JP2013045082A - Musical piece generation device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate musically natural and various melodies.SOLUTION: A musical piece generation device includes: a first acquisition unit 42 for generating section designation data GA corresponding to one of a plurality of pieces of section designation data DA designating a time series of a sound production section; a second acquisition unit 44 for selecting one of a plurality of pieces of pitch designation data DB designating a time series of a pitch, as pitch designation data GB; and a melody generation unit for generating melody data which designates a time series of a note corresponding to each sound production section designated by the section designation data GA acquired by the first acquisition unit and a pitch designated by the pitch designation data GB acquired by the second acquisition unit 44 at a start point of the sound production section.

Description

  The present invention relates to a technique for generating music (melody).

  A technique (automatic music technique) that automatically generates a melody of music has been proposed. For example, Patent Literature 1 discloses a technique for generating a melody by assigning a pitch to each note value (sounding section) specified in time series by note value string data. The pitch given to each note value is selected at random. Patent Document 2 discloses a technique for generating a melody by changing each pitch specified by melody material data in time series according to background information such as a key and a scale.

Japanese Patent Laid-Open No. 9-081141 Japanese Patent Laid-Open No. 9-081143

  In the technique of Patent Document 1, since a pitch given to each tone value is selected at random, it is actually difficult to generate a musically natural melody. In the technique of Patent Document 2, it is difficult to generate various melodies because only the pitches specified by the melody material data are changed. In view of the above circumstances, an object of the present invention is to generate musically natural and diverse melodies.

  Means employed by the present invention to solve the above problems will be described. In order to facilitate understanding of the present invention, in the following description, the correspondence between each element of the present invention and the element of each of the embodiments described later is indicated in parentheses, but the scope of the present invention is not limited to the embodiment. It is not intended to limit the example.

  The music generation device of the present invention acquires first section acquisition data (for example, section specification data GA) according to any of a plurality of section specification data (for example, section specification data DA) for specifying a time series of pronunciation sections. Pitch designation data (for example, pitch designation data GB) corresponding to any of the means (for example, the first acquisition unit 42) and a plurality of pitch designation data (for example, pitch designation data DB) for designating a time series of pitches. ) Is acquired by the second acquisition means (for example, the second acquisition unit 44), each sound generation section specified by the section specification data acquired by the first acquisition means, and the pitch specification data acquired by the second acquisition means is the sound generation. Melody generating means (for example, melody generating unit 522) for generating melody data (for example, melody data DM) for specifying a time series of notes corresponding to a pitch specified for a reference time point (for example, a starting point of a pronunciation section). Have . In the above configuration, the melody data is generated using the section designation data according to any of the plurality of section designation data and the pitch designation data according to any of the plurality of pitch designation data. There is an advantage that a variety of music can be generated as compared with the technique of Patent Document 2 in which only the pitches specified by the melody material data are changed. Also, since melody data is generated using the section specification data and pitch specification data as templates, for example, a musically natural melody is generated compared to the technique of Patent Document 1 in which pitches are selected at random. Is possible.

  In a preferred aspect of the present invention, the melody generating means includes a knitted musical note sequence in which notes having each sounding section designated by the section designation data and the pitch designated by the pitch designation data at the reference time of the sounding section are arranged ( For example, a pitch extraction unit (for example, pitch extraction unit 62) that generates a knitted note sequence M) and a pitch that generates melody data by adjusting the pitch of each note of the knitted note sequence generated by the pitch extraction unit. Adjusting means (for example, pitch adjusting unit 64). In the above aspect, since the pitch of each note of the knitted note sequence generated from the section specifying data and the pitch specifying data is adjusted, the musically natural melody is compared with the configuration in which the knitted note sequence is not adjusted. Can be generated.

  The content of the adjustment processing by the pitch adjustment means is arbitrary, but for example, third acquisition means (for example, chord progression data for specifying chord progression (for example, chord progression data corresponding to any of a plurality of chord progression data)) In the configuration including the third acquisition unit 46), the pitch adjusting means calculates the pitch of at least some of the notes (for example, the notes designated as the important sound and the end sound) among the plurality of notes of the knitting note string. A first adjustment process (for example, a first adjustment process SB1) is executed to change the pitch of the constituent sound of the chord designated by the chord progression data for the note production period. In the above aspect, the pitch of at least some of the notes in the knitted note sequence is changed to the constituent sound of the chord specified by the chord progression data, so that the pitch of each note transitions musically naturally. Can be generated. In addition, the pitch of at least some of the notes (for example, the notes specified as the elapsed sound) of the plurality of notes in the knitting note string is set within a predetermined range corresponding to the pitch of the note immediately before or after the note. Even in a configuration in which the pitch adjustment means executes the second adjustment process (for example, the second adjustment process SB2) for changing to a pitch, there is an advantage that a musical piece in which the pitch of each note transitions naturally can be generated.

  The music generation device according to a preferred aspect of the present invention includes a character string setting unit (for example, a character string setting unit 34) that acquires a designated character string in which a plurality of speech units (for example, syllables and phonemes) are arranged in time series, and a melody. Voice synthesizing means (for example, a voice synthesizing unit 524) that generates a voice signal of a voice that is generated by a plurality of notes designated by the melody data generated by the generating means. According to the above aspect, there exists an advantage that the audio | voice signal of the sound which sounded musically natural and various melodies can be produced | generated.

  In a preferred example of the configuration including the speech synthesizing unit, the first acquisition unit specifies the selection unit (for example, the selection unit 422) that selects any one of the plurality of section specifying data and the section specifying data selected by the selecting unit specifies. A musical score section (for example, a musical score section 424) that associates each sound generation section with each sound unit of the designated character string, and the speech synthesis means includes a sound generation section for each of the sound generation sections of the plurality of notes specified by the melody data. Then, the musical score generating means generates a voice signal for voice that pronounces a voice unit corresponding to the sound generation section. In the above aspect, there is an advantage that a sound signal of a natural sound in which the sound unit of the designated character string is pronounced in the sounding interval of each note designated by the melody data is generated.

