摘要
Abstract
In this paper, a method is proposed to calculate pitch period which is used for emotional recognition. Speech signal endpoints - detection is presented by means of combining with the average short - term energy and average short - term zero - crossing rate, and then the pitch period of voiced frame is calculated by autocorrelation function. In this way, male and female voices can be distinguished accurately,and people's different moods can be identified combined with pronunciation characteristics. This method is simple and convenient, providing a good way of speech emotion recognition.关键词
语音情绪识别/基音周期/基音检测/语音端点检测/自相关函数Key words
speech emotion recognition/ pitch period/ pitch detection/ speech signal endpoints -detection/ autocorrelation function分类
信息技术与安全科学
