The Individual and the System: Assessing the Stability of the Output of a Semi-automatic Forensic Voice Comparison System

Abstract

Semi-automatic systems based on traditional linguisticphonetic features are increasingly being used for forensic voice comparison (FVC) casework. In this paper, we examine the stability of the output of a semi-automatic system, based on the long-term formant distributions (LTFDs) of F1, F2, and F3, as the channel quality of the input recordings decreases. Crossvalidated, calibrated GMM-UBM log likelihood-ratios (LLRs) were computed for 97 Standard Southern British English speakers under four conditions. In each condition the same speech material was used, but the technical properties of the recordings changed (high quality studio recording, landline telephone recording, high bit-rate GSM mobile telephone recording and low bit-rate GSM mobile telephone recording). Equal error rate (EER) and the log LR cost function (C llr) were compared across conditions. System validity was found to decrease with poorer technical quality, with the largest differences in EER (21.66%) and C llr (0.46) found between the studio and the low bit-rate GSM conditions. However, importantly, performance for individual speakers was affected differently by channel quality. Speakers that produced stronger evidence overall were found to be more variable. Mean F3 was also found to be a predictor of LLR variability, however no effects were found based on speakersÕ voice quality profiles.

The effect of sampling variability on systems and individual speakers in likelihood ratio-based forensic voice comparison

“…Other studies have looked into the speaker-discriminatory power using suprasegmental variables, e.g. Cantonese tonal f0 (Rose & Wang, 2016), long-term f0 distribution (Kinoshita et al, 2008), voice quality (Hughes et al, 2019), and articulation rate (Lennon et al, 2019). In general, there is the tendency in the literature that higher formants outperform lower formants.…”

Analysis of Voice Changes in Anti Forensic Activities Case Study: Voice Changer with Telephone Effect

“…Hence, for the Anti-Forensic tool, it cannot be concluded that the Anti-Forensic tool is more effective on male gender. Additionally, system performance is dependent on the voice of the person being tested, because several people have different variations of the voice [21].…”

ARVMex: A web application for obtaining acoustic reference values for forensic studies in Spanish

“…Typically, recordings of offenders are taken over telephone lines, while samples of suspects are obtained in face-to-face recordings. Several studies have shown the effects on fundamental frequency and vocalic formants caused by telephone transmission [9][10][11][12][13][14][15][16]. However, none of these studies have reported the level of impact of telephone transmission on the spectral characteristics of the speech signal in particular in Spanish vowels.…”