KR20040013071A - Voice mail service method for voice imitation of famous men in the entertainment business - Google Patents

Abstract

PURPOSE: A voice mail service method and system for mimicking the voices of stars is provided so that a receiver can receive a voice mail, imitated into the voice of a star, when a user sends a short message. CONSTITUTION: A voice mail service system for mimicking the voices of stars is comprised of a server(12), the first user terminal(14), and the second user terminal(16). The server(12) is connected with a communication network(10). The first user terminal(14) transmits a message to the server(12). The second user terminal(16) receives a message from the server(12). The server(12) comprises a formant extraction part(122), a database(124), and a voice synthesis part(126). The formant extraction part(122) analyzes the voice of a desired star and extracts a feature parameter including the fourth formant and the fifth formant at least. The database(124) allocates a serial number for each star's extracted feature parameter and stores them together. The first user terminal(14) comprises a transmission control part(142) and a message transmission part(144). If the message transmission part(144) sends desired star information with a short message, the formant extraction part(122) searches the database(124) and obtains a feature parameter for the star. The voice synthesis part(126) executes voice synthesis using the searched feature parameter and stores the voice-synthesized message in a voice mail box(128).

Description

Voice mail service method for voice imitation of famous men in the entertainment business}

The present invention relates to a voice mail service method and a system thereof, and more particularly, to a voice mail service method and system for simulating a celebrity voice.

The use of text messages on mobile terminals such as cell phones and personal digital assistants (PDAs) is commonplace. Young people who get sick of everyday messages are using emoticons to feed their emotions and humor, but they are also getting tired of their time.

On the other hand, attempts to transmit voice using text messages are not known. This can be expected due to the following reasons. That is, in the conventional text message service, since the amount of data to be transmitted is very small, it is judged that there is no improvement in transmitting various information in the form of text message in the text message of the mobile environment.

On the other hand, speech analysis is an essential and very important process for speech recognition or speech synthesis. Voice parameters are classified into a method of extracting in the time domain and a method of extracting in the frequency domain. In the time domain, energy, zero-crossing, pitch period, pitch frequency, linear prediction coefficients, LPC cepstrum coefficients, and mel frequency (MFCC) Cepstral Coefficient is used, and Spectrum Envelope and Formant are used in the frequency domain.

The technical problem to be achieved by the present invention is to send a text message based on a text message in a mobile environment, the receiver is a voice mail service method and system simulated by the voice of a celebrity receiving a voice mail simulated by the voice of a celebrity To provide.

1 is a block diagram showing the structure of a system in which a voice mail service method for simulating celebrity voices according to an embodiment of the present invention is performed;

2A and 2B are flow charts illustrating the transmitting and receiving steps of the voice mail service method for simulating celebrity voices according to an embodiment of the present invention, respectively.

3 is a diagram illustrating an example of a menu displaying celebrities classified by category.

Voice mail service method for simulating the celebrity voice according to an aspect of the present invention for achieving the above technical problem,

A feature parameter extracting unit for extracting feature parameters by analyzing voices of celebrities who are connected to a communication network and simulating voices, and assigning a serial number to the extracted feature parameters to store a serial number for each entertainer and the feature parameters A voice mail service method performed in a voice mail service system that simulates a celebrity voice including a server, a first user terminal for transmitting a message to the server, and a second user terminal for receiving a voice mail from the server. To

(a) accepting a user's selection of whether to apply voice simulation when the first user presses a send button of a text message to be sent from the first user terminal;

(b) presenting a selection menu in which the celebrities are categorized by category when the voice simulation is selected in step (a);

(c) sending, as a message, the first user's entertainer information to be synthesized by voice along with a text message to be sent based on a menu displayed on the screen of the mobile terminal;

(d) extracting, by the server, the celebrity serial number using the celebrity information in the message sent in step (c) and searching for the feature parameter by referring to the extracted celebrity serial number;

(e) synthesizing the voice using the feature parameter extracted from the server;

(f) storing the synthesized message in a voice mailbox; And

(g) confirming that the voice mail has arrived at the second user terminal; And

(h) accessing the voice mailbox from the second user terminal and receiving a voice mail sent by the first user and simulated by the voice of a famous entertainer.

