JPH09251373A - Sound synthesis method/device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To set only a necessary part in an electronic mail to be synthesized to sound by recognizing prescribed indication information contained in an input sentence and extracting a part which is sound-synthesized from the input sentence in accordance with the indication information so as to generate synthesis sound. SOLUTION: A command recognition processing part 1 recognizes prescribed indication information contained in the input sentence and extracts the part which is synthesized to sound from the input sentence in accordance with indication information. Namely, the command recognition processing part 1 recognizes (retrieves)a sound command '∥speech' mentioned at the head of a line when a text file is inputted. When the command recognition processing part 1 recognizes the sound command '∥speech', it extracts the line where the sound command '∥speech' is mentioned on the head from the text file and supplies it to the text sound synthesis part 2 of a poststage.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a voice synthesizing method and a voice synthesizing apparatus. In particular, the present invention relates to a voice synthesizing method and a voice synthesizing device that enable to obtain a synthesized voice that is easy to understand.

[0002]

2. Description of the Related Art FIG. 8 shows the configuration of an example of a conventional text-to-speech synthesizer. This voice synthesizer is composed of a text voice synthesizer 2, a voice output unit 5, and a speaker 6.

A sentence (input sentence) mixed with kana and kana, which is a text file of Japanese language, for example, to be speech-synthesized,
It is supplied to the text-to-speech synthesis unit 2. The text-to-speech synthesis unit 2 is composed of a language processing unit 3 and a speech synthesis unit 4, and the text file is input to the language processing unit 3. In the language processing unit 3, first, the kanji reading is added to each word / phrase in the input sentence that forms the text file. Furthermore, the language processing unit 3 analyzes the syntactic structure of the input sentence, and adds accent information regarding the accent based on the analysis result. After that, the voice synthesis unit 4 generates a synthetic sound corresponding to the input sentence while performing prosody control based on the reading and accent information attached to the input sentence.

That is, in the speech synthesis unit 4, for example, phoneme unit data corresponding to the reading attached to the input sentence is matched with the content of the sentence of the input sentence such as emphasis, intonation, and pause based on the accent information and the like. It is connected while performing the prosody control.

Specifically, the voice synthesis unit 4 generates prosodic information for adding appropriate intonation, emphasis, and pause to the synthesized voice based on the accent information and the like, and the phoneme piece data is converted into prosodic information. Are connected based on. That is, when the prosody information includes, for example, the pitch pattern of the input sentence, the duration of each phoneme that constitutes the input sentence, the power of each phoneme, and the like, first, phoneme segment data is obtained based on the pitch pattern. The connection interval is adjusted (the pitch period of the phoneme piece data is adjusted), and the number of times of repeatedly connecting the phoneme piece data corresponding to the phoneme is controlled based on the phoneme duration. Further, the amplitude of the phoneme piece data corresponding to the phoneme is controlled based on the power of the phoneme.

The speech waveform obtained by connecting the phoneme unit data based on the prosody information as described above is supplied to the speech output unit 5. The voice output unit 5 includes, for example, a D / A converter and an amplifier, and D / A converts the voice data from the text voice synthesis unit 2 (voice synthesis unit 4).
Further, the level is adjusted appropriately and supplied to the speaker 6. As a result, the speaker 6 outputs a synthesized sound corresponding to the input sentence.

[0007] By the way, recently, the Internet has spread rapidly, and messages are exchanged by electronic mail (E-ma).
il) is getting more done. E-mail can be sent regardless of whether the other party is absent, and the sent e-mail can be viewed at any time by the other party. There is no such thing as not being able to contact you because you are absent.

However, a terminal such as a computer is required to view the e-mail, and therefore it is difficult to confirm the e-mail addressed to itself from a place where the user is out.

Therefore, NIFTY-Serve (trademark), which provides a so-called personal computer communication service, provides a reading service using a synthesized voice of electronic mail. According to this service, when the user accesses the center station by telephone, the e-mail addressed to himself is read aloud by the synthesized voice,
This allows the user to check the e-mail even without a computer.

The reading of the electronic mail by the synthesized voice is performed by the voice synthesizer as shown in FIG.

[0011]

By the way, in the voice synthesizing apparatus of FIG. 8, the voice synthesizing process is performed with the input text file as a processing unit. That is, when one text file is input to the speech synthesizer, all the text data included in the text file are subject to speech synthesis. Therefore, the speech synthesizer receives an e-mail as a text file, for example. When input, the above-described voice synthesis processing is performed on all text data included in the electronic mail.

