Patent Application
20150074090
The present disclosure relates to a media recommendation service for recommending media content to a user.
Recent years have seen a shift in music listening habits from physical media such as Compact Discs (COs) and cassettes to digital media stored on the user's playback device such as Moving Pictures Experts Group Layer 3 (MP3) files. This transition has made music much more accessible to listeners worldwide. However, the increased accessibility of music has only heightened a long-standing problem for the music industry, which is namely the issue of linking audiophiles with new music that matches their listening preferences.
Many companies, technologies, and approaches have emerged to address this issue of music recommendation. Some companies have taken an analytical approach. They review various attributes of a song, such as melody, harmony, lyrics, orchestration, vocal character, and the like, and assign a rating to each attribute. The ratings for each attribute are then assembled to create a holistic classification for the song that is then used by a recommendation engine. The recommendation engine typically requires that the user first identify a song that he or she likes. The recommendation engine then suggests other songs with similar attributions. Companies using this type of approach include Pandora (http://www.pandora.com), Sound Flavor (http://www.soundflavor.com), MusiciP (http://www.musicip.com), and MongoMusic (purchased by Microsoft in 2000).
Other companies take a communal approach. They make recommendations based on the collective wisdom of a group of users with similar musical tastes. These solutions first profile the listening habits of a particular user and then search similar profiles of other users to determine recommendations. Profiles are generally created in a variety of ways such as looking at a user's complete collection, the playcounts of their songs, their favorite playlists, and the like. Companies using this technology include Last.fm (http://www.last.fm), Music Strands http://www.musicstrands.com). WebJay (http://www.webjay.org), Mercora (http://www.mercora.com), betterPropaganda (http://www.betterpropaganda.com), Loomia (http://www.loomia.com), eMusic (http://www.emusic.com), musicmatch (http://www.mmguide.musicmatch.com), genielab (http://genielab.com/), upto11 (http://www.upto11.net/), Napster (http://www.napster.com), and iTunes (http://www.itunes.com) with its celebrity playlists.
The problem with the prior art is that it fails to incorporate a pop culture historical context to music recommendation. For many users, musical preferences are heavily influenced by the type and nature of music that was popular at the time they became interested in music, as witnessed by the popularity of oldies radio stations (i.e., stations playing music from the 1950s) and classic rock stations (i.e., stations playing music from the 1970s and 1980s). Further, their musical tastes are also influenced by the type and nature of other media (such as movies and television) that were popular at particular points in time. For instance, movies such as Saturday Night Fever, Grease, Chariots of Fire, Top Gun, Footloose, Flashdance, The Breakfast Club, and Pretty Woman and television shows such as MASH, Cheers, Happy Days, Fame, and Moonlighting all had soundtracks that enjoyed great popularity in their respective time periods.
For example, a user may like a particular artist from the early 1980s, such as Michael Jackson, but they might also enjoy songs from the television show Fame and the film Top Gun. However, recommendation systems based on music similarity would fail to make these suggestions to the user. As for communal based recommendation systems, they may perform marginally better in this case. More specifically, since communal based recommendation systems rely primarily on musical relationships set forth by particular users, they will inherently reflect some degree of pop culture association in their music recommendations. In other words, some people will remember artists, songs, and movies that were popular at the same time as another given artist or song or movie, and their collections and profile will reflect these associations. However, memory can be unreliable, and it is dubious that one user or collection of users can make all of the associations.
Therefore, there is a need for a media recommendation system that provides recommendations based on a historical context.
The present invention provides a media recommendation system for recommending media content that is historically related to seed media content. The recommended media content may be songs, television programs, movies, or a combination thereof, and the seed media content may be a song, television program, or movie.
