基于语音情感特征参数的选择和提取
The selection and extraction of emotion characteristic parameters Based on German speech
摘要
情感语音之所以能够被听者区分说话人表达不同的情感，是因为其中包含了表达情感信息的特征参数，情感的变化通过特征参数的差异而体现，因此在语音情感识别的研究中，语音情感特征参数的选择和提取对于识别的效果具有极其重要的意义，如果所提取的特征不具有代表性，则无论识别系统有多完善都不会得到很好的识别效果；反之，则就可以事半功倍。本文主要介绍了共振峰和MFCC的提取方法以及通过Matlab提取了德语情感语音的MFCC和四种情感语音的共振峰。
Abstract
Emotional speech is able to distinguish between people who express different emotions, it is because the expression contains one of the characteristic parameters of emotional information, emotional changes in the parameters through the differences in the characteristics of the embodiment of emotion in speech recognition research, emotional speech characteristic parameters of selection and extraction for the identification of the effect is extremely important, if the extracted characteristics are not representative, no matter how perfect identification system will not be very good recognition results; the other hand, it can be a multiplier. This paper mainly introduces the extraction method of formant and MFCC and extracted MFCC of German emotional speech and formant of four emotional speech by Matlab.关键词
情感语音/特征参数/共振峰/MFCCKey words
emotion speech/characteristic parameters/formant/MFCC引用本文复制引用
颜才柄,王虹.基于语音情感特征参数的选择和提取[EB/OL].(2009-03-12)[2026-04-02].http://www.paper.edu.cn/releasepaper/content/200903-402.学科分类
自动化技术、自动化技术设备/计算技术、计算机技术
评论