  In a preferred aspect of the present invention, the selecting means selects the section specifying data having a larger number of pronunciation sections among the plurality of section specifying data as the number of voice units of the specified character string set by the character string setting section is larger. . In the above aspect, since the section designating data having a larger number of sounding sections is selected as the number of sound units of the designated character string is larger, each sounding section can be associated with each sound unit without difficulty.

  In a preferred aspect of the present invention, the musical notation means includes, for each of a plurality of blocks set on the time axis, the number of pronunciation sections designated in the block by the section designation data and the designated character string. The designated character string is divided into blocks so that the number of voice units is approximated, and each sound generation section and each voice unit are associated with each block. According to the above aspect, it is possible to associate the sounding section and the voice unit without difficulty in each block. Also, for blocks where the number of sounding intervals is less than the number of speech units, the sounding interval in the block is divided, and for blocks where the number of sounding intervals exceeds the number of sound units, a predetermined character string in the block is specified. According to the configuration in which a voice unit (for example, a long sound symbol) is added, each sound generation section and each voice unit can be made to correspond one-to-one.

  The music generating apparatus according to a preferred aspect of the present invention includes an accompaniment sound generating unit (for example, an accompaniment sound generating unit 54) that generates an accompaniment signal indicating an accompaniment sound, an audio signal generated by the voice synthesizing unit, and an accompaniment sound generating unit Mixing means (for example, a mixing unit 56) for mixing the generated accompaniment signal. In the above aspect, since the accompaniment signal is mixed with the audio signal, there is an advantage that a musical composition rich in musicality can be generated as compared with the configuration in which the audio signal is generated alone. Further, if the audio signal generated by the voice synthesizing means and the accompaniment signal generated by the accompaniment sound generating means are set to the same tempo, there is also an advantage that a music in which the melody and the accompaniment naturally match can be generated.

  The music generating device according to each of the above aspects is realized by hardware (electronic circuit) such as a dedicated DSP (Digital Signal Processor), or a general-purpose arithmetic processing device such as a CPU (Central Processing Unit) and a program. It is also realized through collaboration. The program according to the present invention includes a first acquisition process for acquiring section specifying data corresponding to any of a plurality of section specifying data for specifying a time series of sound generation sections, and a plurality of pitches for specifying a time series of pitches. Second acquisition process for acquiring pitch specification data according to any of the specified data, each sounding section specified by the section specification data acquired in the first acquisition process, and pitch specification data acquired in the second acquisition process Causes the computer to execute a melody generation process for generating melody data that specifies a time series of notes corresponding to the pitch specified for the reference time point of the pronunciation interval. According to the above program, the same operation and effect as the music generating device according to the present invention are realized. Note that the program of the present invention is provided in a form stored in a computer-readable recording medium and installed in the computer, or is provided in a form distributed via a communication network and installed in the computer.

  The present invention is also realized as a method for generating music. The music generation method of the present invention obtains section designation data corresponding to any of a plurality of section designation data for designating a time series of pronunciation sections, and any of a plurality of pitch designation data for designating a time series of pitches When the pitch specified data is acquired, and the sound generation interval specified by the acquired interval specification data and the pitch specified by the acquired pitch specification data are specified according to the reference time of the sound generation interval. Generate melody data that specifies the series. According to the above method, the same operation and effect as the music generating device according to the present invention are realized.

It is a block diagram of the music production | generation apparatus which concerns on 1st Embodiment of this invention. It is explanatory drawing of area designation data. It is explanatory drawing of pitch designation | designated data. It is explanatory drawing of the division | segmentation of the designated character string. It is a block diagram of a 1st acquisition part. It is a flowchart of a staff division process. It is a block diagram of a song sound production | generation part. It is explanatory drawing of operation | movement of a melody production | generation part. It is a flowchart of an adjustment process. It is a block diagram of an accompaniment sound production | generation part. It is a block diagram of the music production | generation apparatus which concerns on 2nd Embodiment.

<First Embodiment>
FIG. 1 is a block diagram of a music generation device 100A according to the first embodiment of the present invention. The music generation device 100A is a signal processing device (automatic music device) that generates a music and outputs an acoustic signal V of the performance sound of the music. In 1st Embodiment, the structure which produces | generates the music (song song) which added the accompaniment of the instrument sound to the melody of the song sound is illustrated.

  An input device 12 and a sound emitting device 14 are connected to the music generating device 100A. The input device 12 is a device (for example, a mouse or a keyboard) that receives an instruction from a user. The sound emitting device 14 (for example, a speaker or a headphone) emits a sound wave corresponding to the acoustic signal V supplied from the music generating device 100A. In addition, illustration of the D / A converter etc. which convert the acoustic signal V from digital to analog is abbreviate | omitted for convenience.

  The music generation device 100 </ b> A of the first embodiment is realized by a computer system that includes an arithmetic processing device 22 and a storage device 24. The arithmetic processing device 22 executes a program PGM stored in the storage device 24 to thereby generate a plurality of functions (a variable setting unit 32, a character string setting unit 34, a first acquisition unit 42, and a second acquisition unit). An acquisition unit 44, a third acquisition unit 46, a singing sound generation unit 52, an accompaniment sound generation unit 54, and a mixing unit 56) are realized. Note that a part of the functions of the arithmetic processing unit 22 can be shared by a dedicated electronic circuit (DSP).

  The storage device 24 stores a program PGM executed by the arithmetic processing device 22 and various data used by the arithmetic processing device 22. A known recording medium such as a semiconductor recording medium or a magnetic recording medium or a combination of a plurality of types of recording media can be arbitrarily employed as the storage device 24.

