JPS63249261A - Voice type proofreading supporting system - Google Patents

Abstract

PURPOSE:To proofread an expressed Japanese-language sentence by providing a KANJI(Chinese character) dictionary where both of phonetical reading and reading in Japanese pronunciation are registered for each KANJI and a means which detects a position recognized as a proper noun and reading out each KANJI of a proper noun phonetically as well as its Japanese pronunciation with a synthesized voice. CONSTITUTION:A KANJI dictionary 5a is provided where phonetical reading and reading in Japanese pronunciation are registered for each KANJI. When the mode in which proper nouns are detected from words is set by a mode witching means 6, the KANJI dictionary 5a is used to read out Japanese- language sentences, where each KANJI of a part of the detected proper noun is read phonetically as well as in its Japanese pronunciation, with the synthesized voice, thereby proofreading the Japanese-language sentence. Thus, a word dictionary is not provided with a registration space neither a space of 'reading' for reading conversion having a problem in register work but only the word dictionary where phonetical reading and reading in Japanese pronunciation are registered for each KANJI is provided to cope with any proper noun.

Description

[Detailed description of the invention] [Summary] Analyzes written Japanese sentences and divides them into appropriate words,
Each one is given a reading and is read aloud with a synthesized voice,
This audio proofreading support device supports the proofreading of the Japanese text, and each kanji has an on-yomi reading.

By providing a kanji dictionary that registers both kun readings and a means for detecting parts that are recognized as proper nouns by switching modes, it is possible to use the kanji dictionary for the detected proper noun. This system reads a Japanese sentence with an on-yomi and a kun-yomi for each kanji using a synthesized voice, and proofreads the written Japanese sentence.

[Industrial application field]

The present invention relates to an audio proofreading support device that analyzes a written Japanese sentence, divides it into appropriate words, attaches a pronunciation to each word, reads them out in a synthesized voice, and supports the proofreading of the Japanese sentence. Especially regarding its calibration support method.

With the recent advancements in computer systems, the computer systems have mechanized to some extent the creation of documents within companies, the creation of documents such as articles in newspapers and publishing industries, 9 editing, 1 page layout, etc. However, the proofreading work of documents and articles still requires many hands, and there is a desire to mechanize the proofreading work.

Currently, in the proofreading work of articles in newspapers etc., for example, two people read and proofread together.
One person reads the manuscript aloud, while the other checks the galley document using a red pen.

For example, as a mechanized means representing the above reader side,
There are voice-based proofreading support devices, etc., which allow documents created manually to be stored in a document file on a computer system using a keyboard (KB), and then retrieved and read out using a synthesized voice for proofreading. It's starting to be done.

In this case, one of the characteristics of Japanese is that with certain kanji,
There are proper nouns that are constructed, and an accurate reading method for the proper nouns is required.

[Prior art and problems to be solved by the invention] FIG.
) shows an example of the configuration of a voice-based proofreading support device, and (b) shows an example of a word dictionary.

First, a Japanese document is input from the keyboard (KB) 7b of the terminal 7 and registered in the document file 1a.

In this way, the input document is read out from the document file la based on instructions from the operator and sent to the document analysis section 2.

In the document analysis section 2, for example, the applicant of the present application uses the "Sentence Analysis Method" published in Japanese Patent Application Laid-open No. 59-90167.
The analysis is performed using the document analysis method disclosed in .

Generally speaking, (b) the word notation and reading shown in the figure.

A word dictionary 2a storing accents, grammar, etc. is searched, matching words are extracted, and the words are divided into the most appropriate word strings as Japanese sentences.

For the divided words, refer to the word dictionary 2a and enter "
"Yomi" is added and sent to the next speech synthesis section 3.

The speech synthesis section 3 extracts the corresponding phoneme from the phoneme file 3a, performs speech synthesis using, for example, a well-known "rule synthesis method" and sends it out from the speaker 8.

In such a conventional speech-based proofreading support device, if the divided word is a proper noun, the "reading" shown in parentheses is used as the "reading" for the proper noun, as shown in Figure (b). ” is registered, and the “replacement” of the corresponding “reading” is “
He took out the "Yomi" and provided support for proofreading.

For example, the proper noun ``Kato'' was dealt with by adding the alternative reading ``Daisuketsu Kato.''

With this conventional method, if the number of proper nouns is small, it can be handled, but as the proofreading work for documents becomes more mechanized and the number of proper nouns that become a problem in the proofreading work increases, the word In the dictionary 2a, there is a problem that the registration space for the rereading becomes large and the registration work becomes difficult and impractical.

In view of the above-mentioned conventional drawbacks, the present invention uses a voice-based proofreading support system.
Focusing on the fact that there is a kun-yomi, we will write on-yomi readings for each character that makes up proper nouns.

The aim is to provide a practical audio-based proofreading support method by reading out both the kun-yomi and kun-yomi.

[Means for solving problems]

FIG. 1 is a diagram showing an example of the configuration of the audio-based proofreading support system of the present invention.

In the present invention, - A written Japanese sentence is analyzed based on the word dictionary 2a, divided into words, and a reading is added to each word, which is read aloud with a synthesized voice to support proofreading of the Japanese sentence. In the speech-based proofreading support device, a kanji dictionary 5a is provided in which one kanji has both on-yomi and kun-yomi, and when the mode switching means 6 is set to a mode for detecting proper nouns from the word, the detected proper noun is detected. Using the one-kanji dictionary 5a, a Japanese sentence is generated by giving an on-yomi and a kun-yomi to each kanji, and a synthesized voice reads out the part of the proper noun, and proofreads the Japanese sentence. Proofread.

