The forensic unit analyzes the tape by applying the latest computer analysis software; both the voices of the 911 operator and the caller can be clearly heard. The analysis also reveals a barely detectable third voice.
The audio technician uses the software's leveling module to normalize, or bring all three voices to a level signal. It is now clear the voice in the background is threatening the woman with bodily harm. But who is talking in the background? The officer learns the woman has a boyfriend. Dialogues with other residents also reveal a male tenant has been repeatedly trying to gain her affections.
Both men are brought in for interrogation. Their answers to specific questions are collected by a digital tape recorder similar in quality to the 911 devices. The questions are structured to illicit certain words and phrases found on the 911 tape. The forensic audio expert then compares each man's response to the original tape. In a process called "speaker identification," the technician compares bits from the interrogation tapes to corresponding bits on the original tape. An energy distribution pattern of each voice over the vocal frequency range is developed and analyzed. Each man's energy distribution pattern is unique to the individual in a manner similar to the uniqueness of a fingerprint.
At this point the audio lab technician is stumped. He lacks the appropriate training to accurately interpret this type of pattern. An outside expert in speech identification is called in. The speech identification expert studies the tapes and makes a number of adjustments to the software to highlight certain parts of the recording. After his analysis, he determines the voice on the tape is the other tenant and not the boyfriend.
The use of an outside expert is critical in this type of case unless the local analyst has many years of experience with voice/speech analysis. In addition to years of experience, an audio speech expert also should be a board certified forensic audio/video examiner certified by the American Board of Recording Experts. On the witness stand, an expert's credentials will be a strong point with a jury and cannot easily be discredited by the defense. The expert also has had experience in presenting data to a jury in a clear and easily understood manner.
The future sounds clear for forensic audio analysis. New software applications continue to be developed to further enhance audio analysis techniques. Databases of background sounds and other useful audio information are being created. And a variety of training opportunities in forensic audio analysis have sprung up across the country.
DNA isn't the only forensic science making it tougher for criminals to escape the long arm of the law. Today's criminals need to watch what they say because law enforcement may be listening.
Establishing the audio/video forensic lab
The following factors should be considered when establishing or increasing the capability of a modern audio/video forensic laboratory:
- Identifying potential employees
- Requiring a lengthy apprenticeship or equivalent experience of personnel in certain audio and video analyses fields
- Obtaining specialized training in such areas as digital signal analysis, recording theory, sound measurement and video imaging
- Seeking support and guidance from other established laboratories
- Equipping the laboratory to play back and record in numerous analog and digital audio and video formats, and then providing the capability to improve voice intelligibility, compare voices, identify non-voice signals, authenticate recordings, enhance video images and conduct other related analyses
- Procuring professional audio, video, enhancement, signal analysis, imaging and other related equipment
- Identifying physical space for the laboratory
- Excerpt from "Equipping the Modern Audio-Video Forensic Laboratory" in the April 2003 issue of Forensic Science Communications
Doug Hanson is a Ph.D. biochemist and can be reached at firstname.lastname@example.org.