JP6341050B2 - Performance support device - Google Patents

Description

  The present invention relates to a technique for generating a singing voice by an operation as if playing a musical instrument.

  In recent years, there are electronic musical instruments that generate a singing voice by an operation as if playing a musical instrument. For example, a singing voice having a pitch corresponding to the pressed key is generated in response to a key pressing operation on an input device simulating a piano keyboard. Hereinafter, generating a singing voice in response to a performance operation such as pressing a key is referred to as “playing a singing voice” or “playing a singing voice”. In the performance of such a singing voice, the lyrics of the singing voice to be generated according to the performance operation are often determined in advance, and are not interesting. Accordingly, various techniques have been proposed that allow a performer to perform a singing voice while designating lyrics, and examples thereof include techniques disclosed in each of Non-Patent Documents 1 to 3.

 Non-Patent Document 1 discloses a “VOCALOID keyboard” which is a real-time Japanese singing keyboard instrument that enables singing voice performance using singing synthesis. This VOCALOID keyboard has a Japanese input interface having 16 buttons corresponding to each of consonant, vowel, muddy point, and semi-voiced point, and a pitch input device imitating a 12-scale piano keyboard. The Japanese input interface has an optimized button layout so that it can be operated with one hand, and it can input arbitrary Japanese characters by combining Romaji input and Kana input. The user of this VOCALOID keyboard operates the Japanese input interface with one hand (for example, the left hand) and inputs the lyrics at the same time, and at the same time designates the pitch when the lyrics are pronounced to the pitch input device. Singing can be performed by performing the operation with the other hand. Non-Patent Document 2 discloses a technique in which a left key (for example, a left half key) of a piano keyboard is used for lyrics input such as Japanese input. Non-Patent Document 3 discloses a technique for realizing singing voice performance by presetting lyrics in advance and selecting one of the lyrics to be used for performance.

[online], Internet <URL: http://www.interaction-ipsj.org/2012/catalogue/catalogue_interactive3.html> [online], Internet <URL: http://www.nicovideo.jp/watch/sm22254895> [online], Internet <URL: http://otonanokagaku.net/nsx39/>

  However, the techniques disclosed in each of Non-Patent Documents 1 to 3 have the following problems. First, the techniques disclosed in each of Non-Patent Document 1 and Non-Patent Document 2 require that vowels and consonants be simultaneously keyed and difficult to play. There was a problem that it was extremely difficult. The technique disclosed in Non-Patent Document 3 has problems such as pre-preparation of preset lyrics, and lyrics can only be written within the preset lyrics range, and the degree of freedom of writing is low. It was. In addition, Non-Patent Document 3 has a problem that preset lyrics cannot be browsed in a lump, and the selection of lyrics is hindered.

  The present invention has been made in view of the above-described problems, and provides a technique that makes it possible to easily perform a singing voice with a high degree of freedom of writing without requiring preliminary preparation such as presetting lyrics. With the goal.

  In order to solve the above problems, the present invention provides an assigning unit that assigns a candidate character string composed of one or more characters representing the pronunciation of lyrics to each of a plurality of keys to which different pitches are assigned, and When a selection operation is performed on one of a plurality of keys, the sound represented by the candidate character string assigned to the key that has been selected is sounded at the pitch corresponding to the key. A pronunciation control means for giving a voice synthesis instruction to the singing voice synthesizing apparatus, and a candidate character string assigned to each of the plurality of keys by the assigning means to a candidate character string assigned to the key subjected to the selection operation There is provided a performance support apparatus characterized by having selection means for reselecting in response.

  The candidate character string may be a character string composed of phonetic characters such as hiragana, katakana, and alphabet, or may be a character string composed of phonetic symbols. In the performance support apparatus of the present invention, a selection operation (if a user interface having a plurality of keys and physically mimicking a piano keyboard is pressed, a piano keyboard is virtually displayed by an image representing the piano keyboard and a touch panel, etc. The sound represented by the candidate character string assigned to the key that has been subjected to a touch operation on the key in the case of a simulated user interface is the pronunciation of the lyrics, and the pitch corresponding to the key is the pitch It is considered as the pitch when sounding. As described above, according to the present invention, the pronunciation and pitch of the lyrics can be designated by a selection operation for one key, and therefore, it is necessary to designate each by a separate operation. Compared with the technique disclosed in each document of Patent Document 2, it becomes possible to perform a singing voice easily. In addition, in the present invention, although there is a restriction that the pronunciation of the lyrics must be selected from each candidate character string assigned to a plurality of keys, each time a key selection operation is performed, Since the assignment of candidate character strings is updated, the degree of freedom of writing is higher than the technique disclosed in Non-Patent Document 3, and there is no need to prepare in advance, such as presetting lyrics. That is, according to the present invention, it is possible to easily perform a performance of a singing voice with a high degree of freedom of writing without preparing in advance such as presetting lyrics.