[Effect]

That is, according to the present invention, a voice-based proofreading system analyzes a written Japanese sentence, divides it into appropriate words, adds a reading to each word, and reads out the words with a synthesized voice to support the proofreading of the Japanese sentence. The support device is equipped with a kanji dictionary that registers both on-yomi and kun-yomi for each kanji, and a means for detecting locations that are recognized as proper nouns by switching modes. Using a single kanji dictionary, a Japanese sentence with on-yomi and kun-yomi given for each kanji is read out in synthesized voice, and the written Japanese sentence is proofread. , registration space, and "reading" space for transliteration, which is problematic in the registration process, by simply providing a one-word dictionary with on-yomi and kun-yomi for each kanji character. This has the effect of realizing a speech-based proofreading support device that can also handle proper nouns.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

The above-mentioned FIG. 1 is a diagram showing an example of the configuration of the audio-based proofreading support system of the present invention, and FIG. 2 is a diagram showing an example of the configuration of a kanji dictionary. , proper noun detection unit 4. Onkun reading batting club 5. The one kanji dictionary 5a is a necessary means to implement the present invention. Note that the same reference numerals indicate the same objects throughout the figures.

Hereinafter, the audio-based proofreading support system of the present invention will be explained with reference to FIGS. 1 and 2.

Even if the present invention is implemented, the basic operation of the voice-based proofreading support system will not change in particular, so it will be omitted, and here, the explanation will focus on the reading method for proper nouns of the present invention.

First, in the normal mode, the main control unit 1°→text analysis unit 2. →Mode switching unit 6. →Speech synthesis section 3. →The same calibration support operation as before is performed on the route of the speaker 8.

Here, when the operator presses a specific key on the keyboard (KB) 7b to switch the mode to a specific mode, the words divided into six by the sentence analysis section 2 are transferred to the next proper noun detection section 4. , the grammar definition in the word dictionary 2a mentioned above is referred to, and the word is a person's name.

When it is recognized that it is a proper noun such as a place name, -, etc., the onkun-yomi typing section 5 refers to the one-kanji dictionary 5a provided for carrying out the present invention.

As shown in Figure 2, in the Kanji dictionary 5a, the on-yomi and kun-yomi readings are specified for each kanji, so for each kanji that makes up the proper noun, On-yomi and kun-yomi are given and sent to the speech synthesis section via the mode switching section 6.

For example, if "appetizer" is recognized as a person's name, "ka"
It is divided into each kanji for ``east'', and for each, the kanji dictionary 5a is referred to, and
Like 'Higashi', On-yomi and Kun-yomi are given,
By transmitting a synthesized voice saying "Kaffeltou Higashi" from the speaker 8, proofreading of documents in which "characters" are important, such as proper nouns, can be effectively performed.

In this way, the present invention provides a voice-based proofreading support system that
When proofreading a text that includes a large number of proper nouns such as people's names and place names, the mode of the audio proofreading support device is set to a specific mode, and the kanji that are recognized as proper nouns in the text are corrected.
By dividing each kanji into kanji and referring to a kanji dictionary and reading them aloud in both on-yomi and kun-yomi, the text can be proofread without having to give each proper noun a "reading" as in the conventional method. It is unique in that it has been made possible.

〔Effect of the invention〕

As explained above in detail, the audio-based proofreading support method of the present invention analyzes written Japanese sentences, divides them into appropriate words, adds readings to each word, and reads out the words using synthesized speech. An audio-based proofreading support device that supports the proofreading of Japanese texts includes a kanji dictionary that registers both on-yomi and kun-yomi for each kanji, and means for detecting parts that are recognized as proper nouns by switching modes. By providing the above-mentioned proper noun, a Japanese sentence is read out using a synthesized voice with on-yomi and kun-yomi given for each kanji using a kanji dictionary, and the written Japanese sentence is proofread. Therefore, the word dictionary has a registration space, and [[
Each kanji character has an on-yomi reading, without a space for the reading.

By simply providing a one-word dictionary with kun-yomi, it is possible to realize a speech-based proofreading support device that can deal with any proper noun.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of the configuration of the audio-based proofreading support system of the present invention. FIG. 2 is a diagram showing an example of the structure of a kanji dictionary. FIG. 3 is a diagram illustrating a conventional audio-based proofreading support system. In the drawing, ■ is the main control unit + Ia is the document file. 2 is a sentence analysis section + 2a is a word dictionary. 3 is a speech synthesis section + 3a is a phoneme file. 4 is a proper noun detection section, and 5 is an onkun reading section. 5a is a kanji dictionary, 6 is a mode switching section. 7 is a terminal, 7b keyboard (KB). 8 is a speaker. are shown respectively. /L
Meko 1 diagram 2 diagrams [ ] are the hooks 2 enemy f needles 9,000 (F) no Utsu-t/￥'S::- Cane Masafumi I 1 Eight hooks Yomei 1 Ri Zuko 3
figure

Claims (1)

[Claims] Analyze the written Japanese sentence based on the word dictionary (2a), divide it into words, add readings to each word, and read out the resulting words with a synthesized voice. In an audio proofreading support device that supports proofreading, a kanji dictionary (5a) containing both on-yomi and kun-yomi for each kanji is provided, and a mode switching means (6) is used to detect a proper noun from the word. Then, using the above-mentioned Kanji dictionary (5a), a Japanese sentence is generated by giving on-yomi and kun-yomi to each kanji, and the detected proper noun part is read aloud with a synthesized voice. An audio-based proofreading support method characterized by proofreading Japanese sentences.