Balamurali B T Nair, Esam A S Alzqhoul and Bernard J Guillemin
It is becoming increasingly common for mobile phone recordings to be presented as evidence in a court of law. In such situations a forensic scientist is frequently engaged to analyse suspect and offender voice samples with a view to determining the strength-of-evidence, a process called Forensic Voice Comparison (FVC). This paper investigates the extent to which an FVC analysis is negatively impacted by the network through which the speech has passed. Our investigation focuses on the GSM and CDMA networks as these are the ones in common usage currently. Our experimental findings suggest that both networks negatively impact the accuracy of an FVC analysis and that this is worse for the CDMA network. We present strategies for mitigating this impact to some extent.
Compartilhe este artigo