Marksman in position to perform shots with test recording. Note the positions of the microphones.
Dr. Robert Maher remembers it vividly. He received a cold call asking if he could match a gun that was allegedly used in a crime to an audio recording from the crime scene of a gun being fired. Sitting at his desk in his Montana State University office, his initial reaction was, “I really don’t think that is possible.” After hanging up, he thought back to what he had read over the years and realized that there wasn’t enough data to answer that question properly.
It’s not that it might not be possible, he thought; it’s simply that we just don’t know.
Thus began his research into gunshot acoustics that has spanned eight years and three grants from the National Institute of Justice (NIJ). Maher’s initial work in this area began with understanding the acoustical characteristics of gunshots by obtaining high-quality, repeated recordings under controlled conditions. This effort was in line with a strategic goal of NIJ’s Office of Investigative and Forensic Sciences to support foundational forensic science research.
To do this, Maher created an apparatus and a methodology to collect gunshot recordings (photo above). He gathered data from a variety of firearms—five handguns, one revolver, one shotgun, and two rifles—to measure consistency, reliability, and shot-to-shot variability.
He found that although there are similarities when one fires the same gun 10 times, there are also appreciable shot-to-shot differences. The duration of the muzzle blast varies from firearm to firearm, but a given firearm also varies from one shot to another. Although the explanation for the duration variability is not yet known, Maher suggests that this variation will have an impact on forensic analysis of recordings that include gunshots of unknown origin.
Once he had established a repeatable method for accurately recording gunshot acoustics under ideal conditions, Maher was ready to study the limitations of forensic interpretation from common recording devices. This could include cellphones, land-mobile radios, personal audio recorders, audio data collected by emergency call centers, and dispatch center recording systems.
He compared signals at 11 different microphone and personal recording device locations, in addition to a body camera worn by the marksman and an internal recording system in a law enforcement vehicle. This allowed him to verify geometric predictions regarding time of arrival and level at each recording location. For verification purposes, he also compared the timings to a recording made by a cellphone call to a corporate voicemail system. Then, he examined multiple, simultaneous gunshot recordings to see if one could obtain forensically relevant information, despite reflections, distortion, coding artifacts, and other nonideal features. Based on analyses, Maher created a processing method to localize the source of gunshots and reduce incoherent background noise, as well as a method to identify the most likely synchronization point for multiple audio recordings.
Ultimately, Maher’s research deepened the audio forensic knowledge base by developing new and innovative techniques to synchronize and process concurrent, ad hoc audio recordings from a crime scene obtained from body cameras, cellphone videos, surveillance cameras, dashboard cameras, and other recording devices.
Thanks to this research, the field has a better understanding of the limitations of forensic interpretation of user-generated recordings of gunshots, especially with regards to audio bandwidth, recording quality, and questions of authenticity.
How can this research be used?
For audio forensic analysis, it is increasingly likely that multiple user-generated recordings may be presented as evidence in a criminal investigation. Audio evidence may come from handheld smartphones, private surveillance systems, body-worn cameras, and other unsynchronized recording devices. When multiple user-generated recordings are available, audio analysis of the recordings could provide spatial and temporal information about the location and orientation of sound sources, including but not limited to gunshots.
Maher’s gunshot audio analysis was used in the high-profile trial of Cleveland police officer Michael Brelo in 2015, where Maher concluded that 15 of 18 rapid succession shots were fired from Brelo’s gun. Independent FBI investigations corroborated his conclusions.
“Audio recordings can be crucial evidence to reconstruct shooting events,” said Gregory Dutton, a program manager at NIJ. “But to make this most useful to the justice system, we need a foundation of research to help the courts assess the value and limits of an expert’s testimony.”
Republished courtesy of NIJ.