摘要
音阶识别方法不能结合乐音的物理和音乐特性,造成乐音处理信号量杂乱无序。针对该问题,通过复音音高确定转录音符集合,引入频谱图、谱平滑性和调和性估计建立HMM模型。分析对比实验结果表明,HMM模型可以完成乐音物理与音乐特性结合,改善乐音处理信号量杂乱无序现象。
In the musical scale recognition method,physical and musical features of the musical tone cannot be combined,which causes randomness of music tone processing signals.To resolve this problem,the notes set is determined and transcribed by using complex tone pitches.Diagram of spectrum and estimation of spectrum smoothness and harmonicity are introduced to construct the HMM model.The analysis results of comparison experiment show that the HMM model can complete the combina-tion of musical tonesr physical and musical features,and improve the randomness phenomenon of music tone processing signals.
作者
范凌云
周婧
FAN Lingyun;ZHOU Jing(Jiangxi Agricultural University,Nanchang 330045,China)
出处
《现代电子技术》
北大核心
2018年第8期160-162,166,共4页
Modern Electronics Technique
基金
江西省教育厅课题(YS1106)
江西省教育科学规划办公室课题(11YB247)~~
关键词
音阶识别
合理性模式
校对方法
HMM模型
谱平滑性
乐音处理
musical scale recognition
rationality pattern
proofreading method
HMM model
spectrum smoothness
music tone processing
