| 摘要 |
PROBLEM TO BE SOLVED: To provide a device and a method for voice data compression and decompression which secure the real-time properties of voice synthesis by correcting the compression position and decompression position of waveform data and which actualize voice synthesis of high quality. SOLUTION: Voice data including waveform data are extracted from an existent voice waveform dictionary and inputted, part in the waveform data to be used for the voice synthesis specified, the start point and end point of compression are set before and after it; and the waveform data in the set compression section are compressed and the compressed waveform data are decompressed, a compression section specified with the start point and end point of compression giving the highest quality of the decompression result of the compressed waveform data is determined as a compression and decompression point, and the waveform data compressed as waveform data to be used for the voice synthesis and the start point and end point of the compression are registered in a database. |
