The present invention relates to a method and an apparatus for discriminating silent and unvoiced sounds of a voice signal and, in order to easily distinguish and separate silent and unvoiced sounds in a voice signal including both silent and unvoiced sounds mixed with a noise component, An arbitrary value between the voltage level of the silence and the voltage level of the unvoiced sound in the audio signal is set as the reference voltage level, the pitch component of the audio signal waveform is detected, and the absolute value of the detected pitch component voltage level is compared with the reference level, According to the result, the voice signal is separated and output.;A waveform divider for dividing a waveform of the audio signal into units of each period; A level modulator for modulating a level of a voice signal waveform of each cycle unit divided by the waveform divider to remove a DC component included in a voice signal waveform; A pitch detector for detecting a voltage level corresponding to each pitch component of the audio signal waveform modulated by the level modulator; A comparator for comparing the absolute value of the pitch component voltage level detected by the pitch detector with an initial reference voltage level; And a switch for selectively switching the voice signal waveform of each cycle unit divided by the waveform divider according to the comparison result of the comparator.;(Fig. 5)
展开▼