摘要 |
<p>PURPOSE:To reduce the capacity required by a speech element file by registering plural kinds of speech element parameters having different parameter degrees in the speech element file with the same degrees. CONSTITUTION:A speech element file 1 stores speech element parameters for male voice of a parameter degree of 25th degree and another speech element parameter file 2 stores phoneme parameters for female voice of a parameter degree of 18th degree. A character string input section 4 inputs speeches to be voice-synthesized as a character code string. A phoneme sequence accent detecting section 5 analyzes the inputted character code string and finds a phoneme sequence indicating the reading information sequence of the character code string and, at the same time, detects the accent type. A synthesizing parameter producing section 6 successively extracts element parameters corresponding to the phoneme from the files 1 and 2 in accordance with the found phoneme information and produces a phoneme parameter sequence. Simultaneously, the section 6 produces a rhythm parameter sequence. A voice synthesizer 7 produces synthesized voices in accordance with the parameter sequence thus produced.</p> |