Skip to main content
. Author manuscript; available in PMC: 2018 Aug 13.
Published in final edited form as: IEEE/ACM Trans Audio Speech Lang Process. 2017 Apr 20;25(7):1492–1501. doi: 10.1109/TASLP.2017.2696307

TABLE IV.

Average PESQ Scores for Each Approach When the Systems are Trained and Tested Using Multiple Speakers. Bold Score Identifies the Best Performing System, While Bold Room Indicates the One not Seen During Training

Set 1
Set 2
Room A Room B Room C Room D Room A Room B Room C Room D
Mixture 2.93 2.64 2.92 2.39 2.93 2.64 2.92 2.39
RM 3.27 2.98 3.21 2.64 3.27 3.04 3.26 2.74
cRM 3.41 3.12 3.38 2.79 3.38 3.20 3.43 2.93
PSM 3.42 3.13 3.37 2.80 3.39 3.20 3.43 2.94
WPE 3.28 2.95 3.36 2.61 3.28 2.95 3.36 2.61

Set 3
Set 4
Room A Room B Room C Room D Room A Room B Room C Room D

Mixture 2.93 2.64 2.92 2.39 2.93 2.64 2.92 2.39
RM 3.28 2.99 3.25 2.71 3.31 3.02 3.23 2.74
cRM 3.42 3.14 3.40 2.88 3.42 3.15 3.32 2.88
PSM 3.42 3.15 3.39 2.90 3.43 3.16 3.33 2.90
WPE 3.28 2.95 3.36 2.61 3.28 2.95 3.36 2.61