JP2002108378A - Document reading-aloud device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a document reading-aloud device which actualizes the improvement of listener's amusement and convenience by changing sound quality, document contents, intonation, and a voice synthesizing engine according to personal information on a document creator, area information using the reading- aloud device, etc. SOLUTION: The document reading-aloud device which converts a text document into a voice and reads an electronic document aloud is equipped with a creator information extraction part 3 which extracts at least one of pieces of information representing the creator of the object document to be read aloud, the sex, the hometown, address, and the language of the document, a sound quality determination part 4 which determines at least one of the pitch, strength, and vocalizing speed of a synthesized voice and a synthesized voice dictionary when the object document 2 is read aloud, a voice synthesizing engine 1 which generates the synthesized voice of the object document 2 according to the sound quality determined by the sound quantity determination part 4, and a synthesized voice output part 5 outputs the synthesized voice of the voice synthesizing engine 1 to a speaker 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a text-to-speech apparatus that converts a digitized text document into voice and reads it out.

[0002]

2. Description of the Related Art An electronic document reading apparatus using a text-to-speech synthesis technique has attracted attention as a function to be implemented in a personal computer. According to this device, the user can confirm the content of the electronic document by voice. For this reason, it is possible to directly check the information that had to be checked on the display directly from the ear, and the use form of the user's personal computer can be expanded. For example, if this function is realized by a mobile computer, the user can check the contents of the electronic mail only by wearing the inner headphones to the ears without expanding the display outdoors.

[0003]

However, in general, in a device for reproducing a text-synthesized sound, the inflection of the read-out voice is not sufficient, and the read-out voice is always read in the same sound quality, which causes a loss of the entertainment provided to the listener. I was Furthermore, if a text in a language (for example, English) that is not compatible with the synthesis engine is read aloud, the reading is erroneously performed, so that there is a problem that convenience for a listener is impaired.

[0004] As means for enhancing the entertainment, the creator of the document uses a sound quality editing work system of the synthesized voice, and the sound quality of the synthesized voice (the pitch of the sound, the speech speed, the strength of the sound), the intonation, or a part of the speech content. There is a way to edit. For example, the sound quality editing work system can enhance the sound quality to make it feminine, add inflections to add emotion to sentences, and change the utterance content to speak a dialect to make the listeners entertain It is possible to add sex. However, in the current text-to-speech apparatus, it is necessary for the listener himself to change the sound quality of each text to be read. Such a complicated operation significantly lowers the convenience that the document reading device should originally have for the listener, and thus cannot be a general means for improving the entertainment of the listener.

Further, in general, in a device that reads out a text-synthesized sound, it is not possible to determine the language (for example, English) of the text-to-speech document and change the text-to-speech synthesis engine. For example, when a sentence written in English is read out on a Japanese-language synthesis engine, the reading is not performed correctly, which causes a decrease in the convenience of the listener.

As described above, the existing text-to-speech apparatus has a problem that the tone of the text-to-speech does not change and the entertainment for the listener is significantly impaired. Further, even if the document is in English, the Japanese synthesis engine is applied as it is, and thus there is a problem that the reading is not performed correctly.

[0007] An object of the present invention is to provide, for example, any one of the following information: personal information of a document creator, regional information using a reading device, and personal information of a listener of a document.
It is an object of the present invention to provide a text-to-speech apparatus that changes the sound quality, the content of a document, the inflection, and the speech synthesis engine, thereby improving the entertainment and convenience of the listener.

[0008]

According to a first aspect of the present invention, there is provided a document reading apparatus for converting a text document into voice and reading an electronic document.
For the document to be read out, the creator, gender, hometown, address, creator information extracting means for extracting at least one of information indicating the language of the document, and information obtained from the creator information extracting means, Sound quality determining means for determining at least one of the pitch, strength, utterance speed, and utterance speed of the sound in the synthesized speech when reading the document to be read out, and the read-out target based on the sound quality determined by the sound quality determining means. It has a speech synthesis engine for generating a synthesized speech of a document, and synthesized speech output means for outputting a synthesized speech of the speech synthesis engine to a speech device.

