复杂声音信号基频检测和处理的一种方法及实例

周期性信号基频的检测具有重要的意义，通常的硬件检测方法是采用过零比较器或施密特触发器等电路对输入信号进行整形，从而获得基频脉冲。当信号波形复杂时，这种检测方法就会失准。本文提出了一种新的硬件检测方法有效转变点实时基频检测法，它用两个适当电平切割被测信号，从而获得有效的零点，解决了信号复杂时过零点增多的现象，文内并给出了一个检测人声基频进而实现人声向乐器声转化的应用实例。

Detection and processing of the fundamental frequency of complicated signal and an illustration of its application

The fundamental frequency detection of a periodic signal is of great significance. Often, after reshaping the input signal using a zero comparator, the frequency count is accomplished. When the signal waveform is complicated, however, this kind of detection will be inaccurate. The paper proposes a new hardware detection method-real- time detection by an effective change point. The method cuts the signal examined with two appropriate levels and thus gets the effective zero point. It solves the problem of measuring of the zero points when the signal is complicated. An example of application is provided that realizes the transformation from human voice to the sound of musical instrument after detecting the fundamental frequency of the human voice.