JP2015011080A - Karaoke system presenting piece of music sung by person having similar tendency of music selection as recommended piece of music - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide effective musical piece recommendation means for a karaoke user.SOLUTION: A singer having a similar tendency of music selection to an actual singer is identified from among singing history of a plurality of singers and notifies the actual singer of a piece of music that the identified singer has sung but the actual singer has not sung yet so as to tell the actual singer that "this is one you have not sung yet among pieces of music sung by a person having a similar tendency of music selection to you".

Description

  The present invention relates to a technique for recommending requested music to a karaoke user in a communication karaoke system.

  Conventionally, in a karaoke apparatus, a music search using a booklet (a quick reference book) in which a music name, a singer name, and a music identifier (music number) are displayed in association with each other has been common. In the music search using this booklet, the user searches for the music to be sung from the booklet, inputs the music identifier of the music to the remote control device, and transmits it to the karaoke device to reserve the music.

  In recent years, with the spread of online karaoke, the number of songs handled by karaoke devices has risen to tens of thousands. It is difficult to find a desired piece of music from among such a large number of pieces of music. Therefore, it is possible to search for a piece of music through a user interface using a database from a remote control device using a conventional numeric keypad. The remote control device is changing.

  Furthermore, as the number of songs increases, the user may be wondering which song to sing. Patent Literature 1 discloses a communication karaoke system capable of recommending music based on a performance order history collected from a karaoke performance terminal, that is, a history of what music was played after the previously selected music. ing.

Japanese Patent No. 4033944 Japanese Patent No. 3920204

  Patent Document 1 is a performance order totaling information file provided in a host device. Each music piece is associated with a music group list played next to the music, and the performance frequency of each music on each list is unified. Aggregated. The host device collects the performance order history data of each karaoke performance terminal in a timely manner, updates each performance frequency of each music group list in the file based on these data, and then distributes this file to each karaoke performance terminal in a timely manner To do. Each karaoke performance terminal refers to a file every time the accompaniment music of each song is played, extracts a predetermined number of songs in descending order of performance based on the song group list corresponding to the song, and plays these songs for the next performance. In order to recommend it as a candidate, it is displayed on the display device together with the song name and song selection number.

  Patent Document 2 discloses a communication karaoke system in which a karaoke host device and a large number of karaoke performance devices are coupled via a communication network. The karaoke host device collects performance history information in each karaoke performance device and A ranking list in which performance music pieces in the karaoke performance device are tabulated based on the performance frequency is created and distributed to each karaoke performance device. On the basis of the performance history information in, the identification information of a predetermined number of music performed following the music A is extracted, and the identification information of the predetermined number of higher-order music in the ranking list is excluded from the extracted information, The other music identification information is aggregated based on the number of performances, a recommended music list is created, and presented to the user as appropriate.

  Both Patent Literature 1 and Patent Literature 2 extract recommended songs from the history of songs sung, but both place importance on the order of songs, and the number of times a song is sung and then sung as the next song The logic that recommends the most songs is adopted.

In the karaoke store, new songs are coming in one after another, and the number of songs continues to increase. On the other hand, some users constantly pioneer new songs, but on the other hand, there are not a few people who continue to sing the same song as the repertoire. If you leave such a user bored with karaoke, there will be no development of karaoke.
Under such circumstances, Patent Document 1 and Patent Document 2 are proposals for recommending new songs, but both of them continue to sing the “next song” that was sung from the performance history of the karaoke device as “recommended songs”. It only proposes.
Certainly, it is difficult to recommend music that suits individual tastes, but the present invention aggregates the singing data of individual users and identifies “persons who tend to select songs similar to you” anonymously. The purpose is to present music that has been sung by "that person" that "you" has not yet sung and recommend recommended music that is personalized.

In order to solve the above problem, in claim 1 of the present invention, another singer having a tendency similar to the singing history is extracted based on the singing history. The song being sung is identified and notified as a recommended song.
In claim 2, a certain number of singing histories of a certain user are compared with a certain number of singing histories of another user, a user who has many overlapping songs is identified as a user with a similar singing tendency, and the user overlaps with that user. This is to notify a song that is not recommended as a recommended song.
In Claim 3, it has a singer music history database which manages member's information, and a music singer history database which registers the singer for every music, and the music is stored in the singer music history database every time it sings. While registering, the singer is registered in the music singer history database.
After singing a certain number (or for a certain period of time), from the database above, identify the individual who has the most similar music selection tendency to the singer, and select the song that the person sings and does not sing "Is displayed on the login screen of the singer.
In claim 4, the karaoke user is notified by registering the recommended music specified by the recommended music specifying means in the user database.

