|
Audio
and Video
Forensic Filtering Overview
WHAT
IS FORENSIC FILTERING?
According to dictionary.com, the definition of the word "forensics"
is "the use of science and technology to investigate and establish
facts in criminal or civil courts of law." A "filter", according
to the same source, is defined as "any of various electric,
electronic, acoustic, or optical devices used to reject signals,
vibrations, or radiations of certain frequencies while allowing
others to pass." Combining these two definitions conveys the
general meaning of audio and video forensic filtering - using
technical devices to reject noise (or conversely, to restore
or enhance the audio or image) to assist a law enforcement agency's
(LEA) criminal investigation and the court process.
|
.
Admittedly,
this definition broadly encompasses filtering at any stage of
the investigative or court process - whether it is accomplished
by a forensic examiner in a laboratory, a special agent or technical
surveillance specialist in the field, a linguist at a workstation,
or is simply a built-in feature of the collection, recording,
or transcription equipment. Regardless of whether one considers
filtering that happens at the time of collection to be "forensic"
filtering or not, it is useful for the purposes of this discussion
to include it as many of the issues are the same or similar.
In fact, to properly address filtering one has to talk about
"noise" and "interference" and they are introduced into the
evidence, among other places, at the time of collection.
|
| |
NOISE
& INTERFERENCE
The noise and interference that make it onto the evidence recording
may occur in the acoustic or visual environment that the microphone
or camera, respectively, are located in. Some common examples
of acoustic noise and interference include:
•••
resonance
•••
echo
•••
engines
and other machinery
•••
air
conditioning and fan hum
•••
live
music
•••
radio
and TV
•••
wind
•••
other
talkers
•••
aircraft
•••
vehicular
traffic and road noise
|
In
addition, the problems of noise and interference may enter via
the surveillance, recording, and playback equipment, which might
include anything from a body wire on a police informant or undercover
officer, a telephone tap (or "intercept"), a night vision camera,
a wireless transmitter, a closed circuit television (CCTV) camera,
a 911 (or other emergency services) phone recorder, a memo recorder,
security video time lapse recorder or multiplexer, thermal imager
in a helicopter, shotgun microphone or parabolic dish, video tape
recorder (VTR), or the like. All are potential entry points for
noise and interference, whether self-generated (e.g. tape wear,
multiplexing, and compression effects), conducted (e.g. AC mains
hum), or inducted (e.g. cell phone interference).
|
| |
USING
FORENSIC FILTERING
Once the evidence recording has been made, with all its accompanying
problems, forensic filtering can be used in the laboratory to
reduce or eliminate the noise and interference, as well as to
clarify or enhance the audio and video information - the process
otherwise known as restoration and enhancement. Of course, a technician
or forensic examiner, sometimes in conjunction with other professionals
such as linguists or phoneticians, may also perform other tasks,
such as evaluation, repair (of the tape cassette or the tape itself),
recording analysis and authentication, transcription, voice identification,
duplication, still printing, and, eventually, testimony. |
As
mentioned previously, the types of noises and interferences that
can be present in the evidence recording can be strong, subtle,
and/or varied, the filtering techniques must be powerful, precise,
and adaptive. Various devices (processors) for filtering are available,
including consumer and professional grade varieties of hardware
boxes and PC (personal computer) software packages or plug-ins.
The professionals operating them must also have appropriate training
in their function and use, which varies depending on the task
to be performed and the complexity of the processor. The goal,
of course, is to remove the noise and interference and then recover
the needed information, while in the process not introducing any
new problems (processing artifacts) which either cover up information,
annoy the viewer/listener, or otherwise make the evidence seem
unnatural (or tampered with).
|
| |
FORENSIC
FILTERING METHODS
The types of audio
filtering that may be employed in the restoration and enhancement
may include, but are not limited to, the following:
•••
bandlimiting
•••
(e.g.
highpass, lowpass,
•••
bandpass,
etc.)
•••
deconvolution
•••
(a.k.a. "one channel adaptive"
•••
or "LMS" filtering)
•••
equalization
•••
(i.e. graphic, parametric,
•••
spectral)
•••
gain
•••
mixing
and beamforming
•••
(e.g. lobing, summing, combining,
•••
mic steering, array processing)
•••
dynamic
range control
•••
(i.e. noise gate, compression,
•••
expansion, limit)
•••
combing
|
Similarly, for video (to include still images, by definition):
•••
bandlimiting
•••
(e.g. highpass, lowpass,
•••
bandpass, etc.)
•••
brightness
•••
contrast
•••
sharpening
(vertical and horizontal)
•••
frame
averaging
•••
image
stabilization
•••
median
•••
security
demultiplexing
•••
deinterlacing
•••
deblur
•••
(motion and focus,
•••
a.k.a. deconvolution)
•••
equalization
(histogram and
•••
homomorphic) |
| |
|
LOOKING FORWARD
It should be noted that for video processing, recent advances
in PC computing power and storage capabilities now allow live
filtering of not only audio, but also of full motion video.
This has opened the door for many new and powerful techniques
for restoring and enhancing video (hence, the references to
frame averaging and image stabilization above). No longer must
one select a single frame to process in isolation of all the
other information on the surrounding frames. The increase in
computing power has also made advanced spatial filtering techniques
possible (i.e. beamforming) for suppressing noise sources and
recovering remote audio.
|
In
addition, improvements in algorithms and our understanding of
human perception have also led to more powerful and effective
filters, such as adaptive spectral equalization/inversion and
(true) motion and focus deblur, and more embedded intelligence,
thereby making them easier to employ. The field of forensic filtering
continues to grow along with the advancements in technology, becoming
ever more powerful, complex, and subtle.
|
| |
|
If you have any questions, comments, or requests concerning
this exciting and interesting field, feel free to contact us.
We would enjoy hearing from you.
|
| |
|
