KR20030059503A - User made music service system and method in accordance with degree of preference of user's - Google Patents

Abstract

PURPOSE: A system and a method for a customized music service according to the user's preference are provided to suggest music preferred to the user and the music having high similarity on the web by constructing a database of the music files provided from a music service system through classification and storing the characteristic information of each genre. CONSTITUTION: A client terminal(100) accesses the system through the Internet(106), sends a music file play request to the system, and plays the music file provided by responding to the request. A music service server(108) stores diverse music files and the characteristic information value of each music in the database(110), searches the characteristic information value of the music preferred by each user and the music files having the high similarity from the database(110), and offers it as the preferable music file of the corresponding user. The music service server(108) includes an audio database module, an audio signal extraction module, and an audio quantization module.

Description

USER MADE MUSIC SERVICE SYSTEM AND METHOD IN ACCORDANCE WITH DEGREE OF PREFERENCE OF USER'S}

The present invention relates to an online music service system, and more particularly, to a user-customized music service system and method for enabling a customized music service according to a user's preference analysis.

In general, the Internet is a computer communication network that enables businesses, research institutes, libraries, schools, and individuals around the world to search or exchange information with each other using computers. With the increasing popularity of high-speed communication networks, the use trend is rapidly spreading, such that the Internet can be used on a regular basis in ordinary homes. In response to this, a variety of online commercialization and entertainment through the Internet, which is not restricted by the real-world space, is being activated, and the Internet is now becoming a culture of our society.

Therefore, a number of entertainment websites that provide services such as games, movies, and music online have appeared and are in operation. Among them, online music service web that provides playing services for music desired by users connected online, in particular, Sites are also actively operating through various value-added services, backed by explosive popularity from users.

However, a website that provides a conventional music merely performs a function of playing a song for a song selected by a user, and does not provide a more differentiated personalized service considering the characteristics of each user.

Therefore, if the user's preference characteristics for music are analyzed and more differentiated supplementary services are implemented for each user, the online music service system is expected to be activated more.

On the other hand, such a user-specific preference song characteristics analysis system as described in the US Patent No. 6,317,710 "Multimedia Search Apparatus and method for searching multimedia content using speaker detection by audio data" and published on May 4, 2001 Audio analysis techniques such as "Audio content analysis for online audio visual data segmentation and classification" disclosed in Transaction on Speech and Audio processing, Vol. 9, No. 4, pages 441 to 457, are disclosed, but the "Multimedia Search Apparatus and method for searching multimedia content using speaker detection by audio data "mentions only suggestions for speaker search by audio search in multimedia, and" Audio content analysis for online audio visual data segmentation and classification " And only discloses algorithms for classification and use Personalized service by the user by measuring the specific preferred audio track has got a partial interruption of the establishment of music theory as difficulties in processing the audio signal there is still the difficulty of implementation.

Accordingly, an object of the present invention is to provide a user-customized music service system and method for enabling a customized music service according to a user's preference analysis.

The present invention for achieving the above object, in the customized music service system according to the user preferences, can be connected to the system via the Internet, authorizing the music file playback request from the user to the system, A client terminal device for reproducing the corresponding music file provided in response and various music files provided for the online music service to the client terminal device, and database the feature information value for each song, Implement a user-customized music service system including a music service server for retrieving music files highly similar to feature information values from the database and providing the music files as the user's preferred music files.

(a) classifying and storing various music files provided in the server according to instruments and genres, (b) classifying the music files into a plurality of segments, and (c) music sections of the music files among the segments. Registering the feature information value for the corresponding music file from the segment classified as (d) and quantizing the extracted feature information value and storing the feature information of the corresponding music file as registration And (a ') collecting characteristic information values for the user's playback request music file for each user registered in the system, and (b') analyzing the collected characteristic information values for each user's preferred music. Generating a feature information value for (c ') and having a similarity with the preference feature information value of the user when there is a request for a preferred song service from the user; The method may include: searching for a music file having a feature information value; and (d ') providing a music file searched for having a high similarity, and providing list information of a song preferred by the user. Characterized in that.

1 is a network configuration of a user-customized music service system according to an embodiment of the present invention.

2 is a block diagram of a music service server according to an exemplary embodiment of the present invention.

3 is a flowchart illustrating a music file feature information value registration process according to an embodiment of the present invention;

4A and 4B are flowcharts illustrating a user-customized music service according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail the operation of the preferred embodiment according to the present invention.

