Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing

Table 7 WER (%) of ESniff MMSE (M = 4) method for Aurora 2 task using FE feature vectors

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

SNR (dB)

Set A

Set B

Set C

 

Noise type

Noise type

Noise type

 

Sub.

Bab.

Car

Exh.

Ave.

Res.

Str.

Air.

Sta.

Ave.

Sub.

Str.

Ave.

20

1.54

1.65

1.61

1.98

1.69

1.61

2.38

1.67

2.11

1.94

2.16

2.64

2.40

15

2.65

2.51

1.94

2.88

2.49

3.73

3.43

2.73

3.97

3.46

3.12

4.68

3.90

10

3.59

4.40

3.77

5.51

4.32

7.92

6.51

5.81

7.20

6.86

5.65

7.61

6.63

5

7.03

11.11

7.46

8.57

8.54

19.89

13.98

13.14

16.32

15.83

11.17

15.62

13.39

0

18.95

34.57

19.46

20.58

23.39

47.02

33.33

31.50

37.86

37.43

26.51

35.26

30.88

Ave.

6.75

10.85

6.85

7.90

8.09

16.03

11.93

10.97

13.49

13.11

9.72

13.16

11.44

Back to article page