According to a second aspect of the present invention, in a document reading device that converts a text document into a voice and reads an electronic document, a correspondence between a specific word and a word when the word is actually read is shown. A word conversion database, a document conversion unit that converts and outputs the read target document by replacing a conversion target word in the read target document using the word conversion database, and a document conversion unit. A speech synthesis engine for generating a synthesized speech of an output document, and synthesized speech output means for outputting a synthesized speech of the speech synthesis engine to an audio device. According to a third aspect of the present invention, in the document reading device according to the second aspect, at least the address or the place of origin of the document creator among the information of the creator, gender, address, hometown, and language of the document to be read. Creator information extracting means for extracting information including information, wherein the document conversion means uses the word conversion database to convert a word to be converted in the text to be read out from the origin or address of the document creator. The read-out target document is converted and output by replacing the read-out target document according to information. According to a fourth aspect of the present invention, in the document reading apparatus according to the second or third aspect, a transliteration data storage for storing a word to be read out according to the address of the document creator or the information on the place of origin and voice intonation added thereto. Means for converting the intonation of synthesized speech with reference to the intonation conversion data storage means. The invention according to claim 5 is the document reading device according to claim 2, further comprising listener information extracting means for extracting at least the address or hometown information of a person who listens to the reading document, and wherein the document converting means comprises Using a word conversion database, the word to be converted in the read target document, by replacing the word according to the birthplace or address information of the person who listens to the document, the read target document is converted and output. I do.
The invention according to claim 6 is the document reading device according to claim 5, further comprising intonation conversion data storage means for storing words to be read according to the address information of the document listener and voice intonation information added thereto. The speech synthesis engine has a function of converting the intonation of synthesized speech with reference to the intonation conversion data storage means.

According to a seventh aspect of the present invention, in a document reading apparatus for converting a text document into a voice and reading an electronic document, information indicating a creator, a gender, an address, a place of origin, and a language of the document with respect to the reading target document. Creator information extracting means for extracting at least one of the above information, and selecting an appropriate speech synthesis engine from a plurality of speech synthesis engines in accordance with the information obtained by the creator information extraction means to generate a synthesized speech. And a synthesized sound output unit that outputs a synthesized sound output from the sound synthesis engine group to a sound device.

According to an eighth aspect of the present invention, there is provided a computer-readable recording medium storing a program to be executed by the document reading apparatus according to any one of the first to seventh aspects.

In the above configuration, the invention according to the first aspect has a function of extracting personal information of a document creator and changing the sound quality of the text-to-speech based on the extracted information. Therefore, for example, when the invention apparatus is used for reading out a mail document, it is possible to perform an operation of lowering the voice of a male mail and raising the voice of a female mail. Such an operation makes it possible to improve the entertainment on the listener's side of the synthesized speech.

[0013] The invention according to claim 2 has a function of converting a specific word contained in the read-out document into another word. For example, "thank you"
Is converted to the word "large". For this reason, it is possible for the listener of the synthesized voice to read a word using a dialect that is in the area where the present apparatus is used, and it is possible to improve the entertainment of the listener. According to the third aspect of the present invention, the personal information of the document creator is extracted in the second aspect of the invention,
It has a function to partially change the created document based on the information. For this reason, for example, it is possible to change the created document according to the dialect of the hometown of the document creator and perform an operation of reading out the created document. Through such an operation, it is possible to improve the entertainment on the listener side of the synthesized voice. According to a fourth aspect of the present invention, in the third aspect of the invention, there is provided a function of changing the inflection of reading aloud based on personal information of a document creator. Therefore, for example, it is possible to change the inflection of the reading aloud according to the dialect of the hometown of the document creator. By such an operation, it is possible to further improve the entertainment on the listener side of the synthesized voice.

