Phil Rose

Expert opinion is being increasingly sought in the legal process as to whether two or more recordings of speech are from the same speaker. Forensic speaker recognition (FSR) can be very effective, contributing to both conviction and elimination of suspects on the basis of their voice. Yet there is still a considerable lack of understanding on the part of law enforcement agencies, legal practitioners, and indeed phoneticians, speech scientists and linguists, as to what it involves; what constitutes appropriate methodology; what it can achieve; and what its limitations are. The aim of this workshop is to show what is necessary in order to address these questions with the appropriate technical precision.

The last few years have seen a major paradigm shift in the evaluation of forensic identification evidence, and Forensic Speaker Recognition has been very much involved. The workshop will demonstrate how the logically correct Likelihood Ratio-based framework can and must be used for evaluating the strength of forensic phonetic evidence. It will be of especial interest and value to members of the legal profession and law-enforcement agencies.

The workshop will be organised around two themes:

(1) The basic ideas Expressing the outcome – Likelihood Ratios and Bayes’ Theorem; FSR parameters traditional and automatic; what makes FSR difficult; characterising FSR: identification, verification or what? Naïve and Technical forensic speaker Recognition. Likelihood Ratio-based forensic discrimination: univariate and multivariate approaches.

(2) A demonstration of the method using Likelihood Ratio-based analysis on forensically realistic speech acoustics.