摘要 |
[Problem] To provide a voice clarification device capable of generating voice that can be heard easily in various environments without increasing the volume unnecessarily. [Solution] This voice clarification device (250) comprises: an envelope surface extraction unit (292) that extracts, for a spectrum of a target voice signal (254), a curve that contacts, or is along, local peaks of spectral envelopes of said spectrum and that represents an approximate shape of the spectral envelope peaks; a noise adaptation processing unit (300) that deforms the spectrum of the voice signal (254) on the basis of the curve extracted by the envelope surface extraction unit (292); and a sine-wave voice synthesis processing unit (305) that generates a converted voice signal (260) for voice that has been clarified on the basis of the spectrum deformed by the noise adaptation processing unit (300). |