing of the D list in this figure, unlike the listing of the N list, pitch intervals are compared instead of pitch concatenation. Therefore, even when the input melody is transposed and played in a different key from the melody registered in the music database, the comparison can be performed using the relative pitch relationship registered in the music database. it can. As a result, there is a possibility that music that is not listed in the N list in FIG. 18B may be listed as the D list.

  Also for the input of FIG. 18 (A), the interval between the first tone (starting tone) and the second tone of all nine songs from song number 1 to song number 9 in the song database of FIG. Since the pitch is “2” and the pitch between the second and third notes is “2”, in step S902, all these nine songs are extracted as candidates.

  Since there is a corresponding song, step S903 becomes Yes, and the process proceeds to step S904 to calculate the correct answer ratio P of the ST ratio for each corresponding song. The calculation of the correct answer rate is performed using Expression (1), similarly to step S704 for the K list and step S804 for the N list.

  When the correct answer rate of each music piece is obtained in this way, in step S905, the D list is listed in descending order of the score. FIG. 18C shows an example of the D list created in this way. For some songs, the correct answer rates are tied. In this ranking, as in the case of the F list, K list, and N list, an example is shown in which songs are ranked in ascending order of song number. , You can also rank in other ways.

  When the D list is created by the processing illustrated in FIG. 9 through step S508 in FIG. 5 in this way, the CPU 21 returns the processing to the flowchart in FIG. 5, advances the processing to step S509, and proceeds to step S507. A comprehensive list is created using the N list and the D list created in S508 (step S509). The details of the process of creating the comprehensive list are shown in FIG. 10 and will be described with reference to FIG.

  In FIG. 10, first, the CPU 21 integrates the N list and the D list, arranges them in order by a predetermined method, and makes a comprehensive list (step S1001).

  Here, as a method of integrating and arranging in order, for example, in accordance with the correct answer rate of the ST ratio of each song in each of the N list and the D list calculated in step S804 and step S904, regardless of the list type, First, there is a method of arranging music pieces in descending order of the correct answer rate of the ST ratio.

  In addition, for example, when the pitch concatenation or interval is compared in step S802 and step S902, the number of pitches in which the pitch concatenation or interval between the input melody history and the music in the music database match. A method of arranging songs from both the N list and the D list with the item having the highest number as the top is possible.

  Further, a method of integrating the N list and the D list by comprehensively evaluating the comparison result of the pitch connection or the interval and the accuracy rate of the ST ratio is also possible. In this case, one of the evaluations, for example, the one having a large number of notes having the same pitch connection or pitch interval is ranked high, and the one having the same number of matched notes is ranked based on the accuracy rate of the ST ratio. A method of prioritizing the correct answer rate of the ST ratio, and a method of performing a predetermined weighting, adding the evaluation points of both items, and arranging the music pieces in descending order of the weighted and added evaluation points, etc. Is possible.

  When a comprehensive list is created from the N list and the D list by such a method, the CPU 21 searches the overlapping list for a duplicate song (step S1002). This is equivalent to searching for duplicate songs that are listed in both the N list and the D list.

  If there is a duplicate song in step S1002, and if such a duplicate song exists in the general list, it is inconvenient for the subsequent processing, so the CPU 21 deletes the duplicate song from the general list (step S1002). S1003), eliminating duplicate songs from the comprehensive list. The number of overlapping songs is not limited to one, and even if a plurality of songs overlap, they are all deleted.

  FIG. 19 shows an example of the general list created by using the N list and the D list by the processing shown in FIG. In the comprehensive list of FIG. 19, each song is arranged in the order of the correct answer rate, and the duplicate song is deleted.

  As described above, in the song candidate search process shown in FIG. 5, a comprehensive list of song candidates according to the status of the key (tonality) determination process described later is created based on the input melody.

  When the song candidate search process of FIG. 5 is completed through step S304 of FIG. 3 in this way, the CPU 21 returns the process to the flowchart of FIG. 3 and performs a key (tonality) determination process (step S305). The key (tonality) determination processing is executed by, for example, having a diatonic register 2000 as shown in FIG. 20 in the RAM 23 and updating each data.