The invention according to claim 5 has a function of extracting personal information of a listener of a document and partially changing a prepared document based on the information. For this reason, for example, it is possible to perform an operation of changing the prepared document and reading it out according to the dialect of the listener's hometown. Such an operation makes it possible to improve the entertainment on the listener's side of the synthesized speech. According to a sixth aspect of the present invention, in the fifth aspect of the invention, there is provided a function of changing the inflection of reading aloud based on the personal information of the listener. For this reason,
For example, it is possible to change the inflection of reading aloud according to the dialect of the listener's hometown. By such an operation, it is possible to further improve the entertainment on the listener side of the synthesized voice.

The invention according to claim 7 has a function of extracting personal information of a document creator and changing a synthesis engine based on the information. For this reason, for example, it is possible to change the speech synthesis engine according to the dialect of the listener's hometown. Such an operation makes it possible to improve the entertainment on the listener's side of the synthesized speech. In addition, when the written language of the document creator is a foreign language (for example, English), it is possible to switch to a synthetic speech engine corresponding to the language and use it. Such an operation also makes it possible to improve the convenience of the listening side.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention include a first mode for changing the sound quality, a second mode for changing the document content, a third mode for changing the inflection of the document content and reading, and a voice synthesis engine. It can be classified into the fourth form. The first form is Claim 1, the second form is Claim 2, the third and fifth forms, the third form is Claims 4 and 6, and the fourth form is Claim 7. Term.

[Embodiment 1] FIG. 1 is a block diagram of a first embodiment (an embodiment in which sound quality is changed) of the present invention. In the present embodiment, an electronic mail reading device will be described. The reading target document 2 is the text of the electronic mail received by the electronic mail software. The creator information extraction unit 3 extracts personal information of the sender of the electronic mail. The name of the sender can be automatically extracted from the header file of the electronic mail. The address information of the sender can be automatically extracted by referring to the signature part of the electronic mail, or can be input manually by the user of the electronic mail in advance in the address book. Other information, for example, gender information, can also be extracted by storing information manually input by the user of the electronic mail in the address book in advance.

The sound quality determining unit 4 determines the pitch of the sound, the speech speed, the sound intensity, and the voice dictionary from the information of the creator information extracting unit 3. For example, based on the gender information, the voice dictionary is switched to the male in the case of the male, and the voice dictionary is switched to the female in the case of the female. Also, the creator information storage unit 3 may accumulate the set values of the sound pitch, speech speed, sound intensity, and sound dictionary manually input by the user in advance in the creator information storage unit 3 and reflect these. It is possible.

The speech synthesis engine 1 sets various information at the time of generating the synthesized speech to the pitch of the sound, the speech speed, the sound intensity, and the speech dictionary determined by the sound quality determination section 4,
The document 2 to be read out is converted into synthetic sound data. The synthesized sound output unit 5 converts the converted synthesized sound data into an audio signal and outputs the audio signal. A speaker 6 as an audio device generates an acoustic signal based on the audio signal.

In this embodiment and the following embodiments, each component that performs data processing such as a document conversion unit and a speech synthesis engine includes one or a plurality of computers (central processing units) and programs executed by the computers. The program can be recorded on a computer-readable recording medium or distributed via a communication line.

[Embodiment 2] FIGS. 2, 3 and 5 are block diagrams showing the structure of the second embodiment (form for changing the contents of a document) of the present invention. Comparing the configurations shown in FIGS. 2, 3, and 5, the configuration shown in FIG. 3 is obtained by adding the creator information extracting unit 3 to the configuration shown in FIG. 2, and the configuration shown in FIG. 5 is shown in FIG. A listener information extracting unit 9 is provided in place of the creator information extracting unit 3. Here, as an embodiment in the case of changing the contents of a document, an electronic mail reading device using the block diagram of FIG. 3 will be mainly described. In the second embodiment and the following embodiments, the same reference numerals are given to components having the same basic function in each embodiment including the first embodiment.