In one embodiment, the media recommendation service provides music recommendations to the user based on a seed song. In general, information identifying the seed media content (e.g., a song) is provided to a server. The server obtains historical information for the seed media content such as the release date of the seed media content, the date the seed media content first appeared on the music charts, the duration of the seed media content on the music charts, the date the seed media content obtained its peak position on the music charts, and the like. Based on the historical information, the server determines a time window for the seed media content. The server then queries a media content database (e.g., a music database) based on the time window to identify historically related songs or albums associated with media content historically related to the seed media content. In an embodiment the server queries a music database to determine songs that were popular during the same time window as the seed media content. The server in an embodiment may also query a television database and a movie database to identify songs or albums associated with television programs or movies that were popular during the same time window as the seed media content. In an embodiment, a list of the historically related songs (or albums) are received by the server. In an embodiment, historically related songs (or albums) obtained from the various databases are aggregated and the list of historically related songs (or albums) are created based on the aggregation. In a similar fashion, other types of historically related media content such as television programs and movies may be determined based on the seed media content.
In another embodiment, the media recommendation service provided by the server may recommend music, television programs, movies, or any combination thereof based on a seed song, seed album, seed television program, or seed movie.
Those skilled in the art will appreciate the scope of the present Invention and realize additional aspects thereof after reading the following detailed description in association with the accompanying drawings.
The accompanying drawings incorporated in and forming a part of this specification illustrate several aspects of the invention, and together with the description serve to explain the principles of the invention.
The embodiments set forth below represent the necessary information to enable those skilled in the art to practice the invention and illustrate the best mode of practicing the invention. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the invention and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the disclosure and the accompanying claims.
[
The central server 12 includes a recommendation engine 18. The recommendation engine 18 may be implemented in software, hardware, or a combination of software and hardware. The recommendation engine 18 includes a music engine 20, a television engine 22, and a movie engine 24. As discussed below, the music engine 20 communicates with recommendation service 26 to query a music database 28 to obtain historical information for a seed song, where in this embodiment the recommendation service 26 and the music database 28 are hosted by a music server 30. The recommendation engine 18 then determines a time window for the seed song based on the historical information, and the music engine 20 sends a request to the recommendation service 26 such that the music database 28 is queried to identify other songs historically related to the seed song. The television engine 22 communicates with a recommendation service 32 to query a television database 34 based on the time window of the seed song to identify songs associated with television programs that are historically related to the seed song. In this embodiment, the recommendation service 32 and the television database 34 are hosted by a television server 36. In a similar fashion, the movie engine 24 operates to communicate with a recommendation service 38 to query a movie database 40 based on the time window of the seed song to identify songs associated with movies that are historically related to the seed song. In this embodiment, the recommendation service 38 and the movie database 40 are hosted by a movie server 42.
While the databases 28, 34, and 40 are illustrated as three separate databases hosted by three separate servers, the present invention is not limited thereto. For example, the television and movie databases 34 and 40 may be hosted by a single server and may be implemented as a single database. Further, each of the databases 28, 34, and 40 may be implemented as a single database or as a number of distributed databases. Existing databases that may be used or modified to be used by the system 10 include, but are not limited to, the Gracenote music database available at http://www.gracenote.com, the Billboard music charts available at http://www.billboard.biz, the Classic TV database available at http://www.classic-tv.com, and the IMDB movie and television database available at http://www.imdb.com.
The user device 14 may be, for example, a personal computer, a laptop computer, a Personal Digital Assistant (PDA), mobile telephone, or the like having access to the network 16. In this embodiment, the user device 14 includes a media player 44, a recommendation queue 46, a media collection 48, and media metadata 50. The media player 44 may be implemented in software, hardware, or a combination of hardware and software and operates to play media content such as, for example, songs and videos. As discussed below, in one embodiment, the recommendation services of the central server 12 are invoked when a song is played by the media player 44. The recommendation queue 46 may be implemented as part of the media player 44 or as part of an application associated with the central server 12 and operates to store recommendations, such as a list of recommended songs, provided by the central server 12. The recommendations may be presented to the user via the media player 44 or, alternatively, a custom application.
