Compensation of SNR and noise type mismatch using an environmental sniffing based speech recognition solution

EURASIP Journal on Audio, Speech, and Music Processing

Table 1 SNR of reference HMM showing lowest WER as the SNR of test speech was varied

SNR of test speech	SNR of reference HMM showing the lowest word error rate
	Babble	Subway	Car	Exhibition
0	6	4	2	2
2	6	6	4	4
4	8	6	6	6
6	10	8	8	8
8	10	8	8	10
10	12	12	12	10
12	16	10	14	12
14	18	14	14	16
16	18	16	20	18
18	18	18	18	18
20	20	22	20	20
22	26	22	26	26
24	28	22	28	28
26	28	22	30	28
28	28	22	30	30
30	30	24	30	30