JP2016110151A - Voice management server device, conversation voice provision method, and conversation voice provision system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a conversation voice provision method capable of providing voice like voice uttered by a human being, and giving a user a pleasure of a conversation.SOLUTION: A conversation voice provision method employs a voice management server device in which plural speech synthetic data items for use in synthesizing human beings' voices are recorded, and a user terminal capable of transmitting or receiving data to or from the server device. The conversation voice provision method includes a step of automatically transmitting user's talk voice from the user terminal to the server device, a step of recording in the server device the contents of the talk voice received from the user terminal, a step of producing a response message, which responds to the received contents of the talk, by means of artificial intelligence of the server device, a step of producing voice response data, which is used to output the response message with voice selected in advance by the user, using the speech synthetic data by means of the artificial intelligence of the server device, and a step of transmitting the voice response data to a user-side conversation terminal.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a voice management server device, a conversation voice providing method, and a conversation voice providing system, and an apparatus, method, and system having a function for conversation using artificial intelligence (AI / Artificial Intelligence). It is about.

  Although a simple conversation system using a computer has been provided, since the conventional conversation system has an artificial (mechanical) voice, it has hardly felt humanity. Further, since the response message to the voice uttered by the user is a fixed phrase determined in advance, there is a problem that it is easy to get bored in a very short period of time, and the conversation is not interesting.

  SUMMARY OF THE INVENTION In view of the above-described problems of the prior art, an object of the present invention is to provide a voice management server apparatus, a conversation voice providing method, a conversation that can provide human-like voice and can provide a user with enjoyment of conversation over a long period of time. It is to provide a voice providing system.

The above object is a plurality of speech synthesis data (speech segment data) created for each of a plurality of different animal (including human) speech, and a plurality of speech synthesis data for synthesizing the speech Is achieved by a voice management server device having voice information recording means (information recording medium) in which is recorded.
The plurality of speech synthesis data are composed of speech segment data derived from different human speech, and are used when synthesizing the human speech.

In the speech management server device, each of the plurality of speech synthesis data is composed of speech segment data derived from human speech, and is used when synthesizing the human speech.
The “speech segment data” is, for example, data of a short section constituting human speech (data of each pronunciation of sounds such as fifty sounds, muddy sounds, semi-muddy sounds). For example, speech (the voice of the person who emitted the speech segment) can be synthesized by connecting the pronunciations (speech segments) such as fifty sounds, muddy sounds, and semi-turbid sounds.

The voice management server device
Receiving means for accepting a voice use request for designating any voice selected by the user among the voices of the plurality of animals;
Information processing means for generating (synthesizing) voice response data for outputting a response message to the user by the voice selected by the user using the voice synthesis data (voice segment data);
Further

In addition, the voice management server device
The voice response data further includes a transmission means for transmitting the voice response data to an arbitrary terminal selected by the user through the Internet.

In addition, the voice management server device
Usage information recording means for recording usage data indicating the usage frequency of the speech synthesis data (speech segment data), wherein the information processing means responds to the voice usage request from the user; Update the usage data of the recording means.

  The voices of the plurality of animals include present human voices, past human voices, artificially created person voices, current real animal voices, past real animal voices, and artificially produced voices. One or more of the voices of the animals.

  The voices of the animals include character voices, idol voices, talent voices, voice actor voices, actress voices, actor voices, announcer voices, singer voices, celebrity and celebrity voices, Includes one or more voices of family and relatives.

The above-described object is a plurality of speech synthesis data (speech segment data) created for each of a plurality of animal (including human) speech, and a plurality of speech synthesizers for synthesizing the speech. A conversational voice providing method using a voice management server device in which data is recorded and a user terminal capable of transmitting and receiving between the server device,
(a) automatically transmitting the user's talk voice from the user terminal to the server device;
(b) recording the content of the talk voice received from the user terminal in the server device;
(c) generating a response message in response to the received talk content by the server device's artificial intelligence;
(d) generating voice response data for outputting the response message in a voice selected in advance by a user using artificial intelligence of the server device using the voice synthesis data (voice segment data);
(e) transmitting the voice response data to a conversation terminal on the user side;
Is achieved by a method for providing conversational speech including:

  In the step (c) of the conversation voice providing method, a response message that responds to the received talk content is generated so that the response content is based on the user's talk content recorded in the server device.

