摘要
提出了一种基于听觉模型的语音特征提取方法 ,从信号的过零点获得频率信息 ,从振幅峰值和非线性压缩获得强度信息 ,并对这种方法进行了误差分析 ,证明其噪声的鲁棒性。
Introduces a method to extract speech feature based on auditory model,the frequency information of the signal is obtained by zero crossing intervals,the intensity information of the signal is obtained by a peak detector and a compressive nonlinearty.This paper also analyzes the error of this method,verifies the noisy robustness.
出处
《计算机与现代化》
2000年第3期9-13,共5页
Computer and Modernization