In FIG. 3, the reading target document 2 is the text of the electronic mail received by the electronic mail software. The creator information extraction unit 3 extracts personal information of the sender of the electronic mail. The name of the sender can be automatically extracted from the header file of the electronic mail. The address information of the sender can be automatically extracted by referring to the signature part of the electronic mail, or can be input manually by the user of the electronic mail in advance in the address book. Other information, such as the birthplace, place of birth, and gender of the document creator, can also be extracted by storing information manually input in advance in the address book by the user of the electronic mail.

The word conversion database 7 is a database for replacing a corresponding word in the reading target document 2. FIG. 7 shows an example of the word conversion database. As shown in this figure, dialects of each region are assigned to a word in the word conversion database.

The document conversion section 8 is based on the information extracted by the creator information extraction section 3 or information set in advance, and the place of origin and location of the word included in the document 2 to be read aloud. Convert words that can be converted to local dialect. For example, if the document to be read is "Thank you for your mail today." Is converted.

The speech synthesis engine 1 reads out the document converted by the document conversion section 8 and converts it into synthesized sound data.
The synthesized sound output unit 5 converts the converted synthesized sound data into an audio signal and outputs the audio signal. A speaker 6 generates an acoustic signal based on the audio signal.

The reading device shown in FIG.
In the above description, the same operation as above is performed except that the creator information extraction unit 3 is not connected in the above description. In this case, for example, it is effective to use a document reading device using an animation character. By converting the reading target document 2 into a spoken word of a specific animation character and reading it, the individuality of the animation can be emphasized, and the entertainment on the listener side can be improved.

Further, the reading device shown in FIG.
In the above description, instead of using the creator's area information extracted from the creator information extraction unit 3 in the above description, the listening side of the reading target 2 extracted by the listener information extraction unit 9 The document is replaced according to the dialect of the place of birth or location. The listener information extracting unit 9 extracts information used for document replacement by referring to the place of birth, place of birth and gender information from the personal information of the listener registered in advance by the user of the reading device. Output to the converter 8.

[Embodiment 3] FIGS. 4 and 6 are block diagrams of a third embodiment (an embodiment in which the contents of a document and the inflection of reading out are changed). An embodiment will be described as an electronic mail reading device using the block diagram of FIG. The third embodiment is the same as the second embodiment except that the intonation conversion data storage unit 10 is connected. Hereinafter, the intonation conversion data storage unit 10 will be described.

The intonation conversion data storage unit 10 stores the correspondence between a specific word and the intonation information of the voice to be added to the word. Here, the correspondence between a specific word and the intonation information of the voice to be added to the word is classified according to regional information or gender information, and is set so that it can be referred to for each classification. Then, the speech synthesis engine 1 performs document conversion based on the information extracted by the creator information extraction unit 3,
Furthermore, it refers to the correspondence described in the intonation conversion data storage unit 10 and adds intonation information to the reading target document. For example, in the intonation conversion data storage unit 10, a correspondence relationship of "Okini → Raise the ending" is indicated, and when the reading target document 2 is "Please give me a mail today", the speech synthesis engine 1 Then, a synthesized speech with the ending of "Ohokuni" raised is generated.

In the reading device of FIG. 6 (corresponding to the invention described in claim 6), instead of using the creator's local information extracted from the creator information extracting unit 3 in the above description, the listener The document is replaced in accordance with the dialect of the place of birth or location of the listening side to be read out, which is extracted by the information extracting unit 9.

[Embodiment 4] FIG. 7 is a block diagram of a fourth embodiment (an embodiment in which a speech synthesis engine is changed) of the present invention. An embodiment will be described as an electronic mail reading device using the block diagram of FIG.

The reading target document 2 is the text of the electronic mail received by the electronic mail software. The creator information extraction unit 3 extracts personal information of the sender of the electronic mail. The name of the sender can be automatically extracted from the header file of the electronic mail. The address information of the sender can be automatically extracted by referring to the signature part of the electronic mail, or can be input manually by the user of the electronic mail in advance in the address book. Furthermore, it is also possible to set up a mechanism for grasping what is written in the language of the text. Other information, for example, gender information, can also be extracted by accumulating manually input information manually in advance in the address book by the user of the electronic mail.

