Can you really solve crimes by listening? Welcome to the world of the forensic audio examiner/analyst – that is, an expert who investigates sound to provide evidence for legal cases. It may seem odd but these specialists are like radio drama listeners, using sound to fill in a visual picture of a world.
Sounds can be visually evocative: a certain kind of police siren, a particular type of rain, a jingle, a station announcement, they all evoke places, moods, and events. Imagine we have the recording from a smart speaker of a woman dying in a locked room; the audio analyst might be called on to use all their training and skilled listening to reconstruct from sound how the death happened. (In fact, this happens in Exemplar…)
Almost any event has sounds that add important layers of meaning to what has happened. Imagine the scene: the police have smashed down a young man’s front door and are confronted by the resident with a gun. A shot goes off and a policeman lies dead. The police say they knocked and announced themselves, the resident says they didn’t and he thought they were burglars breaking in. Who is right?
The dispute hangs on what the resident could have heard at the time. One way of resolving it would be to contact a forensic audio analyst, who will examine the thickness of the door, the power of the police officer’s voice, the resident’s hearing, the acoustic properties of his room, and the layout of the flat to determine whether a warning was effectively given.
Often visual images are themselves traces of sound events: a pavement of people simultaneously jerking round to look for the source of a sudden noise; the lines of interference that gunshots and other loud noises leave on old video recordings; the shudder of windows in response to a nearby explosion. In the digital era, sounds themselves display as waveforms and can be broken down into its frequencies, harmonics, amplitude and more. Looking at sound is a valuable way to see what is smoothly expected and what is strangely anomalous.
Starring Gina McKee as Jess and Shvorne Marks as Maya.
Forensic audio analysts are typically called on to investigate the source and significance of sounds captured in a recording and the way a sound might have been experienced in a particular circumstance. But sound is everywhere so they also have to be experts in the physical properties of walls, doors, floors, etc, in the biology of human hearing, the nature of all and any kind of recording device, the acoustic signatures of guns, knives, fists, and other blunt instruments, every kind of vehicle, and more, and of course the acoustic properties of any sound you can think of in any possible environment.
Sometimes you’ve got a recording but it’s messy; you’ve cleaned it up but there are sounds that might be important evidence but equally might be irrelevant. In that situation what an audio analyst can do is make a recording, reproducing the exact conditions of the original situation, to try to eliminate extraneous sounds and focus on those relevant to your investigation.
Audio analysts investigating the Kennedy assassination wanted to explore whether an additional shot, apparently captured on a police radio, might have been fired from the “grassy knoll” or whether that was from another source, perhaps an echo from the gunshots emanating from the book depository.
To answer this they had to match the speed of the motorcade, the position of the shooter, the calibre of the rifle, the ambient sound, and even the atmospheric air pressure that will have affected the type of echoes in Dealey Plaza, Dallas, where the killing occurred.
Listen to all episodes of Exemplar, a modern-day thriller. Starring Gina McKee as Jess, a lone wolf scientist whose passion for sound makes her the UK’s leading audio forensic analyst, and Shvorne Marks as Maya, Jess’s new trainee.
A generation or two ago, forensic audio analysts spent their time cleaning up analogue tapes, removing background noise, speeding them up or slowing them down to hear what’s being said. Now it’s mostly digital and the techniques are much more sophisticated.
The next generation of tools involve Machine Audio Learning (MAL) – it’s the system Peter Jackson’s team, working on the Get Back documentary, used to clean up the notoriously noisy Nagra audio tapes that the original film-makers used to record The Beatles session. MAL – named by Jackson’s team after the Beatles’ faithful road manager Mal Evans – learned how to pick out guitars, then drums, then bass, then speech, and eventually started to learn the different Beatles’ voices, eventually producing crystal-clear recordings of the band in conversation or rehearsal with all the background stripped away.
This opens up extraordinary possibilities for forensic audiology (as well as challenges for the right to privacy!) being able to lift a quiet conversation out of a crowded room, picking an individual voice out of a crowd, and being able to track an individual vocal signature across numerous rooms, streets, phone networks, and Zoom calls. It’s clear that the future of crime investigation will continue to be audiological.
Exemplar is part of Radio 4's Limelight podcast series – for fiction and drama serials you won't want to miss from Radio 4.
Award-winning drama exploring the psychological impact of murder. Inspired by a real forensic psychologist and her work in secure prisons.
Dr John Olsson talks to Michael Rosen about how he uses forensic linguistics, specifically authorship analysis, to solve crimes including murder, false witness and hate mail.
Seriously is home to the world’s best audio documentaries and podcast recommendations, with host Vanessa Kisuule.