JP4613924B2 - Song editing apparatus and program - Google Patents

Description

  The present invention relates to a music editing apparatus and program for editing music by connecting phoneme pieces.

As a technique related to song editing, there is a technique called audio mosaicing. In this audio mosaicing technology, various pieces of music are divided into phonemes having a short time length, and segment data indicating the waveform of each phoneme is collected to constitute a segment database. Then, desired segment data is selected from the segment database, the selected segment data is connected on the time axis, and a new song is edited. In addition, there exists a nonpatent literature 1 as literature regarding this kind of technique, for example.
Ari Lazier, Perry Cook, “MOSIEVIUS: FEATURE DRIVEN INTERACTIVE AUDIO MOSAICING”, [on line], Proc of the 6th Int. Conference on Digital Audio Effects (DAFx-03), London, UK, September 8-11, 2003, [Heisei Search on March 6, 19], Internet <URL: http: //soundlab.cs.princeton.edu/publications/mosievius_dafx_2003.pdf>

  By the way, in order to obtain expressive song data, it is necessary to prepare many types of segment data having various characteristics in advance, and select and connect appropriate ones of them. . However, there is a problem that it is difficult to find a suitable piece for each position of the edited music from a large amount of segment data.

  The present invention has been made in view of the circumstances described above, and provides a song editing apparatus and a program that make it easy to select segment data when performing song editing by connecting desired segment data. The purpose is that.

The present invention provides a user with information related to music editing, receives a user editing instruction from the user, storage means for storing segment data indicating an audio waveform of a predetermined time length, and storage means Clustering means for clustering the stored segment data, and cluster sequence data for designating the cluster of each phonemic segment constituting the edited song according to the progress of the song in accordance with an instruction given by the user via the user interface In accordance with the cluster sequence data generating means to be generated and an instruction given by the user via the user interface, the cluster of phonemic segments at each position of the edited music is based on the cluster sequence data generated by the cluster sequence data generating means. And store it in the storage means A song editing apparatus and a computer comprising the editing means for selecting a piece belonging to the cluster from the segment data and using the segment data string indicating the edited song. Provide a functioning program.
According to this invention, the user creates cluster sequence data in advance by the cluster sequence data generating means, selects the piece data belonging to the cluster indicated by the cluster sequence data by the editing means, and selects the piece data of the edited music piece. Can be used for columns. Therefore, it is possible to easily find appropriate segment data from a large number of segment data and edit the music.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram showing the configuration of a music editing apparatus according to an embodiment of the present invention. This music editing apparatus is obtained by installing a music editing program according to an embodiment of the present invention on a computer such as a personal computer.

  In FIG. 1, a CPU 1 is a control center that controls each part of the music editing apparatus. The ROM 2 is a read-only memory that stores a control program for controlling basic operations of the music editing device, such as a loader.

  The display unit 3 is a device for displaying an operation state of the device, input data, a message for an operator, and the like, and is configured by, for example, a liquid crystal display panel and a drive circuit thereof. The operation unit 4 is a means for receiving commands and various types of information from the user, and includes various types of operators. In a preferred embodiment, the operation unit 4 includes a keyboard and a pointing device such as a mouse.

  The interface group 5 includes a network interface for performing data communication with other devices via a network, a driver for transmitting / receiving data to / from an external storage medium such as a magnetic disk or a CD-ROM, and the like. It is comprised by.

  The HDD (hard disk device) 6 is a non-volatile storage device for storing information such as various programs and databases. The RAM 7 is a volatile memory used as a work area by the CPU 1. The CPU 1 loads a program in the HDD 6 into the RAM 7 and executes it in accordance with a command given via the operation unit 4.

  The sound system 8 is means for outputting a song edited by the song editing apparatus or a song in the middle of editing as a sound, and a D / A converter for converting a digital voice signal, which is sound sample data, into an analog voice signal; The amplifier includes an amplifier that amplifies the analog audio signal and a speaker that outputs the output signal of the amplifier as sound. In the present embodiment, the sound system 8, the display unit 3 and the operation unit 4 described above serve as a user interface that provides information related to song editing to the user and receives instructions related to song editing from the user. Fulfill.