In addition, the conversation voice providing method is
The method further includes a step of receiving from the user terminal a voice use request designating any voice selected by the user among the voices of the plurality of animals.

  Further, in the step (b) of the conversational voice providing method, the content of the user's talk voice received from the user terminal is recorded in the server device in a state associated with the date / time data related to the reception date / time.

In addition, the conversation voice providing method is
Further comprising: analyzing the user's emotion from the user's talk voice received from the user terminal; and recording the analyzed emotion data in a state associated with the talk content;
In step (c), a response message is generated in response to the received talk content so that the response content is based on the recorded user talk voice content and the corresponding user emotion.

In the above conversation voice providing method,
Usage frequency data indicating the usage frequency of the voice within a certain period is generated and recorded in a state associated with the identification information of the voice.

In the above conversation voice providing method,
The user terminal is installed in a space where the user lives or stays.

In the above conversation voice providing method,
The user terminal is attached to any one of household electric machine appliances, furniture, and toys.

In addition, the above-mentioned purpose is
A voice management server device that is a plurality of voice synthesis data created for each of voices of a plurality of animals (including humans) in which a plurality of voice synthesis data for synthesizing the voices is recorded; and a user A talk voice providing system configured to include a user terminal that transmits a talk voice from the user terminal to the server apparatus and receives a response voice to the talk voice from the server apparatus,
The voice management server device
Means for recording in the server device the content of the user's talk voice received from the user terminal;
Means for generating a response message in response to the received talk content by the server device's artificial intelligence;
Means for generating voice response data for outputting the response message in a voice preselected by the user;
Means for transmitting the voice response data to a conversation terminal on the user side,
The user terminal is
Means for transmitting the user's talk voice from the user terminal to the server device;
And means for receiving voice response data in response to the talk voice from a server device.

  According to the present invention, real and fictitious human voice can be used as a commercial product. In addition, it becomes possible to manage sound as a product with a computer.

  In addition, by managing the sound as a commercial product with a computer in a batch, the profit (sound usage) for the voice owner (in the case of real existence) and the rights holder (in the case of synthesis) according to the amount of voice used and the frequency of use Amount).

  In addition, it is possible to prevent unauthorized use of sound by collectively managing sound as a product with a computer.

  In addition, according to the present invention, the user can listen to all voices with voices that he / she likes (voices such as idols and characters).

  In addition, according to the present invention, it is possible to freely talk with a favorite voice. As a result, for example, it becomes possible to give the user a new feeling (enjoyment) that an idol or character that should be far away is always close to him / her. In addition, it is possible to give the user a sense (safety) that an idol, a character, and the like are always watching themselves nearby.

  Further, since the server device records the user's behavior pattern (behavioral pattern accompanied by pronunciation) at any time, it becomes possible to return the response content (response message) according to the user's behavior pattern. As a result, the user can feel a pseudo conversation partner (actually synthesized speech transmitted from the voice management server) as a closer conversation partner.

  Further, in the present invention, since a response message based on the contents of the user's past talk voice and the like is generated, the response that is not interesting like the fixed phrase is much less than before, and as a result, Can provide conversation enjoyment for a long time.

  In the present invention, since a response is returned to the user's talk based on the contents of the user's past talk, etc., a new conversation partner (pseudo conversation partner or supporter) who has originality is nurtured. Can be provided to the user.

  In addition, by using the present invention, loneliness and loneliness of a person who lives away from his / her family, a person who lives alone without being close, or an elderly person who lives alone can be eliminated and a sense of security can be given. Moreover, it becomes possible to improve the motivation in daily life by giving many such conversation opportunities to those living alone.

  In addition, it becomes possible to provide a place for active conversation for those who are unfamiliar with communication or who are not good at communication.

  In addition, since the present invention monitors at least sound, it can also be used for crime prevention when the user is absent.

It is a figure which shows schematic structure of the conversation audio | voice provision system of this invention. It is a figure which shows schematic structure of the audio | voice management server apparatus of this invention. It is a figure which shows an example of a data structure of the "voice information" recorded on a database. It is a figure which shows an example of a data structure of the "user information" recorded on a database. It is a figure which shows an example of a data structure of the "voice utilization information" recorded on a database. It is a figure which shows an example of a data structure of "talk information" recorded on a database. It is a figure which shows an example of the "registration process of the voice which a user uses" performed with a voice management server apparatus. It is a figure which shows an example of the "voice transmission process in the user terminal for conversation" performed with the user terminal for conversation. It is a figure which shows an example of the "recording process of the talk content in a voice management server apparatus" performed with a voice management server apparatus. It is a figure which shows an example of the "response process in a voice management server apparatus" performed with a voice management server apparatus. It is a figure which shows an example of the "voice output process in the user terminal for conversation" performed with the user terminal for conversation. It is a figure which shows the outline | summary of the method of collecting the data for speech synthesis (speech segment data) used by this invention. It is a flowchart which shows the outline | summary of the method of collecting the data for speech synthesis (speech segment data) used by this invention.

(Voice management server device)
First, an example of an embodiment of a voice management server device according to the present invention will be described.

  The voice management server device is a server device used in a conversation voice providing system as shown in FIG.

As shown in FIG. 2, the voice management server device mainly has
A program storage unit in which various programs such as artificial intelligence are installed;
A database for recording various information to be described later,
A CPU (Central Processing Unit) used in response message generation and voice synthesis;
Memory used for various information processing, etc.
A communication device for transmitting and receiving information to and from the user terminal via the Internet;
It has.

  Various programs such as artificial intelligence (A.I./Artificial Intelligence) are installed in the program storage unit. The artificial intelligence program is a program for establishing a human conversation between a user described later and a pseudo conversation partner (actually, a voice management server device). By using this artificial intelligence program, the conversation between the user and the server device can be handled according to the user's talk contents (including past talk contents) without having to exchange ordinary conversations or boilerplate phrases. A response full of humanity can be returned from the server device.

  The database includes a voice information recording unit, a user information recording unit, a voice usage information recording unit, a talk information recording unit, and a dictionary information recording unit.

  Audio information having a data structure as shown in FIG. 3 is recorded in the audio information recording unit of the database. As shown in FIG. 3, a plurality of speech synthesis data created for each of a plurality of animal (including human) speech, and a plurality of speech synthesis data for synthesizing the speech are recorded. . In the present embodiment, the voice synthesis data is created based on the voice of an actual speaker.

  Note that the data for voice synthesis recorded in the voice information recording unit is not necessarily limited to the current human voice, and may be a human voice that has existed in the past, or an artificially created person's voice. Voice (fictional person's voice) may be used. Moreover, the sound of the animal which exists now, the sound of the animal which existed in the past, and the sound of the animal produced artificially (the sound of the fictitious animal) may be used.

  As specific examples of speech synthesis data related to human speech, for example, as shown in FIG. 3, character speech, idol speech, talent speech, voice actor speech, actress speech, actor speech, announcer speech Data for synthesizing human voices such as voices, voices of singers, voices of celebrities and celebrities, voices of existing families and relatives, voices of deceased family members and relatives, and the like.

  Moreover, the data for speech synthesis related to human speech is not limited to the above-described data, and may be, for example, an easy-to-hear male or female voice, a scary male voice threatening others, a child voice, an elderly voice, and the like.

  The speech synthesis data described above is recorded in the database in a state (linked state) associated with speech identification information (speech ID) and information indicating the speech type, as shown in FIG.

  As shown in FIG. 4, the user information recording unit of the database includes user identification information (user ID), information indicating the type of voice selected by the user, and identification information (voice ID) of the selected voice. , Are recorded in an associated state (linked state).

  In the voice usage information recording unit of the database, voice usage data indicating the frequency of use of voice synthesis data is recorded as shown in FIG. Further, the voice use data is recorded in a state where the voice identification information (voice ID) and the information indicating the type of the voice are associated (linked).

In the talk information recording part of the database, as shown in FIG.
Information indicating the talk speech received from the terminal of the conversation user side,
Identification information (user ID) of the user of the terminal;
Information on the date and time when the talk was received,
Emotion information of the user obtained by analyzing from the talk voice,
Information on the importance of talk obtained by analyzing the content of the talk voice,
Is recorded in a linked state (in a linked state).

  The emotion information is information obtained by digitizing emotions such as “joy”, “relax”, “anger”, and “sadness” in five stages as shown in FIG. Such emotion information can be obtained by analyzing the talk voice received from the user terminal.

  As shown in FIG. 6, the talk importance level is obtained by quantifying the talk importance level in five stages. For example, a very daily talk such as “good morning” or “just now” is determined to have the importance level 1. Also, an unusual and deep talk such as “I will study soon because it is an exam soon” is judged to be of importance 5.