However, as shown in FIG. 9, for example, a figure created by combining symbols in addition to Japanese kanji / kana mixed sentences, and so-called so-called e-mail for expressing the detailed feelings of the sender are included in the e-mail. Face mark (for example, (^
(^) Etc.), a mail header including information indicating the route until the email is transferred, a so-called signature of the sender, etc. are also included. There was a problem that became difficult to do.

The present invention has been made in view of such a situation, and it is possible to obtain a synthesized voice that is easy to understand.

[0014]

A speech synthesis method according to claim 1 recognizes predetermined instruction information included in an input sentence,
It is characterized in that a part to be voice-synthesized is extracted from the input sentence in accordance with the instruction information, and only a synthesized sound corresponding to the part to be voice-synthesized is generated.

According to a fourth aspect of the present invention, a voice synthesizing device synthesizes voice from an input sentence corresponding to the recognizing means for recognizing predetermined instruction information contained in the input sentence and the instruction information recognized by the recognizing means. The present invention is characterized by comprising extraction means for extracting a part and voice synthesis means for generating a synthetic sound corresponding to the part for voice synthesis extracted by the extraction means.

In the voice synthesizing method according to the first aspect, the predetermined instruction information included in the input sentence is recognized, the portion to be voice-synthesized is extracted from the input sentence corresponding to the instruction information, and the voice thereof is extracted. Only the synthesized sound corresponding to the part to be synthesized is generated.

In the speech synthesizer according to the fourth aspect, the recognizing means recognizes the predetermined instruction information included in the input sentence, and the extracting means inputs the instruction information corresponding to the instruction information recognized by the recognizing means. From the sentence, the part to be synthesized by voice is extracted. The voice synthesizing unit is configured to generate a synthetic voice corresponding to the part to be voice-synthesized extracted by the extracting unit.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below, but before that, in order to clarify the correspondence between each means of the invention described in the claims and the following embodiments. The features of the present invention are described as follows by adding a corresponding embodiment (however, an example) in parentheses after each means.

That is, the speech synthesizer according to the fourth aspect is
A voice synthesizing device for generating a synthetic voice corresponding to an input sentence, comprising a recognition means for recognizing predetermined instruction information included in the input sentence (for example, a command recognition processing unit 1 shown in FIGS. 1 and 3), Corresponding to the instruction information recognized by the recognition means, an extraction means (for example, the command recognition processing unit 1 shown in FIG. 1 or FIG. 3) for extracting a part to be voice-synthesized from the input sentence and the extraction means. Voice synthesizing means for generating a synthesized voice corresponding to the part to be voice-synthesized (for example, FIG.
And the text-to-speech synthesis unit 2 shown in FIG. 3).

The speech synthesizing apparatus according to a fifth aspect of the invention further comprises an addition means (for example, the mail header processing section 31 shown in FIG. 3) for adding instruction information to a predetermined portion of the input sentence. And

Of course, this description does not mean that each means is limited to the above.

FIG. 1 shows the configuration of an embodiment of a speech synthesizer to which the present invention is applied. In the figure, parts corresponding to those in FIG. 8 are designated by the same reference numerals, and description thereof will be omitted below as appropriate. That is, this voice synthesizer has the same configuration as the voice synthesizer of FIG. 8 except that the command recognition processing unit 1 is newly provided.

The command recognition processing section 1 recognizes predetermined instruction information included in the input sentence, and in response to the instruction information,
From the input sentence, the part to be speech-synthesized is extracted. Then, the command recognition processing unit 1 is configured to supply only the extracted portion to the text-to-speech synthesis unit 2 in the subsequent stage.

Next, the operation will be described. In addition,
Here, the instruction information is, for example, information for instructing the portion of the text file to be voice-synthesized line by line (such instruction information will be referred to as a voice command hereinafter). In addition, the line is one line up to the line feed code,
Therefore, from the beginning of the text file to the first line feed code constitutes one line, after that, the line feed code to the next line feed code constitutes one line, and from the last line feed code,
It is assumed that the end of the text file constitutes one line. Further, the voice command is represented by, for example, "\ speech", and when a certain line is voice-synthesized, it is described at the beginning of the line.

Now, for example, assume that the following text file TF1 is input to the command recognition processing unit 1 (note that the portion enclosed by the leading and trailing parentheses "," is the content of the text file). Represents).