  The storage device 24 of the first embodiment stores a plurality of section specifying data DA, a plurality of pitch specifying data DB, a plurality of chord progression data DC, and a chord table TBL. Each section designating data DA, each pitch designating data DB, and each chord progression data DC are generated from a predetermined section (hereinafter referred to as “material section”) in each of a plurality of existing music pieces selected as music generation materials. Template.

  As shown in FIGS. 2 and 3, in the first embodiment, a predetermined number of consecutive bars (eight in the example of the first embodiment) are extracted as the material section Q from each piece of music. Each material section Q has a common time signature (for example, 4/4 time signature). Accordingly, the time lengths of the material sections Q when the material sections Q are reproduced at a common tempo are the same. As shown in FIGS. 2 and 3, the material section Q is divided into four blocks QB (QB1 to QB4) in units of two music bars. However, the number of blocks QB in the material section Q and the number of bars constituting the block QB can be arbitrarily changed.

  As shown in FIG. 2, each section designation data DA stored in the storage device 24 is a time series of sections (hereinafter referred to as “sound generation sections”) S in which each note constituting the melody in the material section Q is pronounced. specify. That is, each section designation data DA expresses a rhythm pattern of melody in the material section Q. Each of the plurality of section designation data DA is generated from different material sections Q (material sections Q of different music pieces or different material sections Q in one music piece). Therefore, the time-series manner of the sound generation section S specified by the section specification data DA (position, duration, and number of sound generation sections S on the time axis) is different for each section specification data DA.

  As shown in FIG. 2, one section designation data DA is composed of a plurality of unit data UA corresponding to different sound generation sections S in the material section Q. Each unit data UA includes attribute information UA1 and time information UA2. The time information UA2 designates the position on the time axis of the sounding section S (for example, the time at the start point) and the duration (sound value).

  The attribute information UA1 designates the musical attribute (musical significance within the musical composition) of the notes in the pronunciation period S. Specifically, as shown in FIG. 2, the attribute information UA1 designates any one of an important sound, a termination sound, and a elapsed sound. The important sound means a musically important note in the material section Q. Specifically, notes having a long duration or constituent sounds of music chords are classified as important sounds. The closing sound is a note (note other than the important sound) located at the end of each block QB in the material section Q. The elapsed sound is a note other than the important sound and the end sound. The attribute information UA1 of each unit data UA is manually set by analyzing the contents of each material section Q, for example, by the provider of the music generation device 100A.

  Each section designation data DA can be described as a music file in the SMF (Standard MIDI File) format conforming to the MIDI (Musical Instrument Digital Interface) standard. Specifically, the event data that designates the attribute information UA1 with a numerical value of the note number for convenience, and the timing data that designates the processing interval of each event data as the time information UA2 are arranged in time series in the SMF format. Data is created as section designation data DA and stored in the storage device 24.

  As shown in FIG. 3, each pitch designation data DB stored in the storage device 24 designates a time series of pitches of each note constituting the melody in the material section Q. That is, each pitch designation data DB represents a melody line in the material section Q. The time-series manner of the pitch designated by the pitch designation data DB is different for each pitch designation data DB.

  As shown in FIG. 3, one pitch designation data DB is constituted by a time series of a plurality of unit data UB corresponding to different notes in the material section Q. Each unit data UB includes pitch information UB1 and time information UB2. The pitch information UB1 designates the pitch of the note, and the time information UB2 designates the position and duration of the note on the time axis. Each pitch designation data DB is music in SMF format in which event data that designates pitch information UB1 with a numerical value of note number and timing data that designates the processing interval of each event data as time information UB2 are arranged in time series. Can be described as a file.

  Each chord progression data DC stored in the storage device 24 designates chord progression (code time series) in the material section Q. Specifically, each chord progression data DC designates a chord for each unit time (eg, eighth note time length) in which the material section Q is divided. The time series of chords specified by each chord progression data is different for each chord progression data DC. The chord table TBL in FIG. 1 designates a plurality of constituent sounds of the chord for each chord that can be designated by the chord progression data DC. The above is the main data stored in the storage device 24 of the first embodiment.

  The variable setting unit 32 in FIG. 1 variably sets a variable applied to music generation. Specifically, the variable setting unit 32 sets a music tempo Xt and a musical style (for example, a genre such as rock or jazz) Xs in accordance with an instruction from the user to the input device 12. The tempo Xt is set to an arbitrary numerical value instructed by the user, for example, and the style Xs is set to the content selected by the user among a plurality of options, for example.

  The character string setting unit 34 variably sets a character string (hereinafter referred to as “designated character string”) DY of the lyrics of the music. Specifically, the character string setting unit 34 sets the designated character string DY according to an instruction (character input) from the user to the input device 12. As shown in FIG. 4, the designated character string DY is composed of a plurality of syllables y (sound units).

  The designated character string DY is divided into a character string Ya and a character string Yb. Specifically, in the configuration in which the user designates the designated character string DY for two lines, the designated character string DY is divided into the character string Ya before the line feed and the character string Yb after the line feed. The method for dividing the designated character string DY is arbitrary. In a configuration in which the user designates a designated character string DY composed of two sentences, the designated character string DY is divided into a character string Ya corresponding to the first sentence and a character string Yb corresponding to the second sentence. .

  The first acquisition unit 42 in FIG. 1 generates section designation data GA corresponding to any of the plurality of section designation data DA stored in the storage device 24. FIG. 5 is a block diagram of the first acquisition unit 42 of the first embodiment. As shown in FIG. 5, the first acquisition unit 42 includes a selection unit 422 and a staff division unit 424.

