匹配追踪时频分解算法的端点检测方法

Speech detection using matching pursuits algorithm in time-frequency domain

为了能在无噪音环境下准确地检测语音信号的端点,传统的方法是使用过零方法检测清音,短时能量方法检测浊音,两者相结合便实现了端点检测。通过对语音信号在时频平面中分布的研究,提出了一种基于匹配追踪时频原子分解算法的端点检测方法。该方法利用匹配追踪算法对信号进行分解,使得信号在时频平面上具有较直观明显的魏格纳能量分布,利用这个特点设置一个门限值再进行端点检测,便能准确检测出语音信号端点。实验结果表明,和传统的方法相对比,因为涉及到了信号的分解,所以实时性较差,且门限问题还有待深入研究,但该方法能更加准确地检测出语音信号的端点,亦为端点检测问题提供了一种新的思维方法。

An algorithm of speech detection suit decomposition is proposed. With in the time-frequency domain based on matching pursuits, decomposed time-frequency matching purenergy of the signal has a clear distribution in the time-frequency domain. Using this characteristic, the endpoint can be differentiated by setting a threshold value. Compared with the conventional method that uses zero crossing ratio to detect surd and short-time energy to detect sonant, although the real-time capability is inferior to the conventional method, the endpoint can be detected accurately. The accuracy is much better than the conventional method in an environment without noise.