Method and apparatus for programmatically generating audio file playlists

• US 7,678,984 B1
• Filed: 10/13/2005
• Issued: 03/16/2010
• Est. Priority Date: 10/13/2005
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Lamere, Paul B.
• Primary Examiner(s): Donels; Jeffrey
• Assistant Examiner(s): UHLIR, CHRISTOPHER J
• Application Number:
  US11/250,085
• Time in Patent Office: 1,615 Days
• Field of Search: 846/15
• US Class Current: 84/615
• CPC Class Codes: G10H 1/0033 Recording/reproducing or transmission of music for electrophonic musical instruments
• CPC Class Codes: G10H 2240/131 Library retrieval, i.e. searching a database or selecting a specific musical piece, segment, pattern, rule or parameter set
• CPC Class Codes: G10H 2250/015 Markov chains, e.g. hidden Markov models [HMM], for musical processing, e.g. musical analysis or musical composition
• CPC Class Codes: G11B 27/034 on discs G11B27/036, G11B27/038 take precedence
• CPC Class Codes: G11B 27/322 used signal is digitally coded

Abstract

Method and apparatus for programmatically generating interesting audio file playlists. A playlist generation mechanism may use an N-gram model of audio file ordering patterns found in a collection of human-generated playlists to automatically generate new playlists. Given play histories indicating one or more played audio files as input, statistical methods may be used to look for sequences of audio files that occur a statistically significant number of times in the N-gram model for inclusion in new, interesting playlists that incorporate the human element found in the collection of playlists. In some embodiments, one more backoff probability methods may be used to provide additional candidate audio files for playlists if there is insufficient coverage for an audio file in the N-gram model. In one embodiment, a class-based statistical model incorporating higher-level statistics for the audio files may be used to weight selection of audio file transitions from the N-gram model.

Claims

1. A method for automatically generating playlists of audio files, the method comprising:

collecting a plurality of audio file playlists into a playlist database, wherein each audio file playlist indicates an ordered sequence of audio files;

generating an N-gram statistical model of the audio file playlists in the playlist database, wherein the N-gram statistical model models ordering sequences of the audio files indicated in the playlists;

receiving a request to generate a new playlist; and

in response to receiving the request;

determining an initial ordered sequence of one or more audio files; and

generating a new playlist comprising an ordered sequence including an indication of one or more of the audio files from the playlist database, wherein said generating a new playlist comprises applying the initial ordered sequence to the N-gram statistical model.

2. The method as recited in claim 1, the method further comprising providing the new playlist to an audio file player.

3. The method as recited in claim 2, wherein the audio file player is a portable audio file player.

4. The method as recited in claim 1, wherein the plurality of audio file playlists includes human-generated playlists.

5. The method as recited in claim 1, wherein the request includes information indicating the initial ordered sequence of one or more audio files.

6. The method as recited in claim 1, wherein the request includes audio file class information for determining an initial ordered sequence of one or more audio files, and wherein said determining an initial ordered sequence of one or more audio files comprises selecting one or more audio files as the initial ordered sequence according to the class information in the request.

7. The method as recited in claim 1, wherein the N-gram statistical model is a bi-gram model of the playlist database, and wherein said generating a new playlist comprises applying the initial ordered sequence to the bi-gram model.

8. The method as recited in claim 1, wherein the N-gram statistical model is a tri-gram model, and wherein said generating a new playlist comprises applying the initial ordered sequence to the tri-gram model.

9. The method as recited in claim 1, wherein said generating a new playlist comprises, if there is insufficient coverage in the N-gram statistical model for the initial ordered sequence of one or more audio files, applying the initial ordered sequence to a lower-level N-gram statistical model of the playlist database.

10. The method as recited in claim 1, further comprising generating a class-based statistical model of the playlist database, wherein the class-based statistical model comprises higher-level statistics for the audio files indicated by the playlists in the playlist database, and wherein the higher-level statistics are derived from information about the audio files indicated by the playlists other than ordering sequences.

11. The method as recited in claim 10, wherein the audio files are digital music files, and wherein the information about the audio files indicated by the playlists includes album information, artist information, genre information, and thematic information.

12. The method as recited in claim 10, wherein said generating a new playlist comprises, if there is insufficient coverage in the N-gram statistical model for the initial ordered sequence of one or more audio files, applying the initial ordered sequence to the class-based statistical model of the playlist database.