  When a song is sung, a song sung by a person with a similar music selection tendency is displayed as a recommended song in the singer's account, so that the singer can continue to be a karaoke user without adding a new repertoire and making it unmanageable. You can continue.

Singer song history database. Music singer history database. Aggregated data related to a specific user A. Similar frequency calculation table. Difference extraction result between Mr. A and Mr. E.

A song singer history database for managing member information and a song singer history database for registering singer for each song. Each time a song is sung, the song is registered in the singer song history database and the song is sung. A singer is registered in the person history database. After a certain period, each person's singing history was as follows. In addition, the following 1-8 shall be numbers which show music.
Mr. A: Song 1, Song 6, Song 5, Song 1, Song 6, Song 8
Mr. B: Song 2, Song 7, Song 6, Song 2, Song 7, Song 5
Mr. C: Song 3, Song 8, Song 4, Song 3, Song 8, Song 2
Mr. D: Song 4, Song 6, Song 3, Song 4, Song 7, Song 2
Mr. E: Song 5, Song 7, Song 2, Song 5, Song 6, Song 7
As a result of the registration, the singer music history database is as shown in FIG. 1, and the music singer history database is as shown in FIG.

At this time, for the specific karaoke user A from the singer music history database in FIG. 1 and the music singer history database in FIG. 2, the specific user in FIG. Aggregated data related to A is generated.
In order to extract the most frequently appearing singers while deleting the person A from the aggregated data related to the specific user A, the similar frequency calculation table of FIG. " Here, Mr. E has the highest “degree of similarity”.
Here, it is understood that the user who has selected the music closest to Mr. A is Mr. E.
From the songs that Mr. E has sung and Mr. A has not yet sung, in the order that Mr. E sang more, it is displayed on Mr. A's login screen as a recommendation to Mr. A. From the difference extraction result of Mr. A and Mr. E in FIG. 5, the songs that Mr. E has sung and that Mr. A has not sung are songs 2 and 7, so I recommend it.

Then, after Mr. A sings the next song, the above-described processing is performed using the updated singer song history database and the song singer history database, and the next recommended song is extracted.
As a result of repeating this, it may be assumed that the singing histories of Mr. A and Mr. E are exactly the same. In that case, the user who is most similar at that time other than Mr. E is extracted in the above logic. Then decide the recommended song.

In the above-described embodiment, the recommended song is determined by referring only to the singing history of one user (Mr. E) having the most similar tendency. You may extract the music which does not wear in common from a user's song history, and may show a recommendation music. As an example of the order of presentation,
1. The songs that are not worn by a plurality of users whose “similarity” is equal to or higher than a certain level are aggregated for all users and presented in descending order of frequency.
2. The songs that are not worn by a plurality of users whose “similarity” is equal to or greater than a certain level are presented in the order of the users whose similarity is high.
There is something like this.

Claims (4)

In a karaoke system having a user database that can identify a user who has selected a song by playing a song based on the selection by the user, and having a singing history for each user,
Based on the singing history, another singer whose music selection tendency is similar to that of the user is extracted, and the music that the singer sings is a song that is not sung by the user as a recommended song for the user. A karaoke system characterized by identifying and notifying. In a karaoke system having a user database that can identify a user who has selected a song by playing a song based on the selection by the user, and having a singing history for each user,
Compare a certain number of singing histories of a certain user with a certain number of singing histories of other users, identify a user with many overlapping songs as a user with similar singing tendency, and recommend songs that do not overlap with that user The karaoke system according to claim 1, wherein the karaoke system is notified as a song. In a karaoke system having a user database that can identify a user who has selected a song by playing a song based on the selection by the user, and having a singing history for each user,
Music registration means for registering a song sung by a user each time a song is sung in a singer's song history database configured around a singer,
At the same time, a singer registration means for registering the singer in a music singer history database composed of music,
About a certain karaoke user, the total number of singing songs by other singers is counted from the song singer history database for all songs registered in the singer's song history database of the user, and the degree of overlapping of singing songs with the user is high. Similar singer extraction means for extracting singers of
The karaoke system according to claim 2, further comprising a recommended song specifying unit that aggregates singer song histories of similar singers and creates a song list excluding the corresponding song in the singer song history of the user.   4. The user is notified by registering a recommended song identified by extracting another singer who has a similar music selection tendency to the user based on the singing history. 5. Karaoke system.

Images