JP2009147775A - Program reproduction method, apparatus, program, and medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve a problem that when reproducing only a desired music piece among recorded musical programs, conventional methods require that management information containing the start time, the end time, and meta-information of each music piece is received in advance, but such detailed management information is to be created and opened to the public by program producers or general users and it is not realistic to create the detailed management information to all musical programs to be broadcasted. <P>SOLUTION: First, the start time and the end time of each music piece is detected from recorded program data using signal processing. Then, for every detected music piece, collating with a music fingerprint information database is carried out to recognize the meta-information. On this occasion, the performer/artist information of the relevant program is previously acquired from an EPG (Electronic Program Guide) and a recognition rate is raised by carrying out only collation to those artists. Last, the start time, the end time, and the recognized meta-information are associated and saved. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for efficiently viewing a recorded program, and in particular, improves the searchability of each song by automatically extracting songs from a music program and automatically assigning meta information (song title, artist name, etc.). The present invention relates to a method and an apparatus.

  With the widespread use of large-capacity HDD recorders and PCs with TV recording functions, TV viewing has become easier, so a viewing style is generally available in which a large amount of content is stored on the HDD and viewed together later. It is becoming. On the other hand, the viewing time of the user is limited with respect to the large number of recorded contents, but there is no time to view the recorded contents but the stored contents that are buried in the HDD without being seen once increase. ing. For this reason, there is an increasing demand for a technology for efficiently viewing a large amount of recorded content stored in the HDD within a limited time.

  In particular, in music programs, various methods have been proposed so far, since only a desired piece of music can be played to greatly improve the viewing time. In Patent Document 1, at the same time that program data is recorded, management information including a start time, an end time, a song title / artist name, etc. for each song in the program is stored, and each song and time in the program data is stored. It is realized by associating with information.

  In this method, in order to extract a song section from each song in recorded program data and to add meta information such as a song name / artist name corresponding to the song, the start time / end time and meta information of each song are determined. It is assumed that the included management information is received.

Such detailed management information needs to be created and released by the program producer or general user, but it is a great effort to create detailed management information for all the music programs to be broadcast. Is not realistic.
JP 2007-028226 A

  An object of the present invention is to provide a method for improving the searchability of each piece of music program without separately creating program management information.

  In the present invention, when a program genre determining step for determining a program genre of a recorded TV program moving image and a result determined in the program genre determining step is a music program, a music section is determined from the TV program moving image. A music segment extraction step to extract, a voice waveform extraction step to extract a continuous audio signal in the music segment extracted in the music segment extraction step, a voice waveform extracted from the voice waveform extraction step, In the music fingerprint information collation step for collating the information with the music fingerprint information database stored in association with each other and acquiring the music information corresponding to the music section, and in the music fingerprint information collation step Obtained in the collation range limiting step for limiting the music information to be collated to the music information including the performer name and the collation range limiting step. A song section information generating step for associating the song information with the song section extracted in the song section extracting step, a song section information storing step for storing the song section information, and the song information and the song section stored in the song section information storing step And a playback step for playing back the video / audio data at the recording position searched in the search step. .

  According to the present invention, it is possible to search for the beginning of each song in a music program and to search for meta information using a song name / artist name as a key without manually creating program management information.

(Embodiment)
FIG. 1 shows a system configuration diagram of a program reproducing apparatus according to Embodiment 1 of the present invention.

  The program playback apparatus shown in FIG. 1 is an EPG (Electronic Program Guide) that is superimposed on a TV broadcast wave or input via a network, with video / audio data of a TV program moving image input via a TV broadcast wave. , Electronic program guide) information and audio recording unit 101 to be stored in association with information, genre determining unit 102 for determining whether or not the program is a music program based on program genre information included in the EPG information, and the music program The music section extraction unit 103 that extracts the start time / end time of the music section from the data, the voice waveform extraction unit 104 that extracts the voice waveform data of the music section, and the music fingerprint information database 105 are referred to, and the music section corresponds to the music section. From the music fingerprint information matching unit 106 for recognizing music information (song name, artist name, etc.) and EPG information The collation range limiting unit 107 that acquires the performer name of the music program and collates the music information to be collated by the music fingerprint information collating unit 106 to the music information including the performer name, and the start time / end time of the music A music information storage unit 108 that stores the recognized music information in association with each other, a search unit 109 that searches for the recording position of the desired music using the music information and the music section as keys, and video / audio data at the recording position. It consists of a playback unit for playback.

