JP2003015666A - Play list generating device, audio information providing device, audio information providing system and its method, program, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image of an object to be provided with audio information and audio information matching the preference of a user even when the user previously has knowledge regarding the audio information. SOLUTION: A play list generating device which generates a play list showing the order of one group of pieces of audio information to be reproduced by a reproducing device 70 is equipped with a database 40 which stores pieces of audio information, a condition input part 10 which inputs conditions regarding the object to be provided with audio information, a transition information acquisition part 20 which acquires transition information showing the transition of impressions on the object that the reproduction of audio information of a 2nd group in specified order under the condition makes when the reproduction is performed in the specified order, and a play list generation part 60 which selects audio information giving the transition of impressions close to the transition of impression that the transition information shows from the pieces of audio information stored in the database and generates a play list including selected pieces of audio information as audio information in a 1st group.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a playlist generation device, an audio information providing device, an audio information providing system and a method thereof for providing audio information suitable for an image to be provided with audio information and user's preference. A program and a recording medium.

[0002]

2. Description of the Related Art Conventionally, as a service for providing music,
There are playlist generation service and background music providing service. These services are based on individual requests of playlist users, and specialists select audio information suitable for that request from the music database based on musical knowledge and make a list in consideration of the playback order. Is provided to users.

A system for automatically providing continuous music is proposed in Japanese Patent Laid-Open No. 2000-98950. In this system, sound signal type information is assigned in advance for each food and drink menu product in a restaurant. BGM of the genre suitable for the ordered product from the type information
By automatically selecting and playing in the store, suitable background music is provided.

Also, from the viewpoint of continuously providing music, radio and cable broadcasting can be considered as a kind of playlist.

[0005]

However, in order to realize these services, expert knowledge about music is required. In order to meet various demands of users in a playlist generation service, it is necessary to have a broad knowledge of past audio information and an understanding of human sensitivity to music. Furthermore, considering that it also supports a huge amount of new songs that are created and sold daily, even if playlists related to the same request are requested at different times, specialists can recreate new ones or create new ones in the past. Needs to be fixed. Also, in the system, radio, and cable broadcasting proposed in the above-mentioned Japanese Patent Laid-Open No. 2000-98950, expert knowledge is required for categorizing audio information and selecting methods according to broadcasting time zones.

Further, the system proposed in the above-mentioned Japanese Patent Laid-Open No. 2000-98950 automatically switches the music genre on the system side, so that there is a high possibility that a music piece which is not desired by the drinker will be reproduced. For radio and cable broadcasting,
Since the information is unilaterally provided from the service provider side, it cannot be said that the user has obtained a sufficiently satisfactory playlist. Moreover, these services cannot reflect the user's preference. On the other hand, if you request the service company to create playlists individually, it is possible to incorporate user preferences in addition to the conditions under which the music is provided, so obtain playlists with a high degree of satisfaction. Is possible. However, in this case, the user must continue to use the generated playlist having the same content.

According to the present invention, a playlist generation device and an audio device for providing audio information suitable for an image to be provided with audio information and a user's preference without requiring the user to have prior knowledge about audio information. An object is to provide an information providing device, an audio information providing system and a method thereof, a program, and a recording medium.

[0008]

SUMMARY OF THE INVENTION A playlist generation device of the present invention is a playlist generation device for generating a playlist indicating the order of a first set of audio information to be reproduced by a reproduction device. A first database storing audio information of the second set, a condition input section for inputting a condition relating to an object to which the first set of audio information to be reproduced by the reproducing apparatus is input, and the second set of the second set based on the condition. A transition information acquisition unit that acquires transition information indicating a transition of an impression given to the object by reproduction of the second set of audio information when the audio information is reproduced in a predetermined order, and an impression transition indicated by the transition information. A plurality of audio information that gives a similar impression transition is selected from the plurality of audio information stored in the first database, and the selected plurality of audio information is selected. A playlist generating unit for generating a playlist that includes audio information as the first set of audio information, the object is achieved.

The condition input section includes an audio information number input section for inputting the number of audio information included in the playlist, and the playlist generation section inputs based on a plurality of impressions indicating at least a part of the transition of impressions. The number of audio information pieces may be selected to be equal to the number of generated audio information pieces.

The condition input unit includes a time input unit for inputting the play time of the play list, and the play list generation unit inputs the total time based on a plurality of impressions indicating at least a part of impression transitions. You may select several audio information so that it may become below reproduction time.

The condition input section includes an image input section for inputting an image of a playlist, and the transition information acquisition section stores a plurality of transition information. The transition information corresponding to the image is selected from the plurality of transition information. May be.

[0012] The first database stores a plurality of related information related to each of the plurality of audio information stored in the first database, and each of the plurality of related information defines an impression of the audio information. The transition information indicates a plurality of coordinate values in the predetermined coordinate system, and the playlist generation unit obtains the coordinate value indicating the position of the corresponding audio information in the predetermined coordinate system. From at least one audio information corresponding to at least one related information having a coordinate value within a predetermined range from the reference coordinate value, the distance between the coordinate value included in the at least one related information and the reference coordinate value. Based on this, audio information to be added to the playlist may be selected.

The condition input section includes a selection method input section for inputting a selection method of audio information, and the playlist generation section includes an evaluation expression generation section for generating an evaluation expression based on the selection method. Audio information to add to the playlist may be selected from the at least one audio information based further on the formula.

The condition input unit includes a satisfaction degree input unit for inputting satisfaction information indicating the satisfaction degree of the object regarding the audio information provided by the playlist generation apparatus to the object. The playlist generation apparatus stores the provided audio information. A second that stores time information indicating the time provided to the target and satisfaction information
The playlist generation unit may further include a database, and the playlist generation unit may select audio information to be added to the playlist from the at least one audio information, based on at least one of the time information and the satisfaction information.

The satisfaction information may further indicate the condition and time of the object when the satisfaction is input.

[0016] The image forming apparatus further comprises an impression determining section for inputting audio information from the outside and determining an impression of the input audio information based on a musical feature of the input audio information.
The first database may store the input audio information together with the reproduction time of the input audio information and the determined impression.

An audio information providing apparatus of the present invention comprises the above-mentioned play list generating apparatus and a reproducing apparatus for reproducing the first set of audio information.

The play list generation method of the present invention is a play list generation method for generating a play list indicating the order of the first set of audio information to be reproduced by the reproduction device. Inputting a condition regarding an object to which one set of audio information is provided, and reproducing the second set of audio information when the second set of audio information is reproduced in a predetermined order based on the condition. A step of acquiring transition information indicating a transition of an impression given to a target, and a plurality of audio information stored in the first database for a plurality of audio information giving a transition of an impression close to the transition of the impression indicated by the transition information. A step of selecting from the information and generating a playlist including the plurality of selected audio information as a first set of audio information. Encompasses the door, the objects can be achieved.

The step of inputting the condition includes the step of inputting the number of audio information included in the playlist, and the step of generating the playlist is input based on a plurality of impressions indicating at least a part of impression transitions. The method may also include the step of selecting a number of audio information equal to the number of audio information generated.

The step of inputting the condition includes the step of inputting the reproduction time of the playlist, and the step of generating the playlist inputs the total time based on a plurality of impressions indicating at least a part of the transition of impressions. The method may include the step of selecting a plurality of audio information such that the audio information is equal to or less than the reproduced time.

The step of inputting the condition may include the step of inputting the image of the playlist, and the step of acquiring the transition information may select the transition information corresponding to the image from the plurality of pieces of transition information.

The first database stores a plurality of related information related to each of the plurality of audio information stored in the first database, and each of the plurality of related information defines an impression of the audio information. The transition information has a coordinate value indicating the position of the corresponding audio information in a predetermined coordinate system, the transition information indicates a plurality of coordinate values in the predetermined coordinate system, and the step of generating the playlist is based on the plurality of coordinate values. Between at least one audio information corresponding to at least one related information having a coordinate value within a predetermined range from the obtained reference coordinate value, between the coordinate value included in the at least one related information and the reference coordinate value. Audio information to be added to the playlist may be selected based on the distance.