1 illustrates a network configuration of a system for providing a customized music service for each user using the Internet according to an embodiment of the present invention. Referring to Figure 1 below,

The client terminal 100 is a user input / output interface device that can access the user-customized music service server 108 through the Internet 106, such as a personal computer (PC) 102, a personal digital assistant (PDA) 104, or the like. In other words, the client terminal 100 includes a protocol stack based on TCP / IP to operate in a client (Server) server environment, the processing function module for analysis and measurement of the user's favorite songs And an audio output module for allowing a client side to check the result processed by the user-customized music service server side of the system.

The processing function module may receive an audio signal and perform the processing on the client side or the server side as a function module for extracting signal characteristic information of the corresponding audio signal. In this case, the processing on the server side may be applied when there is a requirement such as real time processing of the performance processing on the input sound, and when processing on the client side, the feature information extracted from the input sound may be applied to the server side. After comparing with the existing audio DB (Data Base) feature value of the method, the result is checked by the output unit. At this time, the client side extracts the feature information value for the input file, compares the result with the feature information DB on the server side, and sets the method of transmitting the processing result to the client side as the basic operation mode, and after signal analysis and feature extraction on the client side The result format is XML based. The purpose of the XML-based input / output format support is to enable interworking support with other applications on the web.

The music service server 108 is a server implemented on the Internet for an online music playback service according to an embodiment of the present invention. The music service server 108 includes a variety of music contents and provides a service for playing music required for playback from a user. In addition, the music service server 108 classifies music contents such as songs, pops, jazz, classical music, etc. according to musical instruments or genres according to an exemplary embodiment of the present invention, and segments (eg, segmentation) for a predetermined length section representing the characteristics of each song. Segmentation) is performed, and feature values on a time axis and frequency values of a music section are extracted from the songs and stored in a DB having feature information of the corresponding song. Thereafter, the server 108 accesses the server 108 and analyzes the feature information of the songs selected by the users for each user who uses the music service, and stores the feature information of each user's favorite songs. Songs corresponding to the feature information of each star are provided as a preferred song service for each user.

2 shows a detailed block diagram of the user-customized music service server. Referring to FIG. 2, the server controller 200 of the customized music service server 108 controls the overall control and communication of the customized music service server 108 according to communication with a plurality of client terminals on the Internet 106. After controlling the overall operation, and in particular, after analyzing the characteristic information of each user's favorite songs to access the system using the music service according to an embodiment of the present invention, among the various music content provided in the server 108 It provides a user-customized service that searches for songs that match user-specific song feature information and provides them as user-specific song information.

ROM (Read Only Memory) 210 stores various operation programs necessary for the operation of the server controller 200 and stores various basic data necessary for driving the operating program. A random access memory (RAM) 212 temporarily stores a program operated by the control of the server controller 200 or temporarily stores data generated when the operating program is executed. The network interface unit 202 interfaces data transmitted / received through the Internet 106 between a plurality of client terminal devices 100 accessing the online system.

The database storage unit 110 is referred to by the user-customized music service server 108 and features personal information of users registered as members in the system and features of users' favorite songs according to an embodiment of the present invention. Member information DB (214) for storing information, music content DB (216) for distinguishing and storing songs of various genres according to the genre, and feature information for each feature stored in the music content DB (216) Various DBs necessary for operating a user-customized music service such as the value DB 218 are provided.

The audio DB classification module 204 performs a function of classifying the existing audio by genre. The audio DB classification module 204 may first classify the types of music by instruments and automatically classify various audio for signal analysis. Or semi-automatically. At this time, the complexity and accuracy of the system is determined according to the audio DB classification information. The audio DB classification module 204 collects music content DBs 216 classified by genres, such as classical, pop, jazz, and song, and segments each of the songs only for a predetermined length of intervals representing the characteristics of the songs. . The segmentation method includes an automatic method and a semi-automatic method, and there is a method of using the mixture appropriately. In the case of automatically performing segmentation, first, time axis information is extracted from music notes. In order to extract time axis information, a music section, a voice section, a silence section, It is classified into segments such as a voice section and a noise section with a background sound.