The speech synthesis engine group 1a is composed of a plurality of speech synthesis engines a, b, c,... Corresponding to respective languages such as Japanese and English. An appropriate speech synthesis engine is set based on the information of the creator information creating unit 3, and the reading target document 2 is read out. For example, when the text of the electronic mail is in English, it is possible to improve the convenience on the listener side of the synthesized voice by selecting a voice synthesis engine that supports English. When a plurality of speech synthesis engines a, b, c,... Corresponding to the same language (for example, Japanese) and having different synthesis parameters and the like are prepared in advance, a plurality of Japanese speech By switching and using the synthesis engine at random for each document creator, it is possible to give variations to the sound quality when reading out, and it is possible to improve the entertainment of the synthesized voice on the listener's side.

As described above, the embodiments of the present invention have been described with reference to the drawings. However, the embodiments of the present invention are not limited to those described above, and can be appropriately changed. For example, the reading target document 2 is not limited to the electronic mail body, but performs a predetermined conversion process even if the application information is document information attached to the electronic mail or information including character data as image information. Anything that can be converted into text information in this way can be similarly used. Also, the speech synthesis engine 1 and the other data processing units of the document conversion unit 8 and the synthesized sound output unit 5 or the word conversion database 7 and the intonation conversion data storage unit 10 are not necessarily one.
It does not need to be provided on one personal computer, and may be distributed and arranged via a wired and wireless network.

[0035]

As described above, according to the present invention,
Personal information of the document creator, the region where the reading device is used,
It is possible to output a synthesized speech by changing the sound quality, the content of the document, the intonation, and the speech synthesis engine in accordance with any of the personal information of the listener who reads out the document. With such a function, it is possible to provide a text-to-speech apparatus that realizes improvement in the entertainment and convenience of the listener.

In the first embodiment of the present invention, it is possible to read out a document by changing the sound quality for each document creator. For this reason, it is possible to improve the entertainment of the listener.

According to the second embodiment of the present invention, it is possible to change a part of the document content and output a synthesized speech. For example, in the case of a text-to-speech apparatus using an animated character, a part of words of the text to be read is changed so that the text is read aloud according to the personality of the character.
It is possible to improve the entertainment of the listener. In addition, it is possible to improve the listener's entertainment by speaking a dialect that matches the user of the text-to-speech device, the area where the text-to-speech device is used, and the place of origin or location of the document creator. Become.

In the third embodiment of the present invention, a function for adjusting the inflection of reading is further added to the second embodiment of the present invention. For this reason, it is possible to improve the entertainment of the listener.

In the fourth embodiment of the present invention, it is possible to read out a document by changing the speech synthesis engine for each document creator. For this reason, for example, when the read-out document is in English, it is possible to improve the convenience on the listening side by switching and using the speech synthesis engine that supports English. In addition, by using a plurality of Japanese-language speech synthesis engines, it is possible to give variations to the sound quality of each document creator, and it is possible to improve the entertainment on the listening side.

Further, by combining and using the inventions of the first to fourth embodiments, it is possible to enhance the effect of improving the convenience and recreational performance of the listener.

The reading device according to the present invention can be applied to, for example, dialect-compatible speech synthesis software, application software of a personal computer using an interface agent, and a game creation tool. Further, if the present invention is applied to an information providing terminal arranged at a local government office in a certain region, it is possible to create a device that provides voice guidance in a local dialect.

[Brief description of the drawings]

FIG. 1 is a diagram showing an embodiment of a reading device according to the present invention (the invention according to claim 1);

FIG. 2 is a diagram showing another embodiment of the reading device according to the present invention (the invention according to claim 2);

FIG. 3 is a diagram showing another embodiment of the reading device according to the present invention (the invention according to claim 3);

FIG. 4 is a diagram showing another embodiment of the reading device according to the present invention (the invention according to claim 4);

FIG. 5 is a diagram showing another embodiment of the reading device according to the present invention (the invention according to claim 5);

FIG. 6 is a diagram showing another embodiment of the reading device according to the present invention (the invention according to claim 6);

FIG. 7 is a diagram showing another embodiment of the reading device according to the present invention (the invention according to claim 7);

8 is a diagram showing an example of the word conversion database 8 shown in FIG. 3 and the like.

