摘要
提出了一种高效的乐曲演奏者识别系统。该系统主要由节拍点时间序列检测和速度节奏模型分析两部分组成。节拍点检测过程首先通过离散高斯模板滤波得到信号包络,进而利用音乐节拍特性快速计算出速度节奏模型。根据不同钢琴家演奏风格的差异性,对速度节奏模型进行统计分析,准确提取出各钢琴家的速度和节奏特点。实验结果表明,在低运算量的前提下,将该算法应用于分析肖邦夜曲,可以对曲目的演奏者进行准确辨识。
An efficient performer identification system is presented in this paper. This system mainly consists of beat detection part and Tempo - Rhythm model analysis part. In the beat detection procedure, a discrete Gaussian filter is a- dopted to obtain the sound envelope, and then the Tempo - Rhythm model is rapidly worked out by utilizing the features of music metres. According to the various playing styles of different pianists, their Tempo - Rhythm chariacteristics are precisely extracted through statistically analyzing the Tempo - Rhythm models. The experimental results show that while maintaining low computational complexity, the proposed system can acurately identify the performer of Chopin＇ s Noc- tunes.
出处
《电声技术》
2013年第12期85-88,共4页
Audio Engineering
关键词
音频分析
节拍检测
速度分析
演奏者识别
audio analysis
beat deteetion
tempo analysis
performer identification
