On phonetic identification of speakers in the forensic context
The paper introduces current trends in speaker identification using phonetic methods. After providing a short historical background, it describes various attempts which have been made at more objective listening analysis in forensic contexts. One of the main aims of the paper is to propose a protocol for listening analysis in the Czech forensic context, called PRIM. The PRIM protocol combines the advantages, and mitigates the drawbacks, of existing approaches to analytical listening and, most importantly, includes all speech and voice dimensions relevant for the auditory comparison of speakers in Czech. Audio examples of some of the described voice qualities are featured on an accompanying webpage. The role of acoustic analyses in speaker comparison is also discussed, and the main parameters used for such comparisons are briefly introduced. These include various uses of vowel formants and spectral characteristics of noise in obstruent sounds at the segmental level, as well as melodic and temporal characteristics and acoustic correlates of various voice quality settings at the suprasegmental level. Automatic approaches to speaker identification are also briefly mentioned. Finally, the paper relates the current situation in forensic phonetics to the ongoing paradigm shift in forensic science generally, and in the Czech Republic specifically.
Key words: acoustic features, analytical listening, forensic phonetics, the PRIM protocol, voice comparison
Klíčová slova: akustické rysy, analytický poslech, forenzní fonetika, protokol PRIM, srovnávání hlasů
Text je on-line k dispozici v databázi CEEOL.
Fonetický ústav FF UK
nám. Jana Palacha 1/2, 116 38 Praha 1
Naše řeč, volume 105 (2022), issue 3
Previous Jan Volín: Individualita uživatele jazyka: Slovo úvodem
Next Jan Volín, Tomáš Bořil: Individualita mluvčího v užití základní hlasové frekvence při přednesu básní