JPH11175307A - Method and device for reading document aloud - Google Patents

Abstract

PROBLEM TO BE SOLVED: To discriminate between the body and quotations of a document into which the quotations are inserted when the document is read aloud by extracting the specific partial documents from the document and reading the extracted partial documents aloud by voice synthesis in a different voice from other document parts. SOLUTION: The document to be read aloud is inputted from a document input means 12. When a code detecting means 13 does not detect a code '>' indicating a quotation at the head of each line of the inputted document, a voice selecting means 16 selects, for example, a male voice as a 1st voice for reading the body aloud and the 1st voice selected by the voice selecting means 16 is synthesized to read the document aloud. The document continues to be read aloud until the code detecting means 13 detects a carriage-return code. When the code detecting means 13 detects the code '>', the voice selecting means 16 selects, for example, a female voice as a 2nd voice for reading the quotation aloud and the 2nd voice selected by the voice selecting means 16 is synthesized to read the document aloud.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for reading a document, and more particularly, to a method and apparatus for reading a document which can perform speech synthesis under different conditions in the same document and increase the recognition rate of the document.

[0002]

2. Description of the Related Art In recent years, a technique for reading a document described in an electronic file by voice synthesis has been widely used. For example, in a system that reads out e-mails by voice synthesis,
When an e-mail is received, a call is made to a telephone number set by the user in advance, and after the telephone line is connected, the received e-mail is voice-synthesized and read out to convey the contents of the e-mail to the user. Has also been done.

[0003] On the other hand, electronic mail differs from general mail in that it is communicated by mail or the like. For this reason, processing such as copying of a document is easy. Therefore, a part or all of a received document is included as a quotation. In many cases, such documents are used as reply mails.

FIG. 7 shows an example of a description of an electronic mail. FIG. 7A is a diagram showing an e-mail from the user B to the user A, and FIG. 7B is a diagram showing a reply from the user A to the user B.

[0005] As shown in FIG. 7 (b), the user A receives the reply to the user B (see FIG.
The lines from line 6 to line 6 are used as quotations, and the text and quotation are distinguished by adding “>” to the beginning of the line.

By the way, when an e-mail in which a quote is described as shown in FIG. 7B is read out by speech synthesis, before and after the quote (in FIG. 7B, the second to third lines, From the sixth line to the seventh line), the connection of sentences may be strange and the meaning may be unclear.

[0007] Further, as shown in FIG. 8, when a document in which Japanese sentences and English sentences are mixed is read out by speech synthesis, the Japanese sentences are read out correctly, and the English sentences are read as "S.U.M.M.R.Y."", And it may be difficult to understand the meaning by listening to the read text.

[0008]

As described above, when a document is read aloud by speech synthesis, if the read document is a document in which a quote is inserted or a document in which different languages are mixed. Has a problem that it is difficult to understand the meaning by listening to the read text.

Therefore, according to the present invention, when reading out a document in which a quotation is inserted, the text is read out so that the text and the quotation can be distinguished. It is an object of the present invention to provide a text-to-speech method and apparatus that can read aloud in response to the following.

[0010]

According to a first aspect of the present invention, there is provided a document reading method for reading a document described in an electronic file by speech synthesis. It is characterized in that the extracted partial document is synthesized by the above-mentioned speech synthesis with a different voice from other document portions, and is read aloud.

According to a second aspect of the present invention, in the first aspect of the present invention, a quotation is extracted from the text of the document, and the extracted quotation is subjected to speech synthesis different from that of the text.

According to a third aspect of the present invention, in the second aspect of the invention, the document is a document described in a received electronic mail, and a code indicating a quote is detected from the document, and the detection result is obtained. The cited text is extracted on the basis of the extracted text, and the extracted quoted text is subjected to voice synthesis with a voice tone different from that of the text.

According to a fourth aspect of the present invention, in the first aspect of the present invention, the document is extracted according to a language in which the document is described, and the extracted document is subjected to speech synthesis corresponding to a description language of the document. And read aloud.

According to a fifth aspect of the present invention, there is provided a document reading apparatus for reading a document described in an electronic file by voice synthesis, comprising: an extracting unit for extracting a partial document from the document; and a part extracted by the extracting unit. Voice synthesis means for synthesizing the document with a different voice from the other partial documents.