  The selection unit 422 selects any one of the plurality of section designation data DA stored in the storage device 24. In the first embodiment, the section designation data DA is selected according to the number of syllables y constituting the designated character string DY set by the character string setting unit 34. Specifically, the selection unit 422 selects the section designation data DA in which the number of pronunciation sections S increases as the number of syllables y in the designated character string DY increases.

For example, the selection unit 422 calculates an index value α of the following formula (1) for each of the plurality of section designation data DA stored in the storage device 24.
α = | Na−NYa | + | Nb−NYb | (1)
The first term | Na−NYa | in the equation (1) is the number Na of the pronunciation sections S designated by the section designation data DA in the first half section (block QB1 and block QB2) of the material section Q and the designated character string DY. Of these, the absolute value of the difference from the number NYa of syllables y of the preceding character string Ya. Similarly, the second term | Nb−NYb | of the equation (1) includes the number Nb of the sound generation sections S specified by the section specifying data DA in the latter half section (block QB3 and block QB4) of the material section Q, and the specified character. This is the absolute value of the difference from the number NYb of syllables y of the rear character string Yb in the string DY. That is, the number Nb of the sounding sections S and the number Nb of the sounding sections S in the latter half of the material section Q, or the number Na of the sounding sections S in the first half section of the material section Q and the number NYa of the syllable y of the character string Ya are approximated. The index value α becomes smaller as the number NYb of syllables y in the column Yb approximates.

  The selection unit 422 specifies the section designating data DA (that is, the number of pronunciation sections S close to the number of syllables y in the designated character string DY) in which the index value α is within a predetermined range (for example, a range of 5 or less). The designation data DA) is retrieved from the storage device 24, and one section designation data DA is selected at random from the retrieved plurality of section designation data DA.

  The musical score section 424 in FIG. 5 performs musical score processing for associating each sound generation section S indicated by the section designation data DA selected by the selection section 422 with each syllable y of the designated character string DY set by the character string setting section 34. Run. Specifically, as shown in FIG. 4, the staff division unit 424 sets the section designation data DA for each of the four blocks QB (QB1 to QB4) in the material section Q so that the block QBi (i = 1 to 1). 4) Each syllable y of the designated character string DY and the material so that the difference between the number Ni of the sound generation sections S designated in the number N and the number NYi of the syllable y in the block QBi of the designated character string DY is minimized. The sound generation sections S in the section Q are associated with each other.

  FIG. 6 is a flowchart of the staff division process. When the musical score processing is started, the musical score section 424 divides the designated character string DY into blocks QB in the material section Q as shown in FIG. 4 (SA1). The dividing boundary of the designated character string DY is selected from the boundaries of the clauses of the designated character string DY. Known natural language processing such as morphological analysis can be arbitrarily used to specify the phrase of the designated character string DY. In FIG. 4, the case where the boundary between the character string Ya and the character string Yb coincides with the boundary between the block QB2 and the block QB3 is illustrated for convenience, but the boundary between the character string Ya and the character string Yb and each block It does not necessarily match the QB boundary.

Specifically, the staff division unit 424 divides the designated character string DY at the boundary of the phrase so that the index value β defined by the following formula (2) is minimized.
β = | N1−NY1 | ² + | N2−NY2 | ² + | N3−NY3 | ² + | N4−NY4 | ² (2)
As understood from the equation (2), the index value β is determined by the number Ni of the sound generation sections S specified by the section specifying data DA in the block QBi and the syllable y existing in the block QBi of the specified character string DY. This is a numerical value obtained by integrating the square | Ni-NYi | ² of the difference from the number NYi for a plurality of blocks QB1 to QB4. That is, the staff division unit 424 divides the designated character string DY for each block QB so that the difference between the number Ni of the sound generation sections S and the number NYi of the syllables y is minimized for the plurality of blocks QB1 to QB4.

  At the stage where the designated character string DY is divided by the above processing, the number Ni of the sounding sections S in each block QBi and the number NYi of the syllable y of the designated character string DY do not necessarily match. Therefore, the musical score section 424 executes a process (SA2, SA3) for matching the number Ni of the pronunciation sections S with the number NYi of the syllable y of the designated character string DY in each block QB. That is, the musical score section 424 associates the sound generation section S and the syllable y on a one-to-one basis within each block QB.

  First, the score division unit 424 appropriately divides each sounding section S for the block QBi (Ni <NYi) in which the number Ni of the sounding sections S is less than the number NYi of the syllable y, thereby generating the sounding sections in the block QBi. The number S of Ni is increased to the number NYi of syllables y (SA2). Specifically, the score division unit 424 repeats the process of dividing (for example, dividing into two equal parts) the longest sounding section S in the block QBi until the number Ni of sounding sections S reaches the number NYi of syllables y. . The attributes (important sound / final sound / elapsed sound) of the plurality of sounding sections S after the division are set to the same attributes as the sounding section S before the division.

  Secondly, the staff division unit 424 appropriately inserts a predetermined syllable y for adjustment into the designated character string DY for the block QBi (Ni> NYi) in which the number Ni of the pronunciation sections S exceeds the number NYi of the syllables y. As a result, the number NYi of syllables y in the block QBi is increased to the number Ni of the sounding sections S (SA3). Specifically, the staff division unit 424 inserts into the designated character string DY in the block QBi a long syllable symbol “-”, which means continuation of the immediately preceding syllable y, as the syllable y for adjustment. Iterates until NYi reaches the number Ni of the sound generation sections S. The position where the adjustment syllable y is inserted is selected at random from the designated character string DY in the block QBi, for example.