Text file TF1: "This text is
A text sentence for explaining the embodiment of the present invention. In conventional text-to-speech synthesis, one text file is used as a unit for synthesis processing.
Even if the writer of the text had some parts that he wanted to synthesize and some parts that he didn't want to synthesize, he couldn't select it and all the text was synthesized. Therefore, for example, the contents written before this line are not synthesized by voice,
Even if there was a request to synthesize speech after this line, it could not be selected and synthesized. However, the method of the present invention makes it possible. "

When the text file TF1 is input, the command recognition processing unit 1 recognizes (searches) the voice command "\ speech" described at the beginning of the line. However, in this case, in the text file TF1, there is no line in which the voiced command “\ speech” is described at the beginning of the line, so the command recognition processing unit 1 does not recognize the voiced command “\ speech”. Therefore, none of the lines forming the text file TF1 is extracted. As a result, nothing is supplied to the text-to-speech synthesis unit 2 in the subsequent stage, and no synthetic sound is output.

Next, for example, assume that the following text file TF2 is input to the command recognition processing unit 1.

Text file TF2: "This text is
A text sentence for explaining the embodiment of the present invention. In conventional text-to-speech synthesis, one text file is used as a unit for synthesis processing.
Even if the writer of the text had some parts that he wanted to synthesize and some parts that he didn't want to synthesize, he couldn't select it and all the text was synthesized. Therefore, for example, even if there is a request to synthesize the contents written before this line without speech synthesis, and \ speech, it was not possible to select this \ speech and synthesize it. . But the \ speech invention method makes it possible. "

When the text file TF2 is input, the command recognition processing unit 1 recognizes (searches) the voice command "\ speech" described at the beginning of the line. In this case, the voice command "\ speech" described at the beginning of the last three lines of the text file TF2 is recognized.
The command recognition processing unit 1 uses the voice command "\ speech".
When the voice recognition command is recognized, the line in which the voice command "\ speech" is described at the beginning of the line is extracted from the text file TF2 and supplied to the text voice synthesis unit 2 in the subsequent stage. That is,
In this case, the text-to-speech synthesizer 2 cannot select and synthesize the text data "even if there is a request to synthesize speech after this line. However, the method of the present invention is used. You can do that with it. "Is supplied as an input sentence.

Therefore, the text-to-speech synthesis unit 2 performs the same speech synthesis processing as in the case of FIG. 8 described above only on the line in which the speech command "\ speech" is described at the beginning of the line. , From the speaker 6, the synthesized sound corresponding to the last three lines of the text file TF2 "Even if there is a request to synthesize speech after this line, it could not be selected and synthesized. , It is possible using the method of the present invention. "Is output.

As described above, the voice command "\ speech" described at the beginning of the line is recognized, and the line having the voice command "\ speech" described at the beginning is extracted from the text file TF2. Since the voice synthesis is performed, the user (text writer) simply describes the voice command "\ speech" in the portion (here, line) that the user desires to perform voice synthesis. It is possible to output only a synthesized sound.

In the above case, the voice command "\ speec" is added to the beginning of every line of the voice synthesis target.
Although "h" is described, for example, when there are two or more lines to be voice-synthesized, the voice command "\ speech" is described only at the beginning of the first line, and the voice-synthesis target is written. It is also possible to enclose the part to be enclosed in parentheses.

That is, for example, the following text file TF3 can be input to the command recognition processing section 1.

Text file TF3: "This text is
A text sentence for explaining the embodiment of the present invention. In conventional text-to-speech synthesis, one text file is used as a unit for synthesis processing.
Even if the writer of the text had some parts that he wanted to synthesize and some parts that he didn't want to synthesize, he couldn't select it and all the text was synthesized. Therefore, for example, even if there is a request to synthesize the contents written before this line without speech synthesis and \ speech {after this line, it could not be selected and synthesized. However, the method of the present invention makes it possible. } "

In this case, in the command recognition processing unit 1, after the voice command "\ speech" and the parenthesis "{" are recognized, only the portion up to the parenthesis "}" is recognized by the text-to-speech synthesis unit 2. Is supplied to.

Therefore, also in this case, the text file T
As in the case of F2, even if there is a request from the speaker 6 that the synthesized voice corresponding to the last three lines of the text file TF3 is “voice-synthesized after this line,
It was not possible to select it and synthesize speech.
However, the method of the present invention makes it possible. Is output.