According to a sixth aspect of the present invention, in the fifth aspect of the present invention, the extracting means extracts a quotation from the text of the document, and the speech synthesizing means extracts the quotation from the text divided by the dividing means. It is characterized by synthesizing a quotation with a different voice.

According to a seventh aspect of the present invention, in the sixth aspect, the document is a document described in a received e-mail, and the dividing means detects a code indicating a quote from the document. The voice synthesizing unit synthesizes the text and the quoted text divided by the dividing unit with different voice colors.

According to an eighth aspect of the present invention, in the fifth aspect of the invention, the extracting means extracts the document corresponding to a language in which the document is described, and the speech synthesizing means outputs Speech synthesis corresponding to the description language of the extracted document is performed.

According to a ninth aspect of the present invention, a specific partial document is extracted from a document on a medium in which a document reading program for reading a document described in an electronic file by speech synthesis is recorded, and the extracted partial document is extracted. Is synthesized by the above-mentioned voice synthesis with a voice different from that of another document part, and is read aloud.

[0019]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of a document reading method and apparatus according to the present invention;

FIG. 1 is a block diagram showing the configuration of the document reading device.

In FIG. 1, a text-to-speech apparatus 10 includes a control unit 11 for controlling each unit, a document input unit 12 for inputting a text to be read, and a code detection unit 13 for detecting a code indicating the start or end of a quoted sentence. Control code detecting means 14 for detecting a line feed code or a code such as EOF (end of file) indicating the end of a sentence; voice synthesizing means 15 for performing voice synthesis to read a document; and voice color selecting means for selecting a voice color for performing voice synthesis. 16 is provided.

There is no limitation on the method of inserting a quotation in a document such as an electronic mail. However, a quotation is inserted in units of lines, and a mark such as ">" is added to the beginning of the quotation, or "" It is common to insert quotes around the quotes with "".

Referring now to FIG. 2, the operation of the text-to-speech apparatus 10 when reading a document in which a quote such as ">" is added at the beginning of a line and a quote is inserted will be described.

FIG. 2 is a flowchart showing the flow of the operation of the document reading device. When the document reading apparatus 10 starts operating (step 101), a document to be read is input from the document input unit 12. Next, the sign detecting means 13 adds a sign “>” indicating a quotation at the beginning of each line of the input document.
Is detected (step 102).
If the sign “>” is not detected (N in step 102)
O), the voice selecting means 16 selects, for example, a male voice as the first voice to read the text (step 103), and the voice synthesizing means 16 performs voice synthesis using the selected first voice and reads the document (step 103). Step 105). When reading a document,
The control is continued until the control code detecting means 14 detects a line feed code (NO in step 106), a line feed code is detected (YES in step 106), and if this is not the end of the document (NO in step 107), the code is detected again. Means 1
3 is detected (step 10).
2).

When the sign detecting means 13 detects a sign ">" (YES in step 102), the timbre selecting means 1
6 selects a female voice, for example, as the second voice to read the quotation (step 104), and the voice synthesizing means 16 performs voice synthesis using the selected second voice and reads the document (step 105). The reading of the document is continued until the control code detecting means 14 detects a line feed code in the same manner as in the case of reading in the first voice (NO in step 106).

As described above, the text to be read is detected for each line by detecting the sign ">". If the sign is not detected, the document is read in the first voice as the text. If the sign is detected, the second sentence is referred to as the quotation. When the text-to-speech is read out to the end of the sentence (YES in step 107), the text-to-speech apparatus 10 ends the text-to-speech (step 108).

By the way, in an electronic mail or the like, a received mail is often used as a quote, but a document in which a quote is inserted is quoted, that is, a quoted sentence is quoted twice. In some cases.

FIG. 3 is a flow chart showing the flow of the operation of the document reading device when reading a double-cited document. When the text-to-speech apparatus 10 starts operating (step 201) and a text to be read is input from the document input means 12, the code detection means 13 adds a symbol ">" indicating a citation to the beginning of the input document. Is detected (steps 202 and 204), but as a result of the detection, if the sign “>” is not added (step 20).
(NO in step 2), the voice selecting means 16 selects, for example, a male voice as the first voice to read the text (step 20).
3) If one sign “>” has been added (YES in step 202, NO in step 204), for example, a female voice is selected as the second voice that the voice color selecting means 16 reads out the quotation. (Step 205), when two symbols “>” are added and the result is “>>” (YES in Step 202, YES in Step 202), the voice selecting means 16 reads the double quote. For example, a child's voice is selected as the third voice to be raised (step 206).

