JP2839488B2 - Speech synthesizer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention provides a speech that can naturally synthesize speech input word information given as a symbolized code by simple processing. The present invention relates to a synthesizer. (Prior Art) Recently, a prosody parameter and a phoneme parameter of input word information converted into a symbol string are obtained, and a speech is rule-synthesized in accordance with the prosody parameter and the phoneme parameter. Such speech synthesis, together with speech recognition processing technology, is an important technology for realizing a highly natural man-machine interface. By the way, in such speech rule synthesis, it is important to correctly reproduce the accent unique to the word in order to obtain a synthesized speech having a high naturalness. Therefore, in a conventional speech synthesizer, for each simple word, the accent information unique to the simple word and the accent information of a compound word composed of a combination of a plurality of simple words are registered as an accent dictionary, and the input is registered in the accent dictionary. And ask for a unique accent to the input. However, such a conventional speech synthesizer has a disadvantage that it cannot perform processing of alphanumeric characters and the like. (Problems to be Solved by the Invention) As described above, there is a problem that a conventional speech synthesizer cannot process a special code representing an alphanumeric character or the like. The present invention has been made in view of such a problem, and an object thereof is to process not only ordinary character codes but also special codes representing alphanumeric characters and the like, so that matching with an accent dictionary can be efficiently performed. The purpose of the present invention is to provide a speech synthesizer capable of outputting a synthesized speech with higher intelligibility and naturalness than before. [Structure of the Invention] (Means for Solving the Problems) In order to achieve the above object, a speech synthesizer according to the present invention comprises:
Input means for inputting symbolized word information comprising a special code and a character code corresponding to each word;
The special code input by the input means is converted into a first character code for reading a corresponding word, and the character code in the word information divided into the special code is converted to a first character code for reading the corresponding word. A second character code, and a silent code data inserting unit for inserting silent code data between the first character code and the second character code. (Operation) According to the speech synthesizer of the present invention, the special code and the character code in the word information divided into the special code are respectively converted into the first and second character codes of the corresponding word reading. In addition, since the silent code data is automatically inserted between the first character code and the second character code by the silent code data inserting means, each of the first and second character codes is converted into voice. In this case, a non-speech part is inserted between them, whereby a synthesized speech with high intelligibility and naturalness can be output. (Example) Hereinafter, an example of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a voice configuration device according to an embodiment of the present invention. As shown in FIG. 1, the speech synthesizer includes a symbol string input unit 1, a code conversion unit 2, a conversion table 3, a word matching unit 4, an accent dictionary 5, a compound word accent test unit 6, a prosodic parameter generation unit 7, It comprises a phoneme sequence test unit 8, a speech unit file 9, a phoneme parameter generation unit 10, and a speech synthesis unit 11. The symbol string input unit 1 inputs word information converted into a symbol string including a special code and a character code. The code converter 2 converts a special code and a character code input from the symbol string input unit 1 into a predetermined character code string. In this case, one special code is converted into one character code string, and the character codes sandwiched between the special codes are converted into one character code string as a group. Furthermore, between the character code string converted from the special code and the character code string converted as a group from the character code sandwiched between the special codes (Space). In the conversion table 3, a character code string corresponding to the special code and the character code is registered. FIG. 2 shows a special code in the conversion table 3 and a character code string corresponding to the special code. For example, when the special code is “28”, the corresponding character code string is “parentheses”. Similarly, when the special code is “80”, the corresponding character code string is “Kabushiki Geisha”. The word collating unit 4 collates the character code string obtained from the code converting unit 2 with the accent dictionary 5 to obtain accent information corresponding to each character code string. At this time, If there is a part delimited by, the comparison with the accent dictionary 5 is performed separately. The accent information obtained by the word matching unit 4 is input to a compound word accent test unit 6, a prosody parameter generation unit 7, and a phoneme sequence test unit 8. The compound word accent tester 6 tests the accent in the compound word according to the type of the word and the accent information detected and collated when the input word whose word has been collated constitutes a compound word. This test provides information on accents that change when a plurality of simple words constitute a compound word. The prosody parameter generation unit 7 obtains a prosody parameter of the character code sequence subjected to the accent test in this manner, and the prosody parameter sequence is provided to the speech synthesis unit 11. Also, the code obtained from the special code conversion unit 2 is In the case of, the silent parameter is given to the speech synthesis unit 11 for a certain period of time. The phoneme sequence test unit 8 tests whether or not the phoneme is nasalized and unvoiced from each phoneme information of the character code sequence obtained by the word matching unit 4. Then, a phoneme sequence corresponding to the character code sequence is generated according to the test result, and this is given to the phoneme parameter generation unit 10. The phoneme parameter generation unit 10 generates a phoneme parameter sequence corresponding to the phoneme sequence with reference to the speech unit file 9, and the code obtained from the special code conversion unit 2 is In the case of, a phoneme parameter string representing silence is created without referring to the speech unit file 9. The speech synthesizer 11 performs speech synthesis of the word indicated by the input symbol string according to the created phoneme parameter string and the above-described prosody parameter string. Next, the operation of this embodiment will be described. Now, as shown in FIG. 3A, "80, C4, B3, BC,