  By the musical score processing described above, the sound generation section S and the syllable y in each block QBi correspond one-to-one. The musical score section 424 generates section designation data GA for designating the time series of each sounding section S after the musical score processing and the designated character string GY for designating the time series of each syllable y after the musical score processing. That is, the section designating data GA is the sounding section in a state where a part of the plurality of sounding sections S designated by the section designating data DA selected by the selection unit 422 is replaced with the plurality of sounding sections S after the division at step SA2. This is time series data for designating a time series of S, and includes a plurality of unit data UA as with the section designation data DA. The designated character string GY is a character string having contents obtained by adding the syllable y for adjustment to the designated character string DY set by the character string setting unit 34 in step SA3. As understood from the above description, each sound generation section S designated by the section designation data GA and each syllable y constituting the designated character string GY have a one-to-one correspondence.

  The second acquisition unit 44 in FIG. 1 selects any of the plurality of pitch designation data DB stored in the storage device 24 as the pitch designation data GB and acquires it from the storage device 24. Although the method for selecting the pitch designation data DB is arbitrary, for example, any one of the plurality of pitch designation data DB can be randomly selected as the pitch designation data GB.

  The third acquisition unit 46 in FIG. 1 selects any one of the plurality of chord progression data DC stored in the storage device 24 as chord progression data GC and obtains it from the storage device 24. For example, the third acquisition unit 46 randomly selects any one of the plurality of chord progression data DC as the chord progression data GC.

  The singing sound generating unit 52 corresponds to the section specifying data GA generated by the first acquiring unit 42, the pitch specifying data GB acquired by the second acquiring unit 44, and the chord progression data GC acquired by the third acquiring unit 46. A voice signal VA of a voice that produces a melody with the designated character string GY generated by the first acquisition unit 42 is generated.

  FIG. 7 is a block diagram of the singing sound generation unit 52. As shown in FIG. 7, the singing sound generation unit 52 includes a melody generation unit 522 and a voice synthesis unit 524. The melody generation unit 522 is an element that generates the melody data DM using the section specification data GA, the pitch specification data GB, and the chord progression data GC as a template. The pitch extraction unit 62, the pitch adjustment unit 64, It is comprised including.

  The pitch extraction unit 62, as shown in FIG. 8, is a musical note defined by each tone generation section S designated by the section designation data GA and a pitch designated by the pitch designation data GB for the start point of the pronunciation section S. A time series of n (hereinafter referred to as “knitted note string”) M is generated. Specifically, the m-th note n (m is a natural number) of the plurality of notes n constituting the knitted note string M is the m-th pronunciation of the plurality of sound generation intervals S specified by the interval specification data GA. From the start point to the end point of the section S, the pitch specification data GB is maintained at the pitch specified at the start point of the sound generation section S. For example, the first note n in the knitting note string M illustrated in FIG. 8 is the pitch designation data for the start point of the sounding section S in the first sounding section S designated by the section designating data GA. The pitch “C #” specified by GB is maintained. As can be understood from the above description, each note n of the knitting note string M and each sounding section specified by the section specifying data GA have a one-to-one correspondence.

  The pitch adjusting unit 64 in FIG. 7 adjusts the pitch of each note n of the knitted note sequence M generated by the pitch extracting unit 62 to generate melody data DM. FIG. 9 is a flowchart of adjustment processing by the pitch adjustment unit 64. As shown in FIG. 9, the pitch adjustment unit 64 sequentially executes a first adjustment process SB1, a second adjustment process SB2, and a third adjustment process SB3. Note that the order of the adjustment processes (SB1 to SB3) can be changed as appropriate.

  In the first adjustment process SB1, the third acquisition unit 46 acquires the pitch of the note n designated as the important sound or the end sound in the section designation data GA among the plurality of notes n constituting the knitting note string M. This is a process of adjusting according to the progress data GC. Specifically, the pitch adjusting unit 64 specifies the pitch of each note n designated as an important note or a final note in the knitted note string M for the sound generation interval S of the note n by the chord progression data GC. Of the constituent sounds of the chord (for example, the chord at the start point of the sounding section S), the pitch is changed to the pitch closest to the pitch. The sound of the chord designated by the chord progression data GC is specified from the chord table TBL stored in the storage device 24.

  However, the following exception processing is executed from the viewpoint of ensuring musical naturalness. First, when the pitch of the note n of the important note is “F” or “B” and the root note of the chord designated by the chord progression data GC for the note n is “C”, The pitch of note n is not changed. That is, the pitch of the note n is maintained as “F” or “B”. Second, if the root note of the chord designated by the chord progression data GC is “C”, “F” or “G” for the note n of the end note, the pitch of the note n is closest to that pitch. The sound is changed to a pentatonic sound (C, D, E, G, A).

  The second adjustment process SB2 is a process of adjusting the pitch of the note n designated by the section designation data GA as the elapsed sound among the plurality of notes n constituting the knitting note string M. Specifically, the pitch adjusting unit 64 includes a predetermined range including the pitch of each note n designated as the elapsed note in the knitted note string M, including the pitch of the note n immediately before or after the note n. Change to the pitch within. For example, the pitch of each note n of the elapsed sound is a stem tone within a predetermined range (for example, a range of four semitones) with the pitch of the note n immediately before or after it as the center or the end point (upper limit or lower limit). C, D, E, F, G, A, B).

  The third adjustment process SB3 is an exception process when the chord progression data GC designates a specific chord for each note n (important sound, end sound, elapsed sound) of the knitting note string M. Specifically, when the pitch of the note n is “A” and the chord designated for the note n is Fm series (for example, Fm, Fm7, etc.), the pitch adjustment unit 64 Is changed from “A” to “G #”. When the pitch of the note n is “B” and the chord designated for the note n is Fm, Gm, A #, or C7, the pitch adjuster 64 determines that the note n The pitch is changed from “B” to “A #”.

