Research Article

Research on Objective Evaluation of Recording Audio Restoration Based on Deep Learning Network

Table 3

The results of each index under different audio characteristics.

Feature typecorrelation coefficient RMean square error RMSE

MFCC (1-2)0.43690.6882

MFCC (1-5dimensions)0.69290.5847

MFCC (1-7dimensions)0.80150.4758

MFCC (1-13dimensions)0.83430.2799

MFCC (1-20dimensions)0.89090.2307

GFCC (13dimensions)0.91170.1924