  In a more preferred aspect, the performance support device has display control means for causing a display device to display candidate character strings assigned to each of the plurality of keys by the assigning means. According to such an aspect, the user of the performance support device grasps the candidate character string assigned to each key through the display content of the display device, and composes the lyrics (or composes in addition to the lyrics) improvised. It becomes possible to play a singing voice with reference. Here, the display device may be a device separate from the plurality of keys, or may be a display frame (small liquid crystal display provided for each key) provided for each key. In the latter case, the display control unit displays the candidate character string assigned by the assigning unit in the display frame of each key. According to such an aspect, the user of the performance support apparatus can grasp the candidate character strings assigned to the keys without looking away from the plurality of keys. On the other hand, in the former mode, when performing a singing voice using a general MIDI keyboard or the like in which the display frame is not provided for each key, a candidate character string assigned to each key is displayed. The performance can be performed while grasping. In addition, as a display device in the former mode, a general liquid crystal display can be used, and a tablet terminal can also be used. As a display mode of the candidate character string in this mode, a virtual keyboard is displayed on the display device, a display frame is set for each key in the virtual keyboard, and the candidate character string is displayed in the display frame as in FIG. An embodiment is mentioned.

In a more preferred aspect, the storage unit stores a history of keys (selection order) for which the selection operation has been performed, and the selection unit includes a plurality of candidate character strings indicated by the history stored in the storage unit. A candidate character string to be assigned to each key is selected. As a specific example of an aspect of selecting a candidate character string to be newly assigned to each key based on a sequence of candidate character strings indicated by the history, the candidate character strings indicated by the history are followed by a character string obtained in the order of selection. In such a case, a character string composed of one or more characters that do not cause lexicographic inconsistency is selected as a candidate character string to be newly assigned to each key. According to such an aspect, although the word unit obtained by linking one or more candidate character strings has a lexical meaning, the whole sentence may be a lyric that does not make a special meaning. On the other hand, as shown in Examples 1 to 3 below, it is possible to come up with lyrics that would not normally occur.
Example 1: Gomez in the left batter box ...
Example 2: If you have a 5 meter cat ...
Example 3: Gome-Ringo-Dame-Ringo ....
In addition, according to this aspect, even a user who can “song a song but cannot compose a song” can simply create a melody by simply selecting the lyrics and typing (that is, , Even people who have no knowledge of composing can easily play the original singing voice). As a more preferable mode, a mode in which the selection unit selects a candidate character string to be assigned to each of a plurality of keys in consideration of the transition of the pitch indicated by the history stored in the storage unit can be considered. For example, from among character strings made up of one or more characters that do not cause inconsistency in lexicographic (or linguistic) terms, those that do not cause inconsistency in terms of pitch transition (intonation) are narrowed. Condition.

  In another preferred aspect, the assigning means assigns a candidate character string to each of the plurality of keys in consideration of a transition of a pitch indicated by a history stored in the storage means. . As will be described later, the melody (that is, the transition of the pitch) that is played by sequentially selecting each key can be either preferable or unfavorable. Therefore, based on the melody played so far, the display mode of the candidate character string for the key corresponding to the pitch that is the preferred melody when selected next is made different from the other (for example, the display color) , Different character fonts, and other character keys do not display candidate character strings, etc., so that the user can be presented with a pitch that is a favorable melody. It is possible to play a singing voice while improvising lyrics and composing. Further, by sequentially performing the key selection operation using the display mode as a clue, it is possible to play a melody with a preferable melody without worrying about the lyrics.

  In yet another preferred aspect, the device further comprises an operator that instructs a user to reset assignment by the assigning unit, and the assigning unit is triggered by an instruction to reset by an operation on the operator. The assignment of candidate character strings to each of the above is changed. According to such an aspect, the range of correspondence between lyrics and pitches is widened, and it becomes possible to perform a singing voice while paying attention to both the linguistic meaning of the lyrics and the natural transition of the pitches.