The step of inputting a condition includes the step of inputting a selection method of audio information, and the step of generating a play list includes the step of generating an evaluation expression based on the selection method. Based on further, audio information to be added to the playlist may be selected from the at least one audio information.

The step of inputting the condition includes the step of inputting satisfaction information indicating the satisfaction of the object regarding the audio information provided to the object, and the playlist generation method indicates the time when the audio information provided is provided to the object. The step of storing the time information and the satisfaction information indicating in the second database further comprises the step of generating a playlist, wherein the step of generating the playlist further includes at least one audio information based on at least one of the time information and the satisfaction information. Audio information to be added to the playlist may be selected from.

The satisfaction level information may further indicate the condition and time of the target when the satisfaction level is input.

[0026] A step of inputting audio information from the outside and determining an impression of the input audio information based on a musical feature of the input audio information, and a reproduction time of the input audio information and the input audio information together with the determined impression. May be stored in the first database.

The method may further include reproducing the first set of audio information.

The program of the present invention is a program for causing a computer to execute a process of generating a playlist showing the order of the first set of audio information to be reproduced by the reproducing device, and the generating process is the reproducing device. Inputting a condition regarding an object for which the first set of audio information to be played by the second set of audio information is played in a predetermined order based on the condition. A step of acquiring transition information indicating a transition of an impression given to the target by the reproduction of the audio information, and a plurality of audio information giving a transition of an impression similar to the transition of the impression indicated by the transition information to the target are stored in the database. A playlist including a plurality of audio information selected from the plurality of audio information and including the selected plurality of audio information as a first set of audio information. Includes the step of generating a preparative, the objects can be achieved.

The recording medium of the present invention is a computer-readable recording medium recording a program for causing a computer to execute a process of generating a play list indicating the order of the first set of audio information to be reproduced by the reproducing device. In the generation processing, the step of inputting a condition regarding an object to which the first set of audio information to be reproduced by the reproducing device is input, and the second set of audio information is arranged in a predetermined order based on the condition. Obtaining transition information indicating a transition of impressions given to the target by the reproduction of the second set of audio information when being played back with, and giving a transition of impressions similar to the transition of the impressions indicated by the transition information to the targets. Select multiple audio information from multiple audio information stored in the database, and select multiple audio information It includes the steps of: generating a playlist that includes a first set of audio information, the object is achieved.

The audio information providing system of the present invention is
What is claimed is: 1. An audio information providing system, comprising: a server that generates a playlist that indicates an order of a first set of audio information to be reproduced by a reproduction device; and a terminal that is connected to the server via a network. A condition input unit for inputting a condition regarding an object to which the first set of audio information to be reproduced by the reproducing device is provided,
The server stores an impression that the reproduction of the second set of audio information gives the target when the second set of audio information is reproduced in a predetermined order based on the database storing the plurality of audio information and the condition. A transition information acquisition unit that acquires transition information indicating transitions of the transition, and a plurality of audio information that gives an impression transition similar to the transition of the impression indicated by the transition information, selected from the plurality of audio information stored in the database. A play list generation unit that generates a play list including a plurality of selected pieces of audio information as a first set of audio information; and an output unit that outputs the generated play list, and the terminal plays the play received from the server. The storage unit for storing the list and the reproducing apparatus for reproducing the first set of audio information are provided, thereby achieving the above object. It is.

[0031]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1A shows an audio information providing apparatus 1000 according to the first embodiment of the present invention. The audio information providing apparatus 1000 includes a playlist generating apparatus 1001 and a reproducing apparatus 70. The playlist generation device 1001 generates a playlist indicating the order of a set of audio information to be reproduced by the reproduction device 70. The playback device 70 plays back a set of audio information in the order shown in the playlist. In the present invention, audio information means information including audio contents such as music, voice, and sound effects.

The play list generating apparatus 1001 has an audio information database 40 in which a lot of audio information is stored, a condition input section 10, a transition information acquisition section 20,
And a playlist generation unit 60 that generates a playlist.

FIG. 1B shows a playlist generation apparatus 1001.
3 is a flowchart showing the operation of FIG. The user inputs to the condition input unit 10 a condition relating to a target to which a set of audio information reproduced by the reproducing device 70 is provided (S11). Next, the transition information acquisition unit 20 indicates the transition of the impression given to the target by the reproduction of the predetermined set of audio information when the predetermined set of audio information is reproduced in the predetermined order based on the input condition. Acquire transition information (S
12). Next, the playlist generation unit 60 selects, from the plurality of audio information stored in the audio information database 40, a plurality of audio information that gives an impression transition that is similar to the impression transition indicated by the transition information. A playlist including the plurality of audio information thus generated as the above-mentioned set of audio information is generated (S1).
3).

The condition input unit 10 includes an ID input unit 11, an image input unit 12, a selection method input unit 13, an audio information number input unit 14, a time input unit 15, and a satisfaction degree input unit 16. .

A user inputs personal information such as a user name and a user number and a password as personal ID information to the ID input unit 11. By using this personal ID information, personal data can be used in the audio information database 40 and the history database 50.

FIG. 2A shows an example of the image input unit 12. In the image input unit 12, a plurality of options such as season, time, place to listen to audio information, purpose, etc. are set. The user inputs a desired playlist image by selecting one of the options.

The image of audio information is uniquely determined by the combination of these options, and the code assigned to each option is expressed by the combination. A combination of such codes is called an image code. The generated image code is the impression coordinate value list acquisition unit 21.
Entered in.

As an option item, "DJ name"
Etc. can also be set. By selecting the name of a famous DJ, it is possible to generate a playlist similar to the selection pattern of that DJ.

FIG. 2B shows another example of the image input section 12. In FIG. 2B, for a playlist desired by the user, adjectives representing the image of audio information (eg, “bright ← → dark”, “intense ← → calm”) and tempo of the audio information (eg, ,
Specify the transition of parameters such as "fast ← → slow").
For example, "beginning with a song that feels dark, gradually changing to a song that feels brighter (for example, code 00 in the light ← → dark choice)" or "beginning with a song that feels intense,
The transition of each parameter, such as a song with a quiet feeling on the way and a song with a strong feeling at the end (for example, chord 10 in the option of intense ← → calm) is selected by a button or the like. A code is assigned to each button,
An image code is generated by a combination of these codes.

Since it is troublesome for the user to set a large number of parameters, a representative parameter among the large number of parameters and a principal component parameter obtained by performing a principal component analysis (for example, the first parameter shown in FIG. 2B). And the second principal component, etc.) as an option to reduce the number of parameters,
The number of user operations can be reduced.

In the example shown in FIG. 2B, there are four variations of each parameter (that is, codes 00, 01, 10, 11).
However, it is possible to set more complicated fluctuations.

FIG. 3 shows an example of the selection method input section 13. A user inputs a selection method of audio information of a playlist desired by the user. The selection method is expressed by “important item”, “preference reflection degree”, and “request reflection degree”.

In the priority item, it is determined which of the user's desired image and taste is to be weighted for the search.
The important items are divided into four stages, and a code is assigned to each stage.

As the degree of reflection of preference, any one of "selection of favorite song", "selection of song not recently listened to" and "selection of song not heard" is selected. A code is assigned to each option.

In the degree of reflection of the request, it is decided whether the audio information that absolutely matches the user's designation should be selected, or whether the audio information should be selected by the device side to some extent. When the reflection degree of the request is set to be high, a playlist desired by the user is likely to be generated, but a playlist including the same audio information for the same image is likely to be generated each time. On the other hand, if the request reflection level is set low, even if playlists are generated with the same image, playlists that are basically in line with the user's wishes and that include different audio information each time are created. It is easy to be generated. The degree of reflection of the request is also divided into four stages, and a code is assigned to each stage.

The selection method is uniquely determined by the combination of the above three codes, and such a combination of codes is called a selection method code. The generated selection method code is input to the evaluation expression generation unit 62, and the evaluation expression is generated based on the selection method code.

