Skip to main content

Table 2 Calculation results in noise mismatching condition when IBM is the training target

From: Research on monaural speech segregation based on feature selection

Feature and feature group

STOI

PESQ

SDR

SNR

AMS

0.7517

2.2208

8.3039

8.2443

AMS + GF

0.8119

2.4371

10.5176

10.2655

AMS + GF + MRCG

0.8148

2.4167

10.6053

10.4037

AMS + MFCC

0.8088

2.4427

10.3002

10.0512

AMS + MFCC + GF

0.8132

2.4169

10.5283

10.3024

AMS + MFCC + GF + MRCG

0.8166

2.4006

10.5186

10.3628

AMS + MFCC + MRCG

0.8149

2.3970

10.6519

10.4322

AMS + MRCG

0.8127

2.4184

10.3728

10.2319

GF

0.8076

2.3951

10.2246

9.8945

GF + MRCG

0.8108

2.3995

10.4422

10.1793

MFCC

0.8067

2.4295

10.1689

9.7345

MFCC + GF

0.8096

2.4206

10.4220

10.0669

MFCC + GF + MRCG

0.8111

2.3881

10.5219

10.2396

MFCC + MRCG

0.8108

2.3769

10.6505

10.2960

MRCG

0.8110

2.3964

10.3754

10.1218

  1. Bold data represents the optimal value of each feature group for this indicator