Phonetics and Speaker Identification

At the present time, phoneticians are receiving more and more requests for advice and assistance from police and solicitors in connection with cases involving tape-recorded speech samples, often with the assumption on the part of the inquirers of the existence of 'voice prints'. Without entering into a lengthy discussion of all the issues involved, it must be placed on record that the relevant specialists, the acoustic phoneticians, in Great Britain and indeed most other countries are totally opposed to any such concept within their discipline. It is certainly true that the speech of every person is unique and individual, but it cannot be argued from that that any visual representation of the measurements of speech made by electronic devices can capture this individual character. At most one can express the belief that the further evolution of electronics may well bring about such a possibility. Speaker identification by the acoustic aspect of speech analysis, i.e. the measurement electronically of the physical characteristics of speech, is thus not possible. There then remains the auditory aspect of speech analysis, this being the description of the sounds of speech by means of the researcher's ear, expressed in terms of a set of conventional symbols or phonetic notation. The auditory phonetician's judgements are, of course, subjective, but against this must be set the very long and intensive training all professional phoneticians undergo and the fact that their conclusions, being framed in terms of widely accepted conventions, can be examined critically by other phoneticians. Defence counsel have been known to complain of the subjec-