论文部分内容阅读
In this paper, a new method for making v/uv decision is developed which uses a multi-feature v/uv classification algorithm based on the analysis of cepstral peak, zero crossing rate, and autocorrelation function (ACF) peak of short-time segments of the sp