Forensic Speaker Comparison
Definition
A systematic examination comparing acoustic and phonetic features of a questioned voice recording against known reference recordings of a named individual, expressed in probabilistic or likelihood-ratio terms, not as a categorical identity statement.
- Compares
- Questioned voice recording vs known reference recordings
- Subject
- A named individual
- Result format
- Probabilistic or likelihood-ratio, not categorical
- Also called
- Forensic voice comparison
Common questions
Why do examiners avoid stating a categorical 'yes, it is this person' conclusion?+
Voice characteristics overlap across speakers and can vary with health, emotion and recording conditions, so the field has moved toward expressing findings as a likelihood ratio, a strength of support for one hypothesis over another, rather than a certain identification.
What reference material does an examiner need from the named individual?+
Comparable known recordings are needed, ideally matching the questioned recording in channel, language and speaking style, since mismatched recording conditions between the known and questioned samples can distort the comparison and weaken the conclusion.
Related terms
- Formant Frequencies
- Resonance frequencies of the vocal tract that shape vowel quality. F1 and F2 (the first and second formants) are the most informative...
- I-Vector / X-Vector
- Fixed-length mathematical representations of a speech utterance used in automatic speaker recognition. I-vectors are derived from Gaussian mixture model statistics; x-vectors are...
- IAFPA
- International Association for Forensic Phonetics and Acoustics: the primary professional body for forensic phoneticians and audio analysts, which publishes guidelines on speaker...
- Likelihood Ratio (LR)
- The ratio of two conditional probabilities: the probability of the observed evidence given the prosecution's hypothesis (same source), divided by the probability...
- PLDA (Probabilistic Linear Discriminant Analysis)
- A statistical back-end model used with i-vector and x-vector systems to compute a similarity score between two utterance representations, normalised for within-speaker...