  In addition to the above-described database, the voice management server device has a CPU (Information Processing Unit / Central Processing Unit) used when generating a response message and synthesizing voice as shown in FIG.

  The CPU performs, for example, information processing such as generating a response message for the user and generating voice response data for outputting the response message using the voice selected by the user by using the voice synthesis data. Have a role to do. Also, for example, it plays a role of executing information processing such as updating the voice usage data of the usage information recording unit in response to a voice usage request from the user.

In addition, as shown in FIG.
A receiving means for accepting a voice usage request for designating an arbitrary voice selected by the user among a plurality of animal voices;
Transmitting means for transmitting the voice response data to an arbitrary terminal selected by the user through the Internet;
It is comprised including.

(Conversation voice providing system)
Next, an example of an embodiment of the conversation voice providing system of the present invention will be described.

As shown in FIG.
A “voice management server device”, which is a plurality of voice synthesis data created for each of a plurality of animal (including human) voices, and in which a plurality of voice synthesis data for synthesizing the voice is recorded; ,
A user's talk voice is transmitted from the user terminal to the server device, and a response voice to the talk voice is received from the server device;
It is comprised including.

  The configuration of the voice management server device is as described above.

  The conversation user terminal includes a microphone and a speaker as shown in FIG. This user terminal is installed in a space where the user resides or stays. As a specific example of the space in which the user lives, for example, the room of the user's residence can be cited. Moreover, as a specific example of the space where the user stays, for example, a passenger car interior or a room such as a hotel.

  Such a user terminal for conversation may be simply attached to a living space or a staying space, or may be attached in a state of being incorporated in any of home electric machine appliances, furniture, and toys.

  Hereinafter, an example of a conversation voice providing method using the conversation voice providing system (voice management server apparatus, setting user terminal, conversation user terminal) of the present invention will be described for each process.

(Voice registration process for users)
The description will be mainly based on FIG.

  First, the voice management server device stands by in a standby state. Subsequently, when a use voice registration request is received from the setting user terminal, selectable voice list data is transmitted to the setting user terminal via the Internet. The selectable audio list is, for example, a list of “audio types” as listed in FIG. The setting user terminal is, for example, a smartphone or a personal computer owned by the user.

  Subsequently, the setting user terminal displays the voice list received from the server device on the display, and the user selects a favorite voice from the display list according to his / her preference.

  For example, when it is desired to receive a conversation voice with a favorite idle voice, the user terminal for setting is operated to select the idle voice, and the selection result is transmitted to the voice management server device. Alternatively, when it is desired to receive a conversation voice with the voice of a favorite character, the idle voice is selected. By selecting the voice in this way, it is possible to hear the response (conversation voice) to the talk voice by the favorite voice selected by the user.

  Subsequently, the setting user terminal transmits the voice identification information (voice ID, see FIG. 3) selected by the user to the server device, and the server device receives the voice ID through the Internet.

  Next, the server device transmits selectable response frequency information to the setting user terminal. The user terminal that has received the selectable response frequency displays the response frequency on the screen in a selectable manner, and then the user selects an optimal response frequency. That is, the user can freely set the frequency of response from the server device to his / her talk voice.

  Subsequently, the setting user terminal transmits response frequency identification information (response frequency ID) selected by the user to the server device, and the server device receives the identification information through the Internet.

  Next, in the server device, the identification information (user ID) of the user terminal that performed communication this time, the identification information (voice ID) of the selected voice received in the previous step, and the identification relating to the reaction frequency received in the previous step Information (response frequency ID) is recorded in the database in a state of being associated with each other (in a state of being associated).

  This completes the registration processing of the voice used by the user, and the computer returns to the standby state again except when the program ends.

(Voice transmission processing at user terminal for conversation)
The description will be mainly based on FIG.

  A user terminal installed in the user's living space or stay space monitors the user's talk voice using a microphone, CPU, or the like. The talk voice includes all the sounds emitted from the mouth by the user. That is, the talk voice includes all voices such as monologue, sigh, cough, and sneeze in addition to the user's conversation voice.

  When the user terminal detects user voice, the voice is converted into voice data. This voice data conversion process is continued until silence is detected. Subsequently, the obtained audio data is transmitted to the server device via the Internet.

  Thus, the voice transmission process in the conversation user terminal is completed, and the apparatus returns to the standby state again except when the program ends.