  In the present embodiment, the CPU 21 stores a value in a series of items of the diatonic register 2000 every time the melody sound is pressed. In the example of FIG. 20, a series of values are stored in time series for each of the five melody sounds (reference numerals 2001 to 2005). In FIG. 20, the value is a key depressed timewise in the direction of arrow t. That is, the keys are pressed in the order of “C”, “D”, “E”, “F”, and “B” as in the melody sound item.

  In the present embodiment, the values of the following items for a plurality of melody sounds are stored in the unit registers 2001 to 2005 of the diatonic register 2000 by the CPU 21. Each of the unit registers 2001 to 2005 has items of a melody sound, a note length, a provisional key, a provisional code, a provisional function, a melody sound history, a key candidate register, and a decision key, and the unit register has a value for each item. Can be stored. The melody sound stores a note name of a pressed key. The key length stores the key depression time of the key. It is to be noted that the ST value may be recorded with the music database together with the music duration.

  Finally, when the key (tonality) is determined, the CPU 21 stores the key name in the item of the determined key in the unit register (for example, stores the unit register 2005). However, in order for the CPU 21 to determine that the key has been determined, a plurality of key presses are required. Therefore, in the present embodiment, a temporary key is specified by the processing of the CPU 21 and the key name is stored in the item of the temporary key in the unit register until it is determined that the key can be determined. Under the provisional key, a provisional chord name appropriate for the melody sound is stored in the provisional chord item by the CPU 21. Further, in the item of the provisional function, the CPU 21 operates the provisional chord function under the provisional key (the chord name when the main tone is I, and the tonic (T), dominant (D), subdominant (S ) Is stored.

  In the melody sound history, note names of keys depressed are accumulated at the start of the performance or at a predetermined timing. For example, the unit register 2001 for the first key press stores only the pressed key C, and the unit register 2002 for the next key press stores two keys CD. The key candidate stores one or more possible key names when the key is pressed.

  The narrowing down of key (tonality) candidates from the melody history data is executed by the CPU 21 referring to the diatonic scale table 2100 in FIG.

  The diatonic scale table 2100 stores key scale notes (sounds corresponding to musical scales) for each of the twelve keys C to B in a distinguishable manner. For example, if the key is C, note names of C, D, E, F, G, A, and B are stored (see reference numeral 2101), and if the key is G, G, A, B, C, D, E, The pitch name of F # is stored (see reference numeral 2102).

  Then, the CPU 21 compares the melody sound history of FIG. 20 with the diatonic scale table 2100, and determines whether there is a key whose tone name included in the melody sound history is entirely included in the diatonic scale of a certain key. Is examined. Such a key may not exist, or a plurality of keys may exist. For example, C, D, and E are stored in the melody sound history of the unit register 2003. Therefore, referring to the diatonic scale table 2100, there are three keys C, G, and F that include all of C, D, and E as diatonic scales. Therefore, in this case, three keys C, D, and E can be key candidates. (Key candidate field of unit register 2003)

  Then, when the number of key (tonality) candidates becomes one, the CPU 21 sets the candidate key (tonality) as a decision key (tonality). On the other hand, when there are two or more key candidates, the CPU 21 stores, for example, the key with the smallest key signature among the candidate keys as the temporary key, as the value of the temporary key in the unit register. When the key signature is the same (for example, F and G, D and B), the key of the ♯ system is preferentially used as the temporary key.

  As described above, when the key (tonality) determination processing is performed in step S305 of FIG. 3, the CPU 21 performs the automatic code determination processing (step S306).

  The automatic chord determination process is a process of automatically determining a chord (chord) to be attached to the current melody based on the melody sound input up to the present time, and can be realized by various methods.

  For example, based on the current melody sound, it is possible to determine a chord using a temporary chord determination map 2200 as shown in FIG. This method can be used, for example, in a case where a sufficient melody history has not been obtained yet, for example, in the case of the first note or about several notes after the start.

  Further, particularly when the key is not determined, the code can be determined using the code database of FIG. As shown in FIG. 23, the chord database 2300 stores, for each chord name, chord constituent sounds and scale notes related to the chord. In FIG. 23, for example, note names indicated by hatching are chord constituent sounds.

  For example, the CPU 21 selects a predetermined number of sounds in descending order of sound length from a melody history of a predetermined length. As the melody history of a predetermined length, for example, a melody history of a predetermined bar or a melody history of a predetermined number of sounds can be determined.

