摘要
时域分析法是一种基本的语音分析方法。由于钢琴乐音模型与语音模型相似,故语音时域分析法可被用于钢琴乐音分析中。提出一种基于频率特征及动态窗长的钢琴基音识别时域算法,该算法能根据检测目标设计更为合适的分帧加窗模式,经过实验,可有效地提高检测精度及速度,在钢琴调音及单乐音基音识别应用中具有推广意义。
Time domain processing is a basic method of speech analysis.Because the model of piano musical tone is similar to the speech model,the speech time-domain analysis can be used in the analysis of piano musical tone.This paper puts forward a time-domain algorithm of piano pitch recognition based on the frequency characteristics and dynamic window,this algorithm can design more appropriate mode of frame and window according to the detect targets,and also can effectively improve the detection accuracy and speed through the experiments,this method has popularization in the applications of piano tuning and pitch detection of single musical tone.
出处
《信息技术》
2011年第5期155-156,160,共3页
Information Technology
关键词
基音检测
短时自相关
动态窗长
频率特征
时域分析
pitch detection
short-time autocorrelation
dynamic window
frequency characteristics
time domain analysis