In addition, the feature parameter includes at least the fourth and fifth formant frequency components, and step (e) includes:

(e ') synthesizing the speech in a connection synthesis scheme based on the LPC synthesizer using the feature parameters extracted from the server.

In addition, a voice mail service system according to another aspect of the present invention for achieving the above technical problem,

A server connected to a communication network, a first user terminal for transmitting a message to the server, and a second user terminal for receiving a voice mail from the server, wherein the text message is synthesized by voices of celebrities and sent as voice mail. In the voice mail service system,

The first user terminal,

When the first user presses a send button of a text message to be sent, the user receives a user's selection of whether to apply the voice simulation, and if the user selects to apply the voice simulation, the first celebrity is presented. Control unit; And

And a message transmitter configured to transmit, as a message, celebrity information to be synthesized together with a text message to be sent by the first user based on a menu displayed on the screen of the mobile terminal.

The server,

A feature parameter extraction unit for extracting voice feature parameters by analyzing voices of celebrities to be simulated and retrieving feature parameters using selected celebrity information in a message sent by the first user;

A database storing a serial number of each entertainer and the feature parameters by assigning a serial number to the extracted feature parameters;

A speech synthesizer for synthesizing speech using the retrieved feature parameter; And

A voice mailbox for storing the voice synthesized message as voice mail;

The second user terminal,

After confirming that the voice mail has arrived, the user accesses the voice mail box and receives the voice mail sent by the first user and simulated by the voice of a famous entertainer.

The feature parameter may include at least a fourth and fifth formant frequency component, and the speech synthesis unit includes:

It is preferable to synthesize speech using a connection synthesis method based on the LPC synthesizer using the feature parameters extracted from the server.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First of all, according to the present invention, in order to transmit a voice based on a text message, only a minimum information necessary for speech synthesis is transmitted in a wireless environment with many restrictions, and the voice can be synthesized using the transmitted information. have.

1 is a block diagram illustrating a structure of a system in which a voice mail service method that simulates a celebrity voice according to an embodiment of the present invention is performed. 2A and 2B are flowcharts illustrating transmission steps and reception steps of a voice mail service method that simulates a celebrity voice according to an embodiment of the present invention. 2A and 2B are often referenced below.

Referring to FIG. 1, a voice mail service system that simulates a celebrity voice according to the present invention, which synthesizes and reads a message sent by a user into voices of celebrities in real time, includes a server 12 connected to a communication network 10. , A first user terminal 14 for transmitting a message to the server 12, and a second user terminal 16 for receiving a message from the server. The server 12 includes a feature parameter extractor 122, a database 124, and a forwarding unit 126. The first user terminal 14 includes a transmission controller 142 and a message transmitter 144.

The operation of the system made as above will be described.

First, the feature parameter extraction unit 122 of the server 12 analyzes the voices of famous entertainers to be simulated, and includes at least a fourth formant F4 and a fifth formant F5. Is extracted (step 202), and the database 124 assigns the celebrity serial number to the extracted feature parameters and stores the celebrity serial number and the feature parameters (step 204).

Next, after the user composes a text message (step 206) in the first user terminal 14, the transmission control unit 142 of the first user terminal 14 transmits a text message button to which the first user wants to send. If is pressed (step 208), the user's selection of whether to apply voice simulation is accepted (step 210), and if not selected to apply voice simulation, a normal text message service (step 211) is performed, If it is selected to apply a replica, a menu displaying the celebrities by category is presented (step 212). 3 shows an example of the menu 30 that displays the famous entertainers by category.

Next, the message transmitter 144 of the first user terminal 14 transmits the celebrity information to be synthesized together with the text message that the first user wants to send based on the menu 30 displayed on the screen of the mobile terminal. (Step 214).

Next, the feature parameter extraction unit 122 of the server 12 searches the first database 124 using the entertainer information in the message sent by the first user to include the fourth and fifth formants. After retrieving feature parameters for the selected celebrity, the voice synthesizer 126 synthesizes the voice using the retrieved feature parameters, and stores the voice synthesized message in the voice mailbox 128 as voice mail (step 216). Where, famous

In the above embodiment, the characteristic parameter of the voice has been described as an example of the formant frequency component. Frequencies having peak values in the spectrum obtained when analyzing a voice feature parameter are called feature parameters or formant frequencies and are represented by F0, F1, F2, F3, F4, and F5. Among the characteristic parameters, F0 (0th formant), F1 (1st formant), F2 (2nd formant), and F3 (3rd formant) are used as information included in the audio signal. It can be obtained to some extent from the text message information, and F4 (fourth formant) and F5 (fifth formant) are preferably used as information on the tone of the speaker. According to the present invention, the server connected to the communication network analyzes the voices of many entertainers in advance to database the feature parameters, and the mail sender who receives the service transmits the text message transmitted by the sender to the voice simulation of the famous entertainers. In the case, a selection menu for celebrities supporting voice simulation is presented and one celebrity is selected from among them. The server voice synthesizes the text message using the selected parameter of the entertainer and stores the voice synthesized message as a voice mail in the voice mailbox.

In this case, the voice synthesis is most preferably a connection synthesis method based on the LPC synthesizer. In the synthesis method, the synthesized sound is generated by connecting the recorded pieces in the spotlight from the LPC. The LPC method is much simpler than the feature parameter synthesis because it represents the feature parameter information of the speech (0th formants to 5th formants) by LPC coefficients. The LPC coefficient is a slow change of the speech characteristic. The coefficient is predicted by mathematical modeling that can predict the current sample using the past speech sample. The LPC coefficient predicts the linear predictive coefficient every 10 msec. Can be restored at low data rates. LPC coefficients can be obtained by analyzing the original speech using Least squares autocorrelation method and Least squarescovariance method.When synthesizing, the synthesized sound is generated by the filter consisting of excitation signal and LPC coefficient according to voice / voice information. do.

In the connection synthesis method, the recorded sound (phoneme, diphone, demysyllable, etc.) is encoded by LPC coefficients and stored in advance, so the performance depends on the detailed preparation of the synthesis unit. You can also easily control the fundamental frequency and duration. However, it is difficult to convert personality and control the fundamental frequency of the original audio, resulting in distortion of the characteristic parameter position or the characteristic parameter bandwidth. In addition, increasing or decreasing the duration of the phone will not save the desired vowel sound, but also cause the distortion of rapid transitions between units. In particular, the LPC-based parameter synthesis method results in loss of information of a speech signal due to encoding, thereby degrading sound quality of the synthesized sound. Nevertheless, the parametric method has the advantages of easy implementation, minimizing memory, and easy rhyme control.

Next, the reception control unit 162 of the second user terminal 16 receives a message indicating that the voice mail has arrived, accesses a voice mail mailbox (step 220), and receives a voice mail produced by the voice of a celebrity. And confirm (step 224). For example, a mail sender may send a text message, but the receiver may check a voice mail synthesized with a humorous voice of a famous entertainer "Oh Ji Myung" or a voice unique to the famous sportsman "Lee Seung Yeop". . The voice mail can of course be stored in the phone (step 226), and a later payment can be made (step 228).

According to the voice mail service method and the system supporting the celebrity voice synthesis as described above, even if the mail sender sends a text message based on the text message of the mobile environment, the receiver is synthesized with the voice of the celebrity "Oji name". There is a different service that will check your mail.

The voice mail service system supporting the celebrity voice synthesis as described above obtains voice information from the 0th formant to the 3rd formant, and obtains the tone information of the speaker from the 4th formant and the 5th formant. You can read text messages by simulating the voices of famous entertainers.

In particular, the voice mail service method and system for simulating the celebrity voice according to the present invention as described above, the text message information without increasing the amount of transmission information based on the text message in a wireless environment where the transmission environment is poor compared to the wired environment It is possible to send a text message while maintaining the traffic of the receiver, and the receiver receives a text message sent by the sender in the form of voice mail, thus providing a different service than the conventional simple text message or voice mail.

