Taxa de elocução e taxa de articulação em corpus utilizado na perícia de Comparação de Locutores

Abstract

Este artigo está licenciado sob forma de uma licença Creative Commons Atribuição 4.0 Internacional, que permite uso irrestrito, distribuição e reprodução em qualquer meio, desde que a publicação original seja corretamente citada.Resumo: Neste estudo verificou-se a taxa de elocução (TE) e de articulação (TA), bem como a adequação do uso dessas medidas temporais na perícia de Comparação de Locutores. Calculou-se a TE/TA global e local média nas amostras de fala espontânea questionada (interceptações telefônicas judicialmente autorizadas realizadas sem a ciência dos locutores) e padrão (entrevista semidirigida gravada com o consentimento dos locutores) de sete falantes nativos do português brasileiro, de ambos os sexos, com idades entre 14 e 33 anos. Definiu-se o potencial individualizante da TE/TA e a relação entre essas e as variáveis independentes idade, sexo, escolaridade, tempo transcorrido entre as amostras (gap temporal), tipo de gravação e tamanho do intervalo de fala. Como resultado obteve-se coeficiente de correlação intraclasse satisfatório associado à TA, diferença significativa quanto à ciência da gravação na TE e correlação significativa entre gap temporal e TA e entre tamanho do intervalo de fala e TE. Palavras-chave:Abstract: This study examined the speech rate (SR) and the articulation rate (AR), as well as the appropriateness of the use of these temporal measures in the forensic Speaker Comparison. The global and local mean SR/AR in spontaneous speech samples -obtained from unaware audio recordings (legally authorized telephone interceptions) and known ones (semistructured interview) -from seven native speakers of Brazilian Portuguese, of both sexes, aged between 14 and 33 years were considered. The speaker-discriminating power of SR and AR and the relationship between rates and the independent variables age, sex, formal education, temporal gap between the recordings, type of recording, length of speech stretch were defined. As a result we obtained satisfactory intraclass correlation coefficient for the AR, significant difference between the factors of the variable type of recording in SR and significant correlation between temporal gap and AR and between length of speech stretch and SR.

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

“…Secondly, for the articulation rate estimation, not solely silent pauses were excluded from the intervals' total duration, but also filled pauses. As such, the articulation rate reported by that author is somewhat equivalent to ARTRATE II in the present experiment, which appears to be the case when comparing the global (6,19 vv/s) and local (6.20 vv/s) median values reported by [13] with the median value observed here (6 vv/s). Thirdly, the minimum silent pause duration threshold in the present study was set at 100 ms in comparison to 130 ms on the other, which may result in different estimates.…”

“…Similar observations of the discriminatory potential regarding the two parameters have been suggested in the available literature. Despite observing a higher variance for speech rate in relation to articulation rate, [13] noted a compatible discriminatory power between speech rate and articulation rate in an experiment with BP speakers. However, the before-mentioned study differs from the present regarding some essential aspects.…”

“…Such a phonetic unit comprises all the segments uttered between two consecutive vowel onsets, with the onset of the following vowel defining the beginning of a new V-V unit. It has been studied and employed among others by [11][12][13][14], with its application tracing back at least to [15,16].…”

“…As for Brazilian Portuguese, a phonetic experiment was conducted by [13] on speech tempo parameters based on a realistic forensic data-set. The analyzed material comprised of spontaneous speech samples derived from intercepted telephone conversations and direct non-contemporaneous recordings of the same speakers.…”

“…In the present study, filled pauses displayed a moderate effect size, being suggested as more explanatory of individual patterns than silent pauses. It must be recognized that, as pointed out by [13], the identification and transcription of filled pauses is a somewhat laborious task often requiring several revisions. In some cases, it is uncertain whether one is dealing with a filled pause, an emphasis, or perhaps both, expressed in the form of a lengthened vowel.…”