  The melody generation unit 522 (pitch adjustment unit 64) in FIG. 7 generates melody data DM that specifies the adjusted knitted note string M described above. As shown in FIG. 8, the melody data DM is composed of time series of a plurality of unit data UM corresponding to each note n of the adjusted musical note string M. Each unit data UM includes pitch information UM1 for designating the pitch of the note n and time information UM2 for designating the sound generation section S (position and duration on the time axis) of the note n. Specifically, the melody data DM is described as a music file in the SMF format, like the pitch designation data DB. The relationship between each note n designated by the melody data DM and each syllable y designated by the designated character string GY is the relationship (the pronunciation interval S of each note n) and the relationship associated with the first acquisition unit 42 (the score division unit 424). Each syllable y of the designated character string GY is maintained in a one-to-one relationship).

  7 generates a speech signal VA corresponding to the melody data DM generated by the melody generation unit 522 and the designated character string GY generated by the first acquisition unit 42. The voice signal VA is an acoustic signal of a singing sound when each syllable y of the designated character string GY is uttered at a pitch specified by the note n corresponding to the syllable y in the melody data DM. That is, the audio signal VA corresponds to the singing sound of singing the organized note string M (melody line) with the designated character string GY as lyrics. For the generation of the audio signal VA, for example, a known unit connection type speech synthesis process can be suitably employed. That is, the speech synthesizer 524 sequentially selects speech segments corresponding to each syllable y designated by the designated character string GY, and each speech segment is selected from the pronunciation interval S of each note n designated by the melody data DM. The audio signal VA is generated by adjusting the pitch and connecting them to each other. The tempo of the audio signal VA is set to the tempo Xt set by the variable setting unit 32.

  1 generates an accompaniment signal VB indicating an accompaniment sound. As shown in FIG. 10, the accompaniment sound generation unit 54 includes an accompaniment data generation unit 542 and an accompaniment signal generation unit 544. The accompaniment data generation unit 542 generates accompaniment data DE corresponding to the chord progression data GC acquired by the third acquisition unit 46. The accompaniment data DE is a music file in the SMF format in which accompaniment sounds corresponding to the chords sequentially designated by the chord progression data GC are designated in time series. The type and rhythm of the accompaniment sound specified by the accompaniment data DE is set according to the style Xs set by the variable setting unit 32. The tempo of the accompaniment data DE is set to the tempo Xt set by the variable setting unit 32. The accompaniment signal generation unit 544 generates an accompaniment signal VB by executing predetermined processing (for example, musical tone generation processing by a MIDI sound source) on the accompaniment data DE. As understood from the above description, the audio signal VA and the accompaniment signal VB are set to the same tempo.

  The mixing unit 56 of FIG. 1 generates the acoustic signal V by mixing (weighted sum) of the audio signal VA generated by the singing sound generating unit 52 and the accompaniment signal VB generated by the accompaniment sound generating unit 54. The acoustic signal V generated by the mixing unit 56 is supplied to the sound emitting device 14 to reproduce the sound wave. As can be understood from the above description, the reproduced sound of the acoustic signal V is a performance sound of a song in which an accompaniment sound is added to a song sound sung by the designated character string GY of the melody indicated by the melody data DM.

  As described above, in the first embodiment, the section specifying data GA corresponding to any of the plurality of section specifying data DA and the pitch specifying data GB corresponding to any of the plurality of pitch specifying data DB are used. Using this, melody data DM is generated. Therefore, even if the section designation data GA is common, if the pitch designation data GB is different, melody data DM having different melody is generated, and even if the pitch designation data GB is common, the section designation data GA is If they are different, melody data DM having different melody is generated. That is, according to the first embodiment, for example, compared to the technique of Patent Document 2 in which only the pitches specified by the melody material data are changed, there are a variety of cases even when there are few songs selected as material for music generation. There is an advantage that music can be generated.

  Further, for example, the melody data DM is generated using the section designation data DA representing the rhythm pattern of the existing music and the pitch designation data DB representing the melody line of the existing music as templates. Compared with the technique of Patent Document 1 that randomly selects a melody, there is also an advantage that a musically natural melody can be generated. That is, according to the first embodiment, it is possible to generate musically natural and various melody.

  In the first embodiment, the pitch adjustment unit 64 adjusts the pitch of each note n of the knitting note string M generated from the section specifying data DA and the pitch specifying data DB. There is an advantage that a musically natural melody can be generated as compared with a configuration without adjustment. Specifically, in the first embodiment, the pitch of the note n of the important sound and the end sound is changed to the pitch of the constituent sound of the chord designated by the chord progression data GC (first adjusting process SB1). ), It is possible to shift musically natural pitches of notes that are particularly important in the melody. There is also an advantage that various melody can be generated according to the combination of the section designation data DA, the pitch designation data DB, and the chord progression data DC. Further, since the pitch of the note n of the elapsed sound is changed to a pitch within a predetermined range with respect to the pitch of the immediately preceding or subsequent note n (second adjustment process SB2), the pitch of the elapsed sound is also music. It is possible to transition naturally.

  In the first embodiment, each pronunciation section S designated by the section designation data GA and each syllable y of the designated character string GY are associated by the musical score section 424. Therefore, the correspondence between each pronunciation section S and each syllable y. There is an advantage that a voice signal VA of a clear and natural singing sound can be generated. In the first embodiment, in particular, as the number of syllables y in the designated character string DY increases, the section designation data DA having a larger number of pronunciation sections S is selected, so the number of pronunciation sections S and the number of syllables y are greatly increased. Compared to a configuration that can be different, each sound generation section S can be associated with each syllable y without difficulty. Further, the designated character string DY is divided for each block QB so that the difference between the number Ni of the sounding sections S and the number NYi of the syllables y is minimized for the blocks QB1 to QB4. Are associated with each other, it is possible to associate each sounding section S with each syllable y without difficulty in each of the blocks QB1 to QB4. In addition, the division of the pronunciation section S (SA2) and the insertion of the syllable y (SA3) to the designated character string DY are executed so that each pronunciation section S and each syllable y have a one-to-one correspondence. There is also an advantage that a voice signal VA of a natural singing sound in which one syllable y is assigned to the note n can be generated.