In this case, two types of classification algorithms may be applied to the music section. The first method is to classify music into classical, pop, song or jazz styles, and by using basic frequency information for each instrument such as piano, violin, and the like. The classification method can be applied, and when classifying smaller unit features for a specific instrument part, pitch information with pitch information or melody information with note characteristics can be used. Therefore, the method of classifying into the smallest processing unit, which is a frame unit, may be applied.

The audio signal extraction module 206 performs a function of obtaining a signal processing characteristic of an audio signal to obtain a feature on a time axis and a feature on a frequency on the audio signal, and a concept of recognizing the human brain in addition to physical signal feature extraction. The feature is also extracted for conventional negative feature elements. The audio component quantization module 208 performs a function of quantizing the extracted audio feature information values without loss of information. In this case, the result value obtained by the quantization module is an operation of allocating bits by extracting only main feature information from feature values obtained by signal extraction and providing minimum information.

3 is a flowchart illustrating a process of extracting and registering feature information on music content in a user-customized music service server according to an exemplary embodiment of the present invention. Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS. 1, 2, and 3.

First, when new music content such as a song, pop, jazz or classic is inputted to the system, the server control unit 200 of the music service server 108 proceeds to step S302 in response to step S302 in step S302. The DB classification module 204 is controlled to classify the input music contents by musical instruments and genres, and then store them in the music contents DB 216. Subsequently, the server controller 200 proceeds to step S304 to perform segmentation on a section of a predetermined length representing the characteristics of the music file. In this case, when the segmentation is automatically performed as described above, first, a music section, a voice section, a silent section, a voice section with a background sound, and a noise section are performed using a signal processing technique such as ZCR on the sound of the music file. After extracting time-base information classified into segments, etc., the music files are classified by genres such as classical, pop, song or jazz for the music section among the above, or instruments such as piano, violin, and the like using basic frequency information. Will be categorized by star. The specific instrument part is classified into frame units, which are the smallest processing units, by using pitch information having sound height information or melody information capable of knowing sound characteristics.

Thereafter, the server controller 200 proceeds to step S306 to control the audio signal extraction module to extract feature information on a time axis and a frequency axis from an audio signal belonging to a feature section of the segmented music file. The feature information of the music file is extracted. Subsequently, the server control unit 200 controls the audio component quantization module in step S308 to quantize the extracted audio feature information values without loss of information, and then proceeds to step S310 to the feature value DB 218. Save registration.

4A and 4B illustrate a control flow of a user-customized music service operation according to user preference song information in a music service server according to an exemplary embodiment of the present invention. Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS. 1, 2, 4A, and 4B.

First, a user of a client terminal device, which is a member of the system, accesses the system by using an internet-enabled terminal device such as a personal computer (PC). Then, the server control unit 200 of the music service server 108 of the system recognizes the user's system connection in step S400 of FIG. 4A and requests input of a member ID and password for system connection.

Accordingly, the member user inputs his or her member ID and password to access the system. Then, the server controller 200 receives an ID and password from the user in step S402. The user information DB 214 in the database storage unit 110 refers to whether the user who requested the access is a registered member subscriber. At this time, if the user ID and password entered by the user is determined to be the registered member ID and password of the server control unit 200 proceeds from step (S404) to (S406) step of the user-customized music service server 108 The initial menu web page screen is displayed on the display unit of the client terminal device 100 of the member subscriber. Accordingly, the user who has been authenticated can request to play various music files provided by the music service server and listen to the desired music online.

That is, when a user connected to the system selects a specific music to request playback, the server controller 200 recognizes the music file requested to be played in step S408, and the music content DB 216 in step S410. Search for the music file requested for reproduction and transmit audio data for reproduction of the music file to the client terminal device 100 so that the music file can be reproduced in the client terminal device 100 of the user.

In operation S412, the server controller 200 extracts and collects characteristic information values of music files selected by the user from the characteristic value DB 218, and collects the characteristic information values in step S414. These data are analyzed to generate feature information values of the music preferred by the user and add them to the corresponding user information table in the member information DB 214. In this case, the generated feature information value for the user-specific preference song is used as an important element when performing a user-specific music service as in the processing flow of FIG. 4B to be described later.

