Audio forensics is, simply, the use of scientific investigation, technology and tools to determine the authenticity of sound for civil or criminal cases. The results of audio forensics can be used as evidence in various legal cases dealing with anything from divorce to industrial espionage. It's also not uncommon to see audio forensic analysis applied to surveillance tapes used in cases against employers or employees of a particular business.
Audio forensics has a long history with the United States military and government. In World War II, the technology was used to identify the voices of targeted enemies that sounded over radios and telephones. The use of a sound spectrograph, which plotted voice patterns' frequencies and amplitudes, helped analyzers identify people of interest. In recent years, audio forensics is used to analyze messages created by terrorists to help pinpoint their locations, the audio's creation time and other originating factors.
Some of the things that are commonly evaluated in an audio clip to determine its authenticity are background noises, voice frequency changes, noises emanating from recording equipment and stop, start and pause signatures. Any discontinuity in these areas may signal to the analyzer that the recording is inauthentic or has been compromised. Sometimes, these discontinuities can be masked behind background sounds and extra work has to be done to uncover them.
One of the most popular techniques used during analysis is to compare one unknown sound against a known one to identify it. This can be done in cases involving voices, where one speaker has been identified but the other has not. With voices, prosody, vowel formant trajectory, pitch striations, breath patterns, nasal resonances, speech pathologies and more can be used to identify speakers. However, voice identification is still a complex and controversial issue, and many find that identification and recognition in forensics analysis unsuitable for court submission and lacking in credibility.
Forensic audio analysis also emphasizes the use of signatures. Many audio forensic analysts rely on audio signatures that come from specific controlled sources. An example of this is when an audio analyzer uses frequencies from electrical power sources in their work. For these kinds of analyzers, an electrical socket can provide all the answers to their audio forensics questions.
While there are now many digital tools and programs at a forensic analyzer's disposal, the work involved in evaluating a piece of audio remains high. Analyzers may find themselves listening to a single piece of audio hundreds of times. Detection of discontinuities can be difficult, but not impossible for the trained and dedicated forensic audio analyzer.