PROBLEM TO BE SOLVED: To detects a voice included in an inputted sound more accurately.;SOLUTION: An acoustic analysis part 120 acoustically analyzes the sound inputted from a microphone 110 and a speech detection part 130 obtains feature quantities (frequency, power, etc.) needed for speech detection etc. Then the speech detection part 130 detects the start point and end point of a speech based upon speech detection standard data stored in a data storage part 170. Here, data which are different by phonemes (or phoneme groups) are stored as the speech detection standard data and the speech detection part 130 detects the start point and end point of each phoneme (phoneme group) by using detection standards suitable for detection of the phonemes etc.;COPYRIGHT: (C)2005,JPO&NCIPI
展开▼