(Recording of talk content in the voice management server device)
The description will be mainly based on FIG.

  First, the voice management server device stands by in a standby state. Subsequently, when talk voice data is received from the conversation user terminal, the contents of the user's talk voice are converted into text data for each phrase (or for each sentence) based on the received talk voice data. A speech recognition program is used to convert talk speech into text data.

  Subsequently, based on the received voice data, the user's emotion in the talk voice and the importance of the talk content are analyzed and converted into data. The user's emotion is, for example, joy, relaxation, anger, sadness, as shown in FIG. 6, and is quantified in five stages. The digitization of the user's emotion is determined based on the wavelength of the user's voice, for example. The importance level of the user's talk content is digitized in, for example, five levels, and the importance level is determined from the received talk content. For example, in the case of daily regular phrases such as “Good morning” and “I will come”, the importance is judged to be low, and in the case of unusual conversations such as “I am coming soon” Is determined to have high importance.

  Next, the voice management server device receives the talk voice reception date and time received in the previous step, the content (text data) of the talk voice identified in the previous step, the user's emotion and importance analyzed in the previous step, Record in the database in a state of being associated with each other (in a state of being linked). See FIG.

  Thus, the recording process of the talk content in the voice management server apparatus is completed, and the apparatus returns to the standby state again except when the program ends.

(Response processing in the voice management server device)
Description will be made mainly based on FIG.

  First, the voice management server device stands by in a standby state. Subsequently, when talk voice data is received from the conversation user terminal, the voice management server device determines whether or not to respond to the talk voice received this time. This response determination is performed based on, for example, a response frequency specified in advance (see FIG. 7), the importance level of the talk voice content (see FIG. 6), or the like.

  When the voice management server device determines that it responds to the current talk voice, it refers to the talk information recorded in the database (such as the user's past talk contents and the date and time of the talk). Then, a response message is generated based on the reception date / time of the referenced talk information, the talk content, the user's emotion, the importance level of the talk, and the like.

  The “response message based on the reception date / time of the talk information, the content of the talk, the emotion of the user, the importance level of the talk”, and the like are as follows, for example.

For example, in the case of the talk information illustrated in FIG. 6, for the talk content “Good night” on January 27, 2011, the talk content “18:10 on the same day” Refer to “No.”, and based on the content of the talk, generate a response message such as “Okay? Is it better to study because it is an exam soon?”
Also, for example, for the talk content of 6:10 on January 28, 2011, “Today we will go to the AKB concert”, refer to the talk content “Tomorrow is raining” at 23:50 the previous day. Based on the contents of the talk, a response message such as “Good morning. Do your best today. Do n’t forget your umbrella because it ’s raining today.” Is generated.

  Subsequently, the voice management server device synthesizes a response voice for outputting a response message with a voice selected in advance by the user, and transmits the synthesized response voice data to the user terminal through the Internet. “Response voice data” is data for outputting the response message generated in the previous step in a voice selected in advance by the user. By outputting the voice using the “response voice data”, it is possible to enjoy the feeling as if the voice person or animal selected in advance by the user is directly speaking to himself / herself.

  Thus, the response processing in the voice management server device is completed, and the standby state is restored again except when the program is terminated.

(Voice output processing in conversation user terminal)
The description will be mainly based on FIG.

  First, the conversation user terminal stands by in a standby state. Subsequently, when the conversation user terminal receives the response voice data from the voice management server device, a response based on a predetermined voice (voice preselected by the user) is output from the speaker based on the received data.

  The voice output at the user terminal is completed as described above, and the apparatus returns to the standby state again except when the program ends.

(First embodiment of voice collection method)
Next, a method for collecting speech synthesis data (speech segment data) used in the present invention, that is, an embodiment of a speech sampling method will be described.

  The “speech collection method” referred to in this application is speech unit data used when synthesizing human (human) speech, and speech unit data derived from the human speech is transmitted to the Internet and a server device. It is a method of collecting using

  The processing flow of this method is as shown in the schematic diagram of FIG. 12 and the flowchart of FIG.