  Further, as another aspect for solving the above-described problem, an aspect in which a program that causes a general computer such as a CPU (Central Processing Unit) to function as the assignment means, the sound generation control means, and the selection means can be considered. . This is because by operating the computer according to such a program, the computer can function as the performance support apparatus of the present invention. As a specific provision mode of such a program, a mode in which the program is written and distributed on a computer-readable recording medium such as a CD-ROM (Compact Disk-Read Only Memory), or an electric communication such as the Internet A mode of distribution by downloading via a line is conceivable.

It is a figure which shows the structural example of the song synthesis system 1 containing the performance assistance apparatus 10 of one Embodiment of this invention. 2 is a diagram illustrating an example of an appearance of a user interface unit 120 of the performance support apparatus 10. FIG. It is a figure which shows the structural example of the performance assistance apparatus. It is a flowchart which shows the flow of the process which the control part 110 of the same performance assistance apparatus 10 performs according to the performance assistance program 134a. It is a figure which shows an example of the spread of the lyrics in a song. It is a figure for demonstrating the operation example of the same performance assistance apparatus. It is a figure for demonstrating the operation example of the same performance assistance apparatus. It is a figure for demonstrating a modification (1). It is a figure for demonstrating a modification (1). It is a figure for demonstrating a modification (1). It is a figure for demonstrating a modification (2). It is a figure for demonstrating a modification (2). It is a figure for demonstrating a modification (3). It is a figure for demonstrating a modification (3). It is a figure for demonstrating a modification (3). It is a figure for demonstrating a modification (3).

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram illustrating a configuration example of a singing voice synthesis system 1 including a performance support device 10 according to an embodiment of the present invention. As shown in FIG. 1, the song synthesis system 1 includes a performance support apparatus 10 and a song synthesis engine 20. The performance support apparatus 10 includes a user interface (hereinafter, “I / F”) unit 120 (see FIG. 2) that includes an input device 122 simulating a piano keyboard and a speaker 124. The performance support apparatus 10 generates a control signal for instructing synthesis of the singing voice in accordance with an operation performed on the input device 122, and supplies the control signal to the singing synthesis engine 20. The singing voice synthesis engine 20 is a device that generates and outputs data representing singing voice (for example, waveform data of singing voice) in accordance with an existing singing voice synthesis algorithm such as a unit connection method. The singing voice synthesis engine 20 generates waveform data representing the singing voice according to the control signal given from the performance support device 10, and gives it to the performance support device 10. The performance support apparatus 10 causes the speaker 124 to output the sound represented by the waveform data given from the song synthesis engine 20.

  In the singing voice synthesizing system 1 of the present embodiment, different pitches are assigned to the plurality of keys included in the input device 122, and character strings representing the pronunciation of lyrics (in this embodiment, as shown in FIG. 2). "One Hiragana character") is assigned to each. The character string assigned to each key is a candidate for the lyrics of the music that the user wants to play a singing voice. Therefore, hereinafter, the character string is referred to as a “candidate character string”. The performance support apparatus 10 is triggered by the selection operation for any key (when the input device 122 is physically a piano keyboard, a key pressing operation) is performed. A control signal for instructing synthesis of a singing voice that generates a sound represented by a candidate character string assigned to the key at a pitch corresponding to the key is generated and provided to the singing synthesis engine 20. Due to such a configuration, according to the present embodiment, a performance of singing voice is realized by performing a performance operation such as sequentially pressing keys. Hereinafter, the description will focus on the performance support apparatus 10 that significantly shows the features of the present embodiment.

  FIG. 3 is a diagram illustrating a configuration example of the performance support apparatus 10. As shown in FIG. 3, the performance support apparatus 10 mediates data exchange between the control unit 110, the storage unit 130, the external device I / F unit 140, and these components in addition to the user I / F unit 120 described above. Bus 150 is included.

  The control unit 110 is a CPU, for example. The control unit 110 functions as a control center of the performance support apparatus 10 by executing the performance support program 134a stored in the storage unit 130 (more precisely, the nonvolatile storage unit 134). Details of processing executed by the control unit 110 in accordance with the performance support program 134a will be clarified later. The external device I / F unit 140 is, for example, a USB interface or a parallel interface, and is connected to the song synthesis engine 20 via a signal line such as a USB cable.