Assume that a symbol string "CA, DE, 28, CA, D7, CO, DE, 29" is input. "80", "2"
"8" and "29" are special codes, and the others are ordinary character codes. The code converter 2 converts such a symbol string into a character code string. FIG. 3 shows the state of the conversion. That is, the code conversion unit 2
Refers to the conversion table 3 and performs conversion into a character code string. That is, "80" is converted to "Kabushiki Geisha", and "C4", "B3", "BC", "CA", and "DE" are "G", "U", "S", "C", and "゛", respectively. Is converted to "28" is converted to "parentheses" and converted to "CA", "D7",
"CO" and "DE" are "ha", "la", "ta",
It is converted to “゛” and “29” is converted to “braces”. In this case, between "Kabushiki Geisha" and "Toshiba" (Space) is inserted. Similarly, between "Toshiba" and "parentheses" Is inserted. As can be seen from the figure, the special code itself is converted into one character code string. In contrast, ordinary character codes sandwiched between special codes are treated as a group of character code strings after conversion. That is, "Toshiba" is treated as a character code string as a group. Next, the character code string obtained by the code conversion unit 2 is input to the word collation unit 4, where it is collated with the accent dictionary 5, and accent information is obtained. Matching with this accent dictionary Is performed for each word separated by. That is, the character string “Kabushiki Geisha” is collated with the accent dictionary 5 to obtain predetermined accent information. Next, the character string "Toshiba" is collated with the accent dictionary 5, and predetermined accent information is obtained. in this way The accent matching process is performed for each word delimited by, so that, for example, even if the character code string "Toshiba" is not registered in the accent dictionary 5, the accent test process for other words is not affected. . The compound word accent tester 6 tests the accent of the compound word according to the character code string and the accent information, and gives information on the accent when a plurality of simple words constitute the compound word. The prosodic parameters and phoneme parameters are generated by the prosodic parameter generation unit 7, the phoneme sequence test unit 8, the speech unit file 9, and the phoneme parameter generation unit 10 based on the character code string and the accent information thus obtained. You. In this case, Because it contains A parameter sequence representing a pause (for example, a silence period of 250 ms) between words adjacent to is also generated. These parameters are input to the voice synthesis unit 11 and a synthesized voice is output. As described above, in the present embodiment, when a special code is included in the input symbol string, the special code is converted into a character code string, and the character codes sandwiched between the special codes are converted into a character code string as a group. Since the code strings are individually subjected to word collation processing, accent information unique to the words indicated by those character code strings can be obtained easily and with high accuracy. Further, by giving a pause between the special code and the adjacent code sequence, a great effect in practical use such as an increase in intelligibility can be obtained. Note that the present invention is not limited to the above-described embodiment. For example, the pose symbol Other symbols may be used. Needless to say, the correspondence between the special code and the character code string is not limited to the above-described embodiment. [Effects of the Invention] As described above in detail, according to the present invention, a special code and a character code divided into special codes are respectively converted into first and second character codes of reading of corresponding words. Moreover, silence code data can be automatically inserted between the first character code and the second character code to generate a silent portion, thereby improving the intelligibility and naturalness of the output synthesized speech. it can.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a configuration of a speech synthesizer according to an embodiment of the present invention, FIG. 2 is a diagram showing a configuration example of a conversion table, and FIG. It is a figure showing an example of a character code string obtained by conversion processing. 1 ... symbol string input unit 2 ... code conversion unit 3 ... conversion table 4 ... word collation unit 5 ... accent dictionary 6 ... compound word accent test unit 7 ... prosodic parameter generation unit 8 ... phonological sequence test Unit 9 Voice unit file 10 Phoneme parameter generation unit 11 Voice synthesis unit

Claims (1)

(57) [Claims] Input means for inputting symbolized word information comprising a special code and a character code corresponding to each word, and converting the special code input by the input means into a first character code for reading the corresponding word Conversion means for converting the character code in the word information divided into the special code into a second character code for reading a corresponding word; and the first character code and the second character code And a silence code data insertion unit for inserting silence code data between the speech synthesis device and the speech synthesis device.