摘要
基于物理模型的乐器仿真技术是运用声学和力学理论,通过偏微分方程来描述乐器的振动物理特性,它不仅可以精确地描述乐器发声及发声过程,而且还可以准确地再现演奏者的弹奏力度、速度和情感,表现方法与控制仿真参数完全与真实乐器一样而无人工合成的痕迹。随着计算机计算能力的快速提高以及MPEG4的问世,这种方法日趋活跃,成为目前乐器仿真、声音合成的研究热点。本文阐述了物理模型方法中常见的五种技术,分析了它们的特性及其缺陷,阐述了各自的基本思想及实现方法,重点论述了两个著名模型的局限性,并相应地提出了我们所推导出的一维和二维物理模型。
Musical simulation technique based on physical model is a theory that is able to describe the characteristic ofobject抯 vibration by partial differential equation. It can not only simulate accurately musical instrument抯 sounds and itssounding process, but also reflect the player抯 controlling force strength and speed and emotion. Its playing methods and controlling parameters are the same as those of real musical instrument, unlike artificial synthesis sounds. With improvement of the computational power of computer, and MPEG4 technique being published, physical model methods is becoming one of the most important research domain about musical instrument simulation and sound synthesis. At first, this paper depicts five methods mostly used over physical model, analyzes their characteristics and defects, and explains the fundamental ideals and algorithms. And then, limitations of the two most famous physical models are discussed here. So, we propose a new one-dimensional and two-dimensional model according to their respective limitation.
出处
《系统仿真学报》
CAS
CSCD
2003年第7期939-943,共5页
Journal of System Simulation
关键词
乐器仿真
物理模型
局限性
偏微分方程
musical instrument simulation
physical model
limitation
PDE