That is, when a music preference feature information value for each user is generated through the processing flow of FIG. 4A, when the user preference song service menu is selected from the initial menu webpage screen of step S406, the server controller 200 is selected. In response to this in step S416, the process proceeds to step S418 to check the preference feature information value of the corresponding user from the member information DB 214. Subsequently, the server controller 200 proceeds to step S420 to search for music files having a high similarity to the user's preference feature information value among the music files stored in the feature value DB 218, and retrieves the searched files in step S422. The music files are displayed as a list of the user's favorite songs. Accordingly, the user can easily select a song of a desired style by receiving list information on a favorite song without searching the entire music file in the server.

Meanwhile, in the above description of the present invention, specific embodiments have been described, but various modifications may be made without departing from the scope of the present invention. Therefore, the scope of the invention should be determined by the claims rather than by the described embodiments.

As described above, in the present invention, a music service system implemented on the Internet classifies the music files provided by the system by genre to construct a database, and then extracts the feature elements of the music signal for each genre. Users can easily search for music of their desired style by performing processing analysis and storing the feature information in a database to enable a customized music service that presents target music with high similarity to the user's favorite music on the web. There is an advantage to this.

Claims (17)

In the customized music service system by analyzing the audio signal, A client terminal device capable of accessing the system via the Internet, authorizing a music file play request from a user to the system, and playing the music file provided in response to the request; Various music files provided for the online music service to the client terminal device, characterized in that the database for the characteristic information value of each song, and the user of the music files having high similarity to the characteristic information value of each song from the database And a music service server that searches for and provides a user's favorite music file. The method of claim 1, The client terminal device supports an XML-based input and output format, the user-customized music service system, characterized in that it comprises a TCP / IP-based protocol stack. The method of claim 1, The music service server may include a member information DB for storing personal information about users registered in the system. A music content DB for storing the various music files by instruments and genres And a feature value DB for storing feature information values for each music file. The method of claim 1, The music service server comprises: an audio DB module for classifying the music files by instruments and genres, and performing segmentation on each music file to extract feature information values of each music file; An audio signal extraction module for extracting feature information values for each music file; And an audio quantization module for performing quantization on the extracted feature information value. The method of claim 4, wherein The audio DB module may segment the music files into a plurality of segments by performing segmentation on each of the music files. The method of claim 4, wherein The audio DB module is a user-customized music service system, characterized in that to classify the music file into segments, such as music section, voice section, silent section, noise section according to the audio signal characteristics. The method of claim 6, The audio DB module is a user-customized music service system, characterized in that to perform segment classification for the music file using a ZCR signal processing technique. The method of claim 4, wherein And the audio signal extraction module extracts feature information on a time axis and a frequency axis of an audio signal existing in segments classified into the music section as feature information values of a corresponding music file. The method of claim 1, The music service server is a user-customized music service system, characterized in that for generating a feature information value of the music preferred by each user through the analysis of the feature information values of the music files selected for each user. The method of claim 9, And the music service server stores the generated music feature information of each user as user song preference information in each user information table in the member information DB. In the customized music service method according to user preferences in a customized music service system by analyzing an audio signal including a plurality of client terminal devices and a music service server, (a) classifying and storing various music files provided in the server according to musical instruments and genres; (b) classifying the music file into a plurality of segments; (c) extracting feature information values of the corresponding music files from the segments classified into the music sections of the music files among the segments; (d) quantizing the extracted feature information value and storing the extracted feature information as feature information of the corresponding music file. The method of claim 11, The step (b) may include (b1) performing segmentation on each of the music files. and (b2) classifying the segmented music file into a plurality of segments according to audio signal characteristics. The method of claim 11, And the music file is classified into segments such as a music section, a voice section, a silent section, and a noise section according to the audio signal characteristics. The method of claim 11, And the feature information value includes feature information on a time axis and a frequency axis of an audio signal existing in the music section. In the customized music service method according to user preferences in a customized music service system by analyzing an audio signal including a plurality of client terminal devices and a music service server, (a ') collecting feature information on the user's playback request music file for each user registered in the system; (b ') generating characteristic information values for music of each user by analyzing the collected characteristic information values; (c ') searching for a music file having a feature information value having a high similarity with the user's preference feature information value when there is a request for a preferred song service from the user; and (d ') providing the music file searched for by the similarity to provide the list information of the songs preferred by the user. The method of claim 15, The music file may be classified into segments such as a music section, a voice section, a silent section, and a noise section according to the audio signal characteristic. The method of claim 15, The feature information value may include feature information on a time axis and a frequency axis of an audio signal existing in the music section.