Next, in the above, as the instruction information,
Although the voice command "\ speech" for instructing a part of the text file to be subjected to the voice synthesis process is used, other instruction information is, for example, a part of the text file not to be subjected to the voice synthesis process is instructed. It is also possible to use a command (hereinafter, referred to as an unvoiced command as appropriate). Here, this non-vocalization command is represented by, for example, "\ mute", and when a certain line is not speech-synthesized, it is assumed that the non-vocalization command "\ mute" is described at the beginning of the line. The operation of the synthesizer will be described.

Now, for example, the above-mentioned text file T
It is assumed that F1 is input to the command recognition processing unit 1.

When the text file TF1 is input, the command recognition processing unit 1 recognizes (searches) the voice-disabled command "\ mute" described at the beginning of the line. However, in this case, in the text file TF1, there is no line at the beginning of which the voice-disabled command “\ mute” is described. Therefore, the command recognition processing unit 1 recognizes the voice-disabled command “\ mute”. Not, therefore all the lines that make up the text file TF1 are extracted.

As a result, the entire text file TF1 is supplied as an input sentence to the text-to-speech synthesis unit 2 in the subsequent stage, whereby the speaker 6 produces synthesized speech corresponding to all the lines forming the text file TF1. "This sentence is a text sentence for explaining the embodiment of the present invention. In the conventional text-to-speech synthesis, since one text file is used as a unit for synthesis processing, the person who wrote the text does not Even if there is a part that you want to synthesize and a part that you do not want to synthesize, you cannot select it, and all the texts have been synthesized, so for example, the contents written before this line are not synthesized. , Even if there was a request to synthesize speech after this line, it was not possible to select it and synthesize speech. Using the method of the present invention it becomes the possible. "It is output.

Next, for example, it is assumed that the following text file TF4 is input to the command recognition processing unit 1.

Text file TF4: "\ mute This sentence is a text sentence for explaining the embodiment of the present invention. \ Mute In the conventional text-to-speech synthesis, one text file is synthesized as a unit \ mute. Even if there is a part in the text file that the author of the text \ mute wants to synthesize and a part that he / she does not want to synthesize, it cannot be selected, and all the text is synthesized. Therefore, \ mute For example, even if there is a request to synthesize speech after this line without synthesizing the content written before this line, it is not possible to select it and synthesize speech. I couldn't, but the method of the invention makes it possible. "

When the text file TF4 is input, the command recognition processing section 1 recognizes (searches) the voice-disabled command "\ mute" described at the beginning of the line. In this case, the voice-disabled command "\ mute" described at the beginning of the first to seventh lines of the text file TF4
Is recognized. When the command recognition processing unit 1 recognizes the non-vocalization command "\ mute", the command recognition processing unit 1 removes the portion except the line where the non-vocalization command "\ mute" is described at the beginning of the line
It is extracted from the text file TF4 and supplied to the text-to-speech synthesis unit 2 in the subsequent stage. In other words, in this case, the text-to-speech synthesizer 2 cannot select and synthesize the text data "even if there is a request to synthesize speech after this line." It is possible using the method. "Is supplied as an input sentence.

Therefore, the text-to-speech synthesis unit 2 performs the speech synthesis process only on the portion (line) excluding the line in which the non-speech command "\ mute" is described at the beginning of the line,
As a result, the text file TF4 is output from the speaker 6.
The synthesized speech corresponding to the last three lines of "If you want to synthesize speech after this one, you could not select it and synthesize it. However, using the method of the present invention, Is possible. ”Is output.

As described above, the voice-disabled command "\ mute" described at the beginning of the line is recognized, and the lines other than the line where the voice-disabled command "\ mute" is described at the beginning of the line are changed to a text file. Since the speech is synthesized by extracting it from the TF4, the user does not have the speech-incapable command “\ m
By simply describing "ute", it is possible to prevent the desired portion from being output as a synthetic sound. That is, in other words, as in the case of using the voice command "\ speech" described above, the user only needs to describe the non-vocalization command "\ mute" in addition to the portion where the voice synthesis is desired. It is possible to output only the desired portion as a synthetic sound.

In the above case, the non-vocalization command "\ mute" is added to the beginning of every line in which voice synthesis is not performed.
However, if there are two or more lines to be speech-synthesized, as in the case of using the voicing command "\ speech", voicing is not possible only at the beginning of the first line. It is also possible to describe the command "\ mute" and enclose the target part of speech synthesis in parentheses.

