SPEECH RECOGNITION METHOD AND DEVICE, SPEECH SYNTHESIS METHOD AND DEVICE, RECORDING MEDIUM

摘要

Provided are a data generating unit (3) that detects as a sample point a point at which a differential value obtained by differentiating an input speech signal satisfies a specified condition, and obtains discrete amplitude data at respective detected sample points and timing data representing time intervals between respective sample points, and a correlation computing unit (4) that computes correlation data by using the amplitude data and the timing data, whereby an input speech is recognized by matching correlation data generated for a certain input speech by the correlation computing unit (4) with correlation data similarly generated in advance for various speeches and stored in a data memory (6).