13. The method as recited in claim 1, further comprising:

receiving user-specific data indicating one or more audio file listening preferences of a user that generated the request; and

in said generating a new playlist comprising an ordered sequence including an indication of one or more of the audio files from the playlist database, weighting selection of the one or more of the audio files from the playlist database according to one or more of the audio file listening preferences of the user.

14. A system, comprising:

one or more processors; and

a memory comprising program instructions executable by the one or more processors to;

receive a request to generate a new playlist; and

in response to receiving the request;

determine an initial ordered sequence of one or more audio files; and

generate a new playlist comprising an ordered sequence including an indication of one or more of the audio files from a playlist database, wherein said generating a new playlist comprises applying the initial ordered sequence to an N-gram statistical model of the playlist database;

wherein the playlist database comprises a plurality of audio file playlists, wherein each audio file playlist indicates an ordered sequence of audio files; and

wherein the N-gram statistical model models ordering sequences of the audio files indicated in the playlists.

15. The system as recited in claim 14, wherein the plurality of audio file playlists includes human-generated playlists.

16. The system as recited in claim 14, wherein the request includes information indicating the initial ordered sequence of one or more audio files.

17. The system as recited in claim 14, wherein the request includes audio file class information for determining an initial ordered sequence of one or more audio files, and wherein, to determine an initial ordered sequence of one or more audio files, wherein the program instructions are further configured to select one or more audio files as the initial ordered sequence according to the class information in the request.

18. The system as recited in claim 14, wherein, to generate a new playlist, the program instructions are further executable by the processor to, if there is insufficient coverage in the N-gram statistical model for the initial ordered sequence of one or more audio files, apply the initial ordered sequence to a lower-level N-gram statistical model of the playlist database.

19. The system as recited in claim 14,

wherein, to generate a new playlist, the program instructions are further executable by the processor to, if there is insufficient coverage in the N-gram statistical model for the initial ordered sequence of one or more audio files, apply the initial ordered sequence to a class-based statistical model of the playlist database;

wherein the class-based statistical model comprises higher-level statistics for the audio files indicated by the playlists in the playlist database, wherein the higher-level statistics are derived from information about the audio files indicated by the playlists other than ordering sequences.

20. A computer-accessible medium, comprising program instructions, wherein the program instructions are computer-executable to implement a playlist generation mechanism configured to:

receive a request to generate a new playlist; and

in response to receiving the request;

determine an initial ordered sequence of one or more audio files; and

generate a new playlist comprising an ordered sequence including an indication of one or more of the audio files from a playlist database, wherein said generating a new playlist comprises applying the initial ordered sequence to an N-gram statistical model of the playlist database;

wherein the playlist database comprises a plurality of audio file playlists, wherein each audio file playlist indicates an ordered sequence of audio files; and

wherein the N-gram statistical model models ordering sequences of the audio files indicated in the playlists.

21. The computer-accessible medium as recited in claim 20, wherein the plurality of audio file playlists includes human-generated playlists.

22. The computer-accessible medium as recited in claim 20, wherein the request includes information indicating the initial ordered sequence of one or more audio files.

23. The computer-accessible medium as recited in claim 20, wherein the request includes audio file class information for determining an initial ordered sequence of one or more audio files, and wherein, to determine an initial ordered sequence of one or more audio files, the playlist generation mechanism is further configured to select one or more audio files as the initial ordered sequence according to the class information in the request.

24. The computer-accessible medium as recited in claim 20, wherein, to generate a new playlist, the playlist generation mechanism is further configured to, if there is insufficient coverage in the N-gram statistical model for the initial ordered sequence of one or more audio files, apply the initial ordered sequence to a lower-level N-gram statistical model of the playlist database.

25. The computer-accessible medium as recited in claim 20,

wherein, to generate a new playlist, the playlist generation mechanism is further configured to, if there is insufficient coverage in the N-gram statistical model for the initial ordered sequence of one or more audio files, apply the initial ordered sequence to a class-based statistical model of the playlist database;

wherein the class-based statistical model comprises higher-level statistics for the audio files indicated by the playlists in the playlist database, wherein the higher-level statistics are derived from information about the audio files indicated by the playlists other than ordering sequences.