低信噪比条件下的一种自适应有声/无声判决算法

An Adaptive Method of Speech/Nonspeech Decision in Low SNR Environments

本文描述了一种利用含噪语音短时能量进行有声／无声判决的自适应算法。通常，利用短时能量进行有声／无声判决时，均采用一固定门限，但是，恰当的判决门限显然是噪声统计特性及信号能量的函数。本文提出了一种估计含噪语音短时能量概率密度函数，并根据所期望的误判率估计判决门限的算法。本算法无需预先给出噪声统计信息，且适用于缓变的非平稳噪声情况。

This paper describes an adaptive method of speech/nopspeech decision using the short-time energyas a threshold. Usually, a comstant threshold is adopted to which the short-time energy is compared forspeech nonspeech decision The optimal thshold is however,a function of noise characteristics and the energy of signal.This paper proposes a method of estimating the probability density function of shor-time energy from nisy speech and also a method of estimating the thresholed based on the expected error rate of the speech nonspeech decision This method needs not a priori information abut noise characteristics ofr noise level,