  Information stored in the HDD 6 includes a song editing program 61, a song database 62, a segment database 63, a clustering database 64, and a cluster sequence database 65.

  The song database 62 is a collection of song data indicating various songs. As shown in FIG. 2, each piece of music data indicating each piece of music includes a song ID, song attribute data indicating the attribute of the song such as a song name, a genre name, and an artist name, and waveform data indicating the audio waveform of the song. Including. Here, the waveform data may be time-series sample data obtained by sampling and digitizing an audio waveform of a musical performance sound, or may be a compression-coded version of this time-series sample data.

  In a preferred embodiment, the song editing program 61 and song data are downloaded from a site in the Internet, for example, via an appropriate one in the interface group 5 and installed in the HDD 6. In another embodiment, the music editing program 61 and the music data are traded in a state stored in a computer-readable storage medium such as a CD-ROM or MD. In this aspect, the music editing program 61 and music data are read from the storage medium via an appropriate interface group 5 and installed in the HDD 6.

  The segment database 63 is a collection of segment data indicating various phonemes obtained by dividing the audio waveform indicated by the song data in the song database 62 into short time lengths such as one beat. As shown in FIG. 3, each piece data includes a piece ID, a piece attribute data, and waveform data indicating a waveform of the phone piece. Here, the segment ID is an identifier that uniquely indicates the segment data, the song ID of the song data that is the source of the phoneme indicated by the waveform data, and the sequence number that indicates the position of the phoneme within the song. Consists of. The element attribute data is, for example, waveform data such as Spectral Centroid that indicates the height of the sound element indicated by the waveform data, Loudness that indicates the volume, Brightness that indicates the brightness of the sound, and Noisiness that indicates the roughness of the sound. Is a collection of various feature quantities related to the phoneme segment indicated by. The song attribute data of the song data that is the source of the phoneme piece does not indicate the attribute existing in the phoneme segment, but can indicate the attribute existing in the phoneme segment. The phoneme segment indicated by the waveform data of the segment data is associated with the song attribute data that is the external attribute data of the phoneme segment via the song ID included in the segment ID of the segment data.

  The clustering database 64 is an aggregate of clustering data indicating the result of clustering each piece of data in the piece database 63 by a known clustering algorithm such as the k-NN algorithm. In the present embodiment, various conditions relating to clustering can be stored in the music editing program 61 in advance. If various conditions relating to this clustering are stored in advance, in the present embodiment, when a predetermined number or more of new piece data is added to the piece database 63, clustering under these conditions is executed, Clustering data corresponding to each condition is stored in the clustering database 64 (means for that will be described later).

  As shown in FIG. 4, each clustering data includes a clustering ID that is an identifier of the clustering data, a clustering condition description part that defines a clustering condition, and a cluster-specific segment ID group. Here, the unit ID group for each cluster is a unit ID group for each unit data belonging to cluster 1, a unit ID group for each unit data belonging to cluster 2, and so on. The segment data belonging to each of the clusters is shown.

  As a clustering condition, one or a plurality of feature amounts constituting the segment attribute data of the segment data in the segment database 63 can be designated. FIG. 6 shows the clustering of the segment data in the segment database 63 on the condition that a certain feature quantity 1 (for example, Loudness) and another feature quantity 2 (for example, Spectral Centroid) in the segment attribute data are used. The result of having performed is illustrated. When clustering is performed under such conditions, a cluster indicating a segment ID group of each segment data belonging to the illustrated cluster 1, a segment ID group indicating each segment data belonging to the cluster 2, etc. Clustering data including another segment ID group is created and stored in the clustering database 64. In the present embodiment, in addition to clustering based on such feature amounts, it is also possible to perform clustering based on music attribute data of music data that is the source of each piece data. In the case of performing such clustering, the music editing program 61 may store in advance the condition for performing clustering based on which data in the music attribute data.

  The cluster sequence database 65 is a collection of cluster sequence data. Here, each cluster sequence data is data indicating a so-called skeleton of various music pieces. As shown in FIG. 5, a cluster sequence ID that is an identifier of the cluster sequence data, a music ID, a clustering condition description section, And a sequence data part. The song ID is optional and can be included in the clustering sequence data or omitted. As illustrated in FIG. 7, the sequence data portion is for designating a cluster of each phoneme piece constituting a song along the progress of the song. The clustering condition description part is a part for designating the clustering condition assumed by the sequence data part, that is, the cluster obtained by each cluster indicated by the sequence data part.