  In the first embodiment, since the accompaniment signal VB generated by the accompaniment sound generation unit 54 is added to the audio signal VA, it is possible to generate a musical composition rich in musicality as compared with the configuration in which the audio signal VA is reproduced alone. There are advantages. In the first embodiment, in particular, since the accompaniment signal VB is generated using chord progression data GC selected from a plurality of chord progression data DC, a variety of music is produced as compared with a configuration in which the accompaniment is fixed to one type. There is an advantage that can be generated. Moreover, in the first embodiment, the accompaniment data DE corresponding to the style Xs corresponding to the instruction from the user is generated, so that it is possible to generate music that matches the user's intention and preference. Further, since the audio signal VA and the accompaniment signal VB are set to a common tempo Xt, there is also an advantage that a musical piece in which the accompaniment sound and the melody sound are naturally matched can be generated.

Second Embodiment
A second embodiment of the present invention will be described below. In the first embodiment, a configuration in which the music generation device 100A is realized by a single device has been exemplified. In the second embodiment, the music generation device 100B having the same function as the music generation device 100A of the first embodiment is realized by the cooperation of a plurality of server devices that can communicate with each other. In addition, about the element which an effect | action and a function are equivalent to 1st Embodiment in each aspect illustrated below, each reference detailed in the above description is diverted and each detailed description is abbreviate | omitted suitably.

  FIG. 11 is a block diagram of a music generation device 100B according to the second embodiment. As shown in FIG. 11, the music generation device 100B of the second embodiment is a communication system (music generation system) that generates an acoustic signal V and provides it to the terminal device 10, and includes a management server device 70 and a melody generation server device. 72, an audio signal synthesis server device 74, an accompaniment generation server device 76, and an accompaniment signal synthesis server device 78. Each server device can communicate with each other via a communication network such as the Internet. The terminal device 10 is a communication terminal such as a mobile phone or a personal computer, and includes an input device 12 and a sound emitting device 14.

  The management server device 70 is a web server that communicates with the terminal device 10, and includes a variable setting unit 32, a character string setting unit 34, a storage unit 242, a third acquisition unit 46, and a mixing unit 56. The variable setting unit 32 sets the tempo Xt and the style Xs according to an instruction from the user to the input device 12 of the terminal device 10, and the character string setting unit 34 corresponds to the instruction from the user to the input device 12. Set the specified character string DY. The storage unit 242 stores a plurality of chord progression data DC. The third acquisition unit 46 acquires any of the plurality of chord progression data DC as chord progression data GC. The designated character string DY, tempo Xt and chord progression data GC are transmitted to the melody generation server device 72, and the chord progression data GC, tempo Xt and style Xs are transmitted to the accompaniment generation server device 76.

  The melody generation server device 72 includes a storage unit 244, a first acquisition unit 42, a second acquisition unit 44, and a melody generation unit 522. The storage unit 244 stores a plurality of section designation data DA and a plurality of pitch designation data DB. The first acquisition unit 42 selects any one of the plurality of section designation data DA stored in the storage unit 244 according to the designated character string DY, and executes the musical score processing of FIG. A designated character string GY is generated. The second acquisition unit 44 selects any one of the plurality of pitch designation data DB stored in the storage unit 244 as the pitch designation data GB. The melody generation unit 522 generates melody data DM according to the section designation data GA, the pitch designation data GB, and the chord progression data GC. The melody data DM, the designated character string GY, and the tempo Xt are transmitted to the audio signal synthesis server device 74.

  The voice signal synthesis server device 74 includes a voice synthesis unit 524 that generates a voice signal VA corresponding to the melody data DM, the designated character string GY, and the tempo Xt. The voice signal VA generated by the voice synthesizer 524 is transmitted to the management server device 70 via the melody generation server device 72.

  The accompaniment generation server device 76 generates accompaniment data DE corresponding to the chord progression data GC, the tempo Xt, and the style Xs. The accompaniment signal synthesis server device 78 includes an accompaniment signal generation unit 544 that generates an accompaniment signal VB according to the accompaniment data DE supplied from the accompaniment generation server device 76. The accompaniment signal VB generated by the accompaniment signal generation unit 544 is transmitted to the management server device 70 via the accompaniment generation server device 76.

  The mixing unit 56 of the management server device 70 generates the acoustic signal V by mixing the audio signal VA generated by the audio signal synthesis server device 74 and the accompaniment signal VB generated by the accompaniment signal synthesis server device 78. The acoustic signal V is transmitted to the terminal device 10 and reproduced as a sound wave from the sound emitting device 14.

  In the second embodiment, the same effect as in the first embodiment is realized. Note that the number of server devices constituting the music generation device 100B and the functions shared by each are appropriately changed from the example of FIG. For example, a single server device shares the functions of the melody generation server device 72 and the audio signal synthesis server device 74 of FIG. 11, and the single server device has the functions of the accompaniment generation server device 76 and the accompaniment signal synthesis server device 78. A shared configuration may also be employed.

<Modification>
Each of the above forms can be variously modified. Specific modifications are exemplified below. Two or more aspects arbitrarily selected from the following examples can be appropriately combined.

