摘要 |
PROBLEM TO BE SOLVED: To provide a speech data processor which prevents generation of noise at the joints of speech data by automatically detecting the beginning ends and terminals or the joints thereof of the speech data and subjecting the data to fade-in or fade-out processing. SOLUTION: Difference between input speech data DTia and the speech data DTd delayed by a speech data delay section 102 is calculated by a front and rear speech data difference calculation section 103, when a plurality of the input speech data DTia of arbitrary lengths are reproduced continuously in a speech data beginning end and terminal detection section, which detects the beginning ends and terminals or the joints of the beginning ends and the terminals of the input speech data DTia. In the speech data beginning end and terminal detection section 104, the detection of the beginning ends and terminals or the joints of the beginning ends and the terminals of the input speech data DTia is executed according to the speech data difference SBia calculated by the front and rear speech data difference calculation section 103. |