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

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