One of the most effective software programs for detecting tampering is called Edit Tracker

Forensic audio analysis

Experts turn poor recordings into evidence

Jun 9, 2008

by James Careless

Extracting intelligible speech from a recorded murk of echoes and murmured words, or determining if a cassette audio tape reputedly made by Osama bin Laden is the real deal: These are just some of the challenges of forensic audio. Like its video cousin, forensic audio is an investigative science that works to clarify, filter, and clean up recorded speech and other sounds.

When done right, forensic audio can reveal oral information that clinches a case, prevents a crime, or reveals attempts at evidence tampering. The caveat is that forensic audio is limited by the quality of the original recording. A forensic audiologist can no more upgrade a telephone-quality cassette recording into crystal-clear full-range audio than a forensic video analyst can turn grainy VHS video into stunning HD.

Tackling Tape

Forensic audio is based on improving the intelligibility of recorded audio, not the recording itself. Its practitioners use tools designed to reduce background noise, filter out specific interfering tones (such as a 60 Hz hum caused by a cassette recorder's electric motor), and enhance any frequency ranges that could make recorded speech easier to understand.

Forensic audiologists also listen for background noise that can provide useful clues to investigators. For example, if there's a radio playing faintly in the distance, the station can be identified and the program lineup for the day in question verified. Environmental noises can also hint at a location. Unfortunately, even in today's digital age, forensic audiologists frequently find themselves working from cassette tape sources. Recordings are distorted by tape hiss, limited frequency response (especially if the microphone was cheap or hidden), and echoes caused by hard surfaces. Forensic audiologists are able to "see" tampering of audio evidence using software tools.

Police interrogation rooms and holding cells are frequent culprits for creating problematic audio. "Their bare walls, concrete floors, and bare-bones furniture can cause lots of echo-based distortions, while hidden microphones are often placed behind the suspect's head," said Gary Zacuto, president of Shoreline Studios, based in Santa Monica, CA. "You end up with very poor, hardly intelligible audio recordings in such environments."

Evidence tampering is a particular concern for forensic audiologists, with people using tape splicing, stop-and-start recording, and now editing on their home computers to create bogus recordings. Fortunately, such deceitful attempts can be detected by experts, using a number of software tools that display suspect waveforms on screen and actively seek irregularities that could indicate tampering.

"One of the most effective software programs for detecting tampering is called EdiTracker," said Tom Owen, president of Owl Investigations Services in Colonia, NJ. "The Russian scientist who originally developed it for the KGB brought it to market after the collapse of the Soviet Union. EdiTracker analyses the recording for telltale signs of tampering, such as changes in background noise, artifacts left by stopping-and-starting recording, and a number of other 'signatures' that can be electronically detected." The results of EdiTracker and other similar audio analysis programs are displayed graphically as waveforms, allowing investigators to actually "see" questionable sections of audio.

Finally, like fingerprints, voice prints are unique to each person. Every person speaks slightly differently from everybody else, using patterns and cadences that can be compared against suspect recordings.

Working with Arabic language experts, Tom Owen used voice prints to verify that a 2002 cassette tape reputedly recorded by Osama bin Laden in 2002 was, in fact, bin Laden. "Ironically, we used another piece of Russian-made software called MultiSpeech, which was also originally developed for the KGB," Owen explained.

Case Files

The forensic audiologists interviewed for Government Video have pretty much seen -- and heard -- it all. For example, rock guitarist-turned-forensic audiologist Arlo West runs his own company, Creative Forensic Services, in Auburn, ME. A student of Owen, West helped police convict 16-year-old Sarah Johnson for the double murder of her parents in Sun Valley, ID. "The sheriff's detective sent me a cassette tape recording done with Miss Johnson the morning after the murder," West explained. "The recorder's built-in microphone had been set closer to the detective than the suspect, and there was also some kind of malfunction that distorted the audio. Before analysis, the 45-minute interview was unintelligible. After filtering and enhancement, we achieved about 90 percent intelligibility." The recovered audio generated 60-70 pages of text, which helped convict the suspect.

West is currently working on a case where his client was imprisoned on a drug charge, based on covert surveillance audio. West's analysis of the audio, which was time-stamped by the investigating officers at its beginning and end, has revealed a few disturbing facts. First, the supposedly-continuous audio is more than five minutes longer than it should be, and the songs played before and after the alleged drug deal took place were aired on different days, according to radio station logs West has obtained. His investigation continues.

Owen played a role in the legal drama of Gambino family crime boss John Gotti (also known as the "Teflon Don") following Gotti's conviction and incarceration in 1992. "Although they were not supposed to, the federal government secretly recorded meetings between Gotti and his lawyer in prison," Owen said. "In these recordings, Gotti gives orders to his lawyer with respect to John Gotti, Jr. and the rest of Gotti's family. "When you listen to the recordings, they sound like typical jail tapes. There's lots of sound bouncing around and echoes, and all kinds of people are talking, because the meetings were held in the prison's Visiting Center. The government claimed that these murky recordings contained incriminating evidence, but we didn't hear any after we cleaned them up. The jury agreed, threw out the government's 'evidence,' and there was a mistrial."

Reduce The Need

So how do I get your department staffed with forensic audiologists pronto? The answer is that there is no fast and easy way to hire forensic audiologists, mainly because there are very few experts training them.

"There is one accredited forensic media course at the University of Indianapolis run by the Law Enforcement & Emergency Services Video Association [LEVA], but it is strictly for video," said Owen. "The New York Institute for Forensic Audio, which is a division of my own company, offers training courses and analyst certification. Graduates of my course have also been training audiologists across the country."

The fact that forensic audiologists are few and far between does not mean that the science of forensic audio cannot be applied by law enforcement agencies large and small. The best way to benefit from forensic audio analysis in the absence of experts is to reduce the need for it.

Case in point: Most police interrogation rooms and cells are acoustically hostile to audio recording. Their surfaces are hard, causing audio to bounce around the room like echoing ping-pong balls. Also, departments often fail to mount their microphones properly, especially if they attempt to conceal them. What audio is captured is then recorded onto noise-prone cassette tapes or recorded digitally using such high compression that a tremendous amount of audio clarity is lost.

In the same vein, some patrol officers equipped with on-body microphones neglect to turn them on when dealing with drivers. Plus, in-car microphones are often improperly mounted to capture suspects speaking among themselves. Meanwhile, some undercover officers do such a good job concealing their microphones that quality audio can't penetrate all the coverings.

In so many instances, the solution to these problems is prevention. Having an audio professional properly place and install microphones -- and taking the expert's advice on proper acoustic surfacing -- can ensure good quality, intelligible audio. If you add a redundant digital recording system, minimal compression, and a properly organized audio archive, most of your recorded audio problems should disappear. In those cases where they still arise, especially when working with third-party audio collected by someone else, forensic audiologists can help.


Creative Forensic

Digital Audio

Recording Tips

Microphones work best when the person being recorded is close to them and speaks into them. Proper microphone placement and minimally-compressed digital recording are the best ways to reduce the need for forensic audio. Here are a few more tips to improve your audio evidence:

  • Test the quality of your audio recording before using it for evidence collection.
  • Rooms with curtains and fabric-covered furniture have fewer echoes than hard-surfaced rooms with metal/wooden furniture.
  • Use a second audio recorder to provide a backup should the first one fail. Use UPS battery systems to keep them running if the main power fails.
  • Properly archive and catalog your recordings, and make copies of valuable recordings.
  • When in doubt, call an audio expert for help. Don't lose a case because you cut corners.


Forensic audio improves the intelligibility of recorded audio, not the recording itself.

"Government Video" magazine

 June, 2008