Principles of Forensic Audio and Video Analysis
To assist in an investigation, forensic experts can repair, recover, enhance and analyze audio and video recordings using an array of scientific tools and techniques.
Repair and Recovery of Evidence
Before audio and video evidence can be analyzed, it may first need to be repaired or recovered from damaged media or a damaged recording device.
Repairing evidence is especially common for analog and digital magnetic tape. It may need to be spliced back together or put into a new audio/video housing in order to recover the audio or video. In today’s digital world, CDs, DVDs, cell phones, portable cameras and other sources of digital media and recording devices can be damaged by heat, misuse, the environmental conditions of a crime scene, or simply on purpose by an offender. Even in these situations, the digital files can be recovered and used for analysis.
Evidence Enhancement
The most common function of forensic video and audio experts is to clarify a recording so that it is more apparent to investigators, attorneys and jurors what the evidence demonstrates.
To enhance a video recording, filters can be used to adjust the brightness and contrast, correct the color, crop and resize an image, enhance edge detail and reduce visual distortion. The speed of playback can also be adjusted to more accurately display the frame rate at which it was recorded.
Before Stabilization (Courtesy of Target© Forensic Service)
After Stabilization (Courtesy of Target© Forensic Service)
To enhance an audio recording, filters can be employed to improve clarity. This may entail removal of unwanted noise or enhancing the intelligibility of speech. Recordings will often be made in less than ideal circumstances, such as when someone is wearing a body wire. Utilizing audio engineering techniques may allow faint voices or events to be heard more clearly on playback.
Analysis, Interpretation and Identification
Authentication of recordings – In many criminal cases, the authenticity of the recording and the content of the recording may be called in to question. Forensic audio and video experts can examine a variety of characteristics of the audio or video recording to determine whether the evidence has been altered. This includes confirming the integrity (verification) of the recording, as well as authenticating that the content of the image or audio is what it purports to be.
If the ambient sound present on an audio recording changes abruptly, this could indicate that the environment where the recording took place suddenly changed. The volume and tone of a voice on the recording can provide clues as to distance and spatial relationships within a scene. Lighting conditions can be examined to estimate the time of day or environmental conditions at the time of the recording.
Technical details may also confirm information about a recording. For instance, an unnatural waveform present in the audio or video signal may indicate that an edit has been made. A physical identifier may be present in the signal on magnetic tape that can identify it as a copy or indicate that it was recorded on a particular device. Sometimes, a perpetrator will try to destroy audio or video evidence; however, using these methods, the recording can be analyzed to determine what occurred.
In the famous Watergate investigation, a great deal of effort was spent examining an 18½-minute gap in an audio recording of President Richard Nixon discussing the Watergate break in with his Chief of Staff. Analysis of the audio signature1 left behind in this erased portion allowed investigators to determine which White House tape recorder made the erasure and how many different erasures were made. Examining the level of AC hum recorded to tape even provided details on whether the recording took place in Nixon’s secretary’s office or in another location.
And new techniques are constantly being developed. A unique approach employed in the United Kingdom examines the low-frequency hum captured when a recorder is plugged into an electrical outlet or near a strong electrical current. This frequency will alternate slightly depending on the power load experienced at that time of day. By examining minute fluctuations of this frequency, analysts can determine whether a recording took place at the stated time and whether the recording is continuous and unaltered. This technique has been in use in the UK for over eight years; in the United States, this technique is still being researched and databases are being built for comparison.
Identifying people or objects on a recording – Identifying a person or object from an image on a video or voice on an audio recording requires training in Image Content Analysis or speech science. These examinations are detailed comparisons of an unknown recording to a known recording, or an unknown object to a known object in an attempt to make a positive identification. For instance, an image of a hat at the crime scene may be compared with a hat found on a suspect. The comparison techniques used in image analysis follow the same detailed comparison techniques as Fingerprint and Document examiners. Learn more about Fingerprints ▸ and Document Examination ▸ The analysis and comparison of voices is an evolving area of practice that can be controversial in criminal cases.