  The CPU 21 compares a predetermined number of sounds selected from the longest ones with the chord database 2300 and determines whether there is a chord having a chord constituent sound including the predetermined number of sounds. If a chord is found, the chord can be determined as a code for the automatic chord determination processing, and if not found, the chord can be determined by a method such as reducing the number of selected sounds and checking again with the chord database.

  It is also possible to determine a code using a code determination table as shown in FIG. 24 according to the function of the previous code and the progress of the next melody.

  Here, the function of the chord refers to “tonic (TO)”, “subdominant (SU)”, and “dominant (DO)” in music theory, and the left end of FIG. Have been.

  In the chord determination table shown in FIG. 24, the function of the previous chord (either tonic (TO), subdominant (SU) or dominant (DO)) and the combination of the previous melody sound PM and the current melody sound CM , The code name is obtained. Note that the code determination table 2400 in FIG. 24 is for the key (C). Therefore, in the case of another key, the pitch between the key tone of the key and C is set as an offset, and the previous melody when the key is set to C in consideration of the offset from the actual previous melody sound PM and the current melody sound CM. The sound PM and the current melody sound CM may be calculated and used. Although not shown in the example of FIG. 24, a chord name may not be obtained depending on the combination of the previous melody sound PM and the current melody sound CM.

  Furthermore, it is also possible to execute the key (tonality) determination processing in step S305 and the automatic code determination processing in step S306 by using another known technique. For example, it is possible to use a method disclosed in JP-A-2011-158855 or JP-A-2012-68548.

  When the automatic code determination process is performed in step S306 in FIG. 3 as described above, the CPU 21 executes a code selection process (step S307). This code selection processing is shown in detail in FIG. 11, and will be described with reference to FIG.

  The chord selection processing executed in FIG. 11 includes the chord progression of the music listed in the music candidate search processing executed in FIG. 5 and subsequent steps through step S304 and the automatic chord judgment processing executed in step S306. This is a process of determining which of the determined chords should be used for performing the automatic accompaniment.

  In FIG. 11, the CPU 21 first determines whether or not there is a song in the comprehensive list that matches a predetermined number of pitches / intervals, for example, 10 or more notes (step S1101). Regarding the number of sounds, the “Music Theme Encyclopedia” (Ongaku Tomosha) allows you to search for songs from a sequence of up to six sounds. May be searched for about 10 songs. From this, for example, if it is checked whether about 6 to 10 sounds match, the music can be searched with a certain degree of accuracy. If the melody input has not yet reached the predetermined number of sounds (for example, 10 sounds), this processing may be skipped.

  If there is no pitch / pitch link in the general list that matches the predetermined number (step S1101: NO), the CPU 21 proceeds to step S1104, outputs a code according to the real-time coding result, and executes this code. The selection process ends (step S1104). It should be noted that even if the songs corresponding to the F list, the K list, and the N list / D list are not searched in the first place and the songs are not picked up in the comprehensive list, it is determined NO in step S1101. At 1104, a code according to the real-time coding result is output.

  On the other hand, when there is a music piece whose pitch / pitch connection is equal to the predetermined number in the comprehensive list (step S1101: YES), the CPU 21 determines whether or not there is a music piece whose sounding timing is matched by 80% or more. It is determined (step S1102). Specifically, it is determined whether or not any of the songs listed in the comprehensive list has a “correct answer rate” of 80 or more. As described above, the correct answer rate is an evaluation value relating to the sounding timing obtained by comparing the value of the “ST ratio with the start sound” between the music database and the input melody history. By selecting a song having a value equal to or greater than a predetermined value, it can be determined that the user is playing a song that is almost the same as the database, and a chord progression can be selected from the song database to be used as an automatic accompaniment code. Note that the reference value of “80 points” can be changed as appropriate and implemented using a different reference value. Further, for example, it is also possible to accept a designation relating to “proficiency” with a button (not shown) or the like, and to change the reference value of the judgment according to the proficiency.

  If there is no song whose sounding timing matches 80% or more (step S1102: NO), the CPU 21 proceeds to step S1104, outputs a code according to the result of the real-time coding, and ends the present code selection process. (Step S1104).