Although each option is divided into four stages in the above description, it may be divided into any number of stages by changing the setting of the evaluation formula generating section 62. Further, even when these options are not input, the evaluation expression generation unit 62 can operate with default settings.

In the audio information number input unit 14, the number of audio information included in the playlist desired by the user is input. The number of audio information designated by the user is input to the selection unit 63. The selection unit 63 can generate a playlist that satisfies the specified number of audio information and does not break the transition of the image of the entire playlist indicated by the transition information.

In the time input section 15, the reproduction time of the playlist desired by the user is input. The time designated by the user is input to the selection unit 63. The selection unit 63 can generate a playlist that satisfies the designated reproduction time and does not break the transition of the image of the entire playlist indicated by the transition information.

In the present embodiment, usually, the number of audio information and the reproduction time are not input at the same time.
In addition, even when neither the number of audio information nor the reproduction time is input, the playlist generation device 1001 operates.

The satisfaction degree input section 16 has a skip button. The user clicks the skip button when the currently selected audio information is different from the image desired by the user. By clicking the skip button, the currently selected audio information, the currently designated image, and the current time can be registered in the history database 50. Such information is called satisfaction information. After that, when the selection unit 63 selects the audio information, if the satisfaction information about the audio information selected when the same image is designated in the past is registered, avoid reselection of the same audio information. Is possible.

Also, by clicking the skip button, it is possible to issue a reproduction stop instruction to the reproduction device 70 and reproduce the next audio information of the play list.

The transition information acquisition unit 20 includes an impression coordinate value list acquisition unit 21 and an image / impression coordinate value list conversion unit 22.
With.

Impression coordinate value list acquisition unit 2 in order to provide continuous audio information suitable for the user's image.
1 acquires from the image / impression coordinate value list conversion unit 22 an impression coordinate value list in which impression coordinate values suitable for the user's image are continuously arranged in time series.

Here, the impression coordinate value and the impression coordinate value list will be described. FIG. 4B shows a specific impression coordinate value and a list of impression coordinate values.

The impression coordinate value is a coordinate value in a coordinate system that represents a subjective impression of audio information such as "bright", "intense", "exhilarating", and "sharp". This coordinate system can define the impression of audio information. Words representing impressions such as "bright" and "dark" are applied to the coordinate axes of the coordinate system. FIG. 4A shows an example of a coordinate system representing an impression. Although the coordinate system shown in FIG. 4A is represented by two dimensions for the sake of simplicity of description, the coordinate system representing an impression may be represented by three or more dimensions. In the coordinate system shown in FIG. 4A, for example, the audio information assigned to the position of the impression coordinate value α indicates that a bright and intense impression is given to the object. In addition, each axis of the coordinate system may represent an impression of being another object such as “sharpness” and “softness” or a single impression of “exhilarity”. Since the impression coordinate value includes a plurality of items related to musical subjectivity, the impression indicated by the impression coordinate value can be expressed by a vector.

Further, FIG. 4B shows transition information showing the transition of impression given to the object by reproduction of one set of audio information.
The transition information is represented by an impression coordinate value list. The impression coordinate value list is a list in which a plurality of impression coordinate values are arranged in time series. 4A shows the transition of the impression indicated by the transition information and the impression coordinate value (indicated by a star) included in the impression coordinate value list in the coordinate system shown in FIG. 4A. As illustrated, the impression coordinate values included in the impression coordinate value list have values along the transition of the impression indicated by the transition information. By reproducing a predetermined set of audio information having each impression coordinate value included in the impression coordinate value list, the transition of the impression indicated by the transition information can be given to the target. The predetermined audio information set may be an ideal audio information set having an impression coordinate value that matches an impression coordinate value included in the impression coordinate value list. The predetermined audio information set may be a virtual audio information set. Even with a set of audio information corresponding to impression coordinate values approximate to the impression coordinate values in the impression coordinate value list, the same impression as that of the predetermined set of audio information can be given to the object. The playlist generation unit 60 selects from the audio information database 40 the audio information corresponding to the coordinate values approximate to the impression coordinate values in the impression coordinate value list,
Generate a playlist.

Conventionally, the audio information included in the playlist generated to suit the image desired by the user has been fixed to one impression. For example, when an image of “bright” is input, a playlist including only audio information that gives a bright impression is generated. However, in the embodiment of the present invention, by arranging impression coordinate values consecutively in a time series in the impression coordinate value list, the set of audio information reproduced in the order of the playlist becomes “gradually intense audio information. It is possible to give an impression such as “enhance the mood” or “make a change by inserting audio information of a different impression after continuing audio information of a similar impression”. The numbers corresponding to the impression coordinate values arranged in time series are called list numbers. The impression coordinate value list is the playlist generation unit 60.
Entered in. The playlist generation unit 60 generates a playlist by selecting, from the audio information database 40, audio information corresponding to coordinate values approximate to impression coordinate values of all list numbers or some list numbers.

The impression coordinate value list acquisition unit 21 displays the image code input from the image input unit 12 as an image.
By inputting into the impression coordinate value list conversion unit 22, it is possible to refer to the pre-registered impression coordinate value list. In the image / impression coordinate value list conversion unit 22, an impression coordinate value list suitable for the image of the image code is registered on a one-to-one basis for all patterns of the image code. These registered impression coordinate value lists can be generated by experts, famous DJs or CDs.
An automatic analysis of the selection of is conceivable.

The generated impression coordinate value list and image code are input to the database search unit 61. The database search unit 61 searches for a plurality of candidates from the audio information registered in the audio information database 40 using the impression coordinate value as a key for each list number.

The impression determining section 30 is used when new audio information is registered in the audio information database 40. First, the reproduction time is calculated from the input audio information. Next, the musical characteristics such as the tempo of the song, the chord progression, and the musical composition are automatically extracted. The impression coordinate value of the input audio information is calculated using these musical characteristics. The input audio information is input to the audio information database 40 together with the related information indicating the impression coordinate value and the reproduction time of the input audio information. The format of the audio information input to the impression determination unit 30 is C
Any format such as D source, WAV file, MP3 file, etc. is possible.

FIG. 5 shows the audio information database 40.
The related information related to the audio information and the audio information stored in is shown. In the audio information database 40, each audio information input from the impression determination unit 30 is added with an audio information number together with related information and stored. Further, by adding the content of the audio information content to the related information, it is possible to restrict the listening of some audio information for each user according to the ID input to the ID input unit 11.

FIG. 6 shows history information and satisfaction information stored in the history database 50. The history database 50 holds the audio information number of the audio information provided to the user in the past, the history information and the satisfaction level information of the user regarding the audio information. The history information and the satisfaction information are the I input to the ID input unit 11.
It is stored for each D. The history information indicates the time (that is, the date and the time) of listening to certain audio information within a predetermined period, and the number of times of listening within a predetermined period can be obtained based on this history information. Satisfaction information indicates an image code and time when certain audio information is selected for an image desired by a user within a predetermined period and the user is not satisfied with the result. The satisfaction information is input to the history database 50 by the user clicking the skip button of the satisfaction input unit 16. The history information and the satisfaction information represent the user's preference, and these pieces of information are reflected in the evaluation value serving as a scale when the audio information is selected by the selection unit 63.

Further, the history information and the satisfaction information in the history database 50 are set with a time limit for holding the information, and the information is automatically deleted when the time elapses.

The play list generating section 60 includes a database searching section 61, an evaluation formula generating section 62, and a selecting section 63.

The database search unit 61 receives the impression coordinate value list acquired by the impression coordinate value list acquisition unit 21. The database search unit 61 searches the audio information database 40 by determining the number of audio information searches to be performed for each impression coordinate value using the impression coordinate value for each list number in the input impression coordinate value list as a search key. . The database search unit 61 acquires, as a search result, the Euclidean distance between the searched audio information search audio information number and the impression coordinate value of the searched audio information and the impression coordinate value used as the search key.
Further, when the user has input the reproduction time with the time input unit 15, the reproduction time is also acquired. A candidate list is obtained by adding the audio information number and the Euclidean distance (further reproduction time) to each of the acquired audio information items in the impression coordinate value list. FIG. 7 shows an example of the candidate list. The candidate list is input to the selection unit 63, and the final result of the play list is generated from this candidate list. The image code is also input to the selection unit 63 via the transition information acquisition unit 20 and the database search unit 61.

