Method and Device for Detecting Music Segment, and Method and Device for Recording Data

• US 20090088878A1
• Filed: 12/25/2006
• Published: 04/02/2009
• Est. Priority Date: 12/27/2005
• Status: Active Grant

• Current Assignee: Mitsubishi Electric Corporation
• Original Assignee: Mitsubishi Electric Corporation
• Inventors: Suginohara, Hidetsugu, Otsuka, Isao
• Granted Patent: US 8,855,796 B2
• Time in Patent Office: Days
• US Class Current: 700/94
• CPC Class Codes: G10H 1/00 Details of electrophonic musical instruments
• CPC Class Codes: G10H 2210/031 Musical analysis, i.e. isolation, extraction or identification of musical elements or musical parameters from a raw acoustic signal or from an encoded audio signal
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: G11B 2020/10546 specifically adapted for audio data
• CPC Class Codes: H04N 5/781 on disks or drums
• CPC Class Codes: H04N 5/85 on discs or drums
• CPC Class Codes: H04N 9/8042 involving data reduction
• CPC Class Codes: H04N 9/8063 using time division multiplex of the PCM audio and PCM video signals
• CPC Class Codes: H04N 9/8205 involving the multiplexing of an additional signal and the colour video signal

Abstract

The present invention is to provide a method of efficiently detecting music scenes from television broadcasts.

A music-segment detection method according to the present invention includes a data input step of inputting data including an audio and video signals; a plurality of candidate-music-segment detection steps of detecting candidate music-segments from the audio signal based on criteria different from each other; and a music-segment detection step of detecting a music segment based on the plurality of detected candidate music-segments. Thereby, high detection accuracy can be ensured for a wide variety of music with various tones, such as rock, popular ballad, and classic.

Claims

1-14. -14. (canceled)

15. A music-segment detection method comprising:

a data input step of inputting audio/video data including an audio signal and a video signal;

a plurality of candidate-music-segment detection steps of detecting, from the audio signal, respective candidate music-segments of the audio/video data, based on a plurality of criteria different from each other; and

a music-segment detection step of detecting a music segment based on the plurality of candidate music-segments each detected by the plurality of candidate-music-segment detection steps,

the method of detecting a music segment being characterized in that

a first criterion included in the plurality of criteria is set so as to detect boundaries between the candidate music-segments and segments other than the candidate music-segments, and

a second criterion included in the plurality of criteria and different from the first criterion is set so as to detect intermediary portions in the candidate music-segments.

16. The music-segment detection method of claim 15, wherein the first criterion is set based on the difference between the left and right channels of the audio signal.

17. The music-segment detection method of claim 16, wherein the second criteria is set based on whether or not the characteristic of the audio signal corresponds to a class model determined to be a music scene among a plurality of predetermined audio class models.

18. The music-segment detection method of claim 17, wherein the difference between the left and right channels of the audio signal or the characteristic of the audio signal is calculated based on orthogonal transform coefficients obtained by orthogonally transforming the audio signal.

19. The music-segment detection method of claim 15, wherein

the plurality of candidate-music-segment detection steps includes a plurality of respective correction steps of correcting the detected candidate music-segments;

and

the music-segment detection step detects a music segment based on candidate music-segments, instead of the plurality of candidate music-segments, each corrected by the plurality of correction steps.

20. The music-segment detection method of claim 15, further comprising a commercial detection step of detecting a commercial segment based on the audio and video signals, wherein in the music-segment detection step the commercial segment is further eliminated from the music segment detected based on the plurality of candidate music-segments.

21. A data recording method comprising:

a data recording step of recording data including that of a music segment detected by the music-segment detection method of any one of claim 15 through claim 20; and

a time-information recording step of recording time information of the music segment.

22. A music-segment detection device comprising:

a plurality of candidate-music-segment detection units for detecting, from an audio signal in audio/video data including the audio signal and a video signal, respective candidate music-segments of the audio/video data, based on a plurality of criteria different from each other; and

a music-segment detection unit for detecting a music segment based on the plurality of candidate music-segments each detected by the plurality of candidate-music-segment detection units,

the music-segment detection device being characterized in that

a first criterion included in the plurality of criteria is set so as to detect boundaries between the candidate music-segments and segments other than the candidate music-segments, and

a second criterion included in the plurality of criteria and different from the first criterion is set so as to detect intermediary portions in the candidate music-segments.

23. The music-segment detection device of claim 22, wherein the first criterion is set based on the difference between the left and right channels of the audio signal.

24. The music-segment detection device of claim 23, wherein the second criteria is set based on whether or not the characteristic of the audio signal corresponds to a class model determined to be a music scene among a plurality of predetermined audio class models.

25. The music-segment detection device of claim 24, wherein the difference between the left and right channels of the audio signal or the characteristic of the audio signal is calculated based on orthogonal transform coefficients obtained by orthogonally transforming the audio signal.

26. The music-segment detection device of claim 22, wherein

the plurality of candidate-music-segment detection units includes a plurality of respective correction units for correcting the detected candidate music-segments;

and

the music-segment detection unit detects a music segment based on candidate music-segments, instead of the plurality of candidate music-segments, each corrected by the plurality of correction units.

27. The music-segment detection device of claim 22, further comprising a commercial detection unit for detecting a commercial segment based on the audio and video signals, wherein the music-segment detection unit further eliminates the commercial segment from the music segment detected based on the plurality of candidate music-segments.

28. A data recording device comprising a data recording unit for recording data including that of a music segment detected by the music-segment detection device of any one of claim 22 through claim 27, wherein the data recording unit records time information of the music segment.