SPEECH WAVEFORM COMPRESSING METHOD BY SIMILARITY OF FUNDAMENTAL FREQUENCY/FIRST FORMANT FREQUENCY RATIO PER PITCH INTERVAL

摘要

PURPOSE: A speech waveform compressing method by similarity of F0/F1 ratio per pitch interval is provided to prevent degradation of tone quality and to obtain lifetime increase of a battery and real time processing by measuring similarity of speech waveform with F0/F1 ratio parameter to remove speech waveform redundancy. CONSTITUTION: A speech waveform compressing method by similarity of F0/F1 ratio per pitch interval comprises the steps of dividing speech waveform into constant size of frames, determining whether the divided speech waveform is voiced sound or not, and compressing the speech waveform depending on F0/F1 ratio when the speech waveform is voiced sound. The compressing is conducted only when F0/F1 ratio at the first interval is different from F0/F1 ratio at next interval by predetermined value or more. The difference does not reach the predetermined value, compressing is not conducted.