In one embodiment, the media collection 48 includes a number of songs forming the user's media, or music, collection. The media collection 48 may additionally include a number of television programs such as episodes of “Lost,” a number of movies, or both television programs and movies. The media metadata 50 is information describing the content of the media collection 48. Note that while the media metadata 50 is illustrated as being separate from the media collection 48, the media metadata 50 may alternatively be stored as part of the media files forming the media collection 48. As an example, for a song, the media metadata 50 may include 103 tags, which include information such as, but not limited to, the title of the song, artist of the song, album from which the song came, date of release, lyrics, and the like. Further, the metadata for the song may be included within the digital song file. Alternatively, the metadata for the song may be stored in an application file associated with, for example, the media player 44.
The central server 12 then identifies the song based on the seed song identification information (step 102). Once identified, the central server 12, and more specifically the music engine 20 of the recommendation engine 18, requests historical information for the seed song from the music server 30 (step 104), and the music server 30 obtains the requested historical information from the music database 28 and returns the historical information to the central server 12 (step 106). The historical information may include information identifying any and all albums on which the seed song was released, a release date for the seed song, the release date for each album on which the seed song was released, a date that the seed song first appeared on the music charts, a duration or range of dates that the seed song remained on the music charts, a date that the seed song first went off of the music charts, the beginning position of the seed song on the music charts, the highest position the seed song obtained on the music charts, the position of the seed song on the music charts immediately before leaving the music charts, and the like. All or a portion of the historical data may be included within the metadata for the song provided from the user device 14. If so, the recommendation engine 18 may obtain the historical information from the metadata rather than the music server 30. Exemplary music charts are the Billboard music charts, the Contemporary Hit Radio music charts, the American Bandstand music charts, the American Top 40 music charts, the VH1 music charts, the MTV music charts, the Hit Parade music charts, the UK Singles Charts, and EuroChart. However, other music charts may additionally or alternatively be used. Note that media popularly charts may be used herein to generally refer to music charts, television charts, and movie charts providing the relative popularity songs, music albums, television programs, and movies.
Once the historical information is obtained, the recommendation engine 18 identifies a time window for the seed song (step 108). While this discussion focuses on a single time window, multiple time windows for the seed song may be identified and used by the recommendation engine 18. Generally, the time window for the song is a time period during which the seed song was popular in society. More specifically, as a first example, the time window may be a predetermined time period around the initial release date of the seed song. Thus, if the seed song was released in June of 1983, the time window may be identified as, for example, the two month period of Jun. 1, 1983 to Jul. 31, 1983. As a second example, the time window may correspond to the time period that the seed song was on the music charts. Thus, if the seed song was on the Billboard music charts from Mar.15, 1985 to Jun. 1, 1985, the time window may be, for example, Mar. 15, 1985 to Jun. 1, 1985 or Mar. 1, 1985 to Jun. 30, 1985. As a third example, the time window may correspond to the time that the seed song first occupied the music charts, the time that the seed song reached its peak position on the music charts, the time that the seed song exited the music charts, a time that the seed song was re-released, or the like.
After identifying the time window for the seed song, the music engine 20 of the central server 12 sends a request to the music server 30, and more specifically the recommendation service 26, for other songs historically related to the seed song (step 110). The request includes the time window for the seed song. In response, the recommendation service 26 queries the music database 28 based on the time window for the seed song and sends a response including a list of songs historically related to the seed song to the central server 12 (step 112). The songs returned by the recommendation service 26 of the music server 30 may include, for example, songs released during the time window for the seed song, songs that were on one or more music charts during the time window for the seed song, songs by artists that had a song on one or more music charts during the time window for the seed song, songs by artists who released a song or album during the time window for the seed song, or the like.
