【摘要】 对大量的汉语普通话的语音波形进行分析之后发现汉语普通话的基音提取不能用简单的中心削波自相关方法 ,在进行基音标记时会存在一种基音标记偏移的现象 ,为此在自相关理论的基础上 ,根据声调语言的语音信号音高变化特征 ,对声调语言的基音标记的偏移现象提出了一种基于自相关法和峰值校正的算法 .经实践检验 ,这种算法对声调语言的基音标记的偏移现象能够得到很好的校正 .更多还原

【Abstract】 After analyzing many speech waveforms of Chinese Putonghua syllables,we find out that only using autocorrelation method cannot detect pitch exactly because of the existence of the deviation of pitch marks on tonal language.In this paper,according to the pitch characteristic of tonal language,we propose a new method of pitch detecting based on autocorrelation method and peak correction to solve the deviation of pitch marks on syllables.The results in practical application imply that the algorithm has advantages in pitch detecting of Chinese syllables.更多还原