Skip to main content

Table 1 Details of recording conditions for impulse response measurement

From: Distant-talking speaker identification by generalized spectral subtraction-based dereverberation and its efficient computation

 

Array no

Room

Array type

RT60 (s)

SRR

CENSREC-4 database for training

1

Japanese style room

Linear

0.40

3.24

 

2

Japanese style bath

Linear

0.60

3.28

 

3

Elevator hall

Linear

0.75

2.98

RWCP database for test

4

Echo room (cylinder)

Circle

0.38

3.45

 

5

Tatami-floored room (S)

Circle

0.47

2.89

 

6

Tatami-floored room (L)

Circle

0.60

3.12

 

7

Conference room

Circle

0.78

3.30

 

8

Echo room (panel)

Linear

1.30

2.88

  1. RT60 (s), reverberation time in room; S, small; L, large; SRR, signal-to-reverberation ratio [42], calculated from artificial reverberant data.