Forensic Sound Lab

A FORENSIC AUDIO PRIMER

©Jay Frigoletto

Most people today understand that audio-visual recordings can be quite compelling at trial, and if compelling enough, may even prevent a case from going to trial at all. A person strenuously denying certain events may not complain quite so loudly when faced with a recording of his own voice contradicting his claims. The strongest evidence can compel a settlement, plea agreement, or acquittal.

Recorded evidence of the audio variety comes in many forms, including but not limited to depositions, wiretaps, 911 emergency calls, telephone answering machines, voice mail, handheld recorders, dictation devices, and the sound track from video recordings. These can be recorded to a variety of media, analog or digital. Analog recordings may range from reel-to-reel analog tape of a variety of sizes, formats and speeds, to cassette, microcassette, or videotape. The prevalence of digital media opens wide the field to CD, DVD, and Blu-Ray discs, Digital video formats like DV and HDV, computer hard drives, and flash memory in various forms from USB keychain devices to card formats like Compact Flash, Smart Media, Memory Stick, and more. Further muddying the waters are multiple digital audio formats that can be stored on these media, from common ones like MP3, AAC, WAV, or AIFF, to more esoteric or even proprietary formats.

A few decades ago, the mere existence of an audio recording did not guarantee that it would be useful. Poor quality recording devices, bad recording practice, improperly stored tapes, excessive background noise, or compromised quality due to the location of surreptitiously placed microphones could render an important piece of recorded evidence useless, or at the very least, greatly diminish its effectiveness. The tools available to enhance recordings at that time were mostly basic filters, EQs, compressors, and noise gates. If the quality was decent to begin with, you may be able to "sweeten" it a bit, but removing severe, interfering noise, or pulling buried events or voices out of the background was essentially impossible.

Starting in the 1990s digital audio technology began to accelerate and audio enhancement tools that had previously been impossible to imagine were becoming reality. One early leader was a company called Sonic Solutions, with a product called NoNoise. The company was founded by former trail-blazing employees from Lucasfilm, the company behind the sound and effects for George Lucas’ Star Wars films. Another early competitor was the UK-based Cedar Technologies. Unsurprisingly, these two companies still are among the most respected in the industry. Still, the tools were rare and expensive, on the order of tens of thousands of dollars.

Flash forward to the present and an additional few decades of development have seen newer, even more impressive tools become available. While the best of them are still priced in the thousands of dollars, they are no longer in the tens of thousands. There are even entry level tools available in the hundreds of dollars that home audio enthusiasts can use. Of course a simple piece of software doesn't help without a powerful computer, professional quality analog to digital and digital to analog converters, high resolution monitor speakers, and a quiet, acoustically designed and treated space for critical listening.

It takes far more than tools to do the job. As in any field, experience counts. One must have not only a solid foundation in audio technology, but also additional study, experience, and/or training in forensic applications. In addition, one must become familiar with related topics, such as chain of custody and rules of evidence.

Audio enhancement, restoration, or “clean-up” is done with one of two basic goals in mind. One enhancement is performed for audibility, and the other, for listenability. An audibility enhancement is done to try to bring out more detail, or buried voices or events to the point that they can be better identified, understood, or transcribed. A listenability enhancement is performed to reduce distracting noise, distortion, or intrusions in a recording that divert your attention from the core event or dialog. A listenability enhancement is "cleaner" and more pleasant to listen to, while an audibility enhancement may still sound somewhat "dirty," though the clarity, quality, or amount of available information able to be pulled from the recording is improved.

While the enhancement and clean-up of audio is important, the analysis of the resulting sound can be just as important. One may be called upon to authenticate a recording, or check for evidence of tampering, editing, erasing, taping over, background discontinuities, or factors otherwise compromising the recorded evidence. Analog tapes can be "developed" with ferrofluid solutions and examined under a microscope to look for record or erase head signatures and non-continuities. Spectral analysis can reveal multiple bias currents in analog recordings and background discontinuities in both analog and digital recordings.

A specialty unto itself, voice identification is another essential area of forensic audio. The most common practice involves T-F-A Spectographic analysis. The Spectograph was developed by Bell Labs for use in World War II intelligence gathering. Today, in addition to a waveform analysis, one uses a spectograph to examine the Time, Frequency, and Amplitude components of the sound to rule out or rule in a particular speaker. In critical situations, this may sometimes be done in conjunction with a speech scientist. In trying to identify a speaker, one takes a known exemplar of the voice of the person in question, and compares it to the recorded evidence to try to establish the likelihood that it is the same speaker. Each person has a unique speech pattern, and one compares elements such as formants, fricatives, plosives, pitch striation, nasal resonance, breathing patterns, or any identifying speech pathology.

Voice Identification evidence, though long practiced, is still controversial as there are many variables to consider, and sometimes subjective expert judgements to weigh. The FBI Forensic Services Handbook states that their voice identification services should be used for investigative purposes only. In court, it can be one factor in corroborating support of other evidence, but it is unlikely to be the linch pin of a case.

Other uses of forensic audio can include audibility analysis, sound identification, establishment of timelines, gunshot analysis, engine or machinery frequency, reverberation characteristic, location, or other elements of sound that may be relevant to a case. With officer body cameras coming into widespread use the amount of available recorded evidence is expanding exponentially.

In addition to lab work, or sometimes even on-site data collection and analysis or experimental recreation, forensic audio experts can be called on to testify. The standard for admissible expert evidence is derived from the Frye and Daubert rulings of 1923 and 1993 respectively. A forensic audio expert may simply need to explain the work done, and the non-destructive nature of the enhancements. Other times one may need to explain conclusions, and defend one's position against opposing counsel or other experts.

The future of forensic audio seems bright. Tools are becoming more powerful, more available, and more affordable. Though a young field by comparison to many others, it is reaching a state of maturity and demand for services is increasing as the proven value is increasingly recognized. Popular entertainment, such as the CSI programs and others, is also contributing to the recognition and awareness of the process, though sometimes is a double-edged sword as it can raise unrealistic expectations. Whether for defense or prosecution, criminal or civil, corporate or individual, the value of recorded evidence is undeniable, and the expert enhancement and/or analysis of such evidence can be the difference between success or failure.