That is, for example, the following text file TF5 can be input to the command recognition processing unit 1.

Text file TF5: "\ mute {This sentence is a text sentence for explaining the embodiment of the present invention. In conventional text-to-speech synthesis, since one text file is synthesized as a unit, the text file Even if there are some parts of the text that the writer of the text wants to synthesize and some parts that he does not want to synthesize, the text cannot be selected, and all the texts have been synthesized. So, for example, before this line Even if there is a request to synthesize speech after this line, it is not possible to select it and synthesize speech. However, using the method of the present invention, It will be possible. "

In this case, the command recognition processing unit 1 recognizes only the part up to the parenthesis "}" after recognition of the non-vocalization command "\ mute" and the parenthesis "{". It is supplied to the voice synthesizer 2.

Therefore, also in this case, the text file T
Similar to the case of F4, even if there is a request from the speaker 6 that the synthesized voice corresponding to the last three lines of the text file TF5 "I want to perform voice synthesis after this line,
It was not possible to select it and synthesize speech.
However, the method of the present invention makes it possible. Is output.

Note that, for example, when a line and another line are designated (for example, by clicking with a mouse), an editor for creating and editing a text file will display everything between the designated lines. It is possible to add the voice command "\ speech" or the non-voice command "\ mute" to each line (or the parentheses ({,}) surrounding all the lines The command "\ speec
It is possible to add the command "h" or the command "\ mute" that cannot be voiced).

Next, FIG. 2 shows the configuration of an embodiment of a network system to which the present invention is applied. The user has a computer 11 and, for example, PSTN (Public
Switched Telephone Network) and ISDN (Integrate
SP (Service Provider) server 13 of a service provider (connector) via a public network 12 such as d Service Digital Network) or a private line (not shown).
It is connected to the. The SP server 13 is connected to the Internet 14. That is, computer 1
1 is connected to the Internet 14 via the SP server 13.

In the embodiment shown in FIG. 2, the computer 1
As a computer of a user other than the one user, only the computer 15 is connected to the Internet 14 via the public network 12 and the SP server 16 (a server of another service provider) configured similarly to the SP server 13. However, the computers of other users are similarly connected to the Internet 14 via the SP server 13, the server of another service provider, or the server (host computer) installed in a university, a company, or the like. Has been done.

Although the user can directly connect to the Internet 14, normally, the user makes a contract with a service provider and, as shown in FIG. 2, the SP server 13 or 16 via the public network 12. Access to the Internet 14 to connect to the Internet 14.

On the Internet 14, TCP /
IP (Transmission Control Protocol / Internet Proto)
According to a protocol called col), it is designed to communicate between computers. A WWW is built on the Internet 14. In this WWW, HTTP (Hyper Text Transfer Protocol) is used.
Data is transferred by a protocol called l), and a screen is described in HTML (Hyper Text Markup Language), so that information can be easily retrieved and displayed. In addition to the WWW, the Internet 14 is provided with services such as so-called electronic mail (E-mail) and net news equivalent to bulletin boards in personal computer communication, and has a terminal (computer). The users can exchange e-mails, write articles on a specific subject, and read the articles. The email is
A protocol called SMTP (Simple Mail Transfer Protocol) and an article in net news are transferred by a protocol called NNTP (Network News Transfer Protocol).

In the network system of FIG. 2, the SP server 13 (or 16) can be accessed via the public network 12 not only by the computers 11 and 15 but also by the telephone (mobile telephone) 17 or the like. With this, for example, the user of the computer 11 operates the push button 17A of the telephone 17 to give a predetermined command to the SP server 13 and send the electronic mail to the user of the computer 11. Can be listened to with a synthetic voice (a service for providing an electronic mail with a synthetic voice in this way is hereinafter referred to as a voice service).

That is, for example, suppose that the user of the computer 15 now creates an electronic mail by the electronic mail application 15A for creating an electronic mail, which is software, and sends the electronic mail to the user of the computer 11. The electronic mail is received by the SP server 16 via the public network 12. SP server 16
When the electronic mail addressed to the user of the computer 11 is received, the electronic mail is transmitted via the Internet 14.
Transfer to the SP server 13 connected to the computer 11. In response to this, the SP server 13 stores an electronic mail addressed to the user of the computer 11.