  The cluster sequence data including the song ID can be generated, for example, in the following case. First, in this embodiment, cluster sequence data can be generated from song data in the song database 62 (details will be described later). In the present embodiment, the generated cluster sequence data can be registered in the cluster sequence database 65, and at that time, the music ID of the original music data can be included in the cluster sequence data. In the present embodiment, the user can generate new cluster sequence data by operating the operation unit 4, or can modify existing cluster sequence data to generate new cluster sequence data. The segment data can be generated by adding the segment data to the new cluster sequence data, and can be registered in the song database 62 as the song data of the new song. In this embodiment, new cluster sequence data that is the origin of this new song data can be registered in the cluster sequence database 65. At this time, the song ID of the song data generated using the cluster sequence data is assigned to the cluster. It can be included in the sequence data.

  The music editing program 61 generates a segment data string of the edited music according to the user's instruction given through the operation unit 4, and emits the data as a music from the sound system 8, or as music data in the music database 62. It is a program to register. The music editing program 61 according to the present embodiment is characterized in that the cluster sequence data described above can be generated in accordance with a user instruction given via the operation unit 4 and an appropriate segment based on the cluster sequence data. The point is that data can be selected and a segment data string of edited music can be generated.

  FIG. 8 is a block diagram showing the configuration of the song editing program 61. As shown in FIG. 8, the music editing program 61 includes a dividing unit 111, an analyzing unit 112, a segment data registering unit 113, a clustering unit 114, a cluster sequence data generating unit 121, and an editing unit. A unit 122, a cluster sequence data display unit 123, a reproduction unit 124, a feature amount curve generation unit 125, and a song data registration unit 126.

  Here, the dividing unit 111 divides the waveform data in the song data read from the song database 62 based on an instruction given via the operation unit 4 into a plurality of segment data having a shorter time length, and the RAM 7 It is a program which makes CPU1 perform the process stored in the element data storage area for the original music. In a preferred embodiment, the time length of the segment data is a length in units of beats, such as 1 beat, 2 beats, (1/2) beat, (1/4) beat, or the like. In this aspect, a known beat detection algorithm can be used as an algorithm for dividing music data into segment data. Alternatively, the waveform indicated by the music data may be analyzed to detect an attack portion, and each piece of data indicating the waveform of each section of 0.5 seconds from each detected attack portion may be used as segment data.

The analysis unit 112 analyzes each piece of piece data in the piece data storage area for the original music, and for each piece of piece data, Spectral
This is a program for causing the CPU 1 to execute processing for obtaining various feature quantities such as Centroid, Loudness, Brightness, Noisiness, etc. and storing them in the feature quantity storage area in the RAM 7. In a preferred embodiment, the analyzing unit 112 is automatically executed even when no instruction is given when the segmenting unit 111 is executed and the segment data string is stored in the original music segment data storage area. The

  When a segment data registration instruction is given via the operation unit 4, the segment data registration unit 113 stores each segment data stored in the segment data storage area for original music and the source of those segment data The piece data in the format shown in FIG. 3 is constructed using the song ID of the song data and the feature quantity of each piece data stored in the feature quantity storage area and registered in the piece database 63. It is a program that causes the CPU 1 to execute processing. More specifically, in the processing by the segment data registration unit 113, for each segment data stored in the feature amount storage area, the segment data is used as waveform data, and the source of the song data that is the source of the segment data is displayed. A piece ID is generated by using a piece ID and a sequence number indicating the position of the piece data in the piece as a piece ID, and using each feature quantity obtained from the piece data as piece attribute data. Register in the database 63.

  The clustering unit 114 is a program for clustering the segment data in the segment database 63 according to various conditions stored in the song editing program 61 and storing the clustering data as the clustering result in the clustering database 64. In the present embodiment, the clustering unit 114 is activated when a predetermined number or more of new piece data is added to the piece database 63.

