Palm Coast
storylines...

Independent Reporting · Palm Coast, Florida
Backgrounders

What Is Voice Stress Analysis?

The tool law enforcement uses to listen to your voice. What it measures, why sheriffs use it, what its critics say, and what we found when we ran it ourselves.


Voice stress analysis is the examination of a speaker's fundamental vocal frequency — the measurable pitch of their voice — for patterns that deviate from their own established baseline under neutral conditions.

The theory is straightforward. When a person experiences psychological stress, the autonomic nervous system responds. Heart rate increases. Muscles tighten — including the muscles of the larynx. Those muscles control the vocal folds. Tighter muscles produce measurable changes in vocal pitch. A trained instrument can find them.

Law enforcement has used voice stress analysis for decades. The instrument used by sheriff's offices and federal agencies is called the Computer Voice Stress Analyzer, or CVSA. More than 1,400 law enforcement agencies in the United States have used it.


The detractors — and we take them seriously

We said we would tell you everything. Here it is.

The National Institute of Justice — the research arm of the U.S. Department of Justice — funded a field evaluation of VSA technology and found it detected only 15 percent of lies about drug use in a real-world jail environment. The executive director of the CVSA manufacturer itself stated in a sworn legal declaration that the instrument “is not capable of lie detection.” Voice stress analysis is not accepted as evidence in Florida courts, or most courts anywhere in the United States.

Critics argue that elevated or compressed vocal pitch can result from ordinary public speaking anxiety, rhetorical emphasis, or background acoustic conditions — and not from deception at all.

We do not dispute any of that.


What we did

We found Praat — an open-source acoustic phonetics program developed at the University of Amsterdam, the same institution whose research underlies the methodology used in professional voice examination. We installed it. We extracted two audio segments from the August 8, 2023 Palm Coast City Council meeting — a neutral baseline from earlier in the same session, and the target segment containing the statement in question. We ran both through the same intra-speaker baseline comparison protocol used by law enforcement examiners. We produced a formal report in exactly the format a sheriff's office would produce.

The full report, methodology document, and statement verification exhibit are all in the Evidence Locker.


What we found

At 14.09 to 14.24 seconds into the target segment — timestamp 2:29:07 in the full meeting video, the precise moment the subject said “these are in no particular order” — fundamental frequency spiked to a sustained cluster of 419 to 475 Hz before immediately returning to the controlled range. In the baseline segment, such spikes were random and scattered. In the target segment, the spike cluster was concentrated at the exact words carrying the denial.

The overall pitch contour of the target segment was also significantly more compressed than the baseline — a spread of 70 Hz versus 174 Hz. Compressed pitch is itself a recognized stress indicator in the forensic phonetics literature.


Our position

We are not saying this person lied. Voice stress analysis does not establish that. What it establishes is that something measurable was happening at that precise moment — and the instrument flagged it.

Combined with the documentary record — the August 10, 2021 minutes in which the same person presented these same priorities as explicitly numbered, categorized, and temporally designated — the finding sends us down an investigative path. We intend to follow it.

One more thing worth knowing: this tool can be run on any audio or video we obtain.


Learn more

Praat — official site — University of Amsterdam
Wikipedia: Praat — background on the software and its academic uses
NIJ: Voice Stress Analysis — Only 15 Percent of Lies Detected — the critical federal study
Criminal Legal News: Guilty Voice — Junk Science or Reliable Evidence? — March 2025


Mr. Gaines practiced law in Manhattan for forty-one years. He considers this significant.