JPS6026997A - Character phoneme converter - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a graphophoneme conversion device that converts a character string of a Japanese sentence into a phoneme sequence.

[Technical background of the invention and its melting point]

A so-called speech synthesizer is known, which converts input Japanese text into speech and outputs it. As a device of this kind, a method has been considered in which a phoneme dictionary stored in the h=position below a phrase is used to extract the phoneme word corresponding to the input character code and output it as a voice. This method had the disadvantage that the output speech did not contain any accents, was monotonous, had no rhythm, and was difficult to interpret. Therefore, it was necessary to analyze the character string and convert it into a phoneme sequence containing accent information, but conventionally this was done by person 1-. In other words, all the work has been done by humans creating phoneme sequences of character strings and inputting them into speech synthesis devices.

Therefore, when converting a character string to a phonetic sequence, you not only change the kanji part of the character string to the reading (phonetic), but also
In order to convert speech into a highly natural sound, it is necessary to accurately test accents. In this case, the hiragana character string has different accent positions depending on whether it is an attached word or an independent word, so even if the same hiragana character string is used, it can be an attached word or an independent word depending on the situation in which it is used. It was important to correctly distinguish between Conventional devices were not capable of correctly identifying whether an input four-character string of hiragana was an attached word or an independent word. Therefore, humans were required to distinguish between these accents and input accent information into a speech synthesis device based on the results, but this method was burdensome for humans in terms of time and effort.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide an apparatus that enables converting character strings of Japanese sentences into phoneme sequences, and in particular, testing the accents of hiragana character strings with high precision.

[Summary of the invention]

The present invention divides a character string containing kanji and kana into phrase units by imparting character types, and when the first character of this phrase is a kanji, the phrase character string is compared with a kanji independent word dictionary, and on the other hand, the first character is a hiragana. When , it is checked against the hiragana independent word dictionary,
If the conjugated ending is a conjunctive or adjunctive form, the remaining hiragana character string after removing the independent word and its conjugated ending from the clause is a specific adjunct character that can be connected to the conjunctive or adjunctive form. Test if the column is at the beginning,
If it is a specific attached word string, the above hiragana string is checked against the attached word dictionary, and if it is not a specific attached word string, it is checked against the above hiragana independent word dictionary, and the first character of the above clause is checked. When the independent word collated from is other than a pragmatic word, or even if it is a pragmatic word, it is not a conjunctive form or an adnominal form, the independent word and its conjugated ending are removed from the clause, and the remaining hiragana character 11 and 11 genitive dictionary are used. By performing the matching, the above-mentioned character string containing kanji and kana is converted into a phonological sequence.

Especially regarding the hiragana character string part, I am wondering if it is an independent word.
Based on correctly identifying the 414 words,
This is a highly accurate test for accents, which are important when converting sentences into speech.

〔Effect of the invention〕

According to the present invention, character strings of Japanese sentences can be converted into phoneme sequences, and in particular, by correctly identifying independent words and attached words for hiragana character strings, it is possible to convert sentences into highly natural sounds. Since important accents can be tested with high accuracy, the quality of the speech synthesizer can be improved, and the burden on humans is reduced in terms of time and effort.

[Embodiments of the invention]

Hereinafter, one embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of the apparatus of this embodiment. First, a character string containing kanji and kana to be converted into speech is stored in the input text storage device (1). The phrase cutting circuit (2) detects a change in character type from the sentences stored in the input sentence storage device (1), such as a change from "kana" to "kanji", and converts the input characters. The column is cut out into phrases and set in the input phrase register (3). For each character string of a phrase set in the input phrase register (3), the following phonetic conversion is performed.

The clause first character verification circuit (4) detects whether the first character of the clause is a kanji or a hiragana, and performs dictionary matching from the beginning of the clause using the kanji independent word matching circuit (5) (signal line αQ). ) or the hiragana independent word matching circuit (II (signal line a7)). The Kanji Independent Word Dictionary (6) stores words starting with kanji along with their parts of speech,
It also stores the phonological sequence and accent type of each word. The Hiragana Independent Word Dictionary (11) is a dictionary that registers words starting with Hiragana, and has the same format as the Kanji Independent Word Dictionary (6).

When the first character of a clause is a kanji, the kanji independent word matching circuit (5) matches the first character string of the clause with the kanji independent word dictionary (6). Among the character strings detected as matches through this matching, the longest character string, that is, the longest matching word, is stored as a candidate word in a temporary storage device provided in the Kanji independent word matching circuit (5). If the independent word of the longest matching word is not a predicate, its phoneme series and accent information are passed through the signal line u heat to the phoneme series register (
141 and set. If the independent word is a predicate, its phoneme sequence and accent information are sequentially sent to the circuits shown below.

