摘要 |
PURPOSE:To accurately decide whether a monophonic or stereophonic acoustic signal which is successively inputted consists principally of a speech or music in a short time. CONSTITUTION:A power calculation part 1 calculates a power value, frame by frame, a voiced sound/voiceless sound decision part 2 decides a voiced or voiceless sound from the power value, frame by frame, and a decision part 3 decides a speech or music from the presence ratio of the numbers of voiced sound and voiceless frames in the time, frame by frame, and holds the decision result of plural precedent frames in case of a failure in decision making. |