The number of songs returned by the music server 30 may be reduced by using additional criteria. As a first example, either the recommendation engine 18 or the recommendation service 26 may randomly select songs from one or more music charts during the time window for the seed song rather than selecting all songs from the music charts during the time window for the seed song. As a second example, either the recommendation engine 18 or the recommendation service 26 may select only those songs from one or more music charts during the time window for the seed song that occupied a position in the charts above a threshold such as the 10th position on the charts. As a third example, songs returned by the music server 30 may be filtered based on criteria such as one or more music attributes of the seed song or based on communal profiles for the seed song.
Next, the television engine 22 of the central server 12 sends a request to the television server 36, and more specifically the recommendation service 32, for songs associated with television programs historically related to the seed song (step 114). The request includes the time window for the seed song. In response, the television server 36 returns a list of songs associated with television programs historically related to the seed song (step 116). More specifically, using the time window for the seed song, the recommendation service 32 queries the television database 34 to identify television programs that were popular or originally released during the time window for the seed song. The recommendation service 32 then identifies any songs, such as soundtracks, associated with the identified television programs. For example, television shows such as “MASH,” “Cheers,” “Happy Days,” “Fame,” and “Moonlighting” had soundtracks that enjoyed great popularity during their respective time periods. Note that once the recommendation service 32 determines that a television program has an associated soundtrack, the recommendation service 32 may query the music server 30 to identify the songs on the soundtrack if the television database 34 does not include this information. Still further, the recommendation service 32 may determine whether the directors, producers, and actors associated with the identified television programs are also associated with other television shows. If so, the recommendation service 32 may further identify songs for those television shows.
Next, the movie engine 24 of the central server 12 sends a request to the movie server 42, and more specifically the recommendation service 38, for songs associated with movies historically related to the seed song (step 118). The request includes the time window for the seed song. In response, the movie server 42 returns a list of songs associated with movies historically related to the seed song (step 120). More specifically, using the time window for the seed song, the recommendation service 38 queries the movie database 40 to identify movies that were popular or originally released during the time window for the seed song. The recommendation service 38 then identifies any songs, such as soundtracks, associated with the identified movies. Note that once the recommendation service 38 determines that a movie has an associated soundtrack, the recommendation service 38 may query the music server 30 to identify the songs on the soundtrack if the movie database 40 does not include this information. Still further, the recommendation service 38 may determine whether the directors, producers, and actors associated with the identified movies are also associated with other movies. If so, the recommendation service 38 may further identify songs for those movies.
The number of songs returned by the music, television, and movie servers 30, 36, and 42 may be reduced by filtering the songs either at the recommendation engine 18 or the respective recommendation services 26, 32, and 38 based on music attributes, where only songs having music attributes similar to those of the seed song will be recommended to the user. In another embodiment, the songs returned by the servers 30, 36, and 42 may be filtered at the user device 14. The music attributes may include, for example, but are not limited to genre, melody, tone, beats-per-minute, and the like. The number of songs may additionally or alternatively be reduced by using communal user music profiles to identify songs that are similar to the seed song. More specifically, the system 10, and more specifically the central server 12, may maintain profiles for each user in the system 10, where the profiles may include information such as, but not limited to, information identifying the songs in the user's music collection, the playcounts of the songs in the user's music collection, favorite playlists, and the like. Using the profiles of other users having the seed song in their music collections or on their favorite playlists, the recommendation engine 18, or alternatively the user device 14, may filter the songs from the servers 30, 36, and 42 based on the profiles of those users. For example, songs from the servers 30, 36, and 42 that do not appear in the music collection of other users who have the seed song in their music collection or on their favorite playlists may be filtered, where the filtered songs are not recommended to the user at the user device 14.
At this point, the central server 12, and more specifically the recommendation engine 18, generates a list of recommended songs based on the responses from the servers 30, 36, and 42 (step 122). The list of recommended songs may be in any desired order. In one embodiment, the list first includes songs from the music server 30, then songs from the television server 36, and then songs from the movie server 42. In another embodiment, the songs from the servers 30, 36, and 42 are randomly sorted to provide the list of recommended songs. In yet another embodiment, the songs from the servers 30, 36, and 42 may be assigned a score or priority based on a comparison of the music attributes of the songs with the music attributes of the seed song. Then, the songs may be sorted by their scores or priorities to provide the list of recommended songs.
