從樂譜到聲音的過濾及音色的製作

Abstract

本論文的目的是利用樂譜作為控制器及數位訊號處理技術來過濾噪音及製 造音色。本論文主要介紹兩大系統：隨著時間不斷變化的噪音過濾系統及 隨著時間不斷變化的音色製作系統。在隨著時間不斷變化的噪音過濾系統 中，本論文提出一新方法對噪音作各種不同方式的過濾。樂譜上的每一個 音對應ㄧ種過濾器，且在其音長的時間流程中會逐漸轉變成其下一個音所 對應的過濾器。此系統不僅可以產生特殊的過濾器，更可產生隨著時間不 斷變化的過濾器。在隨著時間不斷變化的音色製作系統中，本論文提出一 新方法來決定樂譜中每一個音的頻譜或波形。樂譜上不同的音對應不同的 頻譜或波形，且在其音長的時間流程中會逐漸轉變成其下一個音所對應的 頻譜或波形。此系統不僅可以產生特殊的音色，更可產生隨著時間不斷變 化的音色。由此兩系統所產生有趣的聲音及音色證明所提出方法的可行性 。 The aim of this thesis is to use musical scores as controllers and digital signal processing technology to filter noise and create timbre. This thesis has two main parts : (1) the time- varying noise filtering system and (2) the time-varying timbral creation system. In the time-varying noise filtering system, a new way of deriving different kinds of filters for noise is proposed. Every note in a musical score corresponds to one style of filter. Over the duration of each note, its filter is gradually transformed into that corresponding to its following note. This system produces not only unusual filters but also automatic time-varying filters. In the time-varying timbral creation system, a new way of deriving the spectrum or wave for every note in a musical score is proposed. Over the duration of each note, its spectrum or wave is gradually transformed into that for its following note. This system creates not only unusual timbres but also automatic time-varying timbres. Interesting sounds and timbres created by these two systems prove the feasibility of the proposed approaches.