  Hereinafter, the details of the program reproducing apparatus according to the first embodiment of the present invention will be described with reference to FIG. The video / audio recording unit 101 receives a television broadcast wave via the antenna 111, and superimposes the video / audio signal of an arbitrary television program on the television broadcast wave or inputs the program through another network. Save in association with EPG information. The EPG information includes program genre information, performer information, and the like. FIG. 2 shows an example of EPG information. Here, the program genre information represents information used to classify television programs based on the contents such as “music program” and “sports program”, for example, and is predetermined as shown in FIG. It can be identified by the assigned identifier. The performer information stores the names of people who appear in the program. For example, in the case where the program is a music program, the names of performing artists, the names of presenters, and the like are stored.

  The genre determination unit 102 determines whether or not the genre identifier of the television program recorded by the video / audio recording unit 101 corresponds to a music program. If it is determined to be a music program, the genre determination unit 102 is connected to the music segment extraction unit 103. If it is not a music program, the process is immediately terminated.

  The music segment extraction unit 103 first extracts a feature amount from a video signal or an audio signal of program data determined as a music program by the genre determination unit 102, and then the extracted feature amount continuously. A section that matches the characteristics of the music section is detected and automatically extracted as the start time and end time of the music section.

  The voice waveform extraction unit 104 extracts continuous voice waveform data included in each music segment extracted by the music segment extraction unit 103.

  The music fingerprint information matching unit 106 stores the audio waveform data extracted by the audio waveform extraction unit 104 in association with audio waveform fingerprint (music fingerprint) information and music information (music name, artist name, etc.). The music fingerprint information database 105 is collated, and music information corresponding to each music section is acquired.

  Here, the music fingerprint information registered in the music fingerprint information database 105 is usually created by obtaining a sound source from a music CD (Compact Disk) that is generally sold. Music CDs recorded in a stable environment such as studio recording occupy an overwhelming majority. On the other hand, music broadcasted on a music program on a television is often a song with live recording or a unique arrangement. In such a case, the voice waveform data extracted by the voice waveform extraction unit 104 and the music fingerprint information registered in the music fingerprint information database 105 are originally recognized as different music, although they are the same music. The recognition rate will be reduced.

  In the present invention, a collation range limiting unit 107 is provided to increase the recognition rate in such a case. The collation range limiting unit 107 extracts a performer name from the EPG information, and extracts only data whose artist name matches the performer name from the music fingerprint information database 105 as a collation target. In the music fingerprint information collating unit 106, collation narrowed down to only the data extracted by the collation range limiting unit 107 is performed, so that the possibility of erroneous recognition can be suppressed low.

  The music information storage unit 108 stores the music information obtained by the music fingerprint collation unit 106 and the start time / end time of the music section obtained by the music section extraction unit 103 in association with each other. FIG. 4 shows an example of information stored in the music information storage unit 108.

  When arbitrary music information is input as a search key, the search unit 109 acquires the recording position of television program moving image data associated with the music information. The reproduction unit 110 reproduces the moving image data at the recording position acquired by the search unit 109.

  Hereinafter, the flow of processing for reproducing moving image data using this system will be described with reference to the flowchart of FIG.

  First, program genre information is acquired from EPG information transmitted via a broadcast wave or a network. If it is determined to be a music program, the process is continued. If it is determined not to be a music program, the process is immediately terminated. (S501).

  Next, for the moving image data determined to be a music program in S501, feature amounts are extracted using a signal processing technique, and sections in which the feature amounts continuously match the features of the song sections are extracted as song sections. (S502).

  Also, the names of performers of the program are extracted from the EPG information associated with the moving image data determined as the music program in S501 (S505).