The voice mail service method that simulates the celebrity voice according to the present invention as described above may be prepared by a computer program. The computer program is read and executed by a microcomputer to perform a text message service method that mimics a celebrity voice according to the present invention. In addition, the computer program may be stored in semiconductor memories such as ROM, EPROM, EEPROM, and flash memory, and may be downloaded via the Internet and other communication lines.

The above embodiments are illustrative for the purpose of understanding the present invention and various modifications and variations are possible within the scope of the invention as defined by the appended claims. Accordingly, the scope of rights of the text message service method and system for simulating celebrity voices according to the present invention is not limited to the above embodiment.

As described above, according to the present invention, when a message is sent based on a text message in a mobile environment, a receiver and a voice mail service method for simulating a celebrity voice that receives a voice mail simulated by a celebrity voice are provided. Is provided. In particular, the voice mail service method and system for simulating the celebrity voice according to the present invention as described above, the text message information without increasing the amount of transmission information based on the text message in a wireless environment where the transmission environment is poor compared to the wired environment It is possible to send a text message while maintaining the traffic of the receiver, and the receiver receives a simulated mail based on the voice mail of a celebrity, thereby providing a different service from the existing simple text message.

Claims (4)

A feature parameter extracting unit for extracting feature parameters by analyzing voices of celebrities who are connected to a communication network and simulating voices, and assigning a serial number to the extracted feature parameters to store a serial number for each entertainer and the feature parameters A voice mail service method performed in a voice mail service system that simulates a celebrity voice including a server, a first user terminal for transmitting a message to the server, and a second user terminal for receiving a voice mail from the server. To (a) accepting a user's selection of whether to apply voice simulation when the first user presses a send button of a text message to be sent from the first user terminal; (b) presenting a selection menu in which the celebrities are categorized by category when the voice simulation is selected in step (a); (c) sending, as a message, the first user's entertainer information to be synthesized by voice along with a text message to be sent based on a menu displayed on the screen of the mobile terminal; (d) extracting, by the server, the celebrity serial number using the celebrity information in the message sent in step (c) and searching for the feature parameter by referring to the extracted celebrity serial number; (e) synthesizing the voice using the feature parameter extracted from the server; (f) storing the synthesized message in a voice mailbox; (g) confirming that the voice mail has arrived at the second user terminal; And (h) accessing the voice mailbox from the second user terminal to receive a voice mail sent by the first user and simulated by the voice of the celebrity; a voice mail service supporting celebrity voice synthesis, comprising: Way. The method of claim 1, The feature parameter, At least a fourth and fifth formant frequency component, In step (e), (e ') synthesizing the voice in a connection synthesis method based on the LPC synthesizer using the feature parameters extracted from the server. A server connected to a communication network, a first user terminal for transmitting a message to the server, and a second user terminal for receiving a voice mail from the server, wherein the text message is synthesized by voices of celebrities and sent as voice mail. In the voice mail service system, The first user terminal, When the first user presses a send button of a text message to be sent, the user receives a user's selection of whether to apply the voice simulation, and if the user selects to apply the voice simulation, the first celebrity is presented. Control unit; And And a message transmitter configured to transmit, as a message, celebrity information to be synthesized together with a text message to be sent by the first user based on a menu displayed on the screen of the mobile terminal. The server, A feature parameter extraction unit for extracting voice feature parameters by analyzing voices of celebrities to be simulated and retrieving feature parameters using selected celebrity information in a message sent by the first user; A database storing a serial number of each entertainer and the feature parameters by assigning a serial number to the extracted feature parameters; A speech synthesizer for synthesizing speech using the retrieved feature parameter; And A voice mailbox for storing the voice synthesized message as voice mail; The second user terminal, A voice mail service system supporting celebrity voice synthesis, characterized in that a voice mail is received and a voice mail simulated by the first user and received by the first user is received by accessing the voice mailbox. The method of claim 3, The feature parameter, At least a fourth and fifth formant frequency component, The speech synthesis unit, And a voice mail service system supporting celebrity voice synthesis, characterized in that the voice is synthesized using a connection synthesis method based on an LPC synthesizer using the feature parameters extracted from the server.