  On the other hand, if there is a song whose sounding timing matches 80% or more (step S1102: YES), the CPU 21 determines the song selected in step S1101 and step S1102, specifically, the pitch / pitch concatenation. Are referred to a song database (FIG. 12) of songs that match a predetermined number and whose sounding timings match a predetermined rate or more (hereinafter referred to as “applicable songs”). Then, the code of the portion estimated as the current melody in the data of the corresponding song is read from the song database, and it is determined whether or not the code matches the result of the automatic code determination process (real-time code assignment) in step S306. (Step S1103).

  If the result of the automatic chord determination processing (real-time chording) matches the code read from the music database of the corresponding song (step S1103: YES), the CPU 21 proceeds to step S1104, and proceeds to step S1104. The code according to the attachment result is output, and the present code selection process ends (step S1104).

  On the other hand, if the code addition result of the automatic code determination processing (real-time code addition) does not match the code read from the music database of the corresponding music (step S1103: NO), the CPU 21 proceeds to step S1105, The code read from the song database of the corresponding song is output, and the code selection process ends (step S1105).

  When the chord selection process shown in FIG. 11 is thus executed via step S307 in FIG. 3, the CPU 21 returns the process to the flowchart in FIG. 3, and executes the automatic accompaniment process in step S308 (step S308). S308). This automatic accompaniment process is shown in detail in FIG. 25, and will be described with reference to FIG.

  FIG. 25 is a flowchart illustrating an example of the automatic accompaniment process according to the present embodiment. First, the CPU 21 determines whether the electronic musical instrument 10 is operating in the automatic accompaniment mode (step 2501). If it is determined Yes in step 2501, it is determined whether or not the current time has reached the execution timing of the event for the melody sound data in the automatic accompaniment data by referring to a timer (not shown) of the CPU 21. (Step 2502).

  The automatic accompaniment data includes data of three types of musical sounds, namely, melody sounds (including obrigato sounds), chord sounds, and rhythm sounds. The melody sound data and chord sound data include the pitch, sounding timing, and sounding time for each musical sound to be sounded. The rhythm sound data includes the sounding timing of each musical sound (rhythm sound) to be sounded.

  If Yes is determined in the step 2502, the CPU 21 executes a melody sounding / muting process (step 2503). In the melody sound generation / silence processing, it is determined whether or not the event involved in the processing is a note-on event. A note-on event can be determined when the current time substantially coincides with the sounding timing of a predetermined musical tone in the melody sound data. On the other hand, a note-off event can be determined by the fact that the current time substantially matches the time obtained by adding the sounding time to the sounding timing of the musical tone.

  If the event involved in the process is a note-off event, the CPU 21 executes a mute process. On the other hand, if the event involved in the processing is a note-on event, the sound generation processing is executed according to the melody sound data.

  Next, the CPU 21 refers to a timer (not shown) of the CPU 21 to determine whether or not the current time has reached the execution timing of the event for the chord sound data in the automatic accompaniment data (step 2504). If it is determined as Yes in step 2504, the CPU 21 executes chord sounding / muting processing (step 2505). In the chord sound generation / silence processing, the sound generation processing is executed for the chord sound that has reached the sound generation timing, while the mute processing is executed for the chord sound that has reached the mute timing.

  Thereafter, the CPU 21 determines whether or not the current time has reached the execution timing of the event for the rhythm data in the automatic accompaniment data (step 2506). When it is determined Yes in step 2506, the CPU 21 executes a rhythm sound generation process (step 2507). In the rhythm sound generation process, a note-on event is generated for the rhythm sound that has reached the sound generation timing.

  When the automatic accompaniment process (step S308 in FIG. 3) ends, the CPU 21 executes a sound source sounding process (step S309). In the sound source tone generation process, the CPU 21 provides the tone generator 26 with data indicating the tone color and pitch of the musical tone to be pronounced or indicates the tone color and pitch of the musical tone to be silenced based on the generated note-on event. The data is provided to the sound source unit 26. The tone generator 26 reads the waveform data from the ROM 22 in accordance with data indicating a tone color, a pitch, a tone length, and the like, and generates predetermined tone data. As a result, a predetermined musical sound is generated from the speaker 28. Further, based on the note-off event, the CPU 21 instructs the sound source 26 to mute the pitch indicated by the note-off event.

