| 摘要 |
PROBLEM TO BE SOLVED: To provide a voice signal compressing device etc., capable of efficiently compressing the data capacity of data containing components that represent voice that a human utters. SOLUTION: A computer C1 generates a pitch waveform signal by equalizing the time length of a section, corresponding to a unit pitch of voice data to be compressed and generates a sub-band signal, representing time variations in the intensity of fundamental-frequency component and higher-harmonic components of the pitch waveform signal. The sub-band signal is separated into continuous components comprising components having periodicity that nearly matches a prescribed standard and random components comprising other components, and the continuous components are subjected to entropy encoding or linear predictive encoding. COPYRIGHT: (C)2007,JPO&INPIT
|
