SPEECH REPRODUCING METHOD, SPEECH REPRODUCING DEVICE, AND COMPUTER PROGRAM

摘要

A speech reproducing device for reproducing a speech in units of a voxmol while extracting the boundary positions of voxmols. The speech reproducing device comprises a voxmol extracting unit (802) for storing position identification information representing the boundary positions while extracting the boundary positions of two or more speech voxmols and a reproducing unit (803) for reproducing sound information sequence (801) of each speech voxmol from a specified reproduction start point while specifying reproduction start points in the sound information sequence (801) according to the stored position identification information. Especially, the voxmol extracting section (802) extracts a small-amplitude section included in the sound information sequence (801), selects a small-amplitude section sandwiched between two voxmols out of the extracted small-amplitude sections, and specifies the boundary positions of the two voxmols in the selected small-amplitude section as position identification information.