  When the sound source tone generation process (step S309) ends, the CPU 21 executes other processes (for example, displaying an image on the display unit 15, turning on / off an LED (not shown): step S310), and executes step 302. Return to

  As described above, in the present embodiment, an electronic musical instrument that automatically adds an accompaniment to a melody played by a player, includes a music database, and includes a melody performed by the player and each music in the music database. Is compared with. Then, when a song determined to match under the predetermined condition is found, the code registered in the song database is read out, and the accompaniment is performed to the performance melody of the player.

  In this way, the performance melody of the performer is determined not only by the automatic chord identification when the accompaniment is automatically added in real time, but also when the music matches the music registered in the music database. Since the chord matching the music can be read and output, the accuracy of the automatic chording can be improved.

  Furthermore, in the present embodiment, in the comparison processing with the music database, when the key (tonality) determination is not yet determined, an N list for comparing pitches and a D list for comparing pitch intervals are used. , So that comparison and collation can be performed even when the melody is played in a key different from that registered in the music database. Therefore, even when the player plays the melody in a different key from the database, it is possible to effectively use the music database to perform the chording.

  Furthermore, in the present embodiment, the ST ratio between the melody start sound and each melody sound is registered in the music database, and the ST ratio between the start sound and each sound for the played melody is acquired, and The correct answer rate is calculated by comparing with the registered value in the database. Then, the music in the music database is searched preferentially from the one with the highest correct answer rate. Therefore, even if the music is played with a slightly shifted timing of sound generation, it is possible to construct a flexible system that is searched as a candidate music if the characteristics of the music generally match.

  In addition, in parallel with the music database search, automatic chord discrimination processing is also performed, and if the corresponding music is not found in the music database search, chords are automatically assigned based on the result of chord discrimination. ing. Therefore, by performing the coding using the dual structure of the automatic code discrimination and the music database search method, the music registered in the database can be promptly subjected to accurate coding using the database, and the corresponding music can be stored in the database. When there is no such code, when it is impossible to determine, etc., it is possible to attach a code by an automatic chord identification process, and it is possible to perform accurate chording and automatic accompaniment without becoming impossible to code.

  Next, a second embodiment of the present invention will be described. In the second embodiment, the CPU 21 executes a code selection process shown in FIG. 26 instead of the code selection process shown in FIG.

  In the code selection processing of FIG. 26, steps that perform the same processing as in FIG. 11 are given the same step numbers as in FIG. 11, and description thereof will be omitted. In the code selection process in FIG. 26, step S2603 is executed instead of step 1103 in FIG.

  In step S2603, the CPU 21 determines whether the chord added to the performance melody in the automatic chord determination processing in step S306 and the currently played tune in the tune in the tune database determined to satisfy the conditions in S1101 and S1102. Compare the "code function" with the code attached to the melody part.

  Specifically, as the function of the chord attached to the music database, it is assumed that information of the function of the chord is additionally recorded in the music database shown in FIG. 12 for each melody sound. Alternatively, since the key (tonality) of each song is known in the song database of FIG. 12, the function is determined by comparing and matching the tonality with the code for each melody sound. It is good.

  Further, as the function of the chord attached to the melody part currently being played, the function recorded in the provisional function part described in the key (tonality) determination part in step S305 may be used. The function of the chord of the melody being played (temporary function) is compared with the information of the temporary key or the determination key described in step S305 and the information of the chord (chord) attached to the current melody. Then, the function can be determined.

Specifically, it is possible to prepare a table using well-known music theory for various chords and the functions of the chords, and to acquire the function of the chord by referring to this table. For example, the code name corresponding to the tonic, there is a _{"I Maj", "I M7", "III min", "III m7", "VI min", "VI m7".} Code names corresponding to the subdominant include “II _min ”, “II _m7 ”, “II _m7 ⁽⁻⁵⁾ ”, “IV _Maj ”, “IV _M7 ”, “IV _min ”, and [IV _mM7 ]. Further, as code names corresponding to the dominant, there are “III _Maj ”, “III ₇ ”, “III _7sus4 ”, “V _Maj ”, “V ₇ ”, “V _7sus4 ”, and “VII _m7 ⁽⁻⁵⁾ ”. .

  In this manner, in step S2603, the CPU 21 determines whether the code added to the performance melody in the automatic chord determination processing in step S306 and the music in the music database determined to satisfy the conditions in S1101 and S1102. Compare the "chord function" with the chord attached to the melody part currently being played.

