JPH083718B2 - Audio output device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial field of use) The present invention is a voice output device for converting a Japanese sentence displayed in a character code into a voice and outputting the voice, in which proper nouns and common words have different reading accents. The present invention relates to a device for giving proper reading and accent to words in headings.

(Prior Art) An example of the configuration of a conventional audio output device is shown in FIG. 2a.
In FIG. 2, 11 is an input terminal, 12 is a language processing unit, 13 is a dictionary unit, 14 is a voice synthesis unit, and 15 is a voice output unit.

The code data of the Japanese sentence is input to the input terminal 11, and the Japanese sentence input by the language processing unit 12 is divided into words using information such as change points of the character type, and the dictionary of the dictionary unit 13 is referred to. Each word is read and accented, and a pause (voice pause section) is set using information about punctuation marks and particles. In the dictionary section 13, the headline, pronunciation, accent, part of speech, inflectional form, etc. of each word are registered.
The voice synthesis unit 14 combines voice parameters by combining voice parameters based on the information such as reading, accent, and pause provided by the language processing unit 12, and outputs the synthesized voice from the voice output unit 15. .

FIG. 2b is an explanatory diagram of an example of the output of the conventional device. That is, FIG. 2b shows examples 1, 2 and 3 in which the input sentence including the honorific expression and the output are compared, and the output for the Japanese sentence illustrates word boundaries, pauses and the like. .

Incidentally, although it is not the device itself as a publicly known document related to the above-mentioned conventional device, Practical Report on Research and Practical Use, Volume 32 [11
No.], [1983], Nippon Telegraph and Telephone Company, Telecommunications Research Institute, Yoshiyoshi Kamiya, et al. Sentence / sound conversion method applied to data home equipment, 2
There are 5 to 36 pages. This known document discloses a simple and efficient sentence / speech conversion system.

(Problems to be Solved by the Invention) Meanwhile, the conventional device of this type has the above-described configuration.
Since it was an action, in order to improve the naturalness and intelligibility of the voice output from this voice output device, first of all, the correct reading and accent should be applied to the input Japanese sentences of kanji and kana. It is necessary to add, and in Japanese, the same kanji string may be given different readings and cents, so common words and proper nouns such as those realized by conventional devices of this kind are used. It has been difficult to automatically add appropriate readings and accents to words by the method of having a dictionary that does not distinguish them or the device that searches the dictionary in a fixed order even if both have separate dictionaries. Therefore, the conventional output audio device has a drawback that it is often misread, which is one of the major causes of deterioration in naturalness and intelligibility.

In order to solve this drawback, the present invention provides a proper noun dictionary section having means for giving information of proper noun type and a common word dictionary section having means for giving isomorphic information, and makes it unique after a dictionary search. Another object of the present invention is to provide a voice output device capable of performing a noun connection test and adding appropriate reading and accent to the word.

(Means for Solving Problems) A voice output device for converting Japanese sentences including kanji, hiragana, katakana, alphabets, numbers, symbols, etc. written in character codes into voices for outputting to solve the above problems. In, a proper noun dictionary part having means for giving type information indicating types such as "surname", "first name", "place name", "company name" as well as heading, reading, and accent form, and heading, reading, accent form , When a word with a different reading or accent in the same heading along with part-of-speech and inflection is in the proper noun dictionary part, the common word dictionary part has means for giving isomorphic word information indicating that, and as the proper name "name" When the word connected before the word to which the reading is given is searched from the common word dictionary and the isomorphic flag is set, the heading of the word is searched again in the proper noun dictionary, The proper noun connection verification section having means for rewriting the pronunciation and the accent form when it is included in the proper noun dictionary as "surname" is provided, and each section constitutes the present invention by a known technique.

(Operation) Since the present invention is configured as described above, the input character string of the input word sentence is provided by means for giving information on the type of proper noun in the proper noun dictionary section and means for giving isomorphic word information in the general word dictionary section. , A proper noun connection test can be performed after a dictionary search, and appropriate reading and accent can be added to the word.

Embodiment An embodiment of the present invention will be described with reference to the drawings. FIG. 1a is a configuration diagram of an embodiment of the present invention, and FIG. 1b is an explanatory diagram of an example of a general word dictionary and a proper noun dictionary according to an embodiment of the present invention.
The figure is a flow chart of the processing flow of the proper noun connection verification unit according to the embodiment of the present invention. In FIG. 1A, 11 is an input terminal, and a Japanese sentence (code data) is input from the input terminal 11. Reference numeral 12 is a language processing unit, 14 is a voice synthesis unit, 15 is a voice output unit, 16 is a proper noun connection verification unit, 17 is a general word dictionary unit, and 18 is a proper noun dictionary unit. General language dictionary
17 is a general word dictionary 17-1, proper noun dictionary unit 18 is a proper noun dictionary
18-1 is provided.

