基于短时傅里叶算法的行人警示控制器音效合成

Pedestrian Warning Controller Sound Synthesis Based on Short-time Fourier Algorithm

文章首先介绍了美国、日本、欧盟、中国等国对于安静行驶车辆低速提示音的研究情况及相关标准法规的发展历程,并在简述音效应用背景的基础上,指出声源提取的必要性。提出了使用短时傅里叶分析与合成技术以及使用线性预测分析与合成技术从发动机怠速的原始音源文件中提取声源的方法。讨论了该方法的主要理论及提取过程,并分别用实例阐述如何使用短时傅里叶分析与合成技术从原始音源文件中提取发动机怠速音效。通过实例以及实际开发表明,该技术适用于大部分声源的提取,可推广到其他类型音效系统的开发中,具备较大的工程应用价值。

In this paper the research situation of low-speed sounds of quiet vehicles and the development of relevant standards and regulations in the United States,Japan,the European Union,China and other countries are first introduced. On the basis of a brief background of sound effects,the necessity of sound source extraction is pointed out. A method for extracting a sound source from an original sound source file of an engine idle speed using short-time Fourier analysis and synthesis techniques and using linear predictive analysis and synthesis techniques is proposed. The main theory and extraction process of the method are discussed,and examples are given to illustrate how to extract engine idle speed from the original source files using short-time Fourier analysis and synthesis techniques. The examples and actual development show that the technology is more versatile and suitable for the extraction of most sound sources,which can be extended to the development of other types of sound effects systems,and has great engineering application value.