  As shown in FIG. 2, the user I / F unit 120 includes a wheel 126 in addition to the input device 122 and the speaker 124 described above. The wheel 126 can be rotated in the direction indicated by the arrow in FIG. The user operates the rotation direction and the rotation amount of the wheel 126 to adjust the output volume of the speaker 124 and the intensity of the acoustic effect when an acoustic effect such as pitch bend is applied to the sound output from the speaker 124. It can be carried out. As shown in FIG. 2, the input device 122 has a plurality of keys, and when a selection operation (for example, key pressing) is performed on any of the keys, the input device 122 indicates the key that has been selected. A signal is given to the control unit 110. Each of the plurality of keys of the input device 122 is assigned a different pitch, and is provided with a display frame 122a regardless of whether it is a white key or a black key (in FIG. 2). The symbol 122a is attached only to the display frame of the leftmost white key). The display frame 122a is a small liquid crystal display, and the candidate character string assigned to the key is displayed on the display frame 122a of each key. For example, in the example shown in FIG. 2, “i” is assigned as a candidate character string to the leftmost white key, and “a” is assigned as a candidate character string to the leftmost black key. A user who performs a singing voice using the singing voice synthesizing system 1 can grasp the candidate character strings assigned to the keys through the display contents of the display frames 122a of the keys.

  The storage unit 130 includes a volatile storage unit 132 and a nonvolatile storage unit 134. The volatile storage unit 132 is, for example, a RAM (Random Access Memory). The volatile storage unit 132 is used by the control unit 110 as a work area when executing the performance support program 134a. The volatile storage unit 132 stores history data indicating a history of keys that have been selected. The history data correlates information indicating each of the key that has been selected, the pitch corresponding to the key, and the candidate character string assigned to the key at the time of the selection operation, Data arranged in the order in which selection operations were performed. This history data is obtained when a selection operation on the input device 122 is used as an opportunity to reselect a lyric candidate (the above-mentioned candidate character string) following the lyric generated by the selection operation and assign it to each key of the input device 122. Used for

  The nonvolatile storage unit 134 is, for example, a hard disk or a flash ROM (Read Only Memory). In addition to the performance support program 134a, the non-volatile storage unit 134 stores a language dictionary database (abbreviated as “DB” in FIG. 3) 134b and a pitch name dictionary DB 134c in advance. In the language dictionary DB 134b, character strings representing various words that are constituent elements of lyrics are registered. The language dictionary DB 134b is used when selecting a candidate character string to be newly assigned to each key when the key selection operation is performed, similarly to the history data described above. In the pitch name dictionary DB 134c, information indicating the pitches corresponding to the keys of the input device 122 (that is, pitch names) is registered in association with the keys.

  The control unit 110 reads the performance support program 134a from the nonvolatile storage unit 134 to the volatile storage unit 132 when the power supply (not shown in FIG. 1) of the performance support device 10 is turned on or reset, and starts its execution. FIG. 4 is a flowchart showing the flow of processing executed by the control unit 110 in accordance with the performance support program 134a. As shown in FIG. 4, the control unit 110 operating according to the performance support program 134a repeatedly executes a candidate character string assignment process 134a1, a pronunciation control process 134a2, and a candidate character string selection process 134a3. The outline of each of these processes is as follows.

  The candidate character string assignment process 134a1 is a process of assigning a candidate character string to each of the plurality of keys of the input device 122 and writing information indicating the correspondence between them (hereinafter, assignment information) to the volatile storage unit 132. The candidate character strings to be assigned to the keys may be selected at random by referring to the contents stored in the language dictionary DB 134b, for example, immediately after the singing synthesis system 1 is turned on or reset, and the key selection operation is performed once. After the process is performed, selection may be made by a candidate character string selection process 134a3 described later. Further, the assignment of the candidate character string to each key may be performed randomly using a pseudo random number or the like. The candidate character string assignment process 134a1 of the present embodiment includes a display control process for displaying the candidate character string assigned to each key on the display frame 122a of the key. That is, the control unit 110 operating according to the performance support program 134a functions as an assigning unit that executes the candidate character string assigning process 134a1 and also functions as a display control unit that executes the display control process.