When the voice to read the document is selected, the voice synthesizing means 16 synthesizes the voice with the selected voice and reads the document (step 207), and continues until the control code detecting means 14 detects the line feed code. (Step 208
NO).

This reading of the document is repeated until the end of the document (NO in step 209), and when the reading is performed to the end of the sentence (YES in step 209), the document reading apparatus 10
Ends the operation (step 210).

Next, with reference to FIG. 4, the operation of the text-to-speech apparatus 10 when reading a document in which a quote enclosed between "" and "" is inserted will be described.

FIG. 4 is a flow chart showing the flow of the operation of the document reading aloud device when reading out a document in which a quote enclosed by a code is inserted. When the text-to-speech apparatus 10 starts operation (step 301) and a document to be read is input from the document input means 12, first, the voice selection means 16 selects the first voice as the voice to read the text (step 301). 302).

Next, the code ""
And """(N in step 303).
O), the voice synthesizing unit 15 reads out the document in the first voice (step 308). The reading of the document is continued until the control code detecting means 14 detects the EOF indicating the end of the sentence (end of file) (NO in step 309).

Here, the sign detecting means 13 detects the sign "" or "" (YES in step 303), and if the sign is "" (YES in step 304), the starting position of the quote And the second voice is selected as the voice to which the voice selecting means 16 reads the quotation (step 30).
5) The voice synthesizing unit 15 reads out the document in the second voice (step 308).

The reading of the document in the second voice is continued (NO in steps 308 and 309), and the code detecting means 13 detects the code "" or "" again (Y in step 303).
ES), if the sign is "" (NO in step 304), it is determined that the quotation is at the end and the voice color selecting means 16
Selects the first voice as the voice to read the headquarters (step 307), and the voice synthesizing unit 15 reads the document in the first voice (step 308).

In this manner, the reading of the document is continued,
The control code detecting means 14 indicates the end of the sentence (end of file) E
When OF is detected (YES in step 309), document reading-aloud device 10 ends the operation (step 310).

Although the description is omitted here, as in the case of a double quote in which a symbol ">>" is added at the beginning of a line, "["
Within the quotes enclosed by "and"", the double quotes enclosed by""and""can be read out in a different voice.

In this embodiment, reference numerals indicating quotes are described as ">", "", "". However, even when another code is used in the same manner, the voice is changed. Can read aloud a document.

When reading out an e-mail sent or received within a specific group or the like, the text-to-speech apparatus 10 is provided with a voice registering means (not shown) to register the voices of the users in the group. The document can be read aloud in the voice of the user who made the description.

Next, a description will be given of a second embodiment of the document reading method and apparatus according to the present invention.

FIG. 5 is a block diagram showing the configuration of the document reading device according to the second embodiment. In FIG. 5, a document reading aloud device 20 includes a control unit 21 for controlling each unit,
A document input unit 22 for inputting a document to be read, a language information detecting unit 23 for detecting information indicating a language describing the document,
EOF (end of file) indicating line feed code or end of sentence
e) a control code detecting means 24 for detecting codes, a voice synthesizing means 25 for performing voice synthesis and reading a document, and a language selecting means 26 for selecting a language for performing voice synthesis.

In the document reading device 20, the language information detecting means 23 detects in which language the document described in the electronic file is described, and reads the document corresponding to each language. For example, in the case of an electronic mail, the linguistic information detecting means
Multilingual support is provided by a standard called E.
By determining the tag information of the IME, it is determined in which language the document is described.

Here, with reference to FIG.
The operation of 0 will be described. FIG. 6 is a flowchart showing the operation flow of the document reading device 20.

When the text-to-speech apparatus 20 starts operating (step 401) and a text to be read is input from the document input means 22, the language selection means 26 selects an initially set language, for example, Japanese. (Step 402). Next, if the linguistic information detecting means 23 detects linguistic information (YES in step 403), the language selecting means 26 selects a language based on the detected information (step 404). If no information is detected (NO in step 403), the voice synthesizing means 2 uses the language as it is.
5 performs speech synthesis in the language selected by the language selecting means 26 and reads out the document (step 405).

