Skip to main content
. Author manuscript; available in PMC: 2022 Jan 1.
Published in final edited form as: IEEE/ACM Trans Audio Speech Lang Process. 2021 Jan 25;29:2058–2066. doi: 10.1109/taslp.2021.3054302

TABLE II.

Experimental Results for SR models Evaluated on VCTK with Downsampling Factor of 2 and 4

VCTKS VCTKM
Model R SNR LSD PESQ SNR LSD PESQ
Spline 2 19.07 1.99 3.84 18.89 2.08 3.53
DNN-BWE 2 19.04 1.40 3.85 18.80 1.38 3.56
DNN-Cepstral 2 19.89 1.25 3.85 19.09 1.34 3.59
AudioUNet 2 20.82 1.36 3.90 19.94 1.32 3.68
TFNet 2 21.11 1.24 3.91 19.84 0.99 3.72
Proposed 2 22.44 0.94 4.17 22.08 0.88 3.91
Spline 4 15.33 3.13 3.07 13.42 2.99 3.13
DNN-BWE 4 15.30 1.47 3.27 13.53 1.38 3.24
DNN-Cepstral 4 15.47 1.44 3.28 13.87 1.36 3.25
AudioUNet 4 17.29 1.41 3.40 16.65 1.40 3.39
TFNet 4 18.35 1.33 3.49 17.32 1.22 3.48
Proposed 4 18.86 0.94 3.51 18.13 0.95 3.64