  If the functions of the codes are common (step S2603: YES), the code as the real-time coding result is output (step S1104). On the other hand, if the functions of the chords are different (step S2603: NO), the chord read out from the music database of the corresponding music is output (step S1105).

  By doing so, in the second embodiment, when the “code function” of the code read as a result of the real-time coding and the code read from the music database match, the code having the same function is used. As a variety at our company, while outputting the code of the real-time coding result while respecting the result of the real-time coding, if the `` code function '' is different, the code of the real-time coding result In order to avoid the danger of giving outlying misplaced chords, the music database code can be read and output. Therefore, by cooperating with both the music database and real-time coding, the possibility of outputting a large out-of-order code is reduced. It is possible to output code that respects

  Although some embodiments of the present invention have been described above, these embodiments are merely examples and do not limit the technical scope of the present invention. The present invention can take other various embodiments, and various changes such as omissions and substitutions can be made without departing from the gist of the present invention. These embodiments and modifications thereof are included in the scope and gist of the invention described in this specification and the like, and are also included in the invention described in the claims and the equivalents thereof.

  Further, the method described in each of the embodiments described above may be a program that can be executed by a computer, such as a recording medium such as a magnetic disk (floppy disk, hard disk, etc.), an optical disk (CD-ROM, DVD, etc.), a semiconductor memory, etc. And can be applied to various devices, or transmitted by a communication medium and applied to various devices. A computer that realizes the present apparatus reads a program recorded on a recording medium, and executes the above-described processing by controlling the operation of the program.

In this specification, the steps of describing a program recorded on a recording medium are not only performed in chronological order according to the order, but are not necessarily performed in chronological order. This also includes the processing to be executed.
Further, in the present specification, the term “system” means an entire device including a plurality of devices and a plurality of means.

  For example, a part of the processing described in FIG. 3 of the above embodiment may be performed by changing the order of the processing. For example, the order of the song candidate search processing of S304 and the key (tonality) determination processing and the automatic chord determination processing of S305 and S306 can be reversed. By doing so, the latest determination result can be used in determining whether or not “Key fixed?” In step S505.

  According to the present invention described above, based on the history of the melody sound, it is possible to appropriately use both the automatic coding and the coding using the music database to perform the coding. It is possible to perform quick coding.

  That is, according to the present invention, when a corresponding song is not found among the songs registered in the song database means, a code is assigned by automatic coding. Therefore, even when the corresponding music piece is not found in the music piece database, it is possible to avoid a situation in which the chord is not added and no automatic accompaniment is added to the performance of the player.

  On the other hand, according to the present invention, when a corresponding song is searched for among the songs registered in the song database means, an automatic accompaniment is added using the code registered in the song database means. be able to. Therefore, it is possible to prevent a situation in which when a chord is automatically discriminated by the chording based on the automatic chording, a human acoustically apparently strange or wrong chord is added.

  Further, according to the present invention, when a corresponding song is searched for among the songs registered in the song database means, a code between the code by automatic coding and the code registered in the song database is used. The function to be used is compared and the code to be used is selected. Therefore, there is a noticeable error in that the coding by automatic coding is different from the code registered in the music database in terms of the function of the code, and a large loss code is added. Can be prevented.

  Furthermore, if the function is the same as the chord registered in the music database, by selecting automatic chording, after confirming that the chord function is not a big loss, Since it is possible to select a chord that respects the chording algorithm by attaching, it becomes possible to perform automatic accompaniment using a more appropriate chord.

Hereinafter, the inventions described in the claims of the present application will be additionally described.
[Appendix 1]
Song database means in which melody information and a code corresponding to the melody information are stored as song data for each of the plurality of songs,
Performance storage means for storing performance information for sequentially instructing the generation of musical sounds in response to the operation of the performance operator;
Music search means for searching music data having melody information corresponding to the performance information from the music database means based on the performance information stored in the performance storage means;
Code determination means for determining a chord from the performance information stored in the performance storage means,
Code selection means for selecting which code to use, out of the code stored corresponding to the music searched by the music search means and the code determined by the code determination means,
Automatic accompaniment means for instructing the generation of an accompaniment sound based on the chord selected by the chord selection means,
An automatic accompaniment device comprising:

[Appendix 2]
The code selecting means compares the function of each of the codes stored corresponding to the music data searched by the music searching means and the code determined by the code determining means, and a comparison result of the functions of the codes. 2. The automatic accompaniment device according to Supplementary Note 1, wherein a chord is selected based on the chord.

