Voice/Speaker Identification or Elimination via
Aural-Acoustic-Spectrography

Voice/speaker identification or elimination via aural-acoustic-spectrography, including the analyses of the acoustic, phonetic and linguistic characteristics of voices, involves the reflection of vocal characteristics in a spectrogram. Given two adequate exemplars (1 “unknown” exemplar and one “known” exemplar), we can determine whether one particular speaker appears on both exemplars or not. Please note that Dr. Yonovitz, a renown speech and hearing forensic scientist and academician, is a former member of the certification and standards committee of the International Association of Identification [“IAI”].

Spectrographic voice identification was developed at Bell Labs in 1941 and introduced forensically in 1961. The basic premise is that each of us has unique vocal characteristics: sinus cavities, vocal chords and articulators [i.e., lips, teeth, tongue, etc.] confer vocal characteristics, which are individually unique when spectrograms are made. The FBI submitted a report on voice spectrography, which stated a better than 99% accuracy rate when very strict requirements are followed. B. Koenig, “Spectrographic Voice Identification: A Forensic Study,” J. Acoust. Soc. Am. (Letters to the Editor), p. 2088, June 1986. Cf. U.S. vs. Angleton, 269 F.Supp.2d 892 (S.D.TX 2003), but also see State vs. Gary Morrison, No. 03-KK-2790 (LA 2004) (after passing the state’s counterpart to Daubert, Dr. Yonovitz was scheduled to testify in Sept. 2005 about his results from his aural-spectrographic analyses, but the case settled).

Yonovitz & Joe, L.L.P. recently had a peer-reviewed article (“Legal, Scientific and Forensic Controversies Over Spectrographic Voice Analysis for Identification or Elimination”) published in the September 2007 issue (Vol. 7, No. 6) of the Law Enforcement Executive Forum.  To access a copy of this article, please click here.

The forensic procedure is termed aural-acoustic-spectrography.  Aural analyses are an integral part of voice/speaker identification or elimination. Very few examiners have any formal training and research in the speech and hearing sciences. Dr. Yonovitz is an associate professor of the speech and hearing sciences; he was an associate professor at the Speech and Hearing Institute, Graduate School of Biomedical Sciences, The University of Texas Health Science Center at Houston for sixteen years [1973 – 1989]; he was also an associate professor, Chief Audiologist and ASHA Program Director at the Department of Communication Disorders, Conley Speech and Hearing Center at the University of Maine for six years [1989 – 1995].) Examiners trained in spectrogram pattern matching receive little or no formal, clinical or research training in the forensic assessment of speech characteristics.

In one remarkable voice ID case (U.S. vs. Luke Samuels, et al., Crim. No. 99-50016 (W.D. LA 1999)), all 7 co-defendants claimed that they were not parties to the recorded drug transactions. Our analyses revealed 5 eliminations and 2 identifications. We were then ordered to send our exemplars to the FBI. The day after the government’s report was due (~2 weeks before trial), the government dismissed the case against all 7
co-defendants. Please click here to view a letter from that case.

Please call (214-505-TAPE or 703-892-TAPE) or e-mail us to discuss any of our listed services, or for any other matters of audio, video, sound, noise, acoustics, speech or hearing evidence.