In a criminal investigation where there is not enough physical evidence like DNA or fingerprint, a voice recorded on a device could be of immense importance. The voice can be scientifically analyzed and compared with other voice data of persons with a criminal history and can be identified if the person`s voice is on the data. Forensic voice identification involves various steps, like spectrographic analysis, where the person`s acoustic profile is displayed visually. In aural, linguistic analysis, several aspects of the explicit voice, such as expression and rhythm of pronunciation, are measured.
Vocal cords and tract
The basic principle of voice analysis is structured on the vibration and frequency of the sound. Vibration constitutes all sounds, and on frequency, how fast the air vibrates depends.
Lower frequency manifests a low pitch, and higher frequency makes a higher pitch. But most sounds are made up of an assortment of frequencies that gives them unique characteristics. For instance, a piano and violin orchestrated at the same note hear differently from the human ear as it is made of other frequencies and pitches. Vocal cords; two small vibrating muscle located at the throat that generates the sound. The buzzing sound is produced by the vocal cord, and its vibration of it determines the pitch. The vocal tract of a supple tube placed between the vocal cord and lips converts the sound into speech. The jaw and another part of the mouth alter the shape of the vocal tract to make specific speech sounds.
Due to genetic and behavioural differences acoustic profile of a person is unique and distinguished from others. A person`s voice is specific. This makes it recognized by other voices. Voice forensic experts compare two voices and consider the similarity between the two. If the features are way apart, it can be concluded the suspect is not the criminal. Forensic voice analysts are trained in linguistics, a scientific study of language and speech. In auditory analysis, the call made by the suspect is broken into various parts and meticulously scrutinized for specific sounds. Computer software called spectrogram is used to produce an image of the sound, and the whole process is known as acoustic analysis.
There can be two types of voice identification of a suspect; ear witness and forensic speaker identification. The legal validity in the court of law of an ear witness depends on various factors where the witness was hearing the conversation, such as sex, age and race of the witness, duration of the hearing, and time elapsed between the occurrence and identification. There is always ambiguity in such cases, and the court makes a decision on considering many aspects of the testimony.
While forensic speaker identification (SPID) is a more reliable source of identification of voice than of ear witness when a forensic expert analyzes the voice recording and can conclude if the voice matches the voice of the suspect or not, the technology used in SPID is called Aural- Acoustic Speaker Identification that aligns with the Frye and Daubert Standard in US law since the early 1990s. This procedure of voice identification is acknowledged by the International Association of Identification and the National Institute of Standards and Technology.