Forensic Speaker Comparison: A Linguistic–Acoustic Perspective
Paul Foulkes, Peter French · Oxford University Press eBooks · 2012
Abstract Two of the tasks undertaken in forensic voice and speech analysis are speaker profiling and speaker or voice comparison (or speaker/voice identification). The first task involves an examination of features of the voice to ascertain information about the speaker's regional, social, and ethnic background. The second involves comparative analysis against a reference sample known to have been produced by a particular talker. Although analysis of vocal features cannot determine a speaker's identity, it can provide a wealth of information about the speaker, albeit to varying degrees of precision and confidence. It is for this reason that the label ‘speaker comparison’ is now preferred to identification. Globally, speaker comparison is carried out using three general methodologies: ‘voiceprinting’, analysis using automatic systems, and linguistic-acoustic analysis. This article first presents a brief historical review of forensic speaker comparison and then explores the use of the linguistic-acoustic method in forensic speaker comparison.