|
IN-DEPTH
Volumetric Display Technology
INTRODUCTION
Commonly misconstrued as ‘Holograms’, a Volumetric Display
is a device that fulfills all of the following three
criteria: it is capable of displaying an object that has a
physical height, width and length; observers are able to
revolve around the display, in one or more planes, without
losing visual contact with the image being displayed; and
two observers standing simultaneously at any two different
positions equidistant to the display should see different
‘sides’ of the image, relative to the observers’ respective
positions. This article will run through the stereoscopic
methods of producing pseudo-Volumetric Displays and briefly
discuss these display technologies, before presenting our
latest project in this field.
STEREOSCOPIC DISPLAYS
Etymologically, the suffix –scopy originates from the
Greek skopein, ‘to see’; similarly, the prefix
stereo– comes from stereos, ‘solid’[1].
Stereoscopy is the capability of seeing 3-dimensional
solids, as opposed to seeing flat images. As seen in Figure
1, stereoscopic displays give us the illusion of depth by
presenting a slightly different image to each eye.[2]
Human eyes are positioned in a manner that gives us a
bifocal view of our environment. In order to simulate this
effect, companies dealing with Display Technologies have
adopted the same approach to capture images.
Figure
1.
Simulation of a 3D object by using stereoscopy
As we can see in Figure 2, techniques using two lenses to
simulate the human eyes are being used in order to capture a
“stereo” image (containing two angles of the target object).
Figure
2. Capturing an image for a stereoscopic display
AUTO-STEREOSCOPIC DISPLAYS
Auto-Stereoscopic displays differ from regular stereoscopic
displays in their adaptive nature: based on the viewer’s
position, changes in parallax will cause the displayed image
to ‘change angles’. Parallax is often thought of as the
"apparent motion" of an object against a distant background
because of a perspective shift, as seen in Figure 3. When
viewed from Viewpoint A, the object appears to be in front
of the blue square. When the viewpoint is changed to
Viewpoint B, the object appears to have moved to in
front of the red square. [3]
Figure
3. Objects’ perceived relative positioning is
dependent on observer’s positioning [3]
While companies have currently achieved a certain amount of
innovativeness and success, the basic construct of the
aforementioned technologies will do nothing more than
“simulate” a three-dimensional effect: a truly volumetric
display requires a fundamental structural change.
VOLUMETRIC
DISPLAYS
Volumetric 3-D displays embody just one family of 3-D
displays in general. In contrast to the display methods
we’ve just covered, volumetric displays create 3-D imagery
via the emission, scattering, or relaying of illumination
from well-defined regions in (x,y,z) space.
Standard
Static Volumetric Displays
These displays use a three-dimensional array of
predetermined volumetric pixels, known as “voxels”, which
are activated and deactivated on a pre-set manner to display
objects.
Figure
4. Static Volumetric Displays comprise a 3-dimensional array
of static elements [6]
Static volume displays have one notorious advantage over
other displays: they lack moving parts. Devices with moving
parts tend to have a shorter life span due to the
progressive wearing-off of the relevant parts and the
vibrations that cause structural weakening. Static displays
currently suffer from major setbacks, namely voxel blockage
and density. The static constructs like the one shown in
Figure 4 result in voxels closer to the observer blocking
the radiation originating from farther ones in a straight
line of sight. This problem is magnified as the voxel
density (voxels per cubic centimeter) increases, thus
demoting most static volumetric displays to a lower class of
practicality and realism. This problem can be solved by
having voxels that, while de-activated, are invisible to the
human eye.
Laser
Static Volumetric Displays
While Laser Static Volumetric Displays are not usually
considered a sub-category per se, they are important to our
discussion in order to demonstrate the numerous alternative
approaches to achieving Static Volume Rendering. In Figures
5 and 6 below, we can see how two intersecting laser beams
that are invisible to the human eye individually can cause a
localized excitement of the ionized medium once intersecting
at a particular point.
 
Figure
5. Intersecting laser beams radiate light within the
viewable human spectrum in an excitable medium [4]
Figure
6. Diagram of a Laser Static Volumetric Display [4]
It is crucial that the excitation process in the active ion
occur only from the selective absorption of two different
infrared wavelengths, as it is this mechanism that enables a
visible point of light to be "turned on" only where the two
laser beams cross, and nowhere else. By controlling the
spatial co-ordinates of the intersection of the two lasers,
a "voxel", or volumetric pixel, can be addressed at a
specific location inside the bulk imaging medium. [4]
Swept
Surface Volumetric Displays
By far the most common approach to volumetric displays,
Swept Surface Displays use a different approach to generate
the objects within the display region: by using fast-moving
light-emitting or reflective voxels, the human eye is
‘fooled’ into believing that there is an object formed in
mid-air. This ‘Persistence of Vision’ phenomenon is similar
to the approach of current flat screens and television sets,
whereas multiple static frames are quickly displayed to
create the illusion of motion or, in our case, a
‘non-flickery’ object.
Figure 7 depicts the architecture of the most prominent
current Volumetric Display: the Perspecta Spatial 3D, from
Actuality Systems [5]. This device has a
central-axis rotating screen, and a camera that projects
different ‘slices’ of the object to be displayed. The screen
rotates at 900 Revolutions per Minute (15 Revolutions per
Second), thus bordering the ‘flicker’ line.
The
Figure
7. Perspecta Display, from Actuality Systems, Inc. [5]
Perspecta is the leading technology in Volumetric Displays,
and that comes at a high cost: individuals wanting to
acquire this device will need to part with around RM
150,000. This is one of the reasons why the Centre for
Research Excellence at UCSI decided to undertake the quest
to develop a low-cost volumetric display.
MGS GRANT
APPLICATION
Thanks to the infrastructure laid out by the Malaysian
Government, MSC Status companies such as UCSI have access to
the Multimedia Development Corporation R&D Grant Scheme
(MGS). This grant allows enterprising companies to conduct
multimedia-related projects within a maximum period of 2
years. Everything seemed perfect to undertake the
development of Malaysia’s very own Volumetric Display.
After one Introductory Meeting (called ‘Soap Box’), one
Technical and Commercialization Meeting, and one Meeting
with the MGS Management Team, our grant for the project got
approved in July 2005. The total estimated cost for 2 years
of this project