A Hierarchical Framework Approach for Voice Activity Detection and Speech Enhancement

Table 2

Accuracy of the proposed VAD algorithm at different SNR levels.

Noise	SNR
	0 dB	5 dB	10 dB	15 dB	20 dB
White noise	53.3%	61.7%	67.1%	94.5%	97.8%
Buccaneer noise	67.6%	75.4%	84.1%	90.8%	96.5%
Babble noises	84.2%	88.3%	93.2%	94.7%	99.1%
Factory noise	61.9%	69.5%	82.9%	97.5%	98.0%
Average	66.8%	73.7%	81.8%	94.4%	97.8%