论文部分内容阅读
语音激活检测在移动通信中具有重要的意义和广泛的应用,对语音处理的效果也有关键性的影响。本文针对移动通信应用场合的特点,研究了一种计算复杂度很低的语音激活检测算法,使用短时能量、平均过零率和平均幅度差三种时域特征门限进行多级检测。该算法在大信噪比下可以获得较好的检测结果,而且实时性较好。本文介绍了算法的信号处理流程、门限计算方法、部分实现细节以及实验结果。
Voice activity detection in mobile communications has an important meaning and a wide range of applications, the effect of voice processing also has a key impact. In this paper, aiming at the characteristics of mobile communication applications, a speech activity detection algorithm with low computational complexity is studied, and the multi-level detection is performed using three time-domain feature thresholds of short-time energy, average zero-crossing rate and average amplitude difference. The algorithm can get better detection results under large signal-to-noise ratio and better real-time performance. This article describes the algorithm signal processing flow, threshold calculation method, part of the implementation details and experimental results.