基于支持向量机的乐音识别

Music Recognition Based on Support Vector Machine

乐音识别的研究早在20世纪60年代就开始了,但是直到最近,识别系统的性能仍不令人满意。本文通过研究乐音的声音和物理特性,提出了利用支持向量机进行乐音识别的新思路,给出了基本的算法框架,并对钢琴的所有单音进行了识别,实验证明该算法提高了识别的快速性和准确性。

The research for music recognition has been started form 1960s,yet recently,the performance has not been satisfied.The paper investigate the musical and physical characters of the musical sound,and presents new recognition method based on SVM,advances the basic frame of arithmetic,finally recognize the note of piano.Experimental results show the new algorithm is fast and effective.