  The sound generation control process 134a2 is a process executed in response to a key selection operation. More specifically, the control unit 110 enters an input standby state when the execution of the candidate character string assignment process 134a1 is completed, and is triggered by the reception of a signal indicating the selection of any key from the input device 122. 134a2 is executed. In the sound generation control process 134a2, the control unit 110 specifies a candidate character string assigned to the key indicated by the signal output from the input device 122 with reference to the assignment information, and also specifies a pitch corresponding to the key. Is identified with reference to the pitch name dictionary DB 134c. And the control part 110 produces | generates the control signal which instruct | indicates the synthesis | combination of the audio | voice which pronounces the sound which the said candidate character string represents with the said pitch, and while outputting to the song synthesis engine 20, the said candidate character string and the said pitch Is added to the history data. That is, the control unit 110 operating according to the performance support program 134a functions as sound generation control for executing the sound generation control process 134a2.

  The candidate character string selection process 134a3 is a process that is executed upon completion of the execution of the pronunciation control process 134a2. In the candidate character string selection process 134a3, the control unit 110 sets the candidate character string assigned to each key in the candidate character string assignment process 134a1 to the lyrics of the singing voice that has been output so far (that is, the candidate character string indicated by the history data). This is a process of selecting again according to the lyrics expressed by the line. Thus, the control unit 110 operating according to the performance support program 134a functions as a selection unit that executes the candidate character string selection process 134a3.

  The possibility of lyrics in a song spreads in a tree diagram as shown in FIG. From this, it can be said that “song performance” is to select the characters constituting the lyrics and specify the pitch when the characters are pronounced. Some of the lyrics obtained by connecting the characters selected according to the tree diagram shown in FIG. 5 have a linguistic meaning and some have no meaning. The lyrics when performing a song may have a linguistic meaning, or may have no linguistic meaning like lyrics in scat or humming. With lyrics that have no meaning at all, it is easy to get tired of listening, but it is difficult to play while improvising and meaningful lyrics. The purpose of the present embodiment is to assist in “significant lyric generation” for these two contradictory problems, and the candidate character string selection process 134a3 is a process that plays a central role in achieving this purpose. It is.

  Various modes can be considered as a candidate character string selection algorithm in the candidate character string selection processing 134a3. For example, the following five aspects can be considered. The first is a mode in which words are randomly selected from the language dictionary DB 134b. Specifically, words starting from a character string obtained by linking each candidate character string selected up to the immediately preceding key pressing operation in the order of selection are read from the language dictionary DB 134b, and the characters subsequent to the character string in those words This is a mode in which one or a plurality of characters (one character in the present embodiment) is used as a new candidate character string. The second mode is a mode in which a character string consisting only of particles (such as “ga”, “to”, “ni”) is added as a candidate character string to the candidate character string obtained by the first mode. In the first aspect, the lyrics obtained by following the tree diagram shown in FIG. 5 may be biased toward specific parts of speech such as nouns. However, according to the present aspect, such a bias can be avoided. . The third mode is a mode in which candidate character strings are selected by using a writing algorithm such as “associative dictionary” or “synonym dictionary” so that the lyrics become sentences having linguistic meaning. For example, if the lyrics played up to the previous key press operation are the character string “asaha”, words such as “nemui” and “samui” are extracted using “associative dictionary” and “synonym dictionary”. The first character of these words is a candidate character string. The fourth mode is a mode in which lyrics that become meaningless words are mixed at a certain ratio so that extraordinary (meaningless) lyrics can be created. And the 5th mode registers the NG word beforehand so that the performance of the lyric including the specific word (hereinafter referred to as NG word) such as a single number that is contrary to public order and morals such as discrimination words is avoided, When the lyrics played up to the key-pressing operation and the candidate character string obtained by any of the first to fourth modes are concatenated, the concatenated character string matches a pre-registered NG word This is an aspect of excluding character strings.

  For example, in the situation where the candidate character string is assigned to each key of the input device 122 as shown in FIG. 2 by the candidate character string assigning process 134a1, the candidate character string is assumed to indicate the first lyrics and pitch. It is assumed that a key assigned with “go” is pressed. The control unit 110 executes the sound generation control process 134a2 when the key is pressed, and the singing voice sounding the candidate character string (that is, “go”) at the pitch (for example, B3) corresponding to the key. A control signal for instructing synthesis is output to the singing voice synthesis engine 20, and the singing voice is output.

