JPH06289889A - Speech synthesizing device - Google Patents

Abstract

PURPOSE:To easily synthesize a speech which is more likely to Japanese by determining speech information of an English word as Japanese from speech information as English and part-of-speech and speech information on Japanese words before and after it by the speech synthesizing device which converts a KANJE(Chinese character)KANA(Japanese syllabary) mixed sentence of Japanese with which English is mixed into a speech. CONSTITUTION:An English dictionary 6 which contains in pairs of part-of-speech information corresponding to notations of English words and speech information including at least reading and accent information is added to conventional speech synthesizing device, and an English-Japanese speech information conversion ruse 7 which contains rules for generating speech information as Japanese from speech information as English is added. Consequently, the Japanese KANJI- KANA mixed sentence with which English is mixed can be converted into a speech by utilizing an existent English dictionary without structuring a dictionary which contains in pairs of part-of-speech information as Japanese corresponding to notations of English words and the speech information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech synthesizer for converting a Japanese document mixed with English into speech information and outputting speech.

[0002]

2. Description of the Related Art A conventional speech synthesizer for converting a Japanese document mixed with English into speech and outputting the speech includes part-of-speech information as Japanese corresponding to English word notation and at least reading and accent information. It has a dictionary that stores audio information in pairs.

The operation in the case where an input of "draw a guitar" is given to the conventional speech synthesizer will be described below. First, morpheme division is performed using the word dictionary shown in (Table 1), and the generated phonetic symbol generation results are shown in (Table 2). The phonetic symbols in capital letters indicate that there is an accent at that position. The word dictionary stores part-of-speech information and voice information as Japanese for all words regardless of English / Japanese. Therefore, it is possible to directly synthesize a Japanese voice without converting the English part-of-speech information and the voice information into the Japanese part-of-speech information and voice information.

[0004]

[Table 1]

[0005]

[Table 2]

[0006]

However, in the above-mentioned conventional speech synthesizer, a dictionary that stores, for all English words, a pair of part-of-speech information as Japanese for the English word notation and voice information is stored. Difficult to build. Therefore, a method of generating voice information using a Japanese dictionary for Japanese and an English dictionary for English can be considered. However, of the voice information stored in the English dictionary, for example, if the accent information is directly used as the Japanese accent information for voice synthesis, a voice that is very difficult to hear is generated.

It is an object of the present invention to solve such a problem and convert an existing English dictionary into a Japanese kanji / kana mixed sentence in which English is mixed to make it sound like Japanese. .

[0008]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention uses a Japanese word notation in a speech synthesizer for converting a Japanese mixed kanji / kana sentence into a voice. Japanese dictionary storing part-of-speech information corresponding to, and voice information including at least reading and accent information, and English dictionary storing part-of-speech information corresponding to English word notation and voice information including at least reading and accent information And an English-Japanese voice information conversion rule that stores rules for determining the voice information as Japanese of English words from the voice information as English and the part of speech and voice information of the preceding and following Japanese words,
An input unit for inputting a mixed Kanji / Kana sentence of Japanese mixed with English, an input character string storage unit for storing the input character string input by the input unit, and an input character string storage unit A morpheme division result storage unit that stores a morpheme division result including word notation, part-of-speech information, and voice information including at least reading and accent information for the input character string, the Japanese dictionary, the English dictionary, and the input character string In the morpheme division result storage unit corresponding to each morpheme to which the rule can be applied by performing morpheme division while referring to the storage unit, storing the result in the morpheme division result storage unit, and referring to the English-Japanese speech information conversion rule. Voice information generation control unit that changes the voice information of the voice information and stores the voice information in an output voice information storage unit described later, an output voice information storage unit that stores the generated voice information, and the output sound. Characterized by comprising an output means for outputting a sound based on the sound information stored in the information storage unit.

[0009]

According to the present invention, the conventional speech synthesizer is added with a rule for determining the voice information in Japanese of English words from the voice information in English and the part of speech and voice information of the preceding and following Japanese words. Even if you do not build a dictionary that stores Japanese part-of-speech information and voice information for each word notation, you can use the existing English dictionary to Converts kana mixed sentences into voice,
This is to generate more Japanese-like voice information.

[0010]

1 is a functional block diagram of a speech synthesizer according to an embodiment of the present invention. In FIG. 1, 1 is an input means for inputting an input character string, and 2 is an input means 1. An input character string storage unit for storing an input character string, 3 is an output voice information storage unit for storing generated voice information, 4
Is an output means for outputting a voice based on the voice information stored in the output voice information storage unit 3, and 5 is a part of speech information corresponding to the notation of a Japanese word and voice information including at least reading and accent information. Stored in the Japanese dictionary, 6 is an English dictionary stored as a pair of part-of-speech information corresponding to the notation of an English word, and voice information including at least reading and accent information, and 7 is a voice of the English word as Japanese. English-Japanese voice information conversion rule storing rules for determining information from voice information in English and part-of-speech and voice information of Japanese before and after, 8
Is a morpheme division result storage unit that stores a morpheme division result including word notation, part-of-speech information, and voice information including at least reading and accent information for the input character string stored in the input character string storage unit 9, The morpheme division is performed by referring to the Japanese dictionary 5, the English dictionary 6, and the input character string storage unit 2, the result is stored in the morpheme division result storage unit 8, and the English-Japanese speech information conversion rule 7 is referred to. Among the divided morphemes stored in the result storage unit 8, it is a voice information generation control unit that changes the voice information of the divided morphemes to which the rule can be applied and stores the voice information in the output voice information storage unit 3. .

