JPS6382500A - Rule synthesized sound output unit - Google Patents

Abstract

(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Objective of the Invention (Field of Industrial Application)] The present invention relates to a regular synthesized speech output device that receives a character code from the outside and generates a corresponding synthesized speech.

(Prior Art) In recent years, rule-synthesized speech output devices that synthesize and output unit languages such as proper nouns have been used in information service operations in financial institutions and the like.

The rule synthesized sound output device inputs the external character code,
It has the function of combining single sounds + f1 information corresponding to each character code and synthesizing unit language such as proper nouns, but in recent years,
In order to improve the clarity of synthesized sounds, a method is being adopted in which accent information for each unit language is registered in advance.

In this method, the unit language to be output is received as a coded character string, and an accent dictionary is searched based on the character string to determine the accent of the synthesized sound of each unit language.

By the way, since there are an extremely large number of unit languages, including proper nouns such as people's names, product names, and place names, the scale of accent dictionaries is enormous, and it takes a considerable amount of time to search through accent dictionaries.

In the currently developed rule-synthesized speech output devices, the accent dictionary search and speech output processing for each unit language are performed in series, so when outputting the synthesized sounds of the unit language continuously, The problem with this method is that the interval between a certain unit language and the following unit language becomes longer than necessary, making it difficult to separate synthesized sounds.

(Problems to be Solved by the Invention) The present invention was made due to the above-mentioned circumstances, and even when the synthesized sounds of unit languages are output continuously, the interval between unit languages is longer than necessary. The purpose of this invention is to provide a rule-synthesized sound output device that does not require the following methods.

[Effects of the Invention] (Means for Solving the Problems) The rule-synthesized speech output device of the present invention includes an accent information storage means in which accent information of a predetermined unit language is stored, and a method for searching this accent information storage means. an accent information search means for extracting accent information corresponding to a unit language indicated by a character code string inputted from the outside; A regular synthesized sound output device comprising a synthesized speech information creation means that combines information and creates synthesized speech information by taking into account the extracted accent information, and a synthesized speech output means that outputs a synthesized speech according to the synthesized speech information. In the present invention, while the synthesized sound output means is outputting the synthesized sound of a certain unit language, the accent information output means is configured to search for accent information of the next unit subject.

(Function) In the rule synthesized speech output device of the present invention, while the synthesized speech output means is outputting the synthesized speech of a certain unit language, the accent information output means searches for accent information of the next unit subject. , After the output of the synthesized sounds of a certain unit language is finished, the synthesized sounds of the following unit language can be outputted immediately, and the interval between the unit languages does not become longer than necessary.

(Example) Hereinafter, details of an example of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing the configuration of an apparatus according to an embodiment of the present invention.

The apparatus of this embodiment includes a command interpreter 1 that interprets commands input from a host computer, etc., and selects each processing R function, a code interpreter 2 that interprets the code of a character string of a unit language, and An accent dictionary storage unit 3 stores accent information corresponding to a unit language; a dictionary search unit 4 searches for an accent dictionary according to the character code string of each unit language; A reading data storage unit 5 in which the reading data is expanded and the phonetic information corresponding to the character code are held, the phonetic information is combined according to the input character code string, and the phonetic information is combined into the reading data expanded in the reading data storage unit 5. It is comprised of a synthetic sound information creating means 6 that creates synthetic sound information based on the information, and a D/A converter 7 that converts the digital synthetic sound information into analog information and outputs it to the subsequent amplification system.

FIG. 2 is a diagram showing the format of commands input to the device of this embodiment.

Indicated by A in the figure. The first bit of the command specifies the search, the next bit is the number of characters, and the bits after that are JIS.
Indicates the standard character code string.

The command indicated by B consists only of bits indicating an audio output request.

Further, FIG. 3 is a diagram showing a specific example of the above command.

In the figure, AI and A2 indicate examples of dictionary search commands, and B indicates an example of an audio output request command.

The operation of the apparatus of this embodiment will be explained below.

