Background Population Database
Definition
A corpus of speaker recordings used to calibrate the LR in a forensic comparison. Must be matched to the case in language, channel, and population (sex, age, dialect) to produce a valid LR; mismatch is a major source of LR unreliability in multilingual casework.
- Field
- Forensic speaker recognition, LR calibration
- Match criteria
- Language, channel, and population (sex, age, dialect)
- Risk of mismatch
- Major source of LR unreliability
Common questions
Why does channel matching matter as much as language for the background database?+
A background population recorded on landline telephone calls behaves differently from one recorded on mobile or VoIP channels because the compression and bandwidth characteristics alter the acoustic and automatic-system features used for comparison, so a channel mismatch can bias the likelihood ratio even when the language and population match well.
What happens to a speaker comparison when no well-matched background database exists?+
The examiner either has to use a less well-matched database and flag the resulting uncertainty, or decline to produce a numeric likelihood ratio and instead report only a qualitative opinion, since an LR calibrated on a mismatched population can misrepresent the real strength of the evidence.
Is a larger background database always better for LR calibration?+
Size helps statistical stability, but a large database that is poorly matched in language, channel or demographic profile to the case recordings is less useful than a smaller, closely matched one, because the calibration depends on the database reflecting realistic within- and between-speaker variability for that specific case population.
Related terms
- BATVOX
- A commercial automated speaker recognition system (originally Agnitio, later Nuance Communications) widely used in operational forensic casework in Europe and Latin America....
- C_llr (Log-Likelihood Ratio Cost)
- A proper scoring rule measuring the calibration quality of a speaker recognition system's LR outputs; a well-calibrated system has C_llr close to...
- ENFSI BPM
- The ENFSI Best Practice Manual for Forensic Comparison of Speech (2015, rev. 2022); the operational standard for speaker comparison in European forensic...
- ENFSI Verbal Scale
- A standardised table published by the European Network of Forensic Science Institutes that maps ranges of LR values to verbal expressions of...
- Equal Error Rate (EER)
- The point on a classifier's detection error tradeoff curve where the false accept rate equals the false reject rate. Lower EER indicates...
- I-Vector
- A fixed-length low-dimensional speaker representation extracted by projecting a speaker's UBM statistics into a total variability space. Introduced by Dehak et al....
- NIST SRE (Speaker Recognition Evaluation)
- A competitive benchmarking series run by the National Institute of Standards and Technology since 1996, in which research and commercial teams submit...
- 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...
- X-Vector
- A deep-learning speaker embedding produced by the penultimate layer of a time-delay neural network (TDNN) trained for speaker classification. Introduced by Snyder...