  The cluster sequence data generation unit 121 is a program that causes the CPU 1 to execute processing for generating cluster sequence data as illustrated in FIG. 5 in the cluster sequence data storage area in the RAM 7 in accordance with an instruction given via the operation unit 4. is there.

In the process by the cluster sequence data generation unit 121, the user can select one of the following three processes as a process for generating the cluster sequence data.
[Processing a]: A GUI for drawing and inputting cluster sequence data as shown in FIG. 7 is provided by the operation unit 4 and the display unit 3, and the cluster sequence data drawn by the user is acquired via this GUI. .
[Process b]: One of the existing cluster sequence data stored in the cluster sequence database 65 is read. If necessary, modify this cluster sequence data.
[Processing c]: In the state in which the dividing unit 111 and the analyzing unit 112 described above are executed, the clustering condition that is a premise of the sequence data unit is specified (see FIG. 5), and each unit stored in the feature amount storage area Cluster sequence data is generated from data features.

  Further, in the processing by the cluster sequence data generation unit 121, when an instruction is given from the user via the operation unit 4, a cluster sequence ID that does not overlap with other cluster sequence data stored in the cluster sequence data storage area is set. If there is an instruction from the user, the song ID is added and registered in the cluster sequence database 65.

  The editing unit 122 selects the segment data in the segment database 63 in accordance with an instruction given from the user via the operation unit 4, and edits the segment data string of the edited song in the edited song segment data storage area in the RAM 7. This is a program that causes the CPU 1 to execute processing generated in the CPU.

In the processing by the editing unit 122, the user can select one of the following two types of processing as processing for generating the segment data string of the edited music.
[Processing d]: When the user designates a clustering method, the editing unit 122 causes the display unit 3 to display an icon indicating each cluster generated in the clustering method. When the user designates a desired cluster, the editing unit 122 causes the display unit 3 to display an icon indicating the piece data belonging to the cluster. When the user designates one of them, the editing unit 122 uses the designated piece data for the piece data string of the edited music piece.
[Processing e]: The editing unit 122 obtains a cluster at each position of the edited music piece according to the cluster sequence data in the cluster sequence data storage area, and the piece data belonging to the cluster among the piece data in the piece database 63 The segment data to be used for the segment data string of the edited music is selected.

In the process e, it is necessary to select the segment data to be used for the segment data string of the edited music from the plurality of segment data belonging to one cluster. As this selection process, the user can select one of the following two processes.
[Process e1]: The segment data belonging to the cluster specified by the cluster sequence data is displayed on the display unit 3, and one segment data is selected by operating the operation unit 4.
[Process e2]: By operating the operation unit 4, for at least one type of feature amount, a feature amount curve indicating a temporal change of the feature amount for each phoneme segment in a certain section is drawn, and each position in the section is drawn. Then, segment data having the feature quantity closest to the feature quantity indicated by the feature quantity curve is selected from the segment data belonging to the cluster indicated by the cluster sequence data.

  The feature amount curve generation unit 125 is a program that causes the CPU 1 to execute a process for obtaining a feature amount curve necessary for the process e2. The cluster sequence data display unit 123 is a program that causes the CPU 1 to execute a process for displaying the cluster sequence data in the cluster sequence data storage area on the display unit 3 in the manner illustrated in FIG. The playback unit 124 is a program that sends the segment data string in the edited song segment data storage area to the sound system 8 and causes the CPU 1 to execute a process of playing back the edited song. When a song data registration instruction is given via the operation unit 4, the song data registration unit 126 composes song data using the segment data string in the edited song segment data storage area, and stores the song data in the song database 62. This is a program for causing the CPU 1 to execute a process for registration.

  Next, the operation of the present embodiment will be described by taking as an example the case where cluster sequence data is generated, appropriate segment data is built using this cluster sequence data as a framework, and a segment data string of a new song is edited. First, the user gives an instruction to execute the music editing program 61 to the CPU 1 by operating the operation unit 4. When the CPU 1 is executing the song editing program 61, when the user instructs a process of editing a new song based on an existing song by operating the operation unit 4, the song editing program 61 An icon indicating each piece of music data stored in 62 is displayed on the display unit 3 to prompt the user to specify an original music piece as a base. When the user designates an icon indicating the desired original music by operating the operation unit 4, the music editing program 61 reads the music data of the original music from the music database 62 and delivers it to the dividing unit 111.