Thereafter, the user of the computer 11 operates the computer 11 to activate the electronic mail application 11A, thereby accessing the SP server 13 via the public network 12 and requesting an electronic mail addressed to itself. Then, the SP server 13 sends the electronic mail addressed to the user to the computer 11 via the public network 12. The computer 11 receives the electronic mail from the SP server 13 and displays it on a display (not shown), for example. This allows the user to view (read) the email addressed to himself.

In addition, the user of the computer 11 operates the push button 17A of the telephone 17 to dial the telephone number dedicated to the voice conversion service, whereby the public network 1
When the SP server 13 is accessed via 2, a communication link is established between the SP server 13 and the telephone set 17. When the user inputs a command requesting an electronic mail addressed to himself by operating the push button 17A, a push button signal (push tone signal) (or dial pulse) corresponding to the operation is input.
From the telephone 17 through the public network 12 to the SP server 1
3 is sent.

When the SP server 13 receives the push button signal from the telephone set 17, the push button signal is analyzed (decoded). Then, when the push button signal is for requesting an electronic mail addressed to the user of the computer 11, the SP server 13 performs a voice synthesis process on the electronic mail addressed to the user to generate a synthesized sound, and the public network 12 is connected. Via the telephone set 17. As a result, the speaker 17B of the telephone set 17 outputs the synthesized voice read out from the electronic mail.

Therefore, the user can confirm the e-mail addressed to him / herself by the telephone 17 from the place where he / she goes.

Next, FIG. 3 shows a configuration example of the SP server 13 of FIG. The communication unit 21 uses the Internet 14
Via the public network 12
The communication control necessary for communicating with the computer 11 and the telephone 17 is performed. In addition to controlling the entire SP server 13, the command processing unit 22 analyzes the request from the computer 11 and the push button signal transmitted from the telephone 17 and performs processing corresponding to the analysis result. . Text mail storage unit 2
3 stores the e-mail sent to the user who contracts with the connection company having the SP server 13. The text mail storage unit 23 is provided with a storage area for storing an electronic mail addressed to a user who contracts with a connection provider (for each user, a storage area as described below is provided). The electronic mail addressed to each user is stored in the mailbox of the user under the control of the command processing unit 22.

The speech synthesizer 24 has the same structure as the speech synthesizer shown in FIG. 1 except that the mail header processing section 31 is newly provided. The mail header processing unit 31
Instruction information (speech command "\ speech" or non-speech command "\ mute") is added to a predetermined part of the email header (mail header), and output to the command recognition processing unit 1. ing.

In the SP server 13 configured as described above, when an electronic mail is sent from the user of the computer 15 to the user of the computer 11 via the Internet 14, the electronic mail is
The data is received by the communication unit 21, transferred to the text mail storage unit 23 and stored therein under the control of the command processing unit 22.

After that, when the user of the computer 11 operates the computer 11 to access the SP server 13 via the public network 12 and request an e-mail addressed to itself, the command processing section 22 causes the user to operate. The e-mail is read from the mail box (text mail storage unit 23) for use by the communication unit 21. As a result, the electronic mail is transmitted from the communication unit 21 to the computer 11 via the public network 12, and the user can see the electronic mail addressed to himself.

Further, the user of the telephone set 17 (here, also the user of the computer 11) accesses the SP server 13 via the public network 12 by the telephone set 17 and requests the electronic mail addressed to itself. When the push button 17A is operated, a push button signal corresponding to the operation is transmitted from the telephone set 17 to the SP server 13 and received by the communication unit 21. Command processing unit 2
In 2, the push button signal received by the communication unit 21 is decrypted, and an electronic mail addressed to the user of the telephone set 17 is read from the text mail storage unit 23 according to the decryption result, and is read by the voice synthesizer 24. Transferred.

Here, the electronic mail (input sentence) addressed to the user of the computer 11 has a voice command "\ s" at the beginning of a line to be subjected to voice synthesis processing, as shown in FIG.
"peech" is added.

Therefore, in this case, the electronic mail transferred from the command processing unit 22 to the voice synthesizer 24 is
If the command recognition processing unit 1 is directly supplied via the mail header processing unit 31, the voice command "\ speec
Since only the line with "h" added is subject to voice synthesis processing, the mail header, the face mark and figures described in the text, and the part such as the email address in the signature are considered to be the synthesized voice. As a result, it is possible to obtain a synthesized voice that makes it easy to understand the content of the electronic mail.

