Likelihood Ratio Calibration in a Transparent and Testable Forensic Speaker Recognition Framework | Publicación