A Study of Acoustic Features in Arabic Speaker Identification under Noisy Environmental Conditions

TL;DR

This study evaluates the robustness of various acoustic features for Arabic speaker identification in noisy environments, finding GFCC and PNCC outperform traditional MFCC features under different noise conditions.

Contribution

It compares the effectiveness of multiple acoustic features in noisy environments for Arabic speaker identification, highlighting the superior performance of GFCC and PNCC.

Findings

GFCC and PNCC outperform MFCC in noisy conditions

Robust features improve speaker identification accuracy in noise

Performance varies with different noise types and SNR levels

Abstract

One of the major parts of the voice recognition field is the choice of acoustic features which have to be robust against the variability of the speech signal, mismatched conditions, and noisy environments. Thus, different speech feature extraction techniques have been developed. In this paper, we investigate the robustness of several front-end techniques in Arabic speaker identification. We evaluate five different features in babble, factory and subway conditions at the various signal to noise ratios (SNR). The obtained results showed that two of the auditory feature i.e. gammatone frequency cepstral coefficient (GFCC) and power normalization cepstral coefficients (PNCC), unlike their combination performs substantially better than a conventional speaker features i.e. Mel-frequency cepstral coefficients (MFCC).