NAU Professors Fine-tune Forensic Voice ID Analysis

voice id
NAU researchers say the slightest distinctions in dialect may be the determining factor in a criminal case.

“Would you like a soda?” “No, thank you. I’ll take a pop though.”

These variations in American dialect might seem trivial to the casual observer, but to the forensic scientist, the slightest distinction may be the determining factor in a criminal case.

Just ask Northern Arizona University Health Science Professors William Culbertson and Dennis Tanner. “People often mistakenly use the terms accent and dialect interchangeably,” Tanner explains. “Accent refers to how words are pronounced, while a dialect has its own syntax and common expressions that make it unique from other dialects of the same language.”

Professor Bill Culbertson: “My senior students delight when we receive aural method voice identification requests.”

Why does this matter? Consider the New York case People v. Sanchez: A Spanish-speaking Puerto Rican man was indicted for murder in the second degree. The evidence presented at the trial included testimony from an eyewitness who had heard the victim and killer arguing in Spanish prior to the shooting. The witness who had spoken Spanish all of his life stated that the man with the gun was speaking with a Dominican accent, rather than a Puerto Rican accent, and that there is a distinctively Dominican accent and a different, equally distinctive, Puerto Rican accent. In light of these details, the court admitted the testimony regarding the accent of the shooter.

Voice ID as expert evidence

Tanner, who is an experienced expert witness and specialist in aural evaluation, often finds cases involving similar speaker identification issues making its way to his desk. When this occurs, he enlists the aid of Culbertson, and the two begin an auditory perceptual analysis, the Tanner Culbertson Aural Method of Voice Identification. This alternative to the traditional spectrographic voice prints uses expert listeners to identify a suspect voice from one or more audio samples.

 “My senior students delight when we receive aural method voice identification requests,” Culbertson notes, smiling. “Along with Dr. Tanner and me, our trained students serve as expert listeners in these cases. It’s often the highlight of the semester.”

NAU students trained as expert listeners

As expert listeners, the students participate in two aural analyses in a classroom setting. The first examination involves listening to a pair of audio samples and using short-term memory to make a judgment. The second method, duplexing, allows listeners to hear both audio samples in real time simultaneously, providing for instant judgment. Raw data is provided to attorneys and law enforcement to be used as appropriate.

Tanner believes the aural method of voice identification will continue to grow. “Too often voice print results are highly technical and too complicated for juries and lay persons to understand. With the aural method, it’s simple. It’s less expensive and extremely practical.”

--Courtesy of NAU Office of Public Affairs, NAU News