[Description of Signs] 1 ... Speech synthesis engine 1a ... Speech synthesis engine group 2 ... Document to be read out 3 ... Creator information extraction unit 4 ... Sound quality determination unit 5 ... Synthesis sound output unit 6 ... Speaker 7 ... Word conversion database 8 ... Document Conversion unit 9: Listener information extraction unit 10: Inflection conversion data storage unit

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Yasuto Kono 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo East Japan Nippon Telegraph and Telephone Corporation (72) Inventor Mio Hosoya Nishi-Shinjuku, Shinjuku-ku, Tokyo Tome Nippon Telegraph and Telephone Co., Ltd. (72) Inventor Tomoya Kosaka 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo F-term within Nippon Telegraph and Telephone Co., Ltd. AB04 AB26

Claims (8)

[Claims] 1. A document reading device for converting a text document into a voice and reading an electronic document, wherein at least one of information indicating a creator, a gender, a place of origin, an address, and a language of the document is read. Creator information extracting means to be extracted, and information obtained from the creator information extracting means, at least one of pitch, loudness, utterance speed, and synthetic speech dictionary of a synthesized speech when reading out the reading target document is read out. Sound quality determining means for determining; a voice synthesis engine for generating a synthesized voice of the reading target document based on the sound quality determined by the sound quality determining means; and a synthesized voice output for outputting a synthesized voice of the voice synthesis engine to a voice device. Means for reading out a document. 2. A text-to-speech apparatus which converts a text document into a voice and reads an electronic document, comprising: a word conversion database indicating a correspondence between a specific word and a word when the word is actually read aloud; A document conversion unit that converts and outputs the read target document by replacing a conversion target word in the read target document using the word conversion database; and A text-to-speech apparatus comprising: a speech synthesis engine for generating; and a synthesized speech output unit for outputting a synthesized speech of the speech synthesis engine to a speech device. 3. The document reading device according to claim 2, wherein the creator, gender, address,
Origin information, further comprising creator information extraction means for extracting information including at least the address of the creator of the document or information of the origin of the information of the language of the document, the document conversion means, using the word conversion database A text-to-speech apparatus characterized in that the text-to-speech target is converted and output by replacing words to be converted in the text-to-speech according to the origin and address information of the document creator. 4. The document reading device according to claim 2, further comprising an inflection conversion data storage unit that stores a word to be read according to the address or hometown information of the document creator and voice intonation information added thereto. The text-to-speech apparatus according to claim 1, wherein the speech synthesis engine has a function of converting intonation of synthesized speech with reference to the intonation conversion data storage means. 5. The document reading device according to claim 2, further comprising: listener information extracting means for extracting at least the address or birthplace information of a person who listens to the reading document, wherein the document converting means includes the word conversion database. By replacing the words to be converted in the read target document according to the birthplace and address information of the person who listens to the document, thereby converting and outputting the read target document. apparatus. 6. The text-to-speech engine according to claim 5, further comprising: intonation conversion data storage means for storing words to be read according to the address information of the document listener and voice intonation information added thereto. A document reading device having a function of converting the intonation of a synthesized voice with reference to the intonation conversion data storage means. 7. A text-to-speech apparatus for converting a text document into voice and reading an electronic document, wherein a creator, a gender, an address,
Creator information extraction means for extracting at least one of information of information indicating a place of origin and a language of a document; and appropriate speech from a plurality of speech synthesis engines according to the information obtained by the creator information extraction means. A document reading apparatus comprising: a speech synthesis engine group that selects a synthesis engine and generates a synthesized speech; and a synthesized speech output unit that outputs a synthesized speech output from the speech synthesis engine group to a speech device. 8. A computer-readable recording medium which records a program to be executed by the document reading device according to claim 1. Description: