Dean Catoggio
Unit Leader
Audio Visual Unit
Victoria Police Forensic Services Department
Macleod, Victoria, Australia
Bryson Shearwood
Forensic Officer
Audio Visual Unit
Victoria Police Forensic Services Department
Macleod, Victoria, Australia
Jason Ferridge
Forensic Officer
Audio Visual Unit
Victoria Police Forensic Services Department
Macleod, Victoria, Australia
Introduction
| Video-Signal Recording and
Reproduction | Replay
Radio-Frequency Waveform | Replay
Tracking Position | Audio-Track
Configuration | Blind
Validation Trials | Materials
and Methods | Results
and Discussion | Conclusions
| Acknowledgments
| References
Introduction
Forensic audio-recording authenticity analysis became prominent
following a 1974 report prepared by the advisory committee appointed
to examine the erased portion of an audiotape recording between
former U.S. president Richard M. Nixon and his chief of staff, H.
R. Haldeman (Koenig 1990). The tapes became known as the “Watergate
tapes.”
Today, many organizations conduct forensic video authenticity analysis;
however, little documented evidence exists of research conducted
in this area (Aldridge 1995). Although digital communication and
information technologies are both emerging and indeed converging,
many people still own analog systems (Catoggio 2001).
This paper outlines a method for determining whether a particular
videotape was recorded in a particular video recorder. The method
uses three techniques that compare replay radio-frequency waveform,
audio-track configuration, and replay-tracking position. The techniques
require the use of a storage oscilloscope, plotter, video monitor,
and video player.
An investigator submitted to the Victoria, Australia, Police Forensic
Services Department a number of illicit video home system (VHS)
recordings and recorders. The copies had been sold publicly without
classification and without the consent of the copyright owner. In
Australia, films and videotapes must be classified according to
the Commonwealth Classification (Publications, Films and Computer
Games) Act 1995. The Commonwealth Copyright Act 1968 protects copyright
ownership of literary, dramatic, musical, and artistic works. The
investigator requested that the department determine whether any
of the copies had been made in any of the video recorders that had
been seized from the suspects. Before this could be done, it was
necessary to determine whether VHS video recorders impart any electronic
or physical identifying features onto videotapes. A subsequent review
of the available literature and the results of early experiments
confirmed this phenomenon.
When a video recording is replayed, an examination of the video
head-switching transients (Aldridge 1995), signal dropouts, and
picture content (Owen 1989) can help to identify the recorder used.
Traditional audio authentication concepts can also be applied to
the examination of video machines (Koenig 1990).
This paper is principally concerned with three other features that
can be used for identification purposes: replay radio-frequency
waveform (Klasen 1998), audio-track configuration, and replay-tracking
position (Catoggio et al. 2002). To validate and test the robustness
of the examination techniques, the authors conducted two blind trials
using 50 recorded videotapes.
Video-Signal Recording and Reproduction
This section provides a basic explanation of the aspects of video-signal
theory that are important to understanding the analysis techniques
discussed in this paper.
The Video Picture
The transmission of sound and picture information requires a signal
bandwidth of up to 7 megahertz (MHz) (Botto 1992). In Australia
and many countries in Europe, video signals are transmitted using
the Phase-Alternation Line (PAL) standard. A transmission speed
of 25 picture frames per second gives the illusion of continuous
motion. Each frame has 625 horizontal lines, which are divided into
odd and even fields as shown in Figure 1 (Mee and Daniel 1989).
Each line represents the picture information, which varies in color
and brightness along the length of the line. During image capture,
a camera scans one line at a time. A line-sync pulse marks the end
of each line. When the signal is displayed by a television, this
pulse tells its cathode-ray beam to begin tracing the next line.
The Séquential Couleur avec Mémoire (“Sequential
Color with Memory,” SECAM) standard is used in France and
similarly employs 625 lines and 25 frames per second. The National
Television System Committee (NTSC) standard employs 525 lines at
29.97 frames per second and is widely used across the United States.
Figure 1:
Portion of One Picture Frame Consisting of Odd and Even Fields
Two-Head Helical-Scan Recording System
The Japan Victor Company (JVC Group, Yokohama, Japan) introduced
the VHS video-recording format in 1976 (Botto 1992; Mee and Daniel
1989). It uses a two-head helical-scan method to magnetically record
a video signal to tape. Figure 2 (Botto 1992) depicts how, using
this method, tape is transported from a supply reel to a take-up
reel by audio- and video-record heads and tape guides. Tension is
applied to the take-up spool, while a rubber pinch roller and capstan
regulate the tape speed.
Figure 2:
Illustration of VHS Tape Threading and Head Wrap
A high video-head-to-videotape speed is required to record a signal
bandwidth approaching 7 MHz. This speed is achieved by placing two
oppositely set heads into a rotating drum as demonstrated in Figure
2. The drum is tilted by approximately six degrees and rotates in
the same direction as the linear movement of the tape (International
Electrotechnical Commission 1994). The picture information contained
in each field is recorded diagonally as the head rotates across
the tape.
Frequency Modulation
In the context of recording, the Phase-Alternation Line video signal
is transmitted directly through frequency modulation (FM) of a 4.43-MHz
subcarrier (Mee and Daniel 1989). FM is used to transmit all picture
information, including color and brightness.
Audio Recording
Audio information is recorded using a stationary head on a longitudinal
track at the top portion of the tape. In addition, high-fidelity
stereo audio is recorded through careful integration with the picture
information transmitted on the FM subcarrier. Figure 3 (Mee and
Daniel 1989) depicts the audio- and video-track configurations of
a helical-scan system. A stationary full-track erase head precedes
the drum to ensure that the tape is erased prior to recording.
Figure 3:
Illustration of a Two-Head Helical-Scan Video-Recording System and
Recording-Track Configuration
Tape Guides
The tape guides and slant pins cause the tape to wrap more than
180 degrees around the video drum. This allows the second video
head to begin its scan of the tape just as the first video head
is ending its scan. Figures 2 and 3 illustrate how take-up and supply
guides and pins are used to ensure proper alignment of the tape
to the heads.
Tracking
To guarantee optimum reproduction of the video picture between
machines, the tracks must be scanned during playback in precisely
the same way as when they are recorded. Incorrect tracking may cause
picture flicker, loss of color, high-fidelity audio distortion,
and other symptomatic picture and audio degradation.
During recording, the phase of the drum rotation is detected and
adjusted according to the vertical frame sync pulse. This signal
also causes a constant capstan speed. A split from this signal then
becomes a reference signal, which is recorded to a control track
on the bottom portion of the videotape.
During playback, a sync signal is generated and compared with the
detected drum rotation, which, in turn, is controlled so that it
remains constant and in phase. Comparing the sync signal with the
control signal regulates the capstan rotation and, hence, the speed
of the tape. Controlling the tape speed causes the video head to
trace precisely along the recorded track.
Most domestic video players allow a degree of manual or automatic
tracking adjustment to ensure compatibility among machines (Mee
and Daniel 1989). Manual adjustment causes a time delay between
the replay of the control track and the generation of the head-switching
pulses (Aldridge 1995).
Although beyond the scope of this paper, other tracking systems
exist. Some of these systems use digital processes (Mee and Daniel
1989).
Relevance of the Replay Radio-Frequency
Waveform
Frequency modulation of a 4.43-MHz subcarrier transmits the picture
information. This signal is converted to magnetic energy by the
video head, which, in turn, magnetizes the videotape. During playback,
the video heads alternately scan the diagonal signal tracks corresponding
to the odd and even field lines. As they scan, the heads detect
a change in magnetic flux, which induces a weak signal. This weak
signal is directly proportional to the recorded FM 4.43-MHz subcarrier.
Figure 4 (Botto 1992) shows that the signal produced by each video
head is preamplified and eventually combined by a switching amplifier.
The combined signal is often referred to as the replay radio-frequency
waveform or frequency-modulation waveform. Radio frequency, commonly
referred to as RF, is used to indicate the complete range of frequencies
used for the transmission of information by electromagnetic waves
(Spottiswoode 1978). When viewed on a 100-millisecond (ms) oscilloscope
trace, the waveform looks similar to the one depicted in Figure
4.
Figure 4:
Diagram Showing Video-Signal Reproduction
The waveform analysis discussed in this paper is based on a common
technique for servicing video recorders. Examination of the waveform
envelope is used to check the recorder’s tape alignment and
the amount of head wear and tear. Adjusting the tape guides alters
the tape alignment. Assuming that the combination of head wear and
the position of the tape guides might be characteristic of specific
recorders, the authors conducted some early experiments. Figures
5.1 to 5.10 present a summary of findings when comparing the effects
of tape-guide position, tape wear, and head wear on the waveform.
The waveform shapes are exaggerated to aid explanation.
Figures 5.1-5.10:
Comparison of Replay Radio-Frequency Output Waveform Shapes and
Causes
Figure 5.1 depicts the ideal replay waveform of a tape. This assumes
that the recorder has no head wear and that it is properly aligned.
Figures 5.2 to 5.10 depict replay waveforms in which the recorder
is suffering from head wear, misalignment, or other problems. As
expected, the individuality of each recorder becomes more distinct
the further the waveform deviates from the ideal. During these experiments,
all tapes were replayed in a device set for optimum tracking when
playing a National Panasonic VHS alignment tape (No. VFM8180HADH,
Matsushita Electric Industrial, Osaka, Japan).
Experimentation also has shown that although the general waveform
shape is consistent throughout the replay of a recording, fluctuations
can occur depending on changes in image content and television-channel
reception. Examples include moving from a predominantly white scene
to a black scene and intermittent dropouts, respectively.
Over time, machine use can cause head wear and tape misalignment,
altering the waveform. The effects of time and use should be considered
when determining whether a particular recorder was used to make
a tape recording. The effect of dirt buildup on the video heads
and the quality of the videotape also should be considered.
Relevance of Replay Tracking
Tracking is adjusted to ensure compatibility among machines. Typically,
tracking is adjusted using a rotary dial or a keypad system. A measure
of tracking adjustment may be displayed on the screen or on the
machine display. Figure 6 provides an example of an on-screen tracking
indicator. A tape that is recorded and replayed on the same machine
will display a central tracking position. Experimentation has shown
that optimum tracking can vary among machines. In most cases, the
deviation from the central tracking position is consistent for the
duration of the recording. Mechanical faults may cause variations
in tracking over time.
Figure 6:
Example of an On-Screen Tracking Indicator
The very coarse tracking indicator shown in Figure 6 can be useful
in comparing tapes recorded on different machines. It would be expected
that, similar to the replay radio-frequency waveform, the individuality
of each machine will become more distinct the further the tracking
deviates from the ideal central position.
It is important to note that adjusting the replay tracking will
cause a noticeable change in the shape of the replay radio-frequency
waveform. Therefore, it is only valid to compare waveforms displayed
at the same replay tracking (nominally central position) or at their
respective optimum settings.
Relevance of Audio-Track Configuration
Audio information is recorded using a linear track, and high-fidelity
audio is modulated with the video signal along the diagonal tracks.
Examining the audio-track configuration can provide a simple clue
in matching a tape to a recorder.
Recorders fit into one of two audio categories: either they possess
a linear track only, or they possess both a linear and a high-fidelity
track. The authors do not know of any video recorder that records
only high-fidelity audio.
The absence or presence of a high-fidelity track usually is indicated
on the control display of a machine capable of playing back both
linear and high-fidelity tracks. Some video players also allow the
user to select playback of either the linear track or the high-fidelity
track.
Some recorders, particularly professional or high-end consumer
products, allow the user to manually turn off the high-fidelity
recording capability. Therefore, if two tapes are presented for
comparison, the absence of a high-fidelity track on one does not
discount that the tapes share a common origin. Conversely, if a
questioned recorder does not possess high-fidelity recording capabilities,
yet the tape in question contains a high-fidelity recording, then
this recorder could not have been used. Indeed, this finding would
negate the need to conduct tracking and radio-frequency waveform
analyses.
Description of Blind Validation
Trials
A typical forensic case would involve the comparison of an exhibit
tape and a sample recording from an exhibit machine. However, the
validation trials were not based on this case scenario because of
the 50 percent chance of a random match. Instead, the authors decided
to test the robustness of the technique by comparing multiple items
recorded on multiple machines.
The first of two trials was conducted to familiarize the authors
with the limitations of the examination techniques and to determine
accuracy. With the aid of hindsight, the authors conducted a second
trial approximately four years later to refine the examination techniques.
Unique identifier codes were placed on each of 50 TDK VHS videotapes
(No. HS180, TDK Corporation, Tokyo, Japan). An independent assistant
then distributed the tapes to 27 volunteers, as depicted in Table
1. Some volunteers were given one tape, whereas others were given
two or three. Each of the volunteers then recorded approximately
60 minutes of an Australian television broadcast on each of the
tapes. The volunteers were specifically asked to use standard record
speed and not long-play mode. The independent assistant then collated
the recorded tapes and submitted them to the authors, who were unaware
of how the tapes had been distributed.
Table 1: Distribution
of Videotapes to Volunteers
To provide a balance between an adequate sample size and a manageable
number of items for comparison, the trial was limited to 50 tapes
recorded in 27 different machines. The task was to use the techniques
discussed in this paper to determine which of the 50 tapes were
recorded on the same machine. Single tapes recorded on different
machines were introduced so that the authors would not expect to
find a balanced number of matched pairs or groups of three.
In the second trial four years later, the process was repeated
and included some volunteers from the first trial who had retained
their original machines. Trial 2 tapes also were allocated as described
in Table 1. During Trial 2, the authors were again unaware that
the distribution profile was the same as in Trial 1.
Materials and Methods
Initially, an attempt was made to compare only replay radio-frequency
waveform to identify groups. In this case, comparing the radio-frequency
waveform of each tape with the radio-frequency waveform of the other
49 tapes proved very difficult and tedious. Therefore, comparison
of replay-tracking position and audio-track configuration was incorporated
into the testing regime. The testing procedure for Trial 1 is presented
in Table 2.
Table 2: Trial
1 Videotape-Testing Procedure
The effects of the tape brand and age were not considered during
this trial. However, a poor-quality or well-used tape might cause
some instability in the replay radio-frequency waveform.
Replay Machine
The Sony SLV-X827A (Sony Corporation, Tokyo, Japan) is a domestic
VHS Phase-Alternation Line video machine and was used to replay
all tapes. The machine was tested for functionality, ease of access
to circuit-board test points, and ability to consistently reproduce
a waveform approaching the ideal for a National Panasonic VHS Phase-Alternation
Line alignment tape (No. VFM8180HADH). The device was not calibrated
or aligned according to an international standard such as those
published by the International Electro-Technical Commission (Geneva,
Switzerland). The replay machine was used only in these trials;
therefore, the possible waveform-altering effects of time and usage
were minimized. The heads and guides were cleaned before each tape
was tested. As is the case for the comparison of audiotapes, under
these conditions, the precision of the replay device is considered
acceptable (Catoggio 1998).
Determining Audio-Track Configuration and Replay-Tracking
Position
A number of tapes did not possess a high-fidelity audio track.
This was indicated simply by the absence of a “HiFi”
light on the machine display.
The replay machine provided, on a Panasonic BT-H1490Y monitor,
a pictorial indication of replay-tracking position similar to the
one depicted in Figure 6. The automatic replay tracking was activated,
and a pictorial note was made of the tracking display. For the purposes
of this trial, all other picture and audio content was ignored.
Comparing Replay Radio-Frequency Waveform
A Roland DXY-1300 plotter (Roland Corporation, Osaka, Japan) was
connected to a Tektronix 2230 100-MHz digital storage oscilloscope
(Tektronix, Beaverton, Oregon). The replay machine provided a test
point for both the radio-frequency waveform and the switching-head
signal. The origin of these signals is depicted in Figure 4. The
signal from the radio-frequency waveform test point was displayed,
using the switching-head signal as a trigger to stabilize the image.
When displayed at 10 ms per division (i.e., a 100-ms trace), a sequence
of four complete waveform patterns is evident. These correspond
to consecutive odd and even fields. The Tektronix 2230 allows the
user to store and overlay waveforms, which is particularly useful
in comparing small differences in waveforms.
Guidelines for Comparison for Validation Trial 1
Visual comparisons were made of radio-frequency waveforms and replay-tracking
position. To aid this subjective process, the following guidelines
were used to critically compare tapes:
- The two tapes had a common origin if an assessment
was made that there were only minor differences in the features
listed in Table 3. It was expected that radio-frequency waveforms
of the same origin would display some minor differences due to
mechanical and electronic intravariability.
- The two tapes did not have a common origin if an assessment
was made that there was a noticeable difference in at least one
feature listed in Table 3. The noticeable difference was not expected
to be due to mechanical and electronic intravariability.
- It was inconclusive as to whether two tapes had the
same origin if the features of Table 3 were similar but there
was uncertainty as to whether they constituted minor or noticeable
differences.
Table 3: Waveform
Features: Points of Comparison
Examples of minor and noticeable differences in radio-frequency
waveforms are displayed and discussed in Figure 7.
Figure 7:
Comparison of Replay Radio-Frequency Waveforms
Results and Discussion
Results of Trial 1
Based on early experimentation, the authors had an expectation
of the range of differences in waveforms and the ability to distinguish
them. In part, this trial tested these expectations, which turned
out to be incorrect. Based on the guidelines presented in the previous
section, tapes were grouped according to similar waveforms, audio-track
configuration, and replay-tracking position. These groupings were
then compared with the actual groupings listed in Table 1. Table
4 presents a summary of findings for Trial 1, which was conducted
in 1998. Twenty-nine tapes were matched correctly; however, six
tapes were matched incorrectly. A wide variety of waveform patterns
was displayed. It was easy to assess the significance of obvious
differences; however, it was difficult to detect minor differences
and assess their significance.
Table 4: Summary
of Findings for Trial 1 Conducted in 1998
Guidelines for Comparison for Trial 2
After Trial 1, the authors had a far better understanding of the
significance of differences in radio-frequency waveforms. With this
experience and further experimentation, the authors made improvements
to the test procedure in an attempt to decrease the number of incorrect
matches. These improvements are shown in Table 5.
Table 5: Improvements
to the Test Procedure Made for Trial 2
Results of Trial 2
The number of incorrectly matched tapes was reduced to zero in
the second trial. The number of correctly matched tapes was reduced,
whereas the number of inconclusive results increased. Table 6 provides
a summary of findings for Trial 2, which was conducted in 2002.
Table 6: Summary
of Findings for Trial 2 Conducted in 2002
Conclusions
Comparison of replay radio-frequency waveform, audio-track configuration,
and replay-tracking position is a useful technique to aid video
recorder identification. An understanding of the significance of
minor or noticeable differences between waveforms is essential to
the accuracy of this technique. The authors developed an understanding
through an investigation of intra- and inter-machine waveform variations.
Consistency in determining differences can be facilitated by using
a guide, such as the examples in Figure 7. Viewing waveforms at
different replay-tracking settings highlights differences and improves
the robustness of the technique.
An expected improvement occurred in the results of the second of
two blind trials. This was attributed to changes in the test procedure
and interpretations that produced an increase in inconclusive results
but a decrease in false-positive results.
This method can be useful for screening a large quantity of items
for examination. Other video-identification methods can then be
applied to items that display a positive match or inconclusive result.
Other techniques include the examination of head-switching transients,
picture information, and audio-switching transients. This method
also can be adapted for the identification of other helical-scan
recording devices, such as digital videotape, digital audiotape,
and both analog and digital 8-mm video formats. The radio-frequency
waveform of a digital audiotape machine is known to alter according
to guide heights (Issacs 2000).
The authors hope to determine the effects of time and machine usage
on the replay radio-frequency waveform produced by video recorders.
This can be achieved by comparing tapes of the same origin from
Trials 1 and 2, which were recorded four years apart. It also may
be useful to compare the waveforms produced by recorders of the
same brand and model.
Increasingly since the 1960s, law enforcement agencies, business
owners, and even members of the public have used audiovisual devices
to capture evidence of serious crimes, such as murder, assault,
or extortion. Although new technologies continue to emerge, significant
use of analog devices remains. Meanwhile, information and communication
technologies are converging, resulting in the advent of new devices,
such as multifunctional cellular phones. As occurred with the Watergate
tapes, the authenticity of recordings produced by such technologies
may be questioned, giving rise to the need for new methodologies
and, hence, new research opportunities. Given the lack of literature
relating to forensic video analysis, the method presented in this
paper may form a useful part of the forensic practitioner’s
examination tool kit.
Acknowledgments
This paper was adapted from a presentation to the 16th International
Symposium on the Forensic Sciences, Canberra, Australia (Catoggio
et al. 2002).
The authors thank Julia Caluzzi, Peter Woodman, Chantel Marise,
Michael Liddy, and Jim Pearson for their support and suggestions;
Greg Cornel and Layton Moss for assistance in early experiments;
and Bruce Koenig, Doug Lacey, and Graeme Kinraid for their technical
reviews.
References
Aldridge, J. Video Tape Analysis. In: Summary Report of England
Visit. R. Prandolini, Australian Federal Police, August 1-4,
1995, pp. 8-32.
Botto, D. A. Basic guide to colour TV and VCRs. In: Electronics
Australia Publication. Federal Publishing, Alexandria, Australia,
1992, pp. 7-12, 54, 71, 88, 892.
Catoggio, D. Audio Signal Analysis Method Validation.
Presented at the 14th International Symposium on the Forensic Sciences,
Adelaide, Australia, October 12-16, 1998.
Catoggio, D. Big Brother and Trends in Forensic Examination
of Audio Visual Evidence. Presented at the 7th Indo Pacific
Congress on Legal Medicine and Forensic Sciences, Melbourne, Australia,
September 21, 2001 (revised 2002).
Catoggio, D., Shearwood, B., and Ferridge, J. Video Recorder
Identification: Comparison of Tape Replay RF Waveform, Audio Track
Configuration and Replay Tracking. Presented at the 16th International
Symposium on the Forensic Sciences, Canberra, Australia, May 13-17,
2002.
International Electrotechnical Commission. IEC standard 1994-01:
Helical-scan video tape cassette system using 12.65mm (0.5in) magnetic
tape on type VHS-Part 1: VHS and compact VHS video cassette system,
Geneva, Switzerland, 1994.
Isaacs, P. R-DAT. In: Studio Sound. United Business Media
International, London, England, October 2000, pp. 81-83.
Klasen, L. Image analysis. In: Proceedings of the 12th
INTERPOL Forensic Science Symposium. R. S. Frank and H. W.
Peel, eds. Forensic Sciences Foundation Press, Lyon, France, October
20-23, 1998, pp. 261-302.
Koenig, B. E. Authentication of forensic audio recordings, Journal
of the Audio Engineering Society (November 1990) 38:3-33.
Mee, C. D. and Daniel, E. D. Magnetic Recording Handbook: Technology
and Applications, McGraw-Hill, New York, 1989, pp. 91, 886-892,
917.
Owen, T. An introduction to forensic examination of audio and video
tapes, Journal of Forensic Identification (1989) 39:75-85.
Spottiswoode, R., ed. The Focal Encyclopaedia of Film and Television
Techniques, 3rd ed., Focal Press, London, England,
1978, p. 602.
|