FIG. 2 is a circuit block diagram of FIG. 1. 10 is a read only memory (hereinafter abbreviated as ROM), 11 is a keyboard, 12 is a central processing unit (hereinafter abbreviated as CPU), and 13 is a speaker. , 14 are random access memories (hereinafter abbreviated as RAM).

Here, the input means 1 shown in FIG. 1 is the keyboard 11, the Japanese dictionary 5 and the English dictionary 6 are the ROM 10, the input character string storage unit 2, the output voice information storage unit 3, and the morpheme division result storage unit. Reference numeral 8 denotes the RAM 14, and the voice information generation control unit 9 executes the program stored in the ROM 10 while the CPU 12 exchanges data with the ROM 10 and the RAM 14, and the output unit 4 is realized by the speaker 13. .

This operation will be described below with reference to the flow chart of FIG. 3 in the case where the input sentence "guitar is pulled" is given to the speech synthesizer according to the embodiment of the present invention configured as described above.

In step (S1), the voice information generation control unit 9
Performs morpheme division on the input sentence with reference to the Japanese dictionary 5 shown in (Table 3) and the English dictionary 6 shown in (Table 4), and stores the morpheme division in the morpheme division result storage unit 8. In the above example, the reading information for the English word "guitar" at this point is still voice information [gitA: r] in English.
The phonetic symbols shown in capital letters indicate that there is an accent at that position.

[0015]

[Table 3]

[0016]

[Table 4]

In step (S2), the voice information generation control unit 9 has finished checking the applicability of the English-Japanese voice information conversion rule 7 among the divided morphemes stored in the morpheme division result storage unit 8. Check if there are no split morphemes. If the examination of applicability to the English-Japanese speech information conversion rule 7 has not been completed, the process proceeds to step (S3). Otherwise, the process ends.

In step (S3), the voice information generation control unit 9 refers to the notation, the part-of-speech information, and the voice information corresponding to the divided morphemes stored in the morpheme division result storage unit 8 as shown in Table 5. From the English-Japanese speech information conversion rule 7, an applicable rule having the same notation as the divided morpheme and applicable is searched. In the above example, the rule of rule number 1 is the applicable rule. This rule states that "voice information of English words is [* itA: r]
(The accent is at the position of “a”), the Japanese phonetic symbol is [* I ta −] (the accent is converted to the position of “i”) ”. Here, the above [*] represents an arbitrary phonetic symbol. In the above example,
This rule applies to the English word "guitar".

[0019]

[Table 5]

In step (S4), the voice information generation control unit 9 checks whether or not there is an applicable rule that has the same notation as the divided morpheme. If not, step (S2)
Proceed to and detect a divisional morpheme whose rule applicability has not yet been checked. If there is, go to step (S5).

In step (S5), the voice information generation control unit 9 applies the detected applicable symbol voice information rule to the divided morphemes. In the above example, the voice information for the English word “guitar” stored in the morpheme division result storage unit 8 is changed from voice information in English [gitA: r] to voice information in Japanese [gI ta −], It is stored in the output voice information storage unit 3. Table 2 shows the contents of the output voice information storage unit 3 in the above example.

As described above, in the conventional speech synthesizer,
Using the existing English dictionary, a mixture of English words can be created by adding a rule that determines the voice information of English words in Japanese from the English voice information and the part of speech and voice information of the preceding and following Japanese words. It is possible to realize speech synthesis that converts a sentence containing mixed Japanese kanji / kana into speech.

[0023]

As described above, the speech synthesizing device of the present invention is a rule for determining the Japanese voice information of an English word from the voice information of English and the part of speech and voice information of the preceding and following Japanese words. With the addition of, the existing English dictionary can be used to provide Japanese-like voice to the user.

[Brief description of drawings]

FIG. 1 is a functional block diagram of a voice synthesizer according to an embodiment of the present invention.

FIG. 2 is a circuit block diagram of FIG.

3 is a flowchart illustrating the operation of FIG.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Input means, 2 ... Input character string storage part, 3 ... Output voice information storage part, 4 ... Output means, 5 ... Japanese dictionary,
6 ... English dictionary, 7 ... English-Japanese voice information conversion rule, 8 ... Morphological division result storage unit, 9 ... Voice information generation control unit, 10 ...
ROM, 11 ... Keyboard, 12 ... CPU, 13 ... Speaker, 14 ... RAM.

Claims (1)

[Claims] 1. A voice synthesizer for converting a mixed sentence of Japanese kanji / kana mixed with English into voice, and voice information including part-of-speech information corresponding to the notation of a Japanese word and at least reading and accent information. The Japanese dictionary that stores and, the English dictionary that stores the part-of-speech information corresponding to the English word notation, and the voice information including at least the reading and accent information, and the voice information as Japanese of English words Input method for inputting English-Japanese voice information conversion rules that store rules for deciding from the voice information of voice and the part-of-speech and voice information of the preceding and following Japanese, and Japanese kanji / kana mixed sentences in which English is mixed An input character string storage unit that stores an input character string input by the input unit, at least a word notation and a part-of-speech information for the input character string stored in the input character string storage unit. And a morpheme division result storage unit configured to store a morpheme division result including voice information including accent information, morpheme division while referring to the Japanese dictionary, the English dictionary, and the input character string storage unit, and The speech information stored in the morpheme division result storage unit is referred to, the speech information in the morpheme division result storage unit corresponding to each morpheme to which the rule can be applied is changed, and the output speech information storage unit described later. A voice information generation control unit that stores voice information, an output voice information storage unit that stores the generated voice information, and an output unit that outputs a voice based on the voice information stored in the output voice information storage unit. And a voice synthesizing device.