计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (18): 198-200.

• 图形、图像、模式识别 • 上一篇    下一篇

基于CRFs和MPCP特征的和弦识别研究

王 峰,张雪英,李炳男   

  1. 太原理工大学 信息工程学院,太原 030024
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-06-21 发布日期:2011-06-21

Research of chord recognition based on MPCP and CRFs

WANG Feng,ZHANG Xueying,LI Bingnan   

  1. College of Information Engineering,Taiyuan University of Technology,Taiyuan 030024,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-06-21 Published:2011-06-21

摘要: 和弦识别是音乐调式分析和自动标注的基础,同时在分析音乐的结构和旋律方面有着非常重要的作用。结合音乐理论和信号处理知识,提出一种基于MPCP(Mel Pitch Class Profile)特征和CRFs(Conditional Random Fields)模型的和弦识别方法。利用短时傅里叶变换(STFT)对音乐信号进行时频变换,定义了一种新的MPCP特征,最后用CRFs对和弦进行识别。实验结果表明,提出的方法在识别率上优于其他方法,具有一定的潜力。

关键词: 和弦识别, 音级轮廓, Mel音级轮廓, 条件随机场

Abstract: Chord recognition is always an important theoretical construct of analysis key and automatic label,it plays role of understanding the structure,transcribing the note,even to identify melody in music.This paper proposes a new chord recognition approach based MPCP(Mel Pitch Class Profile) and CRFs(Conditional Random Fields) by music theory and signal processing.It applies STFT to get a spectrum from music,and generates MPCP as a new feature,finally recognizes chord by CRFs.The experimental results show that the new method is better than others,and prove our strategy to be quite promising.

Key words: chord recognition, Pitch Class Profile(PCP), Mel PCP, Conditional Random Fields(CRFs)