NIST SRE (Speaker Recognition Evaluation)
Definition
A competitive benchmarking series run by the National Institute of Standards and Technology since 1996, in which research and commercial teams submit speaker recognition decisions on held-out test sets and results are published against ground truth. The primary driver of systematic performance improvement in the field.
- Run by
- US National Institute of Standards and Technology (NIST)
- First held
- 1996
- Format
- Blind evaluation on held-out test sets, results scored against ground truth
- Participants
- Academic, government and commercial speaker recognition teams
Common questions
How do forensic labs use NIST SRE results in casework?+
Labs cite the performance of the specific system and condition matched to their casework to support validation claims, but SRE scores measure population-level accuracy on curated test sets, not the reliability of any single voice comparison in a case.
Does a strong NIST SRE score guarantee a system works on casework audio?+
No. SRE test conditions rarely match real casework audio, such as channel mismatch, background noise or short duration, so a strong SRE ranking shows research capability rather than a guarantee of the same performance on a specific recovered recording.
Related terms
- Background Population Database
- A corpus of speaker recordings used to calibrate the LR in a forensic comparison. Must be matched to the case in language,...
- 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....
- 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...