  In the candidate character string selection process 134a3 subsequent to the pronunciation control process 134a2, the control unit 110 reads words starting from the output lyrics (in this operation example, “go”) from the language dictionary DB 134b, and the candidate character strings are included in those words. Each character that follows is set as a new candidate character string. For example, if words such as “gogo”, “gozen”, “sorry”, “fancy”, “feast”, etc. are read out as words starting with “go”, the control unit 110 reads these words. The second character of the word (ie, “me”, “ze”, “ka”, “chi”, “go”...) Is selected as a new candidate character string. Then, the candidate character string selected in this way is assigned to each key of the input device 122 by the candidate character string assigning process 134a1, and the display content of the display frame 122a of each key is updated as shown in FIG.

  As shown in FIG. 6, in a situation where a candidate character string is assigned to each key of the input device 122, the user presses the key assigned the candidate character string “M” corresponding to the next lyrics. And The control unit 110 executes the sound generation control process 134a2 when the key is pressed, and instructs synthesis of a singing voice that sounds the sound represented by the candidate character string at a pitch (for example, C4) corresponding to the key. A control signal is output to the song synthesis engine 20, and the song voice is output. In the candidate character string selection process 134a3 subsequent to the pronunciation control process 134a2, the control unit 110 reads out words starting from the output lyrics (in this operation example, “gome”) from the language dictionary 134b, and the above words are included in the words. Each character following the candidate character string is set as a new candidate character string. For example, it is assumed that words such as “I'm sorry”, “I'm sorry”, “I'm sorry”, “I'm sorry”, “I'm sorry”, etc. are read out as words that start with “Gome”. Assuming that the appearance rate of “n” is 60%, the appearance rate of “e” is 10%, the appearance rate of “su” is 10, and the others are 20 for the third letter of these words, New candidate character strings are selected so that “N” is 60%, “E” is 10%, “Su” is 10%, and other characters are 20%. The candidate character string selected in this way is assigned to each key of the input device 122 by the candidate character string assignment processing 134a1, so that the display content of the display frame 122a of each key is updated as shown in FIG.

  As described above, according to the present embodiment, there is no need to prepare in advance, such as presetting lyrics as in the technique disclosed in Non-Patent Document 3, and freedom of lyrics such as playing while improvising lyrics It becomes possible to perform a highly sung voice. Here, as a method of selection when the user selects the lyrics (or pitch) to be pronounced next, a mode of selecting the next sound based on the displayed candidate character string (a mode in which lyrics are given priority) ), Select the desired lyrics by selecting the next lyrics (composition priority), select appropriate lyrics and pitches considering both pitch and lyrics, and select randomly A mode is conceivable, and it may be appropriately used according to the user's preference. In the candidate character string assignment processing 134a1 of the present embodiment, the lyrics desired by the user are not necessarily assigned to the keys having the desired pitch, and therefore, the operation of the wheel 126 assigns candidate character strings to the keys. You may enable it to change (for example, change cyclically).

  Further, according to the present embodiment, since the pronunciation of the lyrics and the pitch thereof can be designated by a key pressing operation on one key, both the techniques disclosed in Non-Patent Document 1 and Non-Patent Document 2 are used. It is possible to easily perform a singing voice compared to a technique that requires each to be specified separately. That is, according to the present embodiment, it is possible to easily perform a singing voice with a high degree of freedom in writing lyrics without preparing in advance such as presetting lyrics.

Although one embodiment of the present invention has been described above, it goes without saying that the following modifications may be added to this embodiment.
(1) If a performer who wants to perform a singing voice using the singing synthesizer 1 is not used to composing, it may be difficult to select which pitch to select. Of course, there are preferred (natural) and unfavorable (unnatural) melodies. For example, the score examples (a) to (d) in FIG.
Is a sample on the selection of the seventh sound. In this case, if it is evaluated from the viewpoint of whether the melody is natural, the following score can be given.
Sample A: 100 points Sample B: 10 points Sample C: 15 points Sample D:
30 points Therefore, based on the pitch designated by the user's performance operation, a score (in other words, whether or not a preferable (natural) melody is obtained for each of a plurality of keys following the pitch of the pitch. For example, a value indicating a degree recommended as the next sound) may be calculated, and the display mode of the candidate character string may be different for each key according to the score. Specifically, a candidate character string is not displayed on a key whose score is less than a predetermined threshold (or the candidate character string is displayed in parentheses, the display color of the candidate character is changed, etc.). Conceivable. According to such an aspect, even if it is a beginner of composition, it becomes possible to play a singing voice of a favorable (natural) melody without being confused about which pitch should be selected. The score differs depending on the accompaniment chords (Cm, D7, etc.) and tonality (C major, D minor, etc.). Therefore, it is preferable that additional information for calculating the score, such as information indicating accompaniment chords and information indicating tonality, can be input to the performance support apparatus 10.