[Appendix 3]
The code selecting means, when the functions of the code stored corresponding to the music data searched by the music searching means and the code of the code determined by the code determining means match, the code determining means 3. The automatic accompaniment device according to claim 2, wherein the selected chord is selected, and when the functions of the chord do not match, the chord stored in the music data is selected.

[Appendix 4]
The chord selection means selects a chord determined by the chord determination means when the music search means does not search a music piece having melody information corresponding to the performance information from the music piece database means. 4. The automatic accompaniment device according to any one of supplementary notes 1 to 3.

[Appendix 5]
The music search means, based on the ratio between the length of the first tone and the length of each of the input tones, of the musical tones sequentially instructed by the performance operator, 5. The automatic accompaniment device according to any one of Supplementary notes 1 to 4, wherein a search is made by assigning a ranking to corresponding songs.

[Appendix 6]
The code determination means includes key determination means for determining the tonality of the performance information of the performance operator,
The music search means transposes the music data stored in the music database means based on the key determined by the key determination means, and, based on the transposed music data and the performance information, 6. The automatic accompaniment device according to any one of supplementary notes 1 to 5, wherein a corresponding music piece is searched.

[Appendix 7]
When the key determination by the key determination unit is not determined, the music search unit performs a music search based on pitch data included in melody information corresponding to the music data stored in the music database unit, and 7. The automatic accompaniment apparatus according to appendix 6, wherein music is searched for in combination with music search based on relative interval data between high data.

[Appendix 8]
8. The automatic accompaniment apparatus according to any one of supplementary notes 1 to 7, wherein each music data stored in said music database means is stored in a single key.

[Appendix 9]
8. The automatic accompaniment apparatus according to any one of supplementary notes 1 to 7, wherein each music data stored in the music database means is recorded in a key of the original music.

[Appendix 10]
For each of a plurality of music pieces, music database means in which melody information and a code corresponding to the melody information are stored as music data, and performance information for sequentially instructing generation of musical sounds in response to operation of a performance operator. An automatic accompaniment method used for an automatic accompaniment device having
Based on the performance information stored in the performance storage means, searching the music database means for music data having melody information corresponding to the performance information,
From the performance information stored in the performance storage means, determine a chord,
Selecting which code to use from among the code stored corresponding to the music searched by the music search means and the code determined by the code determination means,
An automatic accompaniment method for instructing generation of an accompaniment sound based on a chord selected by the chord selecting means.

[Appendix 11]
For each of a plurality of music pieces, music database means in which melody information and a code corresponding to the melody information are stored as music data, and performance information for sequentially instructing generation of musical sounds in response to operation of a performance operator. A computer used as an automatic accompaniment device comprising:
A music search step of searching the music database means for music data having melody information corresponding to the performance information, based on the performance information stored in the performance storage means; and performance information stored in the performance storage means. A code determining step of determining a code from
A code selecting step of selecting which code to use from among the stored code and the determined code corresponding to the searched music,
An automatic accompaniment step for instructing the generation of an accompaniment sound based on the selected chord,
Automatic accompaniment program to execute.

DESCRIPTION OF SYMBOLS 10 ... Electronic musical instrument, 11 ... Keyboard, 12, 13 ... Switch, 15 ... Display part, 21 ... CPU, 22 ... ROM, 23 ... RAM, 24 ... Sound system, 25 ... Switch group, 26 ... Sound source part, 27 ... Audio Circuit, 28 speaker, 30 music database, 2000 diatonic register, 2001 to 2005 unit register, 2100 diatonic scale table, 2200 temporary code determination map, 2300 code database

Claims (8)

