以音域為基礎的自動歌曲推薦系統

Abstract

卡拉OK為現代人生活不可或缺的娛樂之一：一般人在KTV中常發生高音會破音、低音下不去的窘況。市面上出現許多評分系統利用音高、音量等資訊幫受試者評分，雖有助於受試者了解該首歌演唱情況，但並未提供受試者有關音域及適合歌曲的資訊。本論文應用於卡拉OK上，主要透過男性受試者的演唱的音高資訊來找出最佳的音域範圍，並從音域資料庫中推薦符合受試者音域的歌曲清單。

Karaoke is one of the indispensable entertainments in modern life. There are many singing performance evaluation systems using the information of pitch and volume to score a singer. These systems may be helpful to let a singer realize his/her vocal performance roughly. However, she or he can not use these information to choose songs suitable for her/his vocal range. In this thesis, an application for Karaoke is implemented to analyze a male singer’s vocal range using pitch information, and then recommend a song list with suitable vocal range. This may be convenient for general karaoke users.