语音端点检测算法研究

Research on Speech Endpoint Detection Algorithm

语音信号端点检测是音频特征提取过程中重要的环节。为了提高噪声环境下语音端点检测的正确率,克服传统的双门限法在低信噪比条件下检测性能急剧下降这一缺陷,提出了一种改进的双门限法进行端点检测。该方法首先对语音信号进行中值滤波处理,然后对双门限法的判别方法进行改进。实验结果表明,在低信噪比的条件下,改进方法比传统方法具有更高的准确率。

Speech signal endpoint detection is an important part in the process of audio feature extraction .In order to improve the accuracy of speech endpoint detection in a noisy environment, and to overcome a defect that the sharply decrease of the detection performance of traditional double threshold method under the condition of low signal to noise ratio, this paper proposes an improved double threshold method to carry out endpoint detection. Firstly, the speech signal is processed by median filter, and then the identification method of double threshold method is improved. The results of the experiment show that the improved method has a higher accuracy than the traditional method under the condition of low signal to noise ratio.