JPH10143352A - Method for converting sentence information, and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently convert an information described by sentence into a voice information by converting voice synthesized information so as to match with an interface, and generating the voice information from thus converted voice synthesized information. SOLUTION: Storage information stored in an information storing device 30 is taken in a sentence information converting device 10 by an information retrieving means 11 via a network part 20. The voice synthesized information A42 converted in a stored information converting means 12 is converted into the voice synthesized information B43 of the interface such that is acceptable by an voice synthesizing means 14 in an interface converting means 13 and transmitted to the voice synthesizing means 14. The voice synthesizing means 14 performs voice synthesizing processing based on the transmitted data, and sends out voice information 44 to a voice information output means 15 to be outputted to the outside therefrom. Thus, sentence information used in www of internet is efficiently synthesized as high-quality voice information.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to an information retrieval apparatus for transmitting desired information from an information storage apparatus between an information storage apparatus and an information retrieval apparatus connected via a network constituted by telecommunication facilities and the like. The present invention relates to a technical field of searching and presenting, and particularly to a technical field of converting and presenting voice.

[0002]

2. Description of the Related Art WWW constructed on the Internet
In a distributed database system such as (World Wide Web) configured by an information storage device and an information search device connected via a network including telecommunications facilities, there is no image display means as information output means. In a simple information retrieval device, the information output means is limited to voice information, vibration, and the like. Under these conditions, the HT used in WWW on the Internet
A sentence described in the ML language is converted into speech information by speech synthesis and output.

[0003]

However, the HTM
Since L was originally devised on the premise of text formatting, to convert HTML text into high-quality audio information,
There has been a problem that a structural analysis of the HTML text is required. Furthermore, even if the HTML text is analyzed, it is difficult to convert the text into voice information that completely reflects the intention of the author.

[0004] An object of the present invention is to provide a method for efficiently retrieving information described by a sentence in the case of performing an information retrieval without using an information retrieval apparatus having no image display means or an image means. It is an object of the present invention to provide a text information conversion device and a conversion method for converting text information.

[0005]

SUMMARY OF THE INVENTION A sentence information conversion apparatus according to the present invention is a sentence information conversion apparatus for efficiently converting stored information into speech information and providing the converted speech information. Information retrieval means for retrieving stored information from a stored information storage device via a network composed of telecommunications facilities, etc., and stored information for analyzing a speech synthesis command from the retrieved stored information and converting it into speech synthesis information Conversion means, interface conversion means for converting the speech synthesis information to match the interface of the speech synthesis means, speech synthesis means for generating speech information from the converted speech synthesis information, and outputting the speech information to the outside Audio information output means.

A sentence information conversion method according to the present invention is a sentence information conversion method using the above-described sentence information conversion device. In the storage information to be converted by the information storage device, a speech synthesis command is previously stored. The information added as the storage information and to which the command for voice synthesis is added is fetched by the search means of the conversion device via the network and sent to the information conversion means. Analyzes the synthesis command, converts it into text data used as speech synthesis information and a command for speech synthesis to generate speech synthesis information, sends the generated speech synthesis information to the interface conversion means, and performs interface conversion. The means converts the speech synthesis information into an interface corresponding to the speech synthesis means and sends it to the speech synthesis means. It generates audio information by performing speech synthesis according to the data from the speech synthesis information, sent out to the audio information output means, voice information output means, for outputting voice information to the outside.

The information stored in the information storage means is preferably described in an HTML language, and the speech synthesis command is preferably described in a command for allowing a search system provided in the HTML to understand it as a comment.

The information stored in the information storage means is HTML
A command described in a language and uniquely provided in HTML may be used to make the search system understand that there is a description using a speech synthesis command.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS HTML, which is a language for describing text information used in the WWW of the Internet and the like, uses a text formatting command called a tag. Although it is possible to generate text-to-speech information by analyzing a sentence using this tag as a clue, there is not enough information to convert it into high-quality voice information that is easy to listen to. In order to solve this, it is necessary to interpose an interface that can be understood by the speech synthesis means, but it is necessary to include commands premised on speech synthesis processing in HTML.

[0010] However, HTML is a specification used worldwide, and arbitrarily expanding it will confuse ordinary search systems that do not have an interface with speech synthesis means, and will cause browsing with ordinary search systems. Will interfere. Therefore, it is necessary to embed a command for speech synthesis processing so as not to affect a search system for analyzing ordinary HTML. Specifically, HTML
There is a command that makes the search system understand as a comment, and the command line of this comment is not output on the screen even if the system recognizes it. In the present invention, using this characteristic, a command for speech synthesis processing is described in a comment line. Therefore, the voice synthesis processing unit may check whether or not a command for voice synthesis processing is described in a comment line that is normally ignored.

Table 1 shows examples of commands for speech synthesis processing. For example,

[0012]

[Table 1] In this case, the utterance speed is set to level 4, the utterance volume is set to level 8, and the voice information is output with the speaker being a woman. As described above, commands such as those shown in Table 1 are described in the HTML document as necessary, so that the information retrieval system that does not support the speech synthesis means is not confused and can be used for the speech synthesis means. It is possible to directly control the speech synthesis means for the search system, and it is possible to provide high-quality, easy-to-listen speech information.

[0013] Further, a command for speech synthesis processing may be described by providing a tag indicating that the command for speech synthesis processing is described in the HTML language. In this description method, it is first shown that information for speech synthesis is described by a tag, and information such as text and commands for speech synthesis is described after the tag.
This example is shown in Table 2.

