说明：  针对低信噪比环境下基于能零比的端点检测方法准确率低的问题,提出了一种结合谱减法和短时零熵法的端点检测方法。首先通过功率谱减法对带噪语音进行降噪处理,然后结合短时平均过零率特征与谱熵特征,构造出一种新的语音参数,即短时零熵,最后运用双门限算法进行端点检测。实验结果表明该方法在复杂噪声环境下具有良好的端点检测性能,降低了检测的误判率。
(To solve the problem of low accuracy of endpoint detection method based on energy zero ratio in low SNR environment, an endpoint detection method combining spectral subtraction and Short-term zero entropy method is proposed. Firstly, the noise is reduced by power spectrum subtraction, then a new speech parameter, short-time zero entropy, is constructed by combining the short-time average zero crossing rate and spectral entropy. Finally, the double threshold algorithm is used for endpoint detection. The experimental results show that this method has good performance of endpoint detection in complex noise environment, and reduces the rate of detection error.)