  When accompaniment sounds are played or played simultaneously according to accompaniment data, the above score is calculated according to the chord progression in the accompaniment sounds in order to avoid selecting sounds that do not match the accompaniment sounds. Also good. For example, as shown in FIG. 9A, when a chord of a certain bar is F (F major, constituent sounds: F, A, C), F #, B, etc. are used in general music. Since the possibility is very small, these scores are set to a sufficiently small value, and as shown in FIG. 9B, the display mode of the candidate character strings assigned to them is different from the others. In the example shown in FIG. 9B, the difference in display mode is expressed by the presence or absence of hatching in the display frame 122a. By implementing these functions, people who are not used to improvisation can enjoy improvisation. It is also possible to learn that each sound has a musical “function”. However, since the “prohibition” in music does not exist as a strict rule, the calculation of the score and the control of the display mode of the candidate character string according to the score are limited to “which sound should be selected. It is an auxiliary tool for avoiding the state of “I don't know”, and it is not necessary to prohibit the pronunciation of a sound with a small score just because it is instructed.

  In the above embodiment, when the user performs the singing voice by freely specifying the lyrics to be pronounced next and the pitch thereof (that is, when the user performs the singing voice while simultaneously composing and writing the lyrics). ) Explained. However, when the musical piece to be played is predetermined (when the lyrics and the melody are predetermined as shown in FIG. 10A), every time a key operation for playing the melody is performed. In addition, the character string of the lyrics to be sounded next may be displayed only on the key corresponding to the pitch to be sounded next, or an underline is added only to the character string as shown in FIG. The aspect may be different.

(2) In the above-described embodiment, a candidate character string that is linguistically natural is selected by referring to the stored contents of the language dictionary and following a played singing voice. Selection and assignment to each key may be performed. Specifically, in candidate character string assignment processing 134a1, candidate character strings may be assigned in consideration of intonation information. FIG. 11 is a diagram illustrating an example of lyrics starting with “ASA” and intonation. FIG. 12 is a diagram showing a candidate character string selected based on the intonation and an example of assignment to each key. In the example illustrated in FIG. 12, the third character when the words “in the morning”, “until morning”, “from morning”, “shallow”, and “agao” are written in hiragana is a plurality of keys of the input device 122. Are assigned in order from the left side. The example shown in FIG. 12 is based on the change from the second tone to the third tone of the intonation shown in FIG. For example, in “Asa-Oga”, the second to third notes tend to increase, so the next lyrics “GA” is assigned to the high pitched part (= the right keyboard).

(3) In the above embodiment, an example of application of the present invention to singing voice performance support for a song whose lyrics are written in Japanese is explained, and a character string consisting of one hiragana character is used as a candidate character string. Of course, a character string consisting of one character may be used as a candidate character string. This is because katakana is a phonetic character as well as hiragana. In addition, as a candidate character string, a character string made up of phonetic symbols defined by X-SAMPA or the like is used as a candidate character string, as shown in FIG. Also good. By using a character string composed of phonetic symbols as a candidate character string, there is an effect that it is possible to express vowels and the like.

In addition, the application object of the present invention is not limited to the performance support of singing voices for music whose lyrics are described in Japanese, but singing voices for music whose lyrics are described in a language other than Japanese such as English. It is also applicable to performance support. 14-16 is a figure for demonstrating the example of application of this invention to the performance support of the singing voice about the music in which the lyrics were described in English. For example, it is assumed that the user selects a key on which the candidate character string “dis-” is displayed from the initial state shown in FIG. In this case, lexicographical candidates for subsequent lyrics are “able” (ie, disable), “card” (ie, discard), “count” (ie, discount), “cuss” (ie, discuss), Words such as “tance” (ie distance), “play” (ie display)... Based on these, the next lyrics are displayed as shown in FIG. Thus, in the case of English, it is desirable to divide the words constituting the lyrics into syllables and display candidate character strings in syllable units. The example shown in FIGS. 14 and 15 is a case where a word of two syllables is selected from the “initial state”, but if another lyrics is selected, a different display is obtained. For example, if you select “I”, I can, I like, I love, I am, I play,
Lyrics such as I got ... are possible. In this case, “can”, “like”, “love”, “am”, “play”, “got”, etc. are candidate character strings assigned to each key, and the key corresponding to “I” is pressed. As shown in FIG. 16, the candidate character string assigned to each key is updated.

