文献类型：期刊文献

中文题名：一种改进的基于短时平均幅度的语音端点检测算法研究

英文题名：An Improved Speech Endpoint Detection Algorithm Analyses Based on Short-term Average Amplitude

作者：柳春[1]

第一作者：柳春

机构：[1]甘肃中医学院

第一机构：甘肃中医药大学

年份：2009

卷号：30

期号：1

起止页码：56

中文期刊名：西北民族大学学报：自然科学版

语种：中文

中文关键词：端点检测;短时平均幅度;噪声环境

外文关键词：Endpoint detection; Short - term average amplitude; Noisy environments

摘要：在噪声环境下,利用短时平均幅度为特征进行语音端点检测.文章在传统端点检测算法的基础上,研究了汉语音节的特点,提出采用短时平均幅度代替短时能量,并为平均幅度引入判决门限.门限值是根据语音信号背景噪声自动计算得到,从而保证了算法在噪声环境下检测的准确性.实验结果表明,与传统的基于短时能量的端点检测算法相比,改进的算法在高信噪比和低信噪比环境下都具有良好的性能.
The speech endpoint detection was analyzed in this paper based on short- term average amplitude feature in the presence of noise. Besides short- term energy feature, to solve the imperfection of traditional endpoint detection algorithm, and with the features of the mandarin, taking into account the amount of computing systems and the fact that short- term energy is too sensitive to high energy singles with a computed threshold to amend endpoint detection algorithm. Experiments show that the proposed algorithm outperforms traditional energy algorithm for speech endpoint detection in noisy environments.