Human error rates for speaker recognition

Link to full Paper

Nifty tech tag lists from Wouter Beeftink