Cantonese forensic voice comparison with higher-level features: likelihood ratio-based validation using F-pattern and tonal F0 trajectories over a disyllabic hexaphone

Abstract

A pilot experiment relating to estimation of strength of evidence in forensic voice comparison is described which explores the use of higher-level features extracted over a disyllabic word as a whole, rather than over individual monosyllables as conventionally practiced. The trajectories of the first three formants and tonal F0 of the hexaphonic disyllabic Cantonese word daihyat 'first' from controlled but natural non-contemporaneous recordings of 23 male speakers are modeled with polynomials, and multivariate likelihood ratios estimated from their coefficients. Evaluation with the log likelihood ratio cost validity metric Cllr shows an optimum performance is obtained, surprisingly, with lower order polynomials, with F2 requiring a cubic fit, and F1 and F3 quadratic. Fusion of F-pattern and tonal F0 results in considerable improvement over the individual features, reducing the Cllr to ca. 0.1. The forensic potential of the daihyat data is demonstrated by fusion with two other higherlevel features: the F-pattern of Cantonese /i/ and short-term F0, which reduces the Cllr still further to 0.03. Important pros and cons of higher-level features and likelihood ratios are discussed, the latter illustrated with data from Japanese, and two varieties of English in real forensic casework.

The effect of sampling variability on overall performance and individual speakers’ behaviour in likelihood ratio-based forensic voice comparison

“…Further, Morrison (2009) explored the speaker-discriminatory power using dynamic vowel formant data. Meanwhile, some other studies examined the speaker-discriminatory power using suprasegmental features, e.g., long-term F0 distribution (Kinoshita et al, 2009), lexical tones (Rose & Wang, 2016), speech tempo (Lennon et al, 2019) and voice quality . Apart from testing different linguistic-phonetic features, many other studies have investigated the effect of non-linguistic factors on LR-based FVC systems, e.g., sample size (Hughes, 2017;Ishihara & Kinoshita, 2008), statistical models (Kinoshita & Wagner, 2014;Morrison, 2011a), calibration methods (Morrison & Poh, 2018), sampling variability (Ali et al, 2015), channel mismatch , reference population mismatch (Watt et al, 2020).…”

“…Several other studies have been carried out investigating the speaker-discriminatory power of vowel trajectories using Cantonese (Chen & Rose, 2012;Li & Rose, 2012;Pang & Rose, 2012;Rose & Wang, 2016).…”

“…Further, the speaker-discriminatory power of other suprasegmental variables have also been studied, e.g., Lennon et al (2019) on speech tempo, Rose & Wang (2016) on tonal F0, and on voice quality.…”

effect of speaker sampling in likelihood ratio based forensic voice comparison

“…Many previous studies have explored the performance of linguistic features such as individual vowels and phonetic sequences using LR-based testing (e.g. Morrison 2009;Zhang et al 2011;Rose and Wang 2016). Typically, a group of speakers (often 60) is selected, split equally into training, test and reference speakers (e.g.…”

Identification of Triphthongs in Pakistani English

“…Later in 2011, Zhang et al has introduced a fourth order DCTs on F2 and F3 which gives best performance results for a forensic voice comparison system which is based on the tokens of Standard Chinese triphthong /iau/. Afterwards, F pattern variations for adjacent sound patterns have been used (Rose & Wang, 2016;Emerich, 2012;Man, 2007;Fant, 2006). But it can be concluded by considering that transition period of gliding segments are not helpful for individual vowel identity but provide crucial information for the identification of a gliding phoneme (Weirich, 2011).…”