U.S. flag

An official website of the United States government, Department of Justice.

Evaluation of the NITV CVSA

NCJ Number
Journal of Forensic Sciences Volume: 53 Issue: 1 Dated: January 2008 Pages: 183-193
Harry Hollien Ph.D.; James D. Harnsberger Ph.D.; Camilo A. Martin M.D.; Kevin A. Hollien B.A.
Date Published
January 2008
11 pages
Using a speech database, this study evaluated a commonly used voice stress analyzer, the National Institute of Truth Verification's (NITV) Computer Voice Stress Analyzer (CVSA).
The evaluation found that the CVSA did not display the expected ability to detect deception, truth, and/or stress level in either the laboratory samples that constitute the Core database or the smaller set of field materials (the SERE database). The CVSA's capacity to detect deception-truth and stress levels was near the chance level, at least for the databases studied. The need for the development of systems that can detect behavioral states from speech and voice analysis is well established, so basic research is warranted. Some modification in research focus is needed. The current research focused on deception and high-stress states. Less emphasis was given to determining when truthful statements were being made, although in many causes it is just as important to discover whether the speaker is telling the truth as it is to determine when he/she is lying. It is also important to explore the basic relationships between speech and deception and develop a richer database of speech samples for the identification of speech/voice cues that will be useful in research, assisting in the development of appropriate new systems, and the evaluation of other and future commercial voice stress analyzers. The speech database used in this evaluation contained material recorded in the laboratory while highly controlled deceptive and shock-induced stress levels were systematically varied, and also during a field procedure. Subjects were 24 males and 24 females drawn from a representative population. All held strong views on an issue and were required to make sharply derogatory statements about it. The CVSA system was then evaluated in a double-blind study using three sets of examiners. 3 tables, 6 figures, and 60 references