Direct Recovery of Clean Speech Using a Hybrid Noise Suppression Algorithm for Robust Speech Recognition System

Table 5

Recognition results for comparison targets.

	Avg 0–20	Relative Imp.	SNR−5	Relative Imp.
MFCC	71.80	19.4%	13.39	108.0%
SS	72.42	18.4%	25.26	10.3%
CMVN	77.78	10.3%	13.90	100.4%
STSA	80.12	7.0%	20.31	37.1%
AFE	82.23	4.3%	24.77	12.4%
MVA	84.15	1.9%	26.24	6.1%