(4) In the above embodiment, the candidate character string assigned to the key is displayed in the display frame 122a of each key, but the display control processing of the candidate character string in the display frame 122a may be omitted. The display frame 122a may be omitted. According to such an aspect, a conventional MIDI keyboard or the like can be used as a keyboard. Even if the display control of the candidate character string is not performed at all, it is not clear what lyrics are pronounced until the key is pressed, and the unexpected lyrics of the user are pronounced. The effect of improving is expected. When a conventional MIDI keyboard or the like is used as a keyboard, each key candidate is displayed on a display device (such as a general liquid crystal display or a tablet terminal) separate from the performance support device 10 instead of the display control process. You may make it make the control part 110 perform the display control process which displays a character string. As a specific example of the display control process, a process of displaying a virtual keyboard on the display device, setting a display frame for each key in the virtual keyboard, and displaying a candidate character string in the display frame as in FIG. Can be mentioned.

(5) In the above embodiment, the case where the performance support apparatus 10 of the embodiment of the present invention is included in the singing voice synthesis system 1 has been described, but the performance support apparatus 10 may be provided alone. This is because it is possible to construct the song synthesis system of the above embodiment by connecting the performance support apparatus 10 to an existing song synthesis engine via a signal line such as a USB cable. Instead of connecting the performance support apparatus 10 to the song synthesis engine via a signal line, the performance support apparatus 10 may be connected via an electric communication line such as the Internet, or both may be connected wirelessly. Moreover, you may provide the performance assistance apparatus 10 of this embodiment incorporated in the song synthesis | combination apparatus which has a song synthesis engine. In the above embodiment, the allocation unit, the display control unit, the sound generation control unit, and the selection unit have been described as being implemented by software modules, but may be implemented by hardware such as an electronic circuit.

  DESCRIPTION OF SYMBOLS 1 ... Singing synthesis system, 10 ... Performance support apparatus, 110 ... Control part, 120 ... User I / F part, 122 ... Input device, 122a ... Display frame, 124 ... Speaker, 126 ... Wheel, 130 ... Memory | storage part, 132 ... Volatile memory unit 134... Nonvolatile memory unit 134 a. Performance support program 134 b. Language dictionary 134 c Sound name dictionary 140 External device I / F unit 150 Bus 20 singing synthesis engine

Claims (5)

Assigning means for assigning a candidate character string consisting of one or more characters representing the pronunciation of the lyrics to each of a plurality of keys to which different pitches are assigned;
When a selection operation is performed on any of the plurality of keys, the sound represented by the candidate character string assigned to the key that has been selected is sounded at a pitch corresponding to the key. Pronunciation control means for giving a voice synthesis instruction to the singing voice synthesis device;
Selection means for reselecting a candidate character string assigned to each of the plurality of keys by the assigning means according to a candidate character string assigned to the key subjected to the selection operation;
A performance support apparatus comprising: The performance support apparatus according to claim 1, further comprising display control means for displaying a candidate character string assigned to each of the plurality of keys by the assigning means on a display device. Storage means for storing a history of keys that have been selected;
The performance support according to claim 2, wherein the selection unit selects a candidate character string to be assigned to each of the plurality of keys in accordance with a sequence of candidate character strings indicated by a history stored in the storage unit. apparatus. The performance support apparatus according to claim 3, wherein the assigning means assigns a candidate character string to each of the plurality of keys in consideration of a transition of a pitch indicated by a history stored in the storage means. . An operator that instructs the user to reset the assignment by the assigning means;
5. The assignment unit according to claim 1, wherein the assignment unit changes assignment of the candidate character string to each of the plurality of keys in response to an instruction to reset by an operation on the operation element. The performance support apparatus described in 1.

Images