摘要 |
PROBLEM TO BE SOLVED: To realize a data receiving unit with a memory of small capacity and a microcomputer of low speed, by generating and transmitting a phonetic representation data comprising a phoneme and a punctuation code, and converting the received phonetic representation data into the composite voice. SOLUTION: In a data transmitting unit 1, a language processing part 4 decomposes a sentence represented by a character data, into the reading syllabary, and adds the punctuation code indicating the punctuation (breather part or the like) in the sentence vocalization, to the parts of punctuation marks and a part between paragraphs, to form the phonetic representation data. In a data receiving unit 2, a pheneme length generating part 8 defines an articulation period of each pheneme (pheneme length), and a soundless period between the phenemes in the received phonetic representation data. A pitch generating part 10 determines an intonation pattern, adds an accent pattern to the voice, and determines a pitch pattern of an input sentence. A waveform connecting part 13 reads out the waveform corresponding to each pheneme, from a pheneme dictionary 14, and synthesizes and outputs a voice waveform, by combining the waveforms on the basis of the pheneme length data and the pitch pattern. |