AN AUTOMATIC SONG TRANSCRIPTION METHOD AND APPARATUS THEREOF

摘要

PURPOSE: An automatic song transcription method and an apparatus thereof are provided to automatically detect a measure after automatically recognizing a sound field and a tune from a voice signal. CONSTITUTION: A pitch showing the height of a voice is extracted from a voice signal inputted through a voice signal input unit(S200). A continuous signal based on the extracted pitch data is divided by a syllable unit(S300). A representative pitch value of syllable is provided. The divided syllable is converted into effective unit sections capable of sound field recognition(S400). The processed sound field data are clustered by each note to recognize a sound field(S500).