(1) The method by which the first acquisition unit 42 acquires the section designation data GA is changed as appropriate. For example, a configuration in which the first acquisition unit 42 randomly selects any one of the plurality of section designation data DA as the section designation data GA (that is, a configuration in which the score section 424 is omitted) is also employed. That is, the configuration for selecting the section designation data DA according to the designated character string DY and the configuration for executing the musical score processing on the section designation data DA can be omitted. As understood from the above description, the first acquisition unit 42 is included as an element for acquiring the section designation data GA corresponding to any of the plurality of section designation data DA. The section designation data GA includes both the section designation data DA and the section designation data DA itself after the musical score processing by the musical score section 424 as illustrated in the first embodiment. The operations of the second acquisition unit 44 and the third acquisition unit 46 are also changed as appropriate. For example, the second acquisition unit 44 selects one of the plurality of pitch designation data DB from the storage device 24 and executes a predetermined process (for example, a process of increasing the pitch of each note by a predetermined amount). A configuration for generating pitch designation data GB is also employed. The same applies to the third acquisition unit 46.

(2) In each of the above-described embodiments, the voice synthesizer 524 generates the voice signal VA corresponding to the designated character string GY, but the character string setting unit 34 and the voice synthesizer 524 may be omitted. That is, the present invention can also be realized as an apparatus for generating melody data DM specifying the knitting note string M, or an apparatus for generating an acoustic signal indicating the musical instrument sound of the knitting note string M specified by the melody data DM. .

(3) The method by which the accompaniment sound generation unit 54 generates the accompaniment signal VB is appropriately changed. For example, a plurality of accompaniment data DE corresponding to different styles can be stored in the storage device 24, and the accompaniment signal VB can be generated by selecting the accompaniment data DE corresponding to the style Xs designated by the user. Further, a configuration in which the accompaniment sound generation unit 54 is omitted (that is, a configuration that generates only a melody) may be employed.

(4) In each of the above-described embodiments, the pitch extraction unit 62 extracts the pitch specified by the pitch specification data GB for the start point of the sound generation section S specified by the section specification data GA. Is not limited to the starting point of the pronunciation period S. For example, the pitch specified by the pitch specification data GB at the time when a predetermined time has elapsed from the start point of the sounding section S or the middle point of the sounding section S is extracted as the pitch of the note n corresponding to the sounding section S. Is also possible. As understood from the above description, the melody generation unit 522 responds to each sound generation section S designated by the section designation data GA and the pitch designated by the pitch designation data GB for the reference time point of the sound generation section S. The reference time of the sound generation section S is a time point that is in a predetermined positional relationship with the sound generation section S (the start point and the middle point of the sound generation section S). Means.

(5) In each of the above-described forms, the first adjustment process SB1 is executed for the note n corresponding to the important sound and the end sound in the knitted note string M, and the second adjustment process SB2 is executed for the note n corresponding to the elapsed sound. However, it is also possible to execute the first adjustment process SB1 and the second adjustment process SB2 on all the notes n of the knitting note string M.

(6) The data format exemplified in each of the above embodiments is arbitrarily changed. For example, the section designation data DA, pitch designation data DB, melody data DM, and accompaniment data DE are not limited to the SMF format and can be described in any format. It is also possible to describe the chord progression data DC in the SMF format. The acoustic signal V, the audio signal VA, and the accompaniment signal VB are not limited to the time waveform sample series, and can be expressed as data specifying the time waveform in time series (for example, performance data compliant with the MIDI standard). .

100A, 100B ... Music generation device, 10 ... Terminal device, 12 ... Input device, 14 ... Sound emission device, 22 ... Arithmetic processing device, 24 ... Storage device, 32 ... Variable setting unit, 34 ... ... Character string setting section, 42 ... First acquisition section, 422 ... Selection section, 424 ... Musical score section, 44 ... Second acquisition section, 46 ... Third acquisition section, 52 ... Singing sound generation section 522 …… Melodic generator, 524 …… Speech synthesizer, 54 …… Accompaniment sound generator, 542 …… Accompaniment data generator, 544 …… Accompaniment signal generator, 56 …… Mixer, 62 …… Pitch Extraction unit 64... Pitch adjustment unit 70... Management server device 72 .. melody generation server device 74 .. speech signal synthesis server device 76 .. accompaniment generation server device 78 .. accompaniment signal synthesis server apparatus.

Claims (5)

First acquisition means for acquiring section specifying data corresponding to any of a plurality of section specifying data specifying a time series of pronunciation sections;
Second acquisition means for acquiring pitch designation data according to any of a plurality of pitch designation data for designating a time series of pitches;
At the time of a note corresponding to each sounding section specified by the section specifying data acquired by the first acquisition means and the pitch specified by the pitch specifying data acquired by the second acquisition means for the reference time of the sounding section A music generation device comprising: melody generation means for generating melody data for designating a series. The melody generation means includes
A pitch extraction means for generating a knitted note sequence in which notes having the respective tone generation sections designated by the section designation data and the pitch designated by the pitch designation data at the reference time of the sound generation section are arranged;
The music generating device according to claim 1, further comprising pitch adjusting means for adjusting the pitch of each note of the knitted note sequence generated by the pitch extracting means to generate the melody data. Character string setting means for acquiring a designated character string in which a plurality of audio units are arranged in time series,
The music generating device according to claim 1, further comprising: a voice synthesizing unit that generates a voice signal of a voice in which the designated character string is pronounced by a plurality of notes designated by the melody data generated by the melody generating unit. The first acquisition means includes
Selecting means for selecting any of the plurality of section designation data;
Including a musical score section that associates each sound generation section designated by the section designation data selected by the selection means with each voice unit of the designated character string,
The voice synthesis unit generates a voice signal of a voice that produces a voice unit corresponding to the sounding section in the sounding section of each of a plurality of notes designated by the melody data. Music generation device. Accompaniment sound generating means for generating an accompaniment signal indicating the accompaniment sound;
The music generating device according to claim 3 or 4, further comprising: a mixing unit that mixes the audio signal generated by the voice synthesizing unit and the accompaniment signal generated by the accompaniment sound generating unit.

Images