论文部分内容阅读
A new method was proposed to identify speech-segment endpoints based on the empirical mode decomposition(EMD) and a new wavelet entropy ratio with improving the accuracy of voice activity detection.With the EMD, the noise signals can be decomposed into se