Forensic Gunshot Acoustic Analysis is Heating Up. Don't Get Burned
Gunshot acoustics hold plenty of investigative promise, but analysis can be difficult even for experts.
If someone fires a weapon in the Federal Reserve Bank in San Francisco, a gunshot detection system instantly locks all security doors, trapping the shooter inside. In metro areas of cities like of Tucson, Los Angeles, and elsewhere, when a shot is fired, within one second of weapon discharge acoustic sensors stationed in strategic urban locations pick up the sound, automatically determine range and bearing, then notify authorities.
Both systems are examples of the role gunshot acoustics are beginning to play in law enforcement and forensics. Forensic gunshot acoustic analysis doesn’t stop with sensor networks. Forensic acoustic experts are looking at ways to exploit the proliferation of mobile recording devices for potential forensic gunshot acoustic evidence that may lurk in the audio files of cell phones and tablets that happen to record a gunshot.
“The ubiquitous availability of audio-capable mobile devices indicates that audio evidence will more frequently be part of forensic investigations,” said gunshot acoustics expert Robert Maher, the head of Montana State University’s Electrical and Computer Engineering department. Maher first became interested in forensic gunshot acoustics after a conversation with a defense attorney regarding a who-shot-first question based on an emergency call center recording. Although that particular case did not hinge on audio forensic analysis in the end, it did pique Maher’s interest enough to perform a series of scientific measurements. He is now a recognized expert in the field.
Maher said that under the right circumstances, recordings of gunshots can potentially tell investigators who shot first or where the weapon was located when discharged. But factors such as microphone quality and angle and direction of the weapon’s muzzle can change the way a gunshot sounds in a recording.
Forensic investigators can reach wrong conclusions if they fail to understand the impact of these variables.
Calling the Shots
For many years, gunshot acoustic research was largely restricted to military, proprietary, or shooting range noise-abatement arenas. Little objective information found its way into open scientific literature for use by the audio forensics community. That seems to be changing. Maher has published several papers on forensic gunshot acoustics, including "Deciphering gunshot recordings" (Proc. Audio Engineering Society 33rd Conference, 2008) and "Acoustical modeling of gunshots including directional information and reflections" (Proc. 131st Audio Engineering Society Convention, 2011).
There is forensic promise, and investigative peril.
A recent FBI-sponsored study (“An introduction to forensic gunshot acoustics”, 162nd Acoustical Society of America, 2011) illustrates the complex nature of forensic gunshot acoustic analysis.
Steven Beck, principal scientist at BAE Systems, along with FBI agents Hirotaka Nakasone and Kenneth Marr, performed gunshot acoustic studies in a controlled environment by placing microphones at a range of angles and distances from each blast in order to capture the sound pattern of a single round from multiple points of view. Gunshots produce two different sounds—a “bang,” caused by rapid expansion of gasses that push the bullet through the barrel, and a “crack,” caused by shockwaves in the air made by supersonic bullets.
The FBI used several different types and models of firearms, along with high-quality recording equipment, and multiple microphone placements relative to the firearm in order to vary range and azimuth angles. While it’s true that gunshot recordings that end up in the hands of forensic analysts are created in far less ideal conditions, Beck said analysts need to understand the underlying acoustics of impulse sounds, and then know what variations can occur and how the recording conditions can affect the signals.
According to Beck, examination of the waveforms showed that the pristine recordings match closely with theoretical models, but there were a number of unexpected findings.
One surprise the Beck team found was how important the angle of the gun was to the sound signature, or waveform, caught by the microphones. According to this research, a shot fired into the air produces a wildly different waveform from one fired in the direction of a recorder, even when both shots came from the same gun. Likewise, a heavy rifle blast fired at a great distance from the microphone can produce a sound signature that is almost indistinguishable from one produced by a smaller, lighter gun fired at close range.
There was another surprise—a bonus “bang” caught by recorders positioned at 90-degree angles to the barrel. The Beck team traced this sound to gasses leaking and expanding out of the side of certain types of firearms. This means investigators could potentially exploit this extra “bang” to identify the direction of the weapon from the recording device.
Maher said current gunshot audio analysis is concerned with three specific issues.
First, as found by the FBI study, there is increasing awareness that the acoustics of gunshots depends significantly upon the orientation of the firearm with respect to the recording device. Maher said that while many investigators assumed muzzle blast from a firearm was an omnidirectional impulse, more recent work has shown that sound level and waveform details vary greatly as a function of azimuth.
Maher said in another paper ("Directional aspects of forensic gunshot recordings", Proc. Audio Engineering Society 39th Conference, 2010), it is important for audio forensic examiners to recognize that the difference in level and waveform details between on-axis and off-axis recordings of the same firearm are often significantly greater than the difference between two firearm types at the same azimuth.
“This can have an important effect upon deducing the firearm type from a recording, especially if the orientation of the firearm with respect to the microphone is not known from some other source of information,” he said.
The second important area of gunshot audio analysis pertains to the effects of mobile device audio systems on the recorded waveform—particularly perceptual audio coders like MP3 and digital speech coding systems used in cell phones.
Maher said if the mobile device is located near the shooting position, the high level sound of the gunshot will typically overload the microphone, input amplifier, and related electronics, resulting in a severely clipped and distorted waveform. Then this distorted gunshot signal is presented to the device’s audio coding system, which is usually a speech coder designed to represent intelligible speech, not gunshots and highly distorted waveforms.
“Thus, interpretation of gunshot acoustic evidence that has been passed through a cell phone’s audio system is an area in need of further research,” he said.
Maher said the third issue is to determine the firearm type and ammunition based on acoustic evidence alone. The big issues are to ascertain the orientation of the gun and dealing with the quality and integrity of the audio recording system.
“In situations where a relatively pristine recording is obtained there may be opportunities to identify useful forensic features from the gunshot recording, but in most real-world cases the quality of the recorded waveform will be poor,” he said.
Still, in situations where the recording device is sufficiently far away from the firearm that the signal does not overload the recorder, Maher said it is feasible to determine the relative timing of the initial gunshot and any significant acoustical echoes caused by reflecting surfaces nearby.
“If more than one recording is available, such as multiple devices in different locations capturing the same event, it may be feasible to reconstruct the most plausible location of the firearm and its orientation,” he said.
Douglas Page writes about forensic science and medicine from Pine Mountain, California. He can be reached at firstname.lastname@example.org.