Skip to main content

Table 11 Denoising and dereverberation performance on test dataset

From: Channel and temporal-frequency attention UNet for monaural speech enhancement

 

WB-PESQ

STOI(%)

SI-SDR

DNSMOS

Noisy

1.295

78.683

3.250

1.384

Uformer

1.996

86.851

7.745

2.698

MTFAA

1.792

84.869

5.391

2.510

CTFUNet

2.164

88.645

8.008

2.805

 

SRMR

CD

LLR

SNR\(_{fw}\)

Noisy

3.149

4.911

0.870

9.814

Uformer

6.340

3.350

0.672

9.407

MTFAA

5.333

3.270

0.476

9.818

CTFUNet

5.948

2.738

0.346

11.717