METHOD FOR ANALYZING AUDIO SIGNAL AND COMPUTER

摘要

PROBLEM TO BE SOLVED: To enable information retrieval, confirmation of the summary of a voice stream, point grasping and summarization by indexing the voice event of an interval including the time sequential segment sequence of an audio signal. SOLUTION: The voice stream is received from a voice engine (B30). The stream is divided into continuous intervals time sequentially for 2 sea or the like (B31). The time sequence of the interval is classified into defined meta- pattern categories (B34). A selected part in the voice stream is processed by a voice recognizing engine (B36). The voice recognizing engine transforms the part of the voice stream processed thereby to a text. The alternative word of a word in the text is referred to by the voice recognizing engine (B38). While using the word and the alternative word for the stream, indexing is performed together with weighting for following retrieval in an information retrieval system or the like (B40).