AN ALGORITHM OF VOICE ENDPOINT DETECTION BASED ON WAVELET ENERGY ENTROPY
Pla Information · Computer Applications and Software · 2011
First,the wavelet transformation is used on vocal signals.Then the wavelet entropy reflecting energy distributing characteristics of the signal is introduced for detecting the point.Finally,the natural weighting factor is improved based on the characteristic of sonant sound.Simulation results indicate that the new algorithm outperforms the adaptive band-partitioning spectral entropy-voice activity detection(BSE-VAD) in detection result in condition of low signal-to-noise ratio.