  The dividing unit 111 divides the waveform data of the music data delivered in this manner to generate a segment data string, stores the segment data string in the segment data storage area for the original music, and the analyzing unit 112 Launch. The analysis unit 112 analyzes each piece of data in the original music piece data storage area, obtains various feature values for each piece data, and stores them in the feature value storage area.

  In this state, the user can activate the cluster sequence data generation unit 121 by operating the operation unit 4. When the cluster sequence data generation unit 121 is activated by the operation of the operation unit 4, the cluster sequence data generation unit 121 displays a selection screen for cluster sequence data generation processing, and allows the user to select one of the above-described processes a, b, and c.

  When the user selects the process c by operating the operation unit 4, the cluster sequence data generation unit 121 refers to the clustering condition description part in each clustering data in the clustering database 64, and describes each clustering condition description part. Clustering conditions are displayed on the display unit 3.

  The user selects a desired clustering condition from the clustering conditions displayed on the display unit 3 by operating the operation unit 4. The cluster sequence data generation unit 121 generates cluster sequence data based on the clustering condition selected by the user in this manner with reference to each feature amount of each piece data in the feature amount storage area.

  For example, when the clustering condition selected by the user is “use feature quantities 1 and 2 for clustering”, the cluster sequence data generation unit 121 generates cluster sequence data according to the following procedure. First, feature amounts 1 and 2 are read out from various feature amounts for the first segment data stored in the feature amount storage area. Next, based on the feature quantities 1 and 2 with reference to the cluster-specific segment ID group of the clustering data corresponding to the clustering condition selected by the user in the clustering database 64 and each segment data in the segment database 63 When the segment data is clustered, it is determined to which cluster the first segment data characterized by the feature quantities 1 and 2 read from the feature quantity storage area belongs. Then, in accordance with this determination result, a portion for specifying a cluster of the first segment data in the cluster sequence data is generated. The cluster sequence data generation unit 121 executes the same for each subsequent segment data, generates cluster sequence data for designating the cluster of each phonemic segment in the edited music in the progression order of the music, and stores it in the cluster sequence data storage area. To store.

  When the cluster sequence data is obtained in the cluster sequence data storage area in this way, the user can activate the editing unit 122 by operating the operation unit 4. When the editing unit 122 is activated by the operation of the operation unit 4, the editing unit 122 causes the display unit 3 to display a selection screen related to the generation processing of the segment data string of the edited music, and selects one of the processes d or e. Here, when the process e is selected, the editing unit 122 activates the cluster sequence data display unit 123, and displays the cluster sequence data in the cluster sequence data storage area in the manner illustrated in FIG. To display. Further, a selection screen for selecting the process e1 or e2 is displayed on the display unit 3 to prompt the user to select.

  FIG. 9 shows the processing content of the editing unit 122 when the user selects the processing e1. For example, when the user wants to determine segment data at a desired time within the duration of the cluster 1 shown in FIG. 9, the cluster 1 on the display screen of the display unit 3 is operated by operating the mouse of the operation unit 4. The mouse pointer is moved to one point on the line segment indicating the duration of time (arrow L1), and the mouse is clicked. As a result, the editing unit 122 reads the segment data belonging to the cluster 1 out of the segment data in the segment database 63 and causes the display unit 3 to display the segment data. In addition, as illustrated in FIG. 7, there may be a case where the user designates one point in a section in which a plurality of clusters coexist, but in such a case, all of the coexisting clusters are assumed. The piece data may be displayed on the display unit 3.

  In a preferred embodiment, when the clustering condition used for generating the cluster sequence data is, for example, “use feature quantities 1 and 2 for clustering”, the editing unit 122, as shown, each unit belonging to the cluster 1 An image obtained by mapping the data to the two-dimensional coordinate system based on the feature amounts 1 and 2 is displayed on the display unit 3.

  In this state, when the user operates the mouse to indicate a desired segment data plot or a nearby point (arrow L2), the editing unit 122 transmits the segment data instructed by the user from the segment database 63. The waveform data in the segment data is read out and stored in the edited song segment data storage area as segment data corresponding to the time point indicated by the user (see arrow L1). Thereafter, the user can generate a segment data string of the edited music piece in the edited music piece data storage area by repeating the same operation.