The addition of the voice command "\ speech" to the electronic mail shown in FIG. 4 is performed by the sender of the electronic mail (here, the user of the computer 15). speec
"h" can be added only to the body and signature of the email. That is, the user who is the sender of the e-mail cannot change the format of the e-mail header part without permission.
The voice command "\ speech" cannot be added.

However, the sender (sender) of the mail in the mail header (the part starting with From :), the title (the part starting with Subject :), the date (the part starting with Date :), etc. are communicated by electronic mail. It is important information above (hereinafter referred to as important information as appropriate), and it is not preferable to exclude such a part from the target of the voice synthesis processing.

Therefore, the mail header processing section 31 of the voice synthesizer 24 is designed to add the voice command "\ speech" to the beginning of the line in which the important information of the mail header is described.

That is, when an electronic mail is transferred from the command processing unit 22 to the voice synthesizer 24, the mail header processing unit 31 writes a line () in which important information is described from the mail header of the electronic mail. For example, Fro
Lines starting with m :, Subject :, and Date) are detected. Furthermore, in the mail header processing unit 31, the voice command "\ speec
"h" is added, so that the electronic mail shown in FIG. 4 is made as shown in FIG. Then, the electronic mail as shown in FIG. 5 is output to the command recognition processing unit 1, and thereafter, the same voice synthesis processing as that described in FIG. 1 is performed.

The synthesized sound obtained by the speech synthesizer 24 (synthesized sound output from the speech output unit 3) is supplied to the communication unit 21 via the command processing unit 22, and the communication unit 21
Then, the synthesized voice is transmitted to the telephone 17 via the public network 12.
Sent to. As a result, the speaker 17B of the telephone 17
From Fig. 5, the voice command "\ speec
Synthetic sound corresponding to the line in which "h" is described "Subject: mail no example Date: Wed, 6 Dec 1995 17:59:46 From: mail_sender@***.***.sony.co.jp Miyazaki @ Sony For each email, the email header, the face mark expressing the sender's emotions, the diagram made by combining the symbols as shown above, and the email sender's signature as shown at the bottom, etc. "Satoshi Miyazaki @ Sony Corporation" is output.

Therefore, in this case, the user of the computer 15 can ask the user of the computer 11 to listen only to the important part, and the user of the computer 11 can make unnecessary synthetic sounds heard. Will be released from. That is, accurate and quick communication can be achieved.

Next, in the above case, the voice command "\ speec" is added at the beginning of the line to be subjected to the voice synthesis processing.
Although the e-mail to which "h" is added is targeted, the e-mail to which the non-voice command "\ mute" is added can be processed in the same manner.

That is, when the electronic mail addressed to the user of the computer 11 is, for example, as shown in FIG.
In the mail header processing unit 31, a line in which important information is described (here, lines starting with From :, Subject :, and Date, respectively) is detected from the mail header of the electronic mail. Further, in the mail header processing unit 31, the non-vocalization command “\ mute” is added to the beginning of the lines other than the line in which the important information is described, whereby the electronic mail shown in FIG. As shown in. Then, the electronic mail as shown in FIG. 7 is output to the command recognition processing unit 1, and thereafter, the same voice synthesis processing as that described in FIG. 1 is performed.

Therefore, in this case, from the voice synthesizer 24, in FIG.
Synthetic sounds corresponding to the lines excluding the line in which `` e '' is written `` Subject: mail no example Date: Wed, 6 Dec 1995 17:59:46 From: mail_sender@***.***.sony.co.jp My name is Miyazaki@Sony.The mail header, the face mark that expresses the sender's emotions, and the figure made up of symbols are combined in one mail, and the mail sender is shown at the bottom. “Satoshi Miyazaki @ Sony Corporation” is output.

Although the speech synthesizer and the network system to which the present invention is applied have been described above, the present invention can be applied to all other systems using speech synthesis.

In this embodiment, the voiced command or the voiced non-voiced command is described at the beginning of the line, and the line-by-line instruction is given as to whether or not to perform voice synthesis.
It is possible to instruct whether to perform voice synthesis in any other unit.

Further, in the present embodiment, "\ speech" is used as the voice command, but other arbitrary notation (for example, "\ voice") is used as the voice command.
Or "\ speech synthesis") can be used. Similarly, as for the non-vocalization command, any notation other than "\ mute" (for example, "\ del" or "\ non-vocalization") can be used. However, it is preferable that the voice command and the non-voice command are described in a text file that is not described (or is less frequently described).

