JPS63189933A - Device for reading sentence aloud - Google Patents

Abstract

PURPOSE:To continue the reading aloud of sentences without interrupting the reading by forming an unknown word reading conversion part and a Chinese character dictionary part and applying phonetic reading or Japanese reading of Chinese characters to an unknown word which is not registered in a sentence analyzing part. CONSTITUTION:A device for reading sentence aloud is provided with the un known word reading conversion part 5 and the KANJI (Chinese character) dictionary part 6 and connected to the sentence analyzing part 21. About an unknown word which is not registered in the analyzing part 21, an unknown word consisting of one KANJI and KANA (Japanese syllabary) is read out by Japanese reading of KANJI by collating the dictionary part 6 and unknown KANJI words consisting of plural KANJI are read out by phonetic reading and outputted to a voice synthesizing part 3. Consequently, the device for read ing sentence aloud can continue reading-aloud of sentences without being inter rupted by unknown words.

Description

[Detailed Description of the Invention] [Summary] The present invention relates to readings to be given to words that are not registered in the text analysis unit in text data input to a text reading device, A pronunciation conversion section and a kanji dictionary section are provided, and the kanji of words consisting of kanji and kana are given kunyomi, and multiple kanji are given onyomi, so that unregistered words can be given readings and read aloud. .

[Industrial application field]

The present invention relates to a text reading device, and particularly relates to a reading given to a word when there is a word in input text data that is not registered in a text analysis section.

A text reading device that reads text data composed of character codes using synthesized speech is used in newspaper companies and the like to proofread a manuscript and text data created from the manuscript.

That is, a text reading device reads text data using electronically synthesized synthesized speech, and a proofreader looks at the manuscript and proofreads it.

In this case, the text data often contains words that are not registered in the text analysis unit, such as coined words and new words.
If the text-reading device stops each time, proofreading of that part is troublesome, and extra man-hours are required for proofreading.

Therefore, there is a need for a text reading device that can read unregistered words by giving them some pronunciation.

[Conventional technology]

FIG. 4 shows a block diagram of a conventional text reading device.

In the figure, text data composed of character codes is input to a text input section l via a file device or line transmission.

The sentence analysis unit 2 classifies the sentence data output from the sentence input unit 1 into words, compares it with the built-in word dictionary, and adds elements necessary for word pronunciation such as pronunciation, accent, and vocal pitch to the classified words. Furthermore, when the word is read out as a sentence, intonation and the like of word pronunciation are added and outputted to the speech synthesis section 3.

The voice synthesis section 3 converts the digital signals of these sound generation elements into analog voice signals, which are output from the speaker 4.

In such a sentence reading device, if there is a word that is not registered in the word dictionary of the sentence analysis unit 2, that is, an unknown word, and it cannot be analyzed into pronunciation elements, speech synthesis cannot be performed and the reading of the sentence is limited to that part. will be stopped.

In that case, print or display the character code to check whether it is correct or incorrect with the original.

[Problem that the invention seeks to solve]

In this conventional method, new words, coined words, abbreviations, etc. frequently appear in texts used by newspaper companies, so the number of unknown words increases, and the reading of the text is interrupted each time.

The reading of a text is interrupted by unknown words, and the work of proofreading the manuscript is significantly hindered by the trouble of checking the unknown words.

The present invention was created in view of these points, and an object of the present invention is to provide a text reading device that gives some kind of pronunciation to unknown words and does not interrupt text reading.

[Means for solving problems]

In order to achieve the above object, a text reading device is provided with an unknown word reading conversion section and a kanji dictionary section, which are connected to a text analysis section.

For unknown words that are not registered in the text analysis department, the kanji in words that are composed of a single kanji or kanji are compared with the kanji dictionary department, and the reading is given using kun-yomi. For unknown kanji, give the on-yomi reading.

[Effect]

Unknown words that cannot be analyzed by the text analysis section are output to the unknown word reading conversion section, which compares them with the kanji registered in the kanji dictionary section and converts them into kanji with kun reading according to the word structure of the unknown word. Alternatively, it gives a reading and outputs it to the speech synthesis section.

By processing this unknown word, the text reading device is not interrupted by the unknown word.

Furthermore, without expecting all words to be registered in the word dictionary in the sentence analysis section, only common words are registered, and unregistered words can be left to the unknown word reading conversion section as unknown words.

By doing so, by reducing the number of words registered in the word dictionary, the average time for word matching of the text analysis text is reduced, and the text can be read out smoothly.