In the present embodiment, the distance between impression coordinate values is defined by the Euclidean distance, but the distance between impression coordinate values may be defined by using another distance calculation formula.

The evaluation formula generating unit 62 includes a selection method input unit 13
When the selection method code described above is input from, the evaluation expression corresponding to the selection method code is generated. The evaluation formula includes an “evaluation function formula” and a “selection rule”. The evaluation function expression is an expression for calculating an evaluation value used when selecting each audio information. The evaluation value is an impression evaluation value obtained from the distance between the impression coordinate value of the audio information stored in the audio information database 40 and the impression coordinate value of the impression coordinate value list, the history information stored in the history database 50, and the satisfaction. It is calculated using the history evaluation value obtained from the degree information. The form and coefficient of the evaluation relational expression are determined according to the selection method code. Also,
The "selection rule" is a rule for determining audio information to be selected using an evaluation value that is a calculation result using an evaluation function formula, and is determined according to a selection method code.
A plurality of types of “evaluation function formula” and “selection rule” may be registered in the evaluation formula generation unit 62 in advance. in this case,
The evaluation formula generation unit 62 selects an appropriate “evaluation function formula” and “selection rule” from the “evaluation function formula” and “selection rule” registered in advance, based on the selection method code.

FIG. 8 shows the correspondence between the selection method code, the evaluation function expression and the selection rule.

The default setting for the evaluation expression generation when the selection method input unit 13 does not input the selection method from the user is, for example, that the emphasis item is image emphasis (code "1").
1 ”), the degree of reflection of preference is“ select mainly favorite music ”(code“ 10 ”), and the degree of reflection of request is low (code“ 0 ”).
0 "). With this setting, audio information that matches the user's image to a certain degree is selected without any other constraint conditions, so even if you repeatedly generate playlists with the same image, a list containing different audio information will be created each time. Can be generated.

An evaluation function formula for calculating the total evaluation value E (x) for a certain audio information x is shown in (Formula 1).

E (x) = wi * Ei (x) + wc * Ec (x) +0.01 (Equation 1) At this time, Ei (x) is an impression evaluation value and Ec (x) is a history evaluation value. Represents Also, wi and wc each represent a weight, and the value is determined by the selection method code. Specifically, wi is set large and wc small when it is close to "image emphasis", and conversely, wi is small and wc is set large when it is close to "preference importance". (Equation 1) “+0.01” on the right side
Indicates that a small amount is added to the evaluation value. This “+0.
01 "indicates that all candidate audio information is Ei (x) =
In order to avoid that the playlist generation unit 60 does not operate because all the evaluation values become 0.0 and there is no audio information to be selected when 0.0 and Ec (x) = 0.0. Is added to.

The impression evaluation value Ei (x) is the distance d (x) between the impression coordinate value that is the key for searching the impression coordinate value list and the impression coordinate value of the candidate audio information, and the maximum distance d allowed during the evaluation process. By using '(constant), Ei (x) = max (d'-d (x), 0) (Expression 2) is expressed. (Equation 2) is an evaluation function expression that increases the evaluation value of audio information close to the search target.

Next, the history evaluation value Ec (x) is defined.
For the history information and the satisfaction information of the history database 50, the last history time is pt (x) and the history count is pc.
(X), the image code i suitable for the user's current wish
Satisfaction information regarding c is ps (x, ic). The satisfaction information ps (x, ic) includes the image code i in the past.
The time when the selection of the audio information x is not satisfied for c is shown. Further, here, all the time information including the final history time pt (x) and the history count pc (x) indicate absolute time. At this time, the history evaluation value E
c (x) is expressed as Ec (x) = h (ps (x, ic)) * (wct * f (pt (x)) + wcc * g (pc (x))) (Equation 3). To be done. wct and wcc represent weights, respectively, and are registered in the evaluation formula generation unit 60 together with the functions f (), g (), and h ().

Here, in the history database 50, the time when the history information is deleted is pt (x) ',
By expressing f (pt (x)) as f (pt (x)) = pt (x) -pt (x) '... (Equation 4), the evaluation value of recently listened audio information becomes large. It is possible to generate various evaluation expressions. Further, by setting f (pt (x)) = t-pt (x) (Equation 5), where t is the current time, an evaluation such that the evaluation value of audio information not recently listened to becomes large. It becomes possible to generate an expression. Furthermore, f (pt (x)) is expressed as if (pt (x) == NULL) then f (pt (x)) = 1 else f (pt (x)) = 0 (Equation 6) Thus, it is possible to generate an evaluation formula that increases the evaluation value of audio information that has not been heard.

Also, let pc 'be the maximum number of times the history generation count of the playlist generation apparatus 1001 can be set, and g (pc
By expressing (x) as g (pc (x)) = min (pc (x), pc ') (Equation 7), an evaluation expression that often increases the evaluation value of audio information. Can be generated. Further, g (pc (x)) is represented by if (pt (x) == NULL) then g (pc (x)) = 0 else g (pc (x)) = max (pc'-pc (x), By expressing 0) ... (Equation 8), it is possible to generate an evaluation equation that increases the evaluation value of audio information that is not listened to so much. Furthermore, g (pc (x)) is expressed as if (pc (x) == 0) then g (pc (x)) = 1 else g (pc (x)) = 0 (Equation 9) As a result, it is possible to generate an evaluation formula that increases the evaluation value of audio information that has never been heard.

Next, the satisfaction information ps (x, ic) is converted into if (ps (x, ic) == NULL) then h (ps (x, ic)) = 1 else h (ps (x, ic) ) = 0 (Equation 10) so that the user has previously received the audio information x
However, if there is a rejected history when searching with the same image, it is possible to avoid the same selection. An evaluation value is calculated according to the evaluation function formula generated as described above.

Next, the selection rule will be described. The selection rules include, for example, “select audio information having the largest evaluation value”, “select audio information using a probability proportional to the evaluation value”, and “randomly select audio information”.

By selecting “audio information having the largest evaluation value” among the candidate audio information, it becomes possible to reflect the user's request to the maximum. In addition, by selecting "audio information using a probability proportional to the evaluation value" in the candidate audio information, it becomes possible to reflect the user's request to some extent, and in the candidate audio information, "randomly select audio information". By doing so, even when a playlist is generated for the same image, different audio information is likely to be included in the playlist each time.

Each item in FIG. 3 will be specifically described below. The "important item" is a coefficient of the "impression evaluation value" and the "history evaluation value" of the evaluation function formula, that is, wi of (Expression 1).
And wc are reflected in the size relationship. Wi with image emphasis
Becomes larger, and wc becomes larger when the preference is emphasized. The “reflection degree of preference” is reflected in the determination of the formula of “history evaluation value”. In other words, regarding the function f () of (Equation 3), if "Selection with favorite song as the center (chord 10)" is used, (Equation 4) is used and "Select with songs not recently listened to (chord 0)
In the case of "1)", (Equation 5) is used, and in the case of "Selecting songs that have not been heard before (code 00)", (Equation 6) is used. Also, the "degree of request reflection" is reflected in the determination of the "selection rule". In the case of "high (code 10)", the selection rule "select audio information with the highest evaluation value" is used, and in the case of "normal (code 01)", the selection rule "probability proportional to the evaluation value" is used. Select audio information using "is used, and in the case of" low (code 00) ", the selection rule" randomly select audio information "is used.

The evaluation function formula and the selection rule generated as described above are input to the selection unit 63 as an evaluation formula, and are used for calculation of an evaluation value and selection of audio information.

The selecting section 63 selects the audio information to be actually reproduced from the candidate list generated by the database searching section 61. And audio information database 40
Load that audio information from and generate a playlist.

