Analysis and Transynthesis of Musical Sound with Application to Erhu

• Main Authors: Yi-Song Siao, 蕭義崧
• Other Authors: Wen-yu Su
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/54860515538310191219

碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 93 === 　Synthesizing musical sound plays an important role on modern music composition for decades. Music composer nowadays can easily take advantages of powerful and user-friendly personal computers to produce his/her desired musical sound. A lot of music synthesis researches focuses on synthesizing a single note of an instrument. However, the bowed stringed instruments are very different from other instruments in their expression. This kind of expression is not achievable by single-tone synthesis. In this article, we try to analyze the correspondent recordings to describe more exquisite synthetic parameters which allow us to reproduce the varying expressions. In our experiments, the instrumental recordings were carefully analyzed and extracted several appropriate synthetic parameters are extracted. Among those parameters, detecting and tracing the fundamental frequency (pitch) is one of the most difficult problems. The pitch detection is a classical issue in the audio processing field. In this article, we consider some characteristics of Erhu and propose an accurate pitch estimation method. 　Since the sound bank recording the characteristics of musical instruments is independent of the melody synthetic parameters, we can synthesize the musical expressions by the melody parameters in cooperated with the sound banks. In other words, it seems that we play the original melody with a different instrument and keep the original musical expressions.