〔Example〕

FIG. 1 shows a block diagram of an embodiment of a text reading device according to the present invention.

Note that the same reference numerals indicate the same objects throughout the figures.

The sentence analysis section 21 is provided with a function of discriminating words that are not registered in the word dictionary as unknown words and outputting them to the unknown word reading conversion section 5.

Therefore, the unknown words in the input text data are output to the unknown word reading conversion unit 5 via the text analysis unit.

The unknown word reading conversion section 5 compares the reading with the kanji dictionary section 6, adds a reading to the unknown word, and outputs it to the speech synthesis section 3.

The kanji dictionary section 6 uses 1 as the first registration method for kanji readings.
Each kanji has one on-yomi and one kun-yomi,
If the unknown word is composed of kanji and kana, the kun-yomi is extracted for that kanji, and if the unknown word is composed of multiple kanji, the on-yomi is given to the kanji.

In this method, we do not necessarily expect the correct reading of the unknown word, but simply the reading that corresponds to the unknown word when reading the sentence aloud.
It is sufficient to specify one kanji.

FIG. 2 is a block diagram of the configuration of another embodiment of the text reading device of the present invention, and FIG. 3 is a diagram illustrating a second registration method for reading kanji.

The sentence analysis section 22 has a function of outputting the unknown word to the unknown word reading conversion section 5 and attaching a mark to the mark type C of the reading of the kanji in the kanji dictionary section 6.

As a second registration method, the kanji reading of the text analysis unit 22 that corresponds to the kanji is read for each kanji as shown in Figure 3.
It has a reading column b for each reading, and a marker type C for each reading.

For example, for ``gyo'', memorize the pronunciations of ``okona (u) [kun], gyou [on], kora [on], etc., and for ``sei'', [
I remember "Marrigoto [Kun], Sei [Sound]".

Then, the marker type C replaces the marker with the on-yomi used every time the text analysis text 22 assigns a reading to a kanji.

The unknown word reading conversion section 5 uses the labeled readings of the kanji dictionary section 6 for the kanji reading of the unknown word.

By doing this, the latest pronunciation used in the sentence analysis sentence 22 is given to the unknown word.

For example, if the unknown word is a word consisting of multiple kanji such as ``Administrative Reform Council'', even if these kanji have multiple phonetic readings, the word ``Administrative Reform Council'' appears at the beginning of the text data, and the word ``Gyosei'' appears at the beginning of the sentence data. In cases where ``Administrative Reform Council'' is read as ``Kaikakushingikai,'' even if it is an unknown word, it will be read as ``Gyoukakushin.''

Note that it is also possible to easily register the pronunciation of an unknown word in the word dictionary of the text analysis section so as to handle the word input for the next text data.

〔Effect of the invention〕

As described above, according to the present invention, even unknown words that are not registered in the text analysis section of the text reading device are given readings, and reading can be continued without interruption, making it practical. is extremely useful.

[BRIEF DESCRIPTION OF THE DRAWINGS] FIG. 1 is a block diagram of the structure of an embodiment of the text reading device of the present invention. FIG. 2 is a block diagram of the structure of another embodiment of the invention. Figure 4 is a diagram explaining the kanji dictionary in Figure 2.
FIG. 2 is a configuration block diagram of a conventional example. In the figure, l is a text data input section, 2, 21, 22 is a text analysis section, 3 is a speech synthesis section, 4 is a speaker, 5 is an unknown word reading conversion section, and 6 is a kanji dictionary section. 9 stones in 1m■t+-fJfJfJ4nA? D・y7
M Fig. 1 Fig. 2 Hancho #cho 4 m Fig. 3 Conventional f'J/1'j-fhf5C fh 7m Fig. 4

Claims (1)

[Claims] The text analysis unit (2) analyzes input text data,
The speech synthesis unit (3) compares the words of the text data with the pronunciations of words registered in advance and outputs the pronunciations for each word.
In a device that synthesizes speech and reads out aloud with synthesized sounds, an unknown word reading conversion unit (5) connected to the sentence analysis unit (21, 22) and a kanji dictionary unit connected to the unknown word reading conversion unit (5). (6), and the text analysis section (21
, 22), the words of the text data that are not registered are output to the unknown word reading conversion unit (5) and processed, and the words that are not registered are words that are composed of a single kanji or kana character. The text reading device is characterized in that the reading of kanji is outputted in kun-yomi, and the reading of kanji of a word consisting of a plurality of kanji is outputted in on-yomi, and the reading is inputted to a speech synthesis unit (3).