摘要 |
PURPOSE:To conduct an efficient data processing by greatly reducing the amount of voice data. CONSTITUTION:Analog voice signals are inputted to a sampling circuit 3 through a microphone 1 and digital voice data are extracted at intervals of a prescribed time. A voice data processing section 4 successively reads the voice data and extracts voice waveforms having a prescribed pitch period. Moreover, the section 4 sets the voice data of voice waveform as a template for the period when similar sound waveforms are periodically repeated and calculates the degree of similarity to the template with respect to the voice data of the succeeding voice waveforms. The templates and each of the degree of similarity are stored in a voice data storage section 5 as the compressed data of voice data for the period above. |