Which list number should the selection unit 63 select from the entire candidate list (list number selection processing)?
And which audio information is selected from among the candidate audio information of the list number (audio information selection processing).

Regarding the list number selection processing, in the condition input section 10, (1) "audio information number input section 14
And "there is no user input to the time input unit 15", (2) "when there is a user input to the audio information number input unit 14", and (3) "use to the time input unit 15". Processing procedure differs depending on the input by the person ”.

Regarding the audio information selection processing,
Basically, according to the evaluation formula generated by the evaluation formula generation unit 61, the above (1) to (3) are all selected according to the same criteria. However, (3) "the user of the time input unit 15 is selected". When there is an input ”, the selection range may be limited due to the relationship between the play time of the playlist and the play time of the audio information.

First, the audio information selection process will be described. The selection unit 63 calculates an evaluation value according to the evaluation function formula input from the evaluation formula generation unit 62. Then, using the evaluation value as a selection criterion, one audio information item is selected from the candidate audio information items in accordance with the selection rule input from the evaluation formula generation unit 62.

The evaluation value is calculated by substituting the distance d (x) between the impression coordinate value, which is a search key in the candidate list, and the impression coordinate value of each audio information into the evaluation function formula. At this time, if the history information regarding the audio information whose evaluation value is to be calculated exists in the history database 50, the final history time pt obtained from the history database 50.
(X), history count pc (x), satisfaction information ps (x, i
c) is also substituted into the evaluation function formula to calculate the evaluation value. The evaluation value is calculated for all candidate audio information belonging to the list number selected in the list number selection processing. The selection unit 63 recognizes the evaluation values of all the candidate audio information belonging to the same list number, and selects 1 from the candidate audio information belonging to the same list number according to the selection rule.
Select one audio information to add to the playlist.

In the audio information selection process,
It is also possible to make a selection such as "the audio information already selected in the playlist this time is not selected again".

Next, a play list generation process in which the list number selection process and the audio information selection process are combined will be described.

(1) When the user has not entered the audio information number input unit 14 and the time input unit 15, audio information selection processing is performed for all list numbers included in the candidate list generated by the database search unit 61. Then, one of the candidate audio information is selected to generate a playlist.

(2) When the user inputs to the audio information number input unit 14, a plurality of list numbers are selected from all the list numbers included in the candidate list generated by the database search unit 61. Further, audio information selection processing is performed to select one from the candidate audio information belonging to the selected list number. FIG. 9 is a flowchart showing a playlist generation process when a user inputs to the audio information number input unit 14.
The flow of the playlist generation processing will be described with reference to FIG.

Let I be the number of remaining reproduced audio information items (the initial value is the number of desired audio information items inputted by the user), and L be the number of candidate list items. First, it is determined whether the number of remaining reproduction audio information is larger than the number of candidate lists (S102). When it is determined in S102 that the number is small, the list numbers in the list number L are selected at equal intervals by the number I of the remaining reproduced audio information. Audio information selection processing is performed, and one is selected from the candidate audio information belonging to the selected list number and added to the playlist (S106). After the processing of S106, the generation of the playlist is ended (S107).

When it is determined in S102 that the number is large, audio information selection processing is performed for all the list numbers of the candidate list, and one is selected from the candidate audio information belonging to all the list numbers and added to the playlist. (S103). The selected audio information number (that is, the list number L) is subtracted from the remaining reproduced audio information number (S104). It is determined whether the number of reproduced audio information has reached the desired number of audio information (S105).
If it has reached in S105, the generation of the playlist is terminated (S107). If it has not reached in S105, the process returns to S102. By generating a playlist in such steps, it is possible to generate a playlist including the same number of audio information items as the desired audio information number, without breaking the transition of the impression of the impression coordinate value list which is the transition information. Is possible.

(3) The play list generation process in the case where the time input unit 15 receives an input from the user will be described.
The input desired reproduction time is Td, and the average reproduction time of one piece of audio information (not the actual average time but an approximate value) is Ta. At this time, the expected reproduction time To of the play list is calculated as in (Equation 11) from the current list number N of the candidate list and the average reproduction time Ta.

To = N * Ta (Equation 11) In the playlist generation process when the user inputs the time input unit 15, the remaining reproduction time to be provided and the remaining list in the candidate list. Is compared with the estimated remaining playback time obtained by multiplying Ta by Ta to change the list number selection method. FIG. 10 is a flowchart showing a playlist generation process when a user inputs to the time input unit 15. The flow of the playlist generation processing will be described with reference to FIG.

At the start of the play list generation process, the current list number N = 0 and the current cumulative reproduction time T = 0 are set (S
201). Next, if the remaining playback time to be provided is larger than the remaining expected playback time in the list, the next list number is selected. If it is smaller, the remaining number of lists is divided into equal intervals so that the expected remaining playback time of the list falls within the remaining playback time to be provided. The smallest list number of the list numbers at the divided positions is selected from the candidate list (S202). Next, it is determined whether or not the list number selected in the process of S202 is larger than the list number L (S20).
3). If it is larger, the process proceeds to S208. If it is smaller or equal, one piece of audio information having a reproduction time Ti shorter than the remaining reproduction time to be provided is selected from the candidate audio information belonging to the selected list number and added to the playlist (S204 to S205). In S204, when the reproduction time Tj of each candidate audio information is longer than the remaining reproduction time to be provided, the generation of the playlist is ended (S210). Next, the reproduction time Ti of the selected audio information is added to the accumulated reproduction time T (S20).
6). Next, it is determined whether the current list number is equal to the list number L (S207). If they are not equal, the process returns to S202. If they are equal, it is determined whether the remaining reproduction time to be provided is shorter than the average reproduction time Ta (S2).
08). If it is long, selection is performed again from the head of the candidate list, so that each variable is initialized (S209). S208
If it is short in, the generation of the playlist is ended (S210). By generating a playlist in such steps, it is possible to generate a playlist in which the total time is equal to or shorter than the desired reproduction time Td without breaking the transition of the entire impression of the impression coordinate value list which is the transition information. Becomes

The audio information included in the play list generated as described above is loaded from the audio information database 40. FIG. 11 shows an example of the generated play list. The audio information included in the playlist is input to the playback device 70 according to the playback order. Also, the audio information number of the audio information included in the playlist and the current image code can be input to the history database 50, and the history information can be updated as the playback device 70 progresses. At this time, if there is an input from the user to the satisfaction degree input unit 16, the satisfaction degree information is also updated.

The reproduction device 70 reproduces the audio information input from the selection unit 63 in the reproduction order.

Not only when any input is made in the condition input section 10, but also when the audio information database 4 is input.
When the audio information stored in 0 is reproduced by the reproducing device 70, the history information in the history database 50 is always updated.

(Embodiment 2) FIG. 12 shows an audio information providing apparatus 2000 according to Embodiment 2 of the present invention. The audio information providing device 2000 is different from the audio information providing device 1000 shown in FIG. The other components are the same as those of the audio information providing apparatus 1000, and therefore their explanations are omitted.

The play list generating section 64 includes an evaluation formula generating section 62 and a selecting section 65. The impression coordinate value list output from the transition information acquisition unit is input to the selection unit 65.

The selection unit 65 reconstructs the impression coordinate value list from the impression coordinate value list generated by the transition information acquisition unit 20 according to the number of audio information or the reproduction time (impression coordinate value list reconstruction process). I do. In addition, the selection unit 65, based on the impression coordinate values registered in the reconstructed impression coordinate value list and the evaluation function formula generated by the evaluation formula generating unit 62, the audio information database 40.
A process of selecting one audio information item (audio information selection process) is performed.

First, the audio information selection process will be described. In the audio information selection processing according to the present embodiment, one piece of audio information regarding the impression coordinate value of a certain list number is selected from the audio information database 40. In the audio information selection processing, the audio information having the shortest Euclidean distance between impression coordinate values may be simply selected, but in the present embodiment as in the first embodiment, the evaluation formula generating unit 62 generates the audio information. Select audio information according to the evaluation function formula. By using the evaluation function formula,
It is possible to perform the audio information selection process in consideration of the past history of the user and the desired selection method. Selector 6
5 calculates the evaluation value according to the evaluation function formula, and selects the audio information according to the selection rule by using the evaluation value as a selection criterion.

