摘要 |
PURPOSE: To resolve a problem related to analysis of the quality of a voice in a voice synthesizer. CONSTITUTION: A voice recognition system is programmed by many persons. This system receives synthesized voices or natural voices from voice synthesizers and persons, which present qualities of different voices, and determines the recognition level of each received voice. The voice to be evaluated from a voice synthesizer is received by the voice recognition system, and one recognition level is assigned to it, and it is ranked by comparison with recognition levels of previously received voices. |