摘要 |
PURPOSE: To provide the phoneme continuance control system which is applied to a speech rule synthesizing device that converts an optional text into a speech and improves the naturalness of speech quality. CONSTITUTION: This system consists of a rhythm unit extraction part 2 which extracts a rhythm unit consisting of one or two syllables from a monosyllable string 1, give information on the kind of the rhythm unit, and outputs a rhythm unit string 3, a rhythm unit length determination part 6 which gives the continuance of each rhythm unit, obtained on the basis of basic continuance from a rhythm unit reference length storage part 4, to the rhythm unit string 3 and outputs a rhythm unit length data string 7, a rhythm unit division part 8 which outputs a monosyllable length data string 9 consisting of the time lengths of respective nonosyllables obtained by dividing the time length of the rhythm unit at a certain ratio and the monosyllable kind, and a phoneme border determination part 14 which determines each syllable continuance 15 as a result of the arrangement of respective consonants on the basis of a rhythm perception point position 11 from a rhythm perception point storage part 10 and consonant length 13 given by a consonant length determination part 12 so that the time interval of the rhythm perception point matches each monosyllable length. |