Preparing a voice lineup

Abstract

This paper will describe the preparation and implementation of a voice lineup or voice parade which formed part of the police investigation into a rape. The police supplied, as a sample of a suspect's voice, the statutory recording of a police interview with the suspect, and also speech samples from eight of their panel of volunteers for identity parades to act as 'foils'. The task was to create a voice parade which would be fair to the suspect. Fragments free of incriminating content were edited from the suspect's sample to provide a 30 second composite sample of his speech, and a similar editing procedure was followed for the 'foils'. A further speaker, for whom an interview recording was available, was also added. To ensure the fairness of the samples, two experiments were carried out in accordance with recommendations formulated at the Forensic Laboratory of the Dutch Justice Ministry. The first provided an estimate of the relative perceived distance of each of the foils from the suspect. The two speakers least like the suspect were eliminated. The second ensured that neither the suspect nor any of the foils sounded stereotypically like a rapist. The problems inherent in such experiments will be discussed, as will the procedure and outcome of the voice parade.