First, when a command AI is input to the apparatus of this embodiment, the command interpreter 1 checks the first bit of the command.

The first bit of the command A1 is "01", that is, it specifies a dictionary search, and this content is transmitted to the code interpreter 2.

The code interpretation unit 2 obtains the number of characters 10 from the second and subsequent bits of the command and the character code strings B6, CC, DE, etc. corresponding to “Kabushiki Gaisha (Co., Ltd.)”, and the dictionary search unit 4 Output to.

The dictionary search section 4 searches the accent dictionary storage section 3 to determine whether or not accent information corresponding to the character code string "r" is stored in the accent dictionary storage section 3.

If there is corresponding accent information, it is extracted, and based on the accent information, reading data consisting of phonetic parameters of "Kabushiki Gaisha" is developed in the reading data storage section 5.

This reading data is sent to the synthesized sound information creation unit 6, and the synthesized sound information creation unit 6 generates a digital synthesized sound from the reading data corresponding to the above-mentioned “kabushikigaisha” at the stage when the command B requesting the output of voice is input. information is created and this synthesized sound information is output to the D/A converter 7.

In the device of this embodiment, following the command A1,
When command A2 that requests the output of a synthesized sound of the character code string "Tokyo (Tokyo)" is input, the command A1
While the synthesized sound of "Kabushiki Gaisha" is being output, the dictionary search section 4 searches for accent information corresponding to the character code string "Tokyo J" and develops the reading data in the reading data storage section 5.

Then, when the command B requesting audio output is input again, the synthesized sound information creation unit 6 uses the reading data to generate the above-mentioned "
Synthetic sound information corresponding to "Tokyo" is created and this synthetic sound information is output to the D/A converter 7.

In this way, in the device of this embodiment, the reading data corresponding to "Tokyo" has already been developed at the stage when the output of the synthesized sound for "Kabushiki Gaisha" has been completed, so the synthesized sound information creation section 6 immediately creates the synthesized sound information. However, the synthesized sound corresponding to "Tokyo" is outputted without unnecessary intervals.

In the above-described embodiment, reading data is expanded with the command A, and synthesized sound information is created with the command B, and the synthesized sound is output. If a means for storing the synthesized sound information is provided, command A may be used to create synthetic sound information, and command B may be used to obtain a synthesized sound from the synthesized sound information.

[Effects of the Invention] As explained above, in the rule synthesized sound output device of the present invention, while the synthesized sound output means outputs the synthesized sound of a certain unit language, the accent information output means outputs the accent information storage means. Since the search is performed, the synthesized sounds of the next unit language can be outputted immediately after outputting the synthesized sounds of a certain unit language, and the interval between the unit languages does not become longer than necessary.

According to the present invention, there is no need to increase the hardware performance of accent dictionary search, so the cost of the device does not increase.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a device according to an embodiment of the present invention, FIG. 2 is a diagram conceptually showing the format of commands input to the device, and FIG. 3 is a concrete example of the same command. FIG. 1...Command interpretation section 2...Code interpretation section 3...Accent dictionary storage section 4...
...Dictionary search section 5...Reading data storage section 6...
...Synthetic sound information creation section 7...D/A converter applicant Toshiba Corporation Agent Patent attorney Satoshi Suyama - Figure 1 Figure 2

Claims (1)

[Claims] (1) An accent information storage means that stores accent information of a predetermined unit language, and an accent that searches this accent information storage means to extract accent information corresponding to the unit language indicated by a character code string input from the outside. an information retrieval means, holding monophonic information corresponding to each character code, combining the monophonic information according to the input character code string, and creating synthesized speech information by adding the extracted accent information; In a rule synthesized sound output device comprising a synthesized sound information creation means and a synthesized sound output means for outputting a synthesized sound according to the synthesized sound information, while the synthesized sound output means outputs a synthesized sound of a certain unit language. 2. A rule synthesized speech output device, characterized in that the accent information output means is configured to search for accent information of the next unit subject.