That is, the outline of the first embodiment of the voice sampling method is as follows.
Access from the voice provider terminal (voice registration terminal) configured to be able to input the voice of the voice provider to the server device used for collecting voice segment data via the Internet,
Sampling the voice of the voice provider via voice input means connected to the voice provider terminal (voice registration terminal),
Audio sampling data related to the sampled audio is transmitted from the audio provider terminal to the server device,
In the server device, voice unit data is generated based on the voice sampling data,
In a state in which the provider identification information related to the voice provider and the voice segment data are associated with each other, the information is stored in the information recording unit of the server device.
That's it.

In this application, “voice provider terminal (voice registration terminal)” includes, for example, a computer that can access the Internet, a portable information terminal (for example, a smartphone or a mobile phone), as shown in FIG. It is done.
Examples of the “voice input means” include a microphone connected to a computer and a microphone built in the portable information terminal.
The “speech segment data” is, for example, data of a short section constituting human speech (data of each sound of pronunciation such as fifty sounds, muddy sounds, and semi-voiced sounds). For example, speech (the voice of the person who emitted the speech segment) can be synthesized by connecting the pronunciations (speech segments) such as fifty sounds, muddy sounds, and semi-turbid sounds.

(Second embodiment of voice collection method)
The outline of the second embodiment of the sound collection method is as follows.
Access from the voice provider terminal configured to be able to input the voice of the voice provider to the server device used for collecting voice segment data via the Internet,
Through the voice input means connected to the voice provider terminal, sample the voice of the voice provider;
In the voice provider terminal, based on the voice sampling data related to the sampled voice, voice unit data is generated,
The generated speech segment data is transmitted from the voice provider terminal to the server device,
In a state in which the provider identification information related to the voice provider and the voice segment data are associated with each other, the information is stored in the information recording unit of the server device.
That's it.

  The functions of the conversation user terminal used in the present invention can be incorporated into, for example, an Android, a robot, a television, a smartphone, a personal computer monitor, a stuffed animal, a doll, a toy, and the like. The conversation user terminal may function in synchronization with the 3D stereoscopic image. It is also possible to function by fusing with image mapping (two-dimensional image), an image on a wall or subject, and the like.

Claims (8)

  A plurality of voice synthesis data created for each of a plurality of different human voices, and having voice information recording means in which a plurality of voice synthesis data for synthesizing the plurality of human voices is recorded Management server device. The plurality of data for speech synthesis are respectively
Consists of speech segment data derived from human speech,
Used to synthesize human speech,
The voice management server apparatus according to claim 1. Receiving means for accepting a voice usage request for designating any voice selected by the user among the plurality of human voices;
Information processing means for generating voice response data for outputting a response message to the user by the voice selected by the user, using the voice synthesis data;
The voice management server apparatus according to claim 1, comprising:   4. The voice management server apparatus according to claim 3, further comprising a transmission means for transmitting the voice response data to an arbitrary terminal selected by a user via the Internet. A plurality of voice synthesis data created for each of a plurality of different human voices, wherein a plurality of voice synthesis data for synthesizing the plurality of human voices are recorded; A conversation voice providing method using a user terminal capable of transmitting and receiving with a server device,
(a) automatically transmitting the user's talk voice from the user terminal to the server device;
(b) recording the content of the talk voice received from the user terminal in the server device;
(c) generating a response message in response to the received talk content by the server device;
(d) generating voice response data for outputting the response message in a voice selected in advance by a user using the voice synthesis data by a server device;
(e) transmitting the voice response data from the server device to the user terminal;
Conversation voice providing method including In the step (c), a response message that responds to the received talk content is generated so that the response content is based on the current and / or past talk content of the user recorded in the server device.
The conversation voice providing method according to claim 5. Receiving a voice use request for designating any voice selected by the user among the plurality of human voices from the user terminal;
The conversation voice providing method according to claim 5. A plurality of voice synthesis data created for each of a plurality of different human voices, wherein the voice management server device stores a plurality of voice synthesis data for synthesizing the plurality of human voices; and a user A talk voice providing system configured to include a user terminal that transmits a talk voice from the user terminal to the server apparatus and receives a response voice to the talk voice from the server apparatus,
The voice management server device
Means for recording in the server device the content of the user's talk voice received from the user terminal;
Means for generating a response message in response to the received talk content;
Means for generating voice response data for outputting the response message in a voice preselected by the user;
Means for transmitting the voice response data to a user terminal,
The user terminal is
Means for transmitting the user's talk voice from the user terminal to the server device;
Means for receiving voice response data in response to the talk voice from a server device;
Conversation voice providing system characterized by including