  FIG. 10 shows the processing content of the editing unit 122 when the user selects the processing e2. In this process e2, the editing unit 122 starts up the feature amount curve generation unit 125. In this state, the user can specify a desired section in the edited music by operating the mouse of the operation unit 4, for example, and can draw a feature amount curve in the section. The feature amount in this case may be within the range of the clustering condition used for generating the cluster sequence data. For example, when the clustering condition is, for example, “use feature amounts 1 and 2 for clustering”, the user can draw a feature amount curve of feature amount 1 or draw a feature amount curve of feature amount 2 It is also possible to draw both feature amount curves of the feature amounts 1 and 2 respectively.

  When a feature amount curve is generated by a drawing process using a mouse for a section having a song, the editing unit 122 selects each piece of piece data belonging to the cluster indicated by the cluster sequence data at each position in the section. The segment data having the feature quantity closest to the feature quantity indicated by the feature quantity curve is selected and stored in the edited song segment data storage area as the segment data of the edited song. For example, as shown in FIG. 10, when a feature amount curve of feature amount 1 is generated for a section T1 within the duration of cluster 1, at each position in this section T1, each piece of piece data belonging to cluster 1 is The segment data having the feature quantity 1 closest to the feature quantity 1 indicated by the feature quantity curve is selected, and the segment data of the edited song is stored in the edited song segment data storage area. The user repeats selection of segment data by drawing such a characteristic amount curve for each section constituting the edited song, and generates a segment data string of the edited song in the edited song segment data storage area. Can do. Further, in the present embodiment, for example, for the section T1, for example, features with feature amount 1 and features 2 may be generated as feature amount curves. In this case, the editing unit 122, at each position in the section T1, from among the piece data belonging to the cluster 1, the feature quantity vector having the feature quantity 1 and the feature quantity 2 indicated by the two feature quantity curves as components. The segment data having the feature vector closest to is selected and stored in the edited song segment data storage area as the segment data of the edited song.

  While performing the operation of generating the segment data string of the edited music piece as described above, the user activates the reproduction unit 124 by operating the operation unit 4 and the piece data in the edited music piece data storage area. The train can be sent to the sound system 8 to play the edited song. The user may repeat the generation of the segment data string of the edited music by using the processing of the editing unit 122 described above until the edited music to be reproduced has a desired content. When the edited song has the desired content, the user can input a song data registration instruction by operating the operation unit 4 if necessary. When this song data registration instruction is given, the song data registration unit 126 constructs new song data using the segment data string in the edited song segment data storage area. At this time, a song ID that does not overlap with the song ID of the song data in the song database 62 is generated and added to the new song data. Also, each item of the song attribute data such as the song name and the genre name is input by the user through the operation of the operation unit 4 and added to the new song data. Then, the new song data configured in this way is registered in the song database 62.

  The operation of this embodiment has been described above by taking as an example the case where cluster sequence data is generated from existing song data. In this embodiment, the user generates cluster sequence data by operating the operation unit 4, or the cluster sequence It is of course possible to read the cluster sequence data in the database 65 and generate a segment data string of the edited music piece using this cluster sequence data. The operation for generating the segment data string of the edited music piece from the cluster sequence data in that case is as described above.

  As described above, according to the present embodiment, the user creates cluster sequence data in advance by the cluster sequence data generation unit 121, and the segment data belonging to the cluster indicated by the cluster sequence data by the editing unit 122. The segment data string of the edited music can be generated by selecting. Therefore, it is possible to easily find appropriate segment data from a large number of segment data and edit the music.

