Comparison of three different VAD algorithms for the robust detection of speech auditory brainstem responses: efficient performance of a novel peak-valley detection based approach
Voice Activity Detection (VAD) problem considers detecting the presence of speech in a noisy signal. The speech/non-speech classification task is not as trivial as it appears, and most of the VAD algorithms fail when the level of background noise increases. In this research we are presenting a new technique for Voice Activity Detection (VAD) in EEG collected brain stem speech evoked potentials data [7, 8, 9]. This one is spectral subtraction method in which we have developed our own mathematical formula for the peak valley detection (PVD) of the frequency spectra to detect the voice activity . The purpose of this research is to compare the performance of this SNR based PVD (SNRPVD) method over Zero-Crossing rate detector  and statistical analysis based algorithms . We have put into application of these three algorithms on these particular data sets of this experiment [7, 8, 9] and VAD is verified and compared the results of these three. MATLAB routines were developed on these particular methodologies. Finally we concluded that the method of SNRPVD surely performing better than the ZCR and statistical algorithms.