Forensic Voice Comparison

WHAT IS BIOMETRIC VOICE COMPARISON?

Forensic voice comparison uses a known voice sample to scientifically compare with an unknown voice sample to determine the speaker’s likelihood ratio. For example, law enforcement may have a suspect and ask for a voice comparison against an unidentified phone call. 

forensic voice comparison

VOICE COMPARISON METHODOLOGY

FVC can be challenging because voices aren’t entirely unique and can change under various circumstances. A voice comparison expert also has to contend with technological limitations such as telephone frequencies and background noises. 

The language a person uses, intoxication and emotional responses all change the sound of a person’s voice. An individual may even try to disguise their voice in certain scenarios. Amidst these circumstances, voice comparison experts have to look at notable patterns and similarities to compare speakers which include format, tone, pitch, dialect, accent etc.


VOICE COMPARISON PROCESS

CRITICAL LISTENING

In any realm of forensic audio, an expert needs to apply critical listening. Voice comparison experts hone this skill with training and experience.  While some people have very distinct, atypical speech patterns, not everyone does. Experts will listen for notable pauses, pronunciations, dialect characteristics and rhythm on both recordings during the critical listening stage.

EXEMPLAR RECORDING

This first step in forensic voice comparison is critical to the voice analysis process. The exemplar recording is audio from the suspected speaker on the presented evidence. Once an expert has the exemplar file, they can begin their comparison. 

BIOMETRIC TESTING

Voice biometrics is a way of translating a voice into a visual fingerprint. With the right technology, experts can create these fingerprints and look for shared physical characteristics between voices. This process applies to the full length of the audio clips and short segments.

FREQUENCY & WAVEFORM

This step in voice comparison involves looking at the audio files of the presented speakers. During this stage, experts can examine the frequencies and waveforms of a person's voice. While behavioral and technological limitations may exist, they can use this analysis to identify similarities in single words or sounds. 

AN UNBIASED APPROACH TO VOICE RECORDING ANALYSIS FOR USE IN COURT

After completing these tests, analysts will determine the likelihood ratio to the unknown voice. This decision is based on a percentage of certainty reached during the analysis.

At Primeau Forensics, we apply voice forensic best practices that the scientific community supports and endorses. We focus on legitimacy and facts — when you contact us for forensic voice comparison, we can use our findings to testify in court with transparency and clarity.

speaker recognition

IF YOU HAVE DIGITAL MEDIA EVIDENCE THAT YOU QUESTION OR NEED HELP UNDERSTANDING,
PLEASE CONTACT US FOR A PRO BONO CONSULTATION.

TALK TO AN EXPERT