ABSTRACT
In the voiceprint method known and unknown voices are compared to determine if they are the same speaker. An unknown voice is recorded over the telephone onto a quality recorder. The known voice is similarly recorded, repeating the questioned message verbatim. An examiner determines, by means of acoustic spectrography and aural analysis, the similarities or differences existing between the known and unknown speakers. A positive conclusion is based on at least 10 pairs of like sounds. Five decisions are available for the examiner to make: positive identification, positive elimination, probable identification, probable elimination, and unable to make a decision with the sample submitted. Seven collaborators, chosen from various parts of the United States, were given recordings of 4 unknown speakers, 2 male and 2 female, to compare with similar recordings of 6 known speakers, 3 male and 3 female. A total of 21 positive identifications and 63 positive eliminations was possible. Twenty correct identifications and 47 correct positive eliminations were made. In 16 tasks, all involving the same unknown sample, collaborators reported they could not make a determination due to the poor quality of the sample. No false identifications were made; however, a trainee examiner, with less than 2 years experience, eliminated a known speaker when in fact a match did exist, thereby falsely eliminating the speaker. Examiners with more than 2 years of experience correctly identified all existing matches. This study indicates that a trained examiner can make very reliable decisions, using the aural and visual methods of comparing known and unknown voices. The voiceprint method has been adopted as official first action.