[0014]

[Table 2] In the above example, <script language = "SoundScript"> and </ s
cript>, information for speech synthesis is described, and <! /-and
-> (H which is interpreted as a comment in the information retrieval system
A command for controlling the voice synthesizing means and information such as text to be voice-synthesized are described between the description and the description in the TML language. This description portion is passed to the voice synthesizing means via the interface section and output by the voice output means. There is an advantage that the load on the information retrieval system can be reduced by adding this method.

On the other hand, in a speech synthesizing unit and an information retrieval system having no interface with the speech synthesizing unit,
This part is ignored because <script language = "SoundScript"> ... </ script> cannot be recognized.

Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a text information conversion device according to a first embodiment of the present invention. In the drawing, reference numeral 10 denotes a text information conversion device, 11 denotes information retrieval means,
2 is stored information conversion means, 13 is interface conversion means, 14 is voice synthesis means, 15 is voice information output means, 2
0 is a network unit, 30 is an information storage device, 41 is stored information, 42 is speech synthesis information A, 43 is speech synthesis information B,
44 and 45 are voice information.

The ten sentence information conversion device is constituted by, for example, a personal computer. The eleventh information retrieval means, the twelve stored information conversion means, the thirteen interface conversion means, and the voice synthesis means 14 operate on a personal computer. It can be realized as application software. The fifteen audio information output means is a speaker device or the like.

Next, the text information conversion method of the present invention will be described. Text information in HTML used in the Internet WWW or the like is stored in the information storage device 30 and is searched and browsed as stored information via the network unit 20. In the present invention, the information storage device 30 stores the storage information described in the normal HTML and the storage information described in the HTML and the voice synthesis command. In the description here, the HTML and the text information described by the voice synthesis command will be regarded as accumulated information.

The information stored in the information storage device 30 is fetched by the information retrieval means 11 into the text information conversion device 10 via the network unit 20. The information retrieval means 11 sequentially sends the acquired accumulated information 41 to the accumulated information converting means 12 while buffering the accumulated information 41 in a necessary amount. The stored information conversion means 12 deciphers the command for speech synthesis from the stored information 41 sent,
The data is converted into text data used as speech synthesis information and commands used during speech synthesis, and unnecessary portions of the speech synthesis information are deleted and converted into speech synthesis information A42.

At this time, commands analyzed as the speech synthesis information A42 include commands as shown in Table 3. The stored information conversion means 12 searches for the commands shown in Table 3 from the stored information and converts them into data in a format that can perform the functions specified in the table.

[0021]

[Table 3] Speech synthesis information A42 converted by stored information conversion means 12
Is the speech synthesis information B of the interface such that the speech synthesis means 14 is accepted by the interface conversion means 13
43 and sent to the speech synthesis means 14. The voice synthesizing unit 14 performs a voice synthesizing process according to the transmitted data, sends voice information 44 to the voice information output unit 16, and outputs the voice information to the outside from the voice information output unit.

With such a configuration, it is possible to efficiently synthesize text information used on the WWW of the Internet as high-quality voice information.

[0023]

According to the present invention, the Internet WW
Efficient and high-quality information for speech synthesis can be provided to the sentence information used in W, and there is an effect that efficient and high-quality speech synthesis can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a text information conversion device according to a first embodiment of this invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Sentence information conversion device 11 Information search means 12 Stored information conversion means 13 Interface conversion means 14 Voice synthesis means 15 Voice information output means 20 Network unit 30 Information storage device 41 Stored information 42 Voice synthesis information A 43 Voice synthesis information B 44, 45 Audio information

Continuation of front page (72) Inventor Noboru Sonehara 3-19-2 Nishishinjuku, Shinjuku-ku, Tokyo Japan Telegraph and Telephone Corporation

Claims (4)

[Claims] 1. A sentence information conversion apparatus for efficiently converting stored information into audio information and providing the converted information, comprising an information storage apparatus storing information to be searched, and a telecommunications facility. Information retrieval means for retrieving stored information via a network, such as: a stored information converting means for analyzing a speech synthesis command from the retrieved stored information and converting the speech synthesis information into speech synthesis information; Interface conversion means for converting to match the interface of the means, voice synthesis means for generating voice information from the converted voice synthesis information, and voice information output means for outputting the voice information to the outside. A sentence information conversion device, characterized in that: 2. A text information conversion method using the text information conversion device according to claim 1, wherein a voice synthesis command of the information storage device includes a voice synthesis command in advance. The information to which the command for voice synthesis is added is taken in by the search means of the conversion device via the network and sent to the information conversion means, and is sent by the stored information conversion means. The voice synthesis command is analyzed from the accumulated information, and converted into text data used as voice synthesis information and a voice synthesis command to generate voice synthesis information, and the generated voice synthesis information is transmitted to the interface. The interface conversion means converts the speech synthesis information into an interface corresponding to the speech synthesis means. The speech synthesis means performs speech synthesis processing in accordance with data from the speech synthesis information to generate speech information, and sends the speech information to the speech information output means. A text information conversion method characterized by outputting information to the outside. 3. The method according to claim 1, wherein the stored information of the information storage means is HT.
The text information according to claim 2, wherein the text information is described in an ML language, and the speech synthesis command is described in a command for causing a search system provided in the HTML to understand the command as a comment. Conversion method. 4. The method according to claim 1, wherein the stored information of the information storage means is HT.
3. The text information conversion method according to claim 2, wherein a description is made in the ML language, and a command uniquely provided in the HTML allows a search system to understand that there is a description by the command for speech synthesis.