The distance d (x) between the impression coordinate value which is the key of the search in the reconstructed impression coordinate value list and the impression coordinate value of each audio information is substituted into the evaluation function formula, and the evaluation value is obtained. calculate. At this time, when the history information about the audio information whose evaluation value is to be calculated exists in the history database 50, the final history time pt (x), the number of times of history pc (x), and the satisfaction information ps obtained from the history database 50. (X, ic) is also substituted into the evaluation function formula to calculate the evaluation value. The evaluation value is calculated for all candidate audio information belonging to the list number selected in the list number selection processing. The selection unit 63 recognizes the evaluation values of all the candidate audio information belonging to the same list number, selects one audio information from the candidate audio information belonging to the same list number according to the selection rule, and adds it to the playlist.

Next, the impression coordinate value list reconstruction processing will be described. Here, the impression coordinate value list output by the transition information acquisition unit 20 is called a basic impression coordinate value list, and the impression coordinate value list generated by the impression coordinate value list reconstruction processing is called a reproduced impression coordinate value list. The reproduction impression coordinate value becomes a reference coordinate value when the audio information selection processing is performed. Also, let the number of lists in the basic impression coordinate value list be L. Regarding the impression coordinate value list reconstruction, in the condition input unit 10, (1 ′) “when the user does not enter the audio information number input unit 14 and the time input unit 15”,
The processing procedure differs depending on (2 ′) “when a user inputs to the audio information number input unit 14” and (3 ′) “when a user inputs to the time input unit 15”.

In the following, in the cases of (1 ') to (3') described above, a play list generation process combining the impression coordinate value list reconstruction process and the audio information selection process will be described.

(1 ') When there is no user's input to the audio information number input unit 14 and the time input unit 15, the reproduction impression coordinate value list is the same as the basic impression coordinate value list. Audio information selection processing is performed according to the basic impression coordinate value list to generate a playlist. That is, a playlist including the same number of audio information items as the list number L of the basic impression coordinate value list is generated.

(2 ') A description will be given of the play list generation process in the case where the user inputs to the audio information number input unit 14. Let I be the desired audio information number designated by the audio information number input unit 14. Here, the basic list number of the basic impression coordinate value list is represented by X, and the impression coordinate value of the basic list number X is limp (X) (basic impression coordinate value).
Express. Based on limp (X), the impression coordinate value simp (Y) (playback impression coordinate value) of the play list number Y in the play impression coordinate value list is generated.

FIG. 13 is a flow chart showing a process of generating a play list when a user inputs to the audio information number input unit 14.

First, the play list number Y is initialized (S
152). Next, the basic list number X corresponding to the current playlist number Y is calculated (S153). The basic list number X is calculated from the formula shown in S153 of FIG. Next, it is determined whether the calculated basic list number X has reached the last list number in the basic impression coordinate value list (S154). If not yet reached, the reproduction impression coordinate value simp (Y) in the current reproduction list number Y is calculated from the Xth and X + 1th basic impression coordinate values in the basic impression coordinate value list (S155). The reproduction impression coordinate value simp (Y) may be an internal division point of the Xth and X + 1th basic impression coordinate values. If it is determined in S154 that the last list number in the basic impression coordinate value list has been reached, the calculated basic list number X
The basic impression coordinate value limp (X) of becomes the reproduction impression coordinate value simp (Y) as it is (S156).

Next, audio information selection processing is performed based on the reproduction impression coordinate value simp (Y) of the reproduction list number Y, and one audio information is selected from the audio information database 40 and added to the playlist (S157).
Next, by incrementing the value of the playlist number Y by one, the process for the next playlist number starts. All playlist numbers Y
If the process is not completed in step S153, the process returns to the process of S153, and if the process is completed, the playlist generation process ends (S158). By generating a playlist in such steps, it is possible to generate a playlist including the same number of audio information items as the desired audio information number, without breaking the transition of the impression of the impression coordinate value list which is the transition information. Is possible.

(3 ') The play list generation process when there is a user input to the time input unit 15 will be described. Let Td be the desired reproduction time designated by the time input unit 15. The basic processing is the same as the above (2 ′), except that the accumulated reproduction time T is used. FIG. 14 is a flowchart showing a playlist generation process in the case where the user has input the time input unit 15.

First, the play list number Y is initialized (S
252). Next, the basic list number X corresponding to the current cumulative reproduction time T is calculated (S253). The basic list number X is calculated from the formula shown in S253 of FIG. next,
It is determined whether the calculated basic list number X has reached the last list number in the basic impression coordinate value list (S2).
54). If not yet reached, from the Xth and X + 1th basic impression coordinate values of the basic impression coordinate value list,
Play impression coordinate value sim for the current play list number Y
p (Y) is calculated (S255). Reproduction impression coordinate value si
mp (Y) may be an internally dividing point of the Xth and X + 1th basic impression coordinate values. If it is determined in S254 that the last list number in the basic impression coordinate value list has been reached, the calculated basic impression coordinate value limp (X) of the basic list number X is the reproduction impression coordinate value sim as it is.
It becomes p (Y) (S256).

Next, among many pieces of audio information stored in the audio information database 40, audio information corresponding to impression coordinate values located within a distance within a threshold value Dt from the reproduction impression coordinate value simp (Y). Then, it is checked whether there is audio information whose reproduction time is shorter than the remaining reproduction time Td-T (S257). If it exists, audio information selection processing is performed based on the reproduction impression coordinate value simp (Y), and one audio information is selected from the audio information satisfying the conditions in S257 and added to the playlist (S258). Then, the reproduction time of the audio information selected in S258 is added to the cumulative reproduction time T. By incrementing the value of the play list number Y by 1, the process goes to the next play list number (S259). S
If there is no corresponding audio information in 257, the playlist generation processing is ended (S260). By generating a playlist in such steps, it is possible to generate a playlist in which the total time is less than or equal to the desired playback time without breaking the transition of the entire impression of the impression coordinate value list that is the transition information. Become.

The play list generation processing in the first and second embodiments of the present invention described above can be recorded in a recording medium in the form of a program. As the recording medium, a computer-readable recording medium of any type such as a flexible disk or an optical disk can be used. By installing the playlist generation processing program read from the recording medium into an arbitrary computer capable of inputting / outputting audio information, the computer can be made to function as an audio information providing device. The playlist generation processing may be executed by a playlist generation processing device built in or connected to the computer, or at least a part of the playlist generation processing may be executed by the computer on software.

FIG. 15 shows one embodiment of a computer that executes such a playlist generation process. The computer 80 includes a disk drive device 81 that reads the program from a recording medium 84 that stores a program for causing the computer 80 to execute the playlist generation process, and a calculation unit 82 that executes the playlist generation process.
A storage unit 83 that functions as the audio information database 40 and the history database 50, and the condition input unit 1
0, a reproducing device 70, and a bus 85. Arithmetic unit 8
Reference numeral 2 includes a CPU, a memory and the like (not shown), and the transition information acquisition unit 20 and the impression determination unit 3 shown in FIGS.
0, the same processing as that of the playlist generation units 60 and 64 is executed.

The program may be provided by a recording medium or may be provided by data distribution via the Internet or the like.

(Embodiment 3) FIG. 16 shows an audio information providing system 3000 according to Embodiment 3 of the present invention.
Indicates.

The audio information providing system 3000 is
The network 100, the server 110, and the terminal 120, such as a personal computer and a music server, are provided. In the audio information providing system 3000, when a condition for providing audio information to the terminal 120 is input, a playlist suitable for the input condition is automatically generated on the server 110 side and distributed to the terminal 120.

A server 110 for delivering audio information and a terminal 120 for receiving the delivered music are connected to the network 100. In addition, the network 10
0 indicates a plurality of terminals 1 having the same function as the terminal 120.
20-1 to 120-n (a natural number is connected to n.

