Skip to main content

Table 4 Calculation results in the noise mismatching condition when IRM is the training target

From: Research on monaural speech segregation based on feature selection

Feature and feature group

STOI

PESQ

SDR

SNR

AMS

0.7538

2.2349

8.2144

8.1006

AMS + GF

0.8117

2.5034

10.4221

10.1191

AMS + GF + MRCG

0.8178

2.5258

10.6274

10.3807

AMS + MFCC

0.8108

2.4820

10.2555

10.0639

AMS + MFCC + GF

0.8160

2.5220

10.5311

10.2775

AMS + MFCC + GF + MRCG

0.8193

2.5349

10.5937

10.3745

AMS + MFCC + MRCG

0.8184

2.5227

10.5991

10.4131

AMS + MRCG

0.8168

2.5169

10.5676

10.3178

GF

0.8092

2.4863

10.2725

9.8783

GF + MRCG

0.8163

2.5130

10.4567

10.2289

MFCC

0.8077

2.4494

10.0376

9.7046

MFCC + GF

0.8133

2.5225

10.5097

10.1582

MFCC + GF + MRCG

0.8164

2.5303

10.5790

10.3323

MFCC + MRCG

0.8151

2.5228

10.5821

10.1904

MRCG

0.8131

2.5036

10.3848

10.0971

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