An automatic accompaniment method used in an automatic accompaniment device that generates an accompaniment sound corresponding to a melody to be played,
For each of a plurality of songs, a song database corresponding to a melody to be played is searched from a song database in which a code corresponding to the melody is stored as song data,
A code included in the searched music data is determined as a code corresponding to the melody,
Generating an accompaniment sound based on the determined chord in correspondence with the generation of a musical tone corresponding to the melody to be played;
In the search of the music data, from the music database, based on a ratio of a tone length of a first tone and a tone length of each tone after the second tone among a plurality of tones constituting the melody. An automatic accompaniment method characterized by searching for corresponding music data. An automatic accompaniment method used in an automatic accompaniment device that generates an accompaniment sound corresponding to a melody to be played,
For each of a plurality of songs, a song database corresponding to a melody to be played is searched from a song database in which a code corresponding to the melody is stored as song data,
A code included in the searched music data is determined as a code corresponding to the melody,
Generating an accompaniment sound based on the determined chord in correspondence with the generation of a musical tone corresponding to the melody to be played;
At a plurality of timings at which generation of a musical tone is repeatedly instructed in response to a continuous operation of a performance operator, control is performed such that music data is searched from the music database based on the operation history of the performance operator at each timing. And
Matching the said performance operator to first start playing is operated timing, and the first note of a tone was first operated, the tone of the first note of the music data in the music database The music data to be played is selected as the music data to be used for the determination of the chord, and at the timing when the performance operator is operated for the second time or more after the performance is started, each of the plurality of musical tones included in the operation history is played. ratio and, based on said comparison result between the plurality of musical respective durations ratio of contained in each musical piece data of the music in the database, automatically you and selects the music data to be used for the determination of the code Accompaniment method. An automatic accompaniment method used for an automatic accompaniment device that generates an accompaniment sound corresponding to a melody to be played,
For each of a plurality of songs, a song database corresponding to a melody to be played is searched from a song database in which a code corresponding to the melody is stored as song data,
A code included in the searched music data is determined as a code corresponding to the melody,
Generating an accompaniment sound based on the determined chord in correspondence with the generation of a musical tone corresponding to the melody to be played;
At a plurality of timings at which generation of a musical tone is repeatedly instructed in response to a continuous operation of a performance operator, control is performed such that music data is searched from the music database based on the operation history of the performance operator at each timing. And
If the tonality based on the operation history is determined at the timing when the performance operator is operated, the music data in the music database is transposed to the determined tonality, and then the operation history is changed. Is compared with the ratio of the length of each of the plurality of musical tones included in each piece of transposed music data, based on the result of the comparison. automatic accompaniment how to and selects the data. An automatic accompaniment method used in an automatic accompaniment device that generates an accompaniment sound corresponding to a melody to be played,
For each of a plurality of songs, a song database corresponding to a melody to be played is searched from a song database in which a code corresponding to the melody is stored as song data,
A code included in the searched music data is determined as a code corresponding to the melody,
Generating an accompaniment sound based on the determined chord in correspondence with the generation of a musical tone corresponding to the melody to be played;
At a plurality of timings at which generation of a musical tone is instructed repeatedly according to a continuous operation of the performance operator, control is performed such that music data is searched from the music database based on the operation history of the performance operator at each timing. And
If the tonality based on the operation history is not determined at the timing when the performance operation element is operated for the second time or more after the performance is started, the pitch or pitch of a plurality of musical tones included in the operation history If, by comparing the plurality of musical respective pitch or interval included in each music data in the music database, based on the comparison result, you and selects the music data to be used for the determination of the code automatic accompaniment method. A comparison result between the pitches of the plurality of musical tones included in the operation history and the respective pitches of the plurality of musical tones included in the music data in the music database, and the pitches of the plurality of musical tones included in the operation history. 5. The automatic accompaniment according to claim 4 , wherein music data to be used for the determination of the chord is selected based on a result of comparison with a plurality of musical tones included in each music data in the music database. Method. Each time a performance operator is operated, an instruction is issued to generate a musical tone corresponding to the operated performance operator,
Searching music data from the music database based on the operation history of the performance operator,
At a plurality of timings at which generation of musical tones is instructed repeatedly according to the continuous operation of the performance operator, music data is searched from the music database based on the operation history of the performance operator at each timing. Control,
The automatic accompaniment method according to any one of claims 1 to 5 , wherein: The automatic accompaniment method according to any one of claims 1 to 6 , wherein the plurality of musical tones constituting the melody are a plurality of musical tones corresponding to the performance of one musical instrument. An automatic accompaniment device for generating an accompaniment sound corresponding to a melody to be played, using the automatic accompaniment method according to any one of claims 1 to 7 .