The audio information providing system 3000 is
Audio information providing apparatus 100 according to the first and second embodiments
The components 0 and 2000 are divided and placed on the network. As a result, even if the user does not have the audio information database 40 at hand, it is possible to download the audio information suitable for the image desired by the user from the server 110 via the network 100.

The server 110 includes a transition information acquisition unit 20 and
Impression determination unit 30 and audio information database 40
And the playlist generation unit 60 (playlist generation unit 64
, And an output unit 115 that outputs the playlist to the terminal 120. The server 110 may include the history database 50 shown in FIG. However, in the audio information distribution service, it is usually unnecessary to redistribute audio information once distributed, so the history database 50 may be omitted from the audio information providing system 3000. The terminal 120 includes the condition input unit 10
A playback device 70, and a storage medium 121 that stores the playlist distributed from the server 110.

In the audio information providing system 3000, the history database 50 is omitted from the audio information providing apparatuses 1000 and 2000, and the output unit 115 and the storage medium 121 are added. The audio information providing system 300 focusing on the difference of such components
The operation of 0 will be described.

The personal ID information input to the condition input section 10 can be stored in the audio information database 40.
By including the personal ID information of the user in the related information related to the audio information included in the playlist distributed to a certain user, it is possible to avoid duplicate distribution of the same audio information in the future. .

When the selection method is input to the condition input section 10, since the history information of the user is not used, "important item"
Do not input the "reflection degree of preference", but only the "reflection degree of request". Since the history information is not used, the playlist generation unit 60 generates the evaluation formula without using the history information. In addition, the satisfaction information is not input to the condition input unit 10.

In the audio information providing system 3000, it is desirable that at least one of the number of audio information and the desired reproduction time is always input to the condition input section 10.

Since the history database 40 does not exist in the audio information providing system 3000, the evaluation function formula for calculating the comprehensive evaluation value E (x) for the audio information x is changed from the above (formula 1) to (formula 12). .

E (x) = Ei (x) +0.01 (Equation 12) The impression evaluation value Ei (x) is calculated by (Equation 2). In (Equation 12), the overall evaluation value E (x) is calculated using only the impression evaluation value Ei (x).

The playlist generation unit 60 substitutes the distance d (x) of each candidate audio information of the candidate list into the evaluation function formula to calculate the comprehensive evaluation value E (x), and the comprehensive evaluation value E
Audio information selection processing is performed based on (x). Audio information database 40 for selected audio information
To create a playlist. The generated playlist is input to the output unit 115.

The output unit 115 includes the playlist generation unit 60.
The playlist input from is distributed via the network 100 to the terminal 120 of the user corresponding to the personal ID information received by the server 110. The terminal 120 stores the received playlist in the storage medium 121.

The storage medium 121 is a hard disk of a personal computer or a music server or the like, and stores the playlist distributed from the server. The audio information included in the play list is input to the playback device 70 and is played back by the playback device 7.
0 reproduces the audio information and is provided to the target.

With the audio information providing system 3000 of the present embodiment, the user inputs the target condition to the terminal 120 connected to the network 100, and the server 110 generates a playlist suitable for the target condition. It is possible to realize a service to be delivered to the terminal 120 by using the above.

[0134]

According to the present invention, a playlist including a plurality of audio information that gives an impression transition similar to the impression transition indicated by the transition information based on the condition of the object to which the audio information is provided is generated. It Accordingly, it is possible to provide a playlist including a plurality of audio information that gives the target an impression transition suitable for the target.

According to the present invention, instead of individually selecting audio information suitable for an image desired by a user from a database, an impression coordinate value list suitable for images in which impression coordinate values are arranged in time series is acquired, Audio information having an impression coordinate value that is similar to the impression coordinate value registered in the impression coordinate value list is selected. As a result, it is possible to provide a playlist satisfying the target condition and having a variety of impressions through a plurality of audio information.

In the present invention, when new audio information is stored in the audio information database, the impression coordinate value of the new audio information is automatically calculated and then stored in the database. Therefore, the user can store the new audio information in the database without requiring any musical knowledge. By storing the new audio information in the database, a playlist including the newly stored audio information can be obtained.

Further, according to the present invention, when selecting one of the plurality of candidate audio information registered in the candidate list, the history information reflecting the user's preference is selected. Thus, it is possible to generate a playlist that matches the taste of the user. Further, by designating a selection rule, it is possible to generate a playlist including different audio information each time, even when a playlist is repeatedly generated with the same image. Also, by specifying not only the image but also the number of audio information to be played and the playback time, the user can obtain a variety of playlists, unlike when listening to a fixed playlist repeatedly. It is possible to obtain a high degree of satisfaction.

[Brief description of drawings]

FIG. 1A is a diagram showing an audio information providing device according to an embodiment of the present invention.

FIG. 1B is a flowchart showing an operation of the playlist generation device according to the embodiment of the present invention.

FIG. 2A is a diagram showing an image input unit according to the embodiment of the present invention.

FIG. 2B is a diagram showing an image input unit according to the embodiment of the present invention.

FIG. 3 is a diagram showing a selection method input unit according to the embodiment of the present invention.

FIG. 4A is a diagram showing a coordinate system representing an impression of the embodiment of the present invention.

FIG. 4B is a diagram showing a relationship between an image code and an impression coordinate value list according to the embodiment of the present invention.

FIG. 5 is a diagram showing an audio information database according to the embodiment of the present invention.

FIG. 6 is a diagram showing a history database according to the embodiment of this invention.

FIG. 7 is a diagram showing a candidate list according to the embodiment of this invention.

FIG. 8 is a diagram showing a relationship between a selection method code, an evaluation function expression, and a selection rule according to the embodiment of this invention.

FIG. 9 is a flowchart showing a playlist generation process in the case where a user inputs to the audio information number input unit according to the embodiment of the present invention.

FIG. 10 is a flowchart showing a playlist generation process in the case where a user inputs to the time input unit according to the embodiment of the present invention.

FIG. 11 is a diagram showing a playlist according to the embodiment of this invention.

FIG. 12 is a diagram showing an audio information providing device according to an embodiment of the present invention.

FIG. 13 is a flowchart showing a playlist generation process in the case where a user inputs to the audio information number input unit according to the embodiment of the present invention.

FIG. 14 is a flowchart showing a playlist generation process when there is a user's input to the time input unit according to the embodiment of the present invention.

FIG. 15 is a diagram showing a computer that executes a playlist generation process according to the embodiment of this invention.

FIG. 16 is a diagram showing an audio information providing system according to an embodiment of the present invention.

[Explanation of symbols]

10 Condition input section 11 ID input section 12 Image input section 13 Selection method input section 14 Audio information number input section 15 hours input section 16 Satisfaction input section 20 Transition information acquisition unit 21 Impression coordinate value list acquisition unit 22 Image / impression coordinate list conversion unit 30 Impression determination section 40 audio information database 50 history database 60 Playlist generator 61 Database Search Department 62 Evaluation Formula Generation Unit 63 Selector 70 Playback device

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor San ▲ Saki ▼ Masayuki             1006 Kadoma, Kadoma-shi, Osaka Matsushita Electric             Sangyo Co., Ltd.

