Multiparametric Analysis of Speaking Fundamental Frequency in Genetically Related Speakers Using Different Speech Materials: Some Forensic Implications

Abstract

Objective: To assess the speaker-discriminatory potential of a set of fundamental frequency estimates in intraidentical twin pair comparisons and cross-pair comparisons (i.e., among all speakers). Participants: A total of 20 Brazilian Portuguese speakers of the same dialect, namely 10 male identical twin pairs aged between 19 and 35, were recruited. Method: the participants were recorded directly through professional microphones while taking part in a spontaneous dialogue over mobile phones. Acoustic measurements were performed in connected speech samples, and in lengthened vowels, at least 160 ms long produced during spontaneous speech. Results: f 0 baseline, central tendency, and extreme values were found mostly discriminatory in intra-twin pair and cross-pair comparisons. These were also the estimates displaying the largest effect sizes. Overall, only three identical twins were found statistically different regarding their f 0 patterns in connected speech, but not for lengthened vowel-based f 0 metrics. Estimates of f 0 variation and modulation were found the least discriminatory across speakers, which may signal the control of speaking style and dialect on dynamic patterns of f 0. Concerning system performance, the base value of f 0 (f 0 baseline) was found the most reliable metric, displaying the lowest equal error rate (EER). Conclusions: the outcomes suggest that, although identical twins were very closely related regarding their f 0 patterns, some pairs could still be differentiated acoustically, only in connected speech. Such findings reinforce the relevance of analyzing long-term f 0 metrics for speaker comparison purposes, with particular consideration to f 0 baseline. Furthermore, f 0 differences across subjects were suggested as more expressive in connected speech than in lengthened vowels.

Assessing the speaker discriminatory power asymmetry of different acoustic-phonetic parameters

“…When comparing previous analyses with data from Brazilian Portuguese [4,5], a relatively similar pattern was also suggested, in which speech temporal parameters performed worse compared to melodic estimates, as revealed by the discrepancy in Cllr and EER values.…”

Multi-parametric analysis of speech timing in inter-talker identical twin pairs and cross-pair comparisons: Some forensic implications

“…This is particularly the case when comparing to other acoustic parameters, such as melodic metrics, cf. [66]. In this instance, previous research has acknowledged the relatively poorer discriminatory potential of speech timing estimates.…”

“…The methodological design adopted in the present study does not allow answering to what extent or to what level the high intra-twin pair temporal congruence may solely be accounted for by environmental influences, since another equally plausible and perhaps overlapping factor is suggested; prosodic entrainment. In addition, the fact that more differences have been found regarding spectral and melodic parameters for the same twins in [29,66] may indicate a possible higher level of prosodic entrainment in speech timing than in formant frequency and melodic patterns. In the experiment on perceptual entrainment conducted by [78] with 42 BP native speakers encouraging evidence was found supporting the assumption of rhythm perception as a listener-speaker entrainment process, in which duration may be considered the main acoustical feature driving the behavior of listeners.…”

Prosodic variation of English in Dominica, Grenada, and Trinidad

“…(2017) and Cavalcanti et al. (forthcoming). Frequency in Hz was additionally converted to semitones (st) with the formula 12 * log2(freq), that is, 12 times the binary logarithm of F0 in Hz, relative to 1 Hz.…”

“…Based on this data, the following measures were computed for each speaker in each speaking style: ‐ Pitch level : Median F0 in Hz, as a measure of overall pitch height that is robust to outliers (Mennen et al., 2012) ‐ Pitch range : 80% range of F0 in st, that is, the difference between the 10th and 90th percentiles in the F0 distribution (Patterson, 2000) as a measure of distance between typical low and high values in the F0 distribution that is robust to outliers ‐ Pitch dynamism : The pitch dynamism quotient (pdq) 11 as a measure of overall variability of F0, defined as the standard deviation of the F0 distribution divided by its mean in Hertz (Cheng, 2020) ‐ Rate of pitch change : The median of the first derivative of F0 as a measure of the typical rate of change in F0 (adapted from Barbosa et al., 2017; Cavalcanti et al., forthcoming) ‐ Variability in rate of pitch change : The 80% range (that is, the difference between the 10th and 90th percentiles) of the first derivative of F0 as a measure of variability in the rate of change in F0 (adapted from Barbosa et al., 2017, Cavalcanti et al., forthcoming) ‐ Tone rate : Number of phonetically identified highs and lows in F0 contour normalized by speaking time (Cavalcanti et al., forthcoming) 12 …”