Next, the operation of this embodiment will be described. The coded Japanese sentence input from the input terminal 11 is divided into words in the language processing unit 12, and for each word, the general word dictionary unit 17 first searches for a general word, and the heading of the word is found. If it is in the general language dictionary 17-1, its reading, accent, etc. are given to the word. In addition, the proper noun dictionary unit 18 provides a proper noun dictionary 18-1
To add illegible or accented words.

Also, when a homomorphic flag is set for the word, a flag indicating that flag is also given to the word. When the heading of the word is not in the general word dictionary 17-1, the proper noun dictionary 18-1 is searched and its reading, accent, and type information are given to the word. A certain noun (word number N) in the proper noun connection verification unit 16
Is searched from the proper noun dictionary 18-1, is read and accented, and if the type information is "name", pay attention to the word immediately before the word (word number N-1). . Then, the word (word number N-1) is searched from the general word dictionary 17-1 and if the isomorphic flag is set, the proper noun dictionary 18-1 for the word (word number N-1) is found.
Search again. As a result, when the type information is searched by "surname", the reading, accent, etc. of the word (word number N-1) are rewritten to those of the proper noun "surname". Appropriate readings, accents, etc. are given to the Japanese sentence input in this way, and these are converted into a voice signal in the voice synthesizer 14 and output from the voice output unit 15.

It should be noted that the respective devices and means of the respective parts for performing the above are all known in the art, and a detailed description thereof will be omitted.

FIG. 1b shows a general word dictionary 17-1 and a proper noun dictionary 18-1.
This is an example of the contents of a common word dictionary, such as plain (Heiya, Hirano), Yamamura (Sanson, Yamamura) etc. Headings are registered in all proper noun dictionaries, and the common word dictionary is flagged with the same word.

With such a processing structure, correct reading and accent can be added to the following sentence examples.

The common word dictionary 17-1 and proper noun dictionary 18-1 shown in FIG. 1b are used for kanji headings and their readings, parts of speech,
It has information such as accent shapes and is stored in memory.

Here, since one-to-one readings are given to the kanji headings in each dictionary, for example, if the word "Hirano" is searched in the order of common words and proper nouns, You can never read "Hirano" because the word "Haya" is given. If you reverse the order of dictionary lookup, you can't read "heya".

Therefore, the common word dictionary 17-1 is given the same word information, the common word is drawn, and when the headline is matched and searched, the information on the presence or absence of the same word is given to the word as a flag. When is set, the word is searched again from the proper noun dictionary 18-1 to give a reading.

Here, the same word means the word having the same heading also in the proper noun dictionary 18-1. Further, the above-mentioned one example will be described in detail.

If the word next to "Hirano" is not searched for as a proper noun name, the dictionary will not be searched again for "Hirano" even if the homonym flag is set.

 The search is performed as follows.

As described above, by using the present invention, it is possible to give appropriate readings and accents to words having different readings and accents in common words and proper nouns while having the same notation, and correct reading rates for input sentences, Since the accent imparting rate is improved, the naturalness and intelligibility of the sentence converted into voice is also improved.

FIG. 3 shows a flowchart of the processing flow of the proper noun connection verification unit according to the embodiment of the present invention, but the detailed description will be omitted.

(Effects of the Invention) As described above, the present invention is a voice output device capable of imparting an appropriate reading and accent to a word having the same notation but different reading and accent. As compared with the conventional device, there is an advantage that the correct reading rate and the accent imparting rate with respect to the input sentence are improved, and thus the intelligibility of the output voice is improved. There is an effect.

[Brief description of drawings]

FIG. 1a is a configuration diagram of an embodiment of the present invention, FIG. 1b is an explanatory diagram of an example of a general word dictionary and a proper noun dictionary of an embodiment of the present invention, and FIG. 2a is a conventional voice output. FIG. 2 is a block diagram of an example of the device, FIG. 2b is an explanatory diagram of an example of the output of the conventional device, and FIG. 11 …… input terminal, 12 …… language processing section, 13 …… dictionary section, 14
...... Speech synthesizer, 15 …… Speech output, 16 …… Proper noun connection tester, 17 …… General word dictionary, 18 …… Proper noun dictionary, 17-1 …… General word dictionary, 18-1 ...... Proper noun dictionary.

Claims (1)

[Claims] 1. A voice output device for converting a Japanese sentence containing a kanji, hiragana, katakana, alphabet, number, symbol, etc., which is written in a character code into a voice and outputs the voice, together with a heading, a reading and an accent form, "surname". ·"Name"·
A proper noun dictionary section having means for giving type information indicating types such as "place name" and "company name", and heading, reading, accent form, part of speech, inflection form, and different reading and accent form words in the same heading In the proper noun dictionary part, the common word dictionary part having means for giving isomorphic information indicating that, and the word connected before the word given as the "name" of the proper noun is the common word If the word is searched from the dictionary and the homonym flag is set, the heading of the word is searched again in the proper noun dictionary, and if it is "surname" in the proper noun dictionary, the pronunciation and accent form are rewritten. A voice output device having a noun connection verification unit.