  After the music sections are extracted in S502, continuous speech waveform data included in each music section is extracted (S503), and the music fingerprint information database in which the music fingerprint information and the music information are stored in association with each other is extracted. By matching, music meta information corresponding to each music section is acquired (S504).

  When collating in S504, by limiting only data whose artist name matches the program performer name extracted in S505 from the music fingerprint information database in advance (S506), erroneous recognition occurs. Reduce the likelihood of doing so.

  When the music information corresponding to each music section is acquired, the start time / end time of the music section obtained in S502 and the music meta information are stored in association with each other (S507).

  When the user reproduces only desired music from the stored moving image data, a list of music meta information stored in the system is first presented. The user selects music information such as artist name and music name from the list as a search key and inputs it to the system (S508). In addition, as a method for the user to input a search key, not only a method for selecting from the presented items, but also a method for the user to freely input a keyword and perform a search may be used.

  When a search key for music information is input, the system acquires a recording position of moving image data associated with the corresponding music information (S509), and reproduces moving image data at the recording position (S510).

  As described above, according to the present invention, it is possible to search for each piece of music in a music program using music information without creating management information in advance.

1 is a system configuration diagram of a program playback device according to Embodiment 1 of the present invention. The figure which shows an example of the EPG information transmitted via a broadcast wave or a network The figure which shows an example of the program genre identification number in EPG The figure which shows the structure of the music information preserve | saved matched with an input moving image Flowchart showing the flow of processing for reproducing moving image data using this system

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Audio | voice recording part 102 Genre determination part 103 Music segment extraction part 104 Audio | voice waveform extraction part 105 Music fingerprint information database 106 Music fingerprint collation part 107 Collation range limitation part 108 Music information storage part 109 Search part 110 Playback part 111 Antenna

Claims (7)

A program genre determination step for determining a program genre of a recorded TV program image;
When the result determined in the program genre determination step is a music program, a music segment extraction step for extracting a music segment from a TV program image;
A voice waveform extraction step for extracting a continuous voice signal in the music section extracted in the music section extraction step;
The music fingerprint information collating step of collating the voice waveform extracted in the voice waveform extracting step with a music fingerprint information database storing the voice waveform and music information in association with each other, and acquiring music information corresponding to the music section When,
A collation range limiting step for inputting a performer name of the music program, and limiting music information to be collated in the music fingerprint information collation step to music information including the performer name;
A program information generation method comprising: a music section information generation step for associating music information obtained in the collation range limiting step with a music section extracted in the music section extraction step. The collation range limiting step extracts a performer name of the program from an electronic program guide sent via a broadcast wave or a network, and from the music information database, only the music information including the performer name is included. The program information generation method according to claim 1, wherein the program information is extracted. 2. The program information generation according to claim 1, wherein the music section extraction step extracts a section in which a video signal or an audio signal of a music program moving image continuously matches a feature of a music scene as a music section. Method. A music section information storage step for storing music section information generated by the program information generation method according to any one of claims 1 to 3,
A search step for searching for a recording position of a desired song based on the song information and the song interval saved in the song interval information saving step;
And a playback step of playing back the video / audio data at the recording position searched in the search step. A program comprising the method according to any one of claims 1 to 4. A medium on which the program according to claim 5 is recorded. A program genre determination unit for determining a program genre of a recorded TV program image;
A music segment extraction unit that extracts a music segment from a TV program image when the result determined by the program genre determination unit is a music program;
A voice waveform extraction unit for extracting a continuous voice signal in the music section extracted by the music section extraction unit;
A music fingerprint information collation unit that collates the voice waveform extracted by the voice waveform extraction unit with a music fingerprint information database stored in association with the voice waveform and music information, and obtains music information corresponding to the music section When,
A collation range limiting unit that inputs the performer name of the music program and limits the music information to be collated by the music fingerprint information collation unit to the music information including the performer name;
A music information storage unit that stores the music information obtained by the collation range limiting unit in association with the music segment extracted by the music segment extraction unit;
A search unit for searching for a recording position of a desired music, using the music information and music section stored in the music information storage unit as keys,
A program reproduction apparatus comprising: a reproduction unit that reproduces video / audio data at the recording position searched by the search unit.

Images