球面波表合成模型之實作與探討

Abstract

本論文介紹的是一個新型態的聲音合成模型-球面波表合成模型,並且結合動態影像,建構一套表現在三維度空間中的即時數位音像互動環境。其運算核心是由即時影音互動程式 Pure data 裡的 pmpd 完成。pmpd 是一個物件,也是一套系統,藉由其內部程式做物理建模的運算,建立出質點和鏈結的互動機制,且以質點彈簧系統做低速的物理運動行為。另外,運用波域合成法(Wave Terrain Synthesis)和掃描合成法(Scanned Synthesis)在掃描波形的特性。以有限元素的概念將質點和鏈結排列成一個球面,表現在球座標上,並且實作出一套即時的控制介面。透過聲音與影像生成機制的模組化,賦予聽覺與視覺同步更直覺的意義。

In this thesis, Sphere Wavetable Synthesis (SWS) model, a new kind of sound synthesis model is introduced. This model provides a real-time digital audio-visual interactive environment in a three-dimensional space. The computational framework of SWS model is based on the pmpd object in Pure Data, a real-time interactive programming language. The pmpd object is able to conduct physical modeling simulations, generating low-speed physical movements of an equivalent mass-spring system. In SWS model, the concepts of Wave Terrain Synthesis and Scanned Synthesis were adopted to obtain dynamic sounds. Masses and links were arranged to construct a virtual sphere, rendering users a real-time interactive interface for sound manipulation. This model aims to modulate the generating mechanism of sounds and images, providing users a more intuitive experience of the audio-visual synchronization.