Abstract: In constructing a speech corpus, many unexpected silence symbols are inserted. They directly affect the results of automatic segmentation. In this paper a automatic silence insertion algorithm is proposed. The effectiveness of the proposed method has been proven through experimental results and the accuracy of silence insertion is more than 95%.

摘要： 在建立语音语料库的过程中经常会碰到语音中加了很多对应文本所没有的静音标记，这些静音的存在直接影响到实际的切分结果。该文提出了一种静音自动添加的方法。在经过粗切分后，通过规则设定，短时能量、短时过零率和算法修正几个步骤，将静音寻找出来，添加到对应文本之中。该方法已运用到实际切分中，准确率在95%以上，取得了较好的效果。

关键词: 静音, 自动切分, 隐马尔可夫模型

CLC Number: 

• TN912.3