  Although one embodiment of the present invention has been described above, other embodiments are possible for the present invention. For example:

(1) The music editing program 61 may replace part or all of the program with an electronic circuit.
(2) When the cluster sequence data in the cluster sequence database 65 is read and displayed on the display unit 3 in accordance with an instruction given via the operation unit 4, the cluster sequence data includes the song ID, and When a reproduction instruction is given through the operation unit 4, the music system corresponding to the music ID may be reproduced by the sound system 8. Thus, the user can understand the relationship between the composition of the music visually confirmed (cluster sequence data) and the composition of the music confirmed visually.
(3) In the process of executing the process e2, for example, when a feature amount curve of feature amount 1 is drawn for a certain section, the upper limit and lower limit of the feature amount 1 of each piece data belonging to the cluster in that section are obtained, and this upper limit The coordinates indicating the upper and lower limit positions may be displayed on the display unit 3 so that the feature amount curve of the feature amount 1 is drawn within the range from to the lower limit.

It is a block diagram which shows the structure of the music editing apparatus which is one Embodiment of this invention. It is a figure which shows the structure of the music data in the embodiment. It is a figure which shows the structure of the segment data registered into the segment database 63 in the embodiment. It is a figure which shows the structure of the clustering data registered into the clustering database 64 in the embodiment. It is a figure which shows the structure of the cluster sequence data registered into the cluster sequence database 65 in the embodiment. It is a figure which shows the condition of the clustering of the segment data performed in the embodiment. It is a figure explaining the content which the cluster sequence data in the embodiment shows. It is a block diagram which shows the structure of the music edit program 61 in the embodiment. It is a figure which shows the 1st operation example of the edit part 122 in the same music edit program 61. FIG. It is a figure which shows the 2nd operation example of the edit part 122 in the music edit program 61. FIG.

Explanation of symbols

1 ... CPU, 2 ... ROM, 3 ... display unit, 4 ... operation unit, 5 ... interface group, 6 ... HDD, 7 ... RAM, 8 ... sound system, 61 ... song editing program 62 …… Song database, 63 …… Unit database, 64 …… Clustering database, 65 …… Cluster sequence database, 111 …… Division section, 112 …… Analysis section, 113 …… Segment data registration section, 114 ... ... clustering section, 121 ... cluster sequence data generation section, 122 ... editing section, 123 ... cluster sequence data display section, 124 ... playback section, 125 ... feature amount curve generation section, 126 ... song data registration section .

Claims (2)

A user interface that provides information about song editing to the user and receives instructions for song editing from the user;
Storage means for storing segment data including waveform data indicating a waveform of a phoneme having a predetermined time length in association with various attribute data including the feature amount of the phoneme;
Clustering means for performing clustering of the segment data stored in the storage means based on attribute data associated with the segment data ;
Cluster sequence data generating means for generating cluster sequence data for designating a cluster of each phonemic piece constituting the edited music along the progress of the music according to an instruction given from the user via the user interface;
Means for determining at least one type of feature quantity at each position in at least a section of the edited music according to an instruction given via the user interface;
In accordance with an instruction given from the user via the user interface, based on the cluster sequence data generated by the cluster sequence data generating means, a cluster of phonemes at each position of the edited music is obtained and stored in the storage means A means for selecting a piece belonging to the cluster from the piece data and generating a piece data string indicating the edited piece , wherein at least one type of feature quantity at each position is determined. A music editing apparatus comprising: an editing unit that selects segment data of a phoneme having a feature quantity closest to the determined feature quantity for at least some sections . Computer
A user interface that provides information about song editing to the user and receives instructions for song editing from the user;
Storage means for storing segment data including waveform data indicating a waveform of a phoneme having a predetermined time length in association with various attribute data including the feature amount of the phoneme;
Clustering means for performing clustering of the segment data stored in the storage means based on attribute data associated with the segment data;
Cluster sequence data generating means for generating cluster sequence data for designating a cluster of each phonemic piece constituting the edited music along the progress of the music according to an instruction given from the user via the user interface;
Means for determining at least one type of feature quantity at each position in at least a section of the edited music according to an instruction given via the user interface;
In accordance with an instruction given from the user via the user interface, based on the cluster sequence data generated by the cluster sequence data generating means, a cluster of phonemes at each position of the edited music is obtained and stored in the storage means A means for selecting a piece belonging to the cluster from the piece data and generating a piece data string indicating the edited piece, wherein at least one type of feature quantity at each position is determined. Editing means for selecting segment data of a phoneme having a feature quantity closest to the determined feature quantity for at least some sections;
A computer program characterized by functioning as a computer program.

Images