Further, in the present embodiment, only one of the voice command and the voice disable command is described in the text file. However, in the text file, for example, voice command and voice disable It is also possible to mix both commands. However, in this case, for the line in which neither the voice command nor the non-voice command is written,
It is necessary to determine in advance whether or not to be the target of the voice synthesis processing.

Although not particularly mentioned in this embodiment, the language for describing the text file is not particularly limited. That is, the text file may be, for example, Japanese, English, French, or a mixture of two or more languages.

Further, in the embodiment shown in FIG. 2, the case where the Internet 14 is used as the computer network has been described, but the present invention can be applied to a network system using other computer networks.

Further, in the embodiment of FIG. 2, the sender of the mail, the title, and the date in the mail header are used as the important information, but the important information can be set arbitrarily. That is, as important information, other than the sender of the mail, the title, and the date, for example, the transfer destination of the copy of the e-mail (the line starting with Cc :), etc. can be set, and all of the mail header is made important information. It is also possible. Furthermore, it is possible not to make all the mail headers important information.

Further, in the embodiment of FIG. 2, the speech synthesizer 24 is provided in the SP server 13, but the speech synthesizer 24 is provided in the computer 11 and the computer 1
It is also possible to make 1 read the received e-mail.

[0087]

According to the voice synthesizing method of the first aspect and the voice synthesizing apparatus of the fourth aspect, the predetermined instruction information included in the input sentence is recognized, and the input is made in correspondence with the instruction information. From the sentence, a part to be voice-synthesized is extracted to generate a synthetic voice. Therefore, it is possible to obtain a synthesized voice that is easy to understand for a necessary portion.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of a speech synthesizer to which the present invention is applied.

FIG. 2 is a diagram showing a configuration of an embodiment of a network system to which the present invention is applied.

FIG. 3 is a block diagram showing a configuration example of an SP server 13 of FIG.

FIG. 4 is a diagram showing an electronic mail.

FIG. 5 is a diagram showing an electronic mail.

FIG. 6 is a diagram showing an electronic mail.

FIG. 7 is a diagram showing an electronic mail.

FIG. 8 is a block diagram showing a configuration of an example of a conventional speech synthesizer.

FIG. 9 is a diagram showing an electronic mail.

[Explanation of symbols]

1 command recognition processor, 2 text-to-speech synthesizer,
3 language processing unit, 4 voice synthesis unit, 5 voice output unit, 6 speaker, 11 computer, 12 public network, 13 SP server, 14 Internet,
15 computer, 16 SP server, 17
Telephone, 17A push button, 17B speaker, 21 communication unit, 22 command processing unit, 23
Text mail storage unit, 31 Mail header processing unit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location H04L 12/54 9466-5K H04L 11/20 101B 12/58

Claims (6)

[Claims] 1. A speech synthesis method for generating a synthetic sound corresponding to an input sentence, comprising recognizing predetermined instruction information included in the input sentence, and corresponding to the instruction information, a voice from the input sentence. A voice synthesizing method characterized in that a part to be synthesized is extracted and only a synthesized voice corresponding to the part to be synthesized is generated. 2. The voice synthesizing method according to claim 1, wherein the instruction information indicates a portion of the input sentence that is to be voice-synthesized or a portion that is not voice-synthesized. 3. The voice synthesizing method according to claim 1, wherein the instruction information indicates, for each line, a portion of the input sentence that is to be voice-synthesized or a portion that is not to be voice-synthesized. 4. A voice synthesis device for generating a synthetic sound corresponding to an input sentence, comprising: a recognition unit that recognizes predetermined instruction information included in the input sentence; and a recognition unit that recognizes the instruction information recognized by the recognition unit. Correspondingly, it is provided with extraction means for extracting a portion to be speech-synthesized from the input sentence, and speech synthesis means for generating a synthesized sound corresponding to the portion to be speech-synthesized extracted by the extraction means. Speech synthesizer. 5. The speech synthesis apparatus according to claim 4, further comprising an addition unit that adds the instruction information to a predetermined portion of the input sentence. 6. The input sentence is an electronic mail transmitted via a computer network, and the adding means adds the instruction information to a predetermined portion of a header of the electronic mail. The speech synthesizer according to claim 5.