The list of recommended songs is then provided to the user device 14 (step 124). The list of recommended songs may include information identifying the recommended songs, Uniform Resource Locators (URLs) enabling download or streaming of the recommended songs from a subscription based music service, URLs enabling purchase and download of the recommended songs from an a-commerce site, URLs enabling playback of samples of the recommended songs, samples of the recommended songs, or the like, or any combination thereof. Once at the user device 14, the recommended songs may be presented to the user of the user device 14 (step 126). For example, the recommended songs may be presented as a list including the information identifying the songs with URLs enabling playback of samples and, if desired, purchase of the recommended songs. The recommended songs may alternatively be presented to the user at the user device 14 as a playlist for the media player 44. This may be particularly beneficial where the user has access to a subscription based music service.
The present invention provides substantial opportunity for variation without departing from the spirit or scope of the present invention. Numerous variations will be apparent to one of ordinary skill in the art upon reading this disclosure. For example, a single server may host all three of the databases 28, 34, and 40. As another example, the databases 28, 34, and 40 may be stored locally at the central server 12. As yet another example, while the time window for the seed song provided to each of the servers 30, 36, and 42 is the same in the discussion above, the present invention is not limited thereto. The narrowness or breadth of the time window for each of the servers 30, 36, and 42 may be determined based on, for example, a priority assigned to each of the servers 30, 36, and 42. For instance, the music server 30 may have a higher priority than the television and music servers 36 and 42. As such, the time window for the request sent to the music server 30 may be broader than the time windows for the requests sent to the television and music servers 36 and 42.
In addition, while the discussion above focuses on recommending songs historically related to a seed song, the present invention is not limited thereto. More specifically, the central server 12 may recommend songs, music albums, television programs, movies, or any combination thereof based on a seed song, seed television program, or seed movie. For example, based on a seed song, the recommendation engine 18 of the central server 12 may query the music server 30 to identify songs historically related to the seed song, query the television server 36 to identify television programs historically related to the seed song and songs associated with the historically related television programs, and query the movie server 42 to identify movies historically related to the seed song and songs associated with the historically related movies. The central server 12 may then recommend songs, television programs, and movies historically related to the seed song.
In a similar fashion, if a seed television program is used, the central server 12 may query the television server 36 to obtain historical information for the seed television program, and then determine a time window for the seed television program. The recommendation engine 18 of the central server 12 may query the music server 30 to identify songs historically related to the seed television program, query the television server 36 to identify television programs historically related to the seed television program and songs associated with the historically related television programs, and query the movie server 42 to identify movies historically related to the seed television program and songs associated with the historically related movies. The central server 12 may then recommend songs, television programs, and movies historically related to the seed television program.
If a seed movie is used, the central server 12 may query the movie server 42 to obtain historical information for the seed movie, and then determine a time window for the seed movie. The recommendation engine 18 of the central server 12 may query the music server 30 to identify songs historically related to the seed movie, query the television server 36 to identify television programs historically related to the seed movie and songs associated with the historically related television programs, and query the movie server 42 to identify movies historically related to the seed movie and songs associated with the historically related movies. The central server 12 may then recommend songs, television programs, and movies historically related to the seed movie.
Those skilled in the art will recognize improvements and modifications to the preferred embodiments of the present invention. All such improvements and modifications are considered within the scope of the concepts disclosed herein and the claims that follow.
This application is a continuation of U.S. patent application Ser. No. 11/471,797, which was filed Jun. 21, 2006, the disclosure of which is hereby incorporated by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11471797 | Jun 2006 | US |
| Child | 14546592 | US |