The change of the language and the reading of the document in steps 403 to 405 are repeated until the control code detecting unit 24 detects the EOF indicating the end of the sentence (end of file) (NO in step 406), and the EOF is detected. (YES in step 406), the document reading apparatus 20 ends the operation (step 407).

In the first embodiment, a text-to-speech apparatus which reads out a text and a quotation in different voices is used. In the second embodiment, a text-to-speech document is read out by performing speech synthesis corresponding to each language for each language. Although the text-to-speech device has been described, it should be added that both may be configured to be realized by the same device.

The document reading apparatus in the first embodiment and the second embodiment are both constituted by a memory or a CPU of a computer, and are realized by operating a program for performing the above operation. You can also.

[0048]

As described above, according to the present invention, a code indicating a quotation is detected from a document to be read, and based on the detection result, the quotation is speech-synthesized with a voice tone different from that of the text and read. Since the user is able to distinguish between the text and the quoted text, the recognition rate of the document is improved.

Further, since the language in which the document to be read is described is determined and the document is read out by speech synthesis corresponding to the language, the recognition rate of the document described in a plurality of languages can be increased. .

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a document reading device.

FIG. 2 is a flowchart showing the flow of the operation of the document reading device.

FIG. 3 is a flowchart showing the flow of the operation of the document reading device when reading a double-cited document.

FIG. 4 is a flowchart showing the flow of the operation of the document reading device when reading a document in which a quote enclosed by a code is inserted.

FIG. 5 is a block diagram illustrating a configuration of a document reading device according to a second embodiment.

FIG. 6 is a flowchart showing the operation flow of the document reading apparatus according to the second embodiment.

FIG. 7 is a diagram showing a description example of an electronic mail.

FIG. 8 is a diagram showing an example of a document in which Japanese sentences and English sentences are mixed.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Document reading-out device 11 Control means 12 Document input means 13 Code detection means 13 14 Control code detection means 14 15 Voice synthesis means 16 Voice color selection means 20 Document reading-out apparatus 21 Control means 22 Document input means 23 Language information detection means 24 Control code detection Means 25 Speech synthesis means 26 Language selection means

Claims (9)

[Claims] 1. A document reading method for reading a document described in an electronic file by speech synthesis, wherein a specific partial document is extracted from the document, and the extracted partial document is different from another document portion by the speech synthesis. A text-to-speech method characterized by speech synthesis. 2. The method according to claim 1, further comprising extracting a quote from the text of the document, and performing a speech synthesis on the extracted quote that is different from that of the text. 3. The document is a document described in a received e-mail. A code indicating a quote is detected from the document, the quote is extracted based on the detection result, and the extracted quote is extracted. 3. The method according to claim 2, wherein speech synthesis is performed on the sentence in a voice different from that of the text. 4. The method according to claim 1, wherein the document is extracted according to a language in which the document is described, and the extracted document is read out by performing speech synthesis corresponding to a description language of the document. How to read documents. 5. A text-to-speech apparatus for reading a document described in an electronic file by speech synthesis, comprising: extracting means for extracting a partial document from the document; and extracting the partial document extracted by the extracting means with another partial document. A text-to-speech apparatus comprising: voice synthesis means for synthesizing with different voices. 6. The method according to claim 1, wherein the extracting means extracts a quote from the text of the document, and the speech synthesis means synthesizes the text and the quote divided by the dividing means with different voices. The document reading device according to claim 5, wherein 7. The document according to claim 1, wherein the document is a document described in a received e-mail, the division unit includes a code detection unit that detects a code indicating a quotation from the document, 7. The text-to-speech apparatus according to claim 6, wherein the text and the quoted text divided by the dividing means are synthesized in different voices. 8. The extracting unit extracts the document corresponding to a language in which the document is described, and the speech synthesizing unit performs speech synthesis corresponding to a description language of the document extracted by the extracting unit. 6. The document reading device according to claim 5, wherein: 9. A medium in which a document reading program for reading a document described in an electronic file by voice synthesis is recorded, and a specific partial document is extracted from the document, and the extracted partial document is converted to another document by the voice synthesis. A medium storing a document reading program characterized by being read out by synthesizing with a different voice from the document part.