| 摘要 |
PURPOSE: A high-speed pitch searching method in a voiced sound section is provided to perform search only around a pitch of a prior frame, if a frame decided as a voiced sound is continued in a voice coder, so as to reduce calculation through the variation of the pitch. CONSTITUTION: A voice codec of a voice coder is initialized(201). If a voice signal of one frame is inputted(202), VAD(Voice Activity Detection) by the frame is performed(203). If the VAD is completed, information on the activity of the current frame and information on a pitch of a prior frame are used to decide a pitch search mode(204,205). The voice signal is coded using the decided pitch search mode(206), and the voice activity and pitch data of the current frame are stored by the coded information(207). Whether the current voice signal is a final signal is decided(208). If so, the steps are ended, and if not, the step of 202 is returned.
|
