Forensic voice comparison with secular shibboleths - A hybrid fused gmm-multivariate likelihood ratio-based approach using alveolo-palatal fricative cepstral spectra

The suitability of voiceless fricative spectra for forensic voice comparison is explored within a Likelihood Ratio-based framework. Non-contemporaneous landline telephone recordings of 99 male Japanese speakers are compared using only tokens of their voiceless alveolo-patalal fricative [ç]. A subset of mean-cepstrally-subtracted LPC CCs from the fricative spectrum from dc to 5 kHz is used. GMM/UBM and multivariate likelihood ratios are extracted for the 99 target and 4851 non-target trials, and fused with logistic regression. An EER of 7.4% and log-LR cost of 0.26 is demonstrated. It is concluded that the [ç] spectrum does have some individualising potential.