Claims (23)

[Claims] 1. A playlist generation device for generating a playlist indicating the order of a first set of audio information to be reproduced by a reproduction device, the first database storing a plurality of audio information, A condition input unit for inputting a condition regarding an object to which the first set of audio information to be reproduced by the reproducing device is input, and a second set of audio information is reproduced in a predetermined order based on the condition. A transition information acquisition unit that acquires transition information indicating a transition of an impression given to the target by reproduction of the second set of audio information; and an impression transition that is similar to the transition of an impression indicated by the transition information. A plurality of audio information to be given to the target is selected from the plurality of audio information stored in the first database, and the selected plurality of audio information is selected. Playlist generator with a play list generation unit for generating a playlist that includes O information as the first set of audio information. 2. The condition input unit includes an audio information number input unit for inputting the number of audio information items included in the playlist, and the playlist generation unit includes a plurality of plurality of transition items of the impression. The playlist generation device according to claim 1, wherein a number of audio information items equal to the number of input audio information items are selected based on an impression. 3. The condition input section includes a time input section for inputting a reproduction time of the playlist, and the playlist generation section, based on a plurality of impressions indicating at least a part of the transition of the impressions, The playlist generation device according to claim 1, wherein a plurality of pieces of audio information are selected so that a total time is equal to or less than the input reproduction time. 4. The condition input section includes an image input section for inputting an image of the playlist, and the transition information acquisition section stores a plurality of transition information.
The playlist generation device according to claim 1, wherein transition information corresponding to the image is selected from the plurality of transition information. 5. The first database stores a plurality of related information related to each of the plurality of audio information stored in the first database, and each of the plurality of related information includes an audio. In a predetermined coordinate system that defines the impression of information, it has a coordinate value indicating the position of the corresponding audio information, the transition information indicates a plurality of coordinate values in the predetermined coordinate system, the playlist generation unit, A coordinate included in the at least one related information from at least one audio information corresponding to at least one related information having a coordinate value within a predetermined range from a reference coordinate value obtained based on the plurality of coordinate values. The play according to claim 1, wherein audio information to be added to the playlist is selected based on a distance between a value and the reference coordinate value. List generator. 6. The condition input section includes a selection method input section for inputting a selection method of the audio information, and the playlist generation section includes an evaluation expression generation section for generating an evaluation expression based on the selection method. The playlist generating device according to claim 5, further comprising: selecting audio information to be added to the playlist from the at least one audio information, further based on the generated evaluation formula. 7. The condition input unit includes a satisfaction degree input unit for inputting satisfaction degree information indicating satisfaction degree of the object regarding audio information provided by the playlist generation apparatus to the object, the playlist generation apparatus Further includes a second database that stores time information indicating a time at which the provided audio information is provided to the target and the satisfaction level information. The playlist generation unit stores the time information and the satisfaction level information. The playlist generation device according to claim 5, wherein audio information to be added to the playlist is selected from the at least one audio information further based on at least one of the. 8. The playlist generation device according to claim 7, wherein the satisfaction level information further indicates a condition and a time of the target when the satisfaction level is input. 9. The apparatus further comprises an impression determination unit that receives audio information from the outside and determines an impression of the input audio information based on a musical feature of the input audio information, wherein the first database includes the input. The playlist generation device according to claim 1, wherein the input audio information is stored together with the reproduction time of the audio information and the determined impression. 10. An audio information providing apparatus, comprising: the playlist generating apparatus according to claim 1; and the reproducing apparatus for reproducing the audio information of the first set. 11. A first device to be reproduced by a reproducing device.
A playlist generation method for generating a playlist showing the order of the audio information of the set, the step of inputting a condition regarding an object to which the first set of audio information reproduced by the reproduction device is provided, Acquiring transition information indicating a transition of an impression given to the object by reproduction of the second set of audio information when the second set of audio information is reproduced in a predetermined order based on the condition; , Selecting from the plurality of audio information stored in the first database a plurality of audio information that gives the target an impression transition similar to the impression transition indicated by the transition information, and selecting the selected plurality of audio information. Generating a playlist included as the first set of audio information. 12. The step of inputting the condition includes the step of inputting the number of audio information included in the playlist, and the step of generating the playlist includes a plurality of steps indicating at least a part of transition of the impression. The playlist generation method according to claim 11, further comprising: selecting a number of audio information pieces equal to the number of input audio information pieces based on an impression. 13. The step of inputting the condition includes the step of inputting a reproduction time of the playlist, and the step of generating the playlist is based on a plurality of impressions indicating at least a part of transition of the impressions. And selecting a plurality of audio information such that the total time is less than or equal to the input playback time.
The playlist generation method described in. 14. The step of inputting the condition includes the step of inputting an image of the playlist, and the step of acquiring the transition information selects transition information corresponding to the image from a plurality of transition information. The playlist generation method according to claim 11. 15. The first database is stored in the first database.
A plurality of related information related to each of the plurality of audio information stored in the database is stored, and each of the plurality of related information corresponds to a corresponding audio in a predetermined coordinate system that defines an impression of the audio information. Having a coordinate value indicating the position of information, the transition information indicates a plurality of coordinate values in the predetermined coordinate system, the step of generating the playlist, the reference coordinates obtained based on the plurality of coordinate values The at least one audio information corresponding to the at least one related information having a coordinate value within a predetermined range from the value;
The playlist generation method according to claim 11, wherein audio information to be added to the playlist is selected based on a distance between a coordinate value included in one piece of related information and the reference coordinate value. 16. The step of inputting the condition includes a step of inputting a selection method of the audio information, and the step of generating the playlist includes a step of generating an evaluation expression based on the selection method, The playlist generation method according to claim 15, wherein audio information to be added to the playlist is selected from the at least one audio information further based on the generated evaluation formula. 17. The step of inputting the condition includes the step of inputting satisfaction information indicating satisfaction of the object with respect to the audio information provided to the object, and the playlist generation method includes the provided audio information. Further including a step of storing time information indicating a time at which the target is provided to the target and the satisfaction degree information in a second database, wherein the step of generating the playlist includes the step of generating the playlist among the time information and the satisfaction degree information. The playlist generation method according to claim 15, further comprising: selecting audio information to be added to the playlist from the at least one audio information, further based on at least one. 18. The satisfaction information further indicates a condition and time of the target when the satisfaction is input,
The playlist generation method according to claim 17. 19. A step of inputting audio information from the outside, determining an impression of the input audio information based on musical characteristics of the input audio information, a reproduction time of the input audio information and the determination The input audio information together with the impression is added to the first
Further comprising the step of storing in a database,
The playlist generation method according to claim 11. 20. The playlist generation method according to claim 11, further comprising playing back the first set of audio information. 21. A first to be reproduced by a reproducing device
Is a program for causing a computer to execute a playlist generation process indicating the order of the audio information of the set, wherein the generation process is provided with the audio information of the first set reproduced by the reproduction device. Inputting a condition relating to the object, and based on the condition, the transition of the impression given to the object by the reproduction of the second set of audio information when the second set of audio information is reproduced in a predetermined order. Acquiring transition information indicating the transition information, and selecting, from the plurality of audio information stored in the database, a plurality of audio information that gives an impression transition similar to the transition of the impression indicated by the transition information to the target, and the selection is performed. Generating a playlist including a plurality of audio information that has been generated as the first set of audio information, A program that includes 22. A first to be reproduced by a reproducing device
Is a computer-readable recording medium having recorded thereon a program for causing a computer to execute a process of generating a play list indicating the order of the sets of audio information, wherein the generation process is performed by the first reproduction device. Inputting a condition regarding an object for which the second set of audio information is provided, and based on the condition, the second set of audio information of the second set of audio information is reproduced when the second set of audio information is reproduced in a predetermined order. A step of obtaining transition information indicating a transition of an impression that reproduction gives to the object; Audio information of the first set, and the plurality of selected audio information are selected from the first set of audio information. And a recording medium including a step of generating a playlist including. 23. A first to be reproduced by a reproducing device
An audio information providing system, comprising: a server that generates a playlist that indicates the order of audio information of a set; and a terminal that is connected to the server via a network, the terminal being reproduced by the reproducing device. The first
A condition input unit for inputting a condition regarding an object to which the set of audio information is provided, and the server stores a database storing a plurality of audio information, and a second set of audio information based on the condition. A transition information acquisition unit that acquires transition information indicating a transition of an impression that the reproduction of the second set of audio information gives to the target when reproduced in a predetermined order; and an impression transition indicated by the transition information. Selecting a plurality of audio information that gives an approximate impression transition to the object from the plurality of audio information stored in the database, and including the selected plurality of audio information as the first set of audio information. A playlist generation unit that generates a playlist; and an output unit that outputs the generated playlist, the terminal Wherein and a reproducing apparatus, audio information providing system for reproducing a storage unit for storing the playlist received from the server, the first set of audio information.