On the other hand, when the first character of a clause is a hiragana, the hiragana independent word matching circuit Ql matches the first character string of the clause with the hiragana independent word dictionary aυ, and performs the same thing as in the case of the above-mentioned kanji independent word matching circuit (5), The phoneme sequence and accent information are sent to the phoneme sequence register a through the signal line acJ.

Next, the above-mentioned kanji independent word dictionary episode 1i! ! (51 or Hiragana independent word matching circuit (II) When the independent word matched by II is a pragmatic (signal line (21) (2υ)) Conjugation ending verification circuit (
7) is a list of conjugated hiragana character strings following independent words (8)
A conjugated ending test is performed by comparing the conjugated endings with ).If it fails, the second candidate word is retrieved from the temporary storage device provided in the Kanji independent word matching circuit (5) or the Hiragana independent word matching circuit α1, and it is Perform the same processing as for matching words.

In the conjugated ending testing circuit (7), after the conjugated ending is verified, the independent word and the conjugated ending are removed from the input phrase character string.

When the conjugated ending that has been verified is in the conjunction form or the conjunction form, the remaining character string is sent to the hiragana character string verification circuit (9) (signal line CI!31). The hiragana character string verification circuit (9) verifies whether the first character string of the remaining character strings is a specific character of two or more hiragana characters as shown below by comparing it with the specific character table a9, and If the characters are specific characters, there is a high possibility that they are adjunct words, so they go to the adjunct word matching circuit 021 through the signal line (24).

After a conjunction form: ``Mazu'', ``Tari'', ``Ttsu''
After the ``tome'' and ``nagara'' adjunctive forms, ``
kara,''``sae,''``shika,''``dakej'' and ``but.''

On the other hand, if it is not a specific character as mentioned above, it goes to the hiragana independent word matching circuit (10) through the signal line (2-wire) and matches the remaining character strings with the hiragana independent word dictionary aυ.

If the matching is successful here, the hiragana independent word matching circuit α0
) is the accent of the above-mentioned adjunctive form or adjunctive form. The accent information is then stored in the phonological sequence register α.
Send it to Yu. (Signal line (I wing) The adjunct word matching circuit α2 verifies the adjunct word by comparing the character string starting with the above specific character with the adjunct word dictionary 0th floor. The adjunct word dictionary (the first floor is Adjuncts/words and compound forms of adjunct words (e.g. "takenade") are registered as headwords, and information on the conditions when they are connected to independent words, and the shift of accent when they are combined with independent words, etc. is registered. has been done.

The adjunct word matching circuit α2 converts the adjunct word that has been matched into a phoneme sequence and sends it to the phoneme sequence register I (signal line (2)
6) ), the accent of a clause in which an independent word and an adjunct are combined is also the adjunct word above! It is verified based on the information registered in 'Q, 3), and the phrase accent position information is also sent to the phoneme series register u.

In addition, when the independent word matched by the Kanji independent matching circuit (5) or the Hiragana independent word matching circuit (10) in matching with the first character string of the clause is other than a pragmatic word, or even if it is a pragmatic word, the conjugated ending test is not performed. Mouth M (When the conjugated ending tested in 71 is other than the adjunctive form or the adjunctive form, the remaining character strings that include the verified independent word and its conjugated ending from the clause are passed through signal lines (27), 08J, and 01. It is sent to the adjunct word matching direction tshaa.Here, the remaining character strings are compared with the adjunct word dictionary 11 (13), and when the adjunct word is matched in the same way as described above, the phoneme sequence and accent information are sent to the rhyme sequence register 04. ). Furthermore, if the phrase character string remains even after removing the matched adjunct, the remaining character string is sent to the hiragana character string verification circuit (9) through the signal line (to) and the same process as above is performed. will be held.

As described above, according to this embodiment, a specific character string is verified in the hiragana character string verification circuit (9), and based on that, the collation order in the hiragana independent word verification circuit α0) and the attached word verification circuit 0 is changed. By doing this, it is possible to correctly distinguish and test whether a hiragana character string following a d-adjunction form or a d-adjunction form is an attached word or an independent word.

For example, in the case of the input character string ``Kashimasu'', when ``Kashimasu'' is verified as a conjunction, the remaining character string ``Masu'' is compared without going through the hiragana character string verification circuit (9). When the hiragana independent word matching circuit 0I performs the test directly, ``Machi (increase)'' is tested as an independent word, so the accent is also tested as ``masu'' (flat type). But needless to say, this is wrong. In this actual htU example, the hiragana character string verification circuit (9) tests ``Mazu'' as a specific character, so in this case, the matching for ``Masu'' is performed in the adjunct matching lOJ path 02); It can be tested as an auxiliary verb, and it can be used as an accent for phrases such as ``speak''.

In the above example, after testing the conjunction form, by going to the attached word matching circuit (1 go) instead of the hiragana independent word matching circuit (00), a correct test result can be obtained for the hiragana character string, but the Even if you decide to always go to the attached word matching circuit after the test, an error will occur.

For example, if the input clause character string is ``easy to recite,'' then when you go to the gold circuit (121), the character ``ya'' is tested as a particle of an adjunct, and then the remaining character string after removing ``ya'' is ``Sui'' is an independent word in the hiragana independent word matching session moo+
The accent is 2 because it is tested as “Sui (sucking)”.
It can be separated into ``I-i'' and ``口〒''. Needless to say, this is a mistake, and the correct answer should be "πτ〒7ii." In this embodiment, after the verification of the adjunctive form or the adjunctive form, the remaining character examples are sent to the hiragana character string verification circuit (9) to identify whether they are attached words or independent words. There is no conditional matching of adjunct words, and there is no unconditional matching of independent words, so the accents can be detected correctly without the errors in the above example.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram showing an apparatus according to an embodiment of the present invention. ■...Input sentence storage device 2...Bunsetsu cutting circuit 3.
...Input clause register 4...Bunsetsu first character verification circuit 5...Kanji independent word collation circuit 6...Kanji independent word dictionary storage device 7...Inflection ending detection circuit 8...Inflection ending table storage Device 9... Hiragana sentence gamma sequence verification circuit 10... Hiragana independent word matching circuit 11... Hiragana independent word dictionary storage device 12... Adjunct word matching circuit 13... Adjunct word dictionary storage device 14...・Phonological series register 15...Specific character table storage device 16-30...Signal line agent Patent attorney Noriyoshi Chika Ken Right (1 idiot) Figure 1 input - (/ Female - Masashitosu 2゛) Self・1 bubble 1. Device Mentop 1 , Poison and overturn!, Zt 7 4 for tt) also Mi all ears ■ 瞠 0 also table test negative circuit i 已・11 turtle, increase L1 73 15 7 'utfi'j-'Jq Chu Gua 7□ test 1L@ro to the town [f/'J -f] to the town fl@ Nishi Reiki 1st ji 13

Claims (1)

[Claims] A bunsetsu cutting means divides a character string of a Japanese sentence into clauses based on changes in character type; a bunsetsu first character verification means tests whether the first character of the clause outputted from the bunsetsu cutting means is a kanji character or a hiragana character; When the first character of the phrase is verified as a kanji by the character verification means, an independent word included in the beginning of the phrase is detected, it is verified whether or not this independent word is a phrasing, and the independent word is removed from the phrase. Kanji independent word matching means outputs the first remaining character string and phonological accent information of this independent word, and when the first character of the clause is verified as hiragana by the clause first character verification means, the beginning part of this clause is verified as hiragana. A hiragana independent word that detects the included independent word, tests whether or not this independent word is a pragmatic word, and outputs the first remaining character string after removing this independent word from the 011 clause and the phonological accent information of this independent word. When the independent word is verified as a term by the collation means and the Kanji independent word collation means or the Hiragana independent word collation means, the conjugated ending of the independent word included in the first remaining character string is an operational form or a combination. A second remaining character string obtained by examining whether or not it is a body shape and removing this conjugated ending from the first remaining character string, phonological accent information of this conjugated ending, and the kanji independent word matching means or hiragana independent word. a conjugated ending testing means for outputting the phonological accent information of the independent word outputted from the matching means as is; and a second remaining character string when the conjugated ending is verified as a conjunctive form or an adjunctive form by the conjugated ending testing means; It is tested whether the first part of the character string is a specific character string that can be connected to the adjunctive form or the adnominal form. A hiragana character string verification means for outputting the independent word and its conjugated ending phonological accent information as is to the hiragana independent word matching means; the first remaining character string and the phonological accent information of the independent word; and the second remaining character string when the conjugated ending is determined not to be a conjunctive or an adjunctive ending by the inflected ending testing means; evening (1 and independent words and phonological accent information at the end of their conjugated words, the second remaining character string when the first part of the second remaining character string is verified to be a specific character string by the hiragana character string verification means) Input the character string and phonological accent information of the conjugated ending of the independent word, detect the adjunct included in the first part or the beginning of the remaining sentence "-T string of 5152, and determine whether this adjunct is an independent word or the conjugated ending. (alliteration accent) of the entire character string consisting of the independent word, the conjugated ending, and the adjunct word, or the independent word and the adjunct word, based on the information when connecting to the adjunct word; A character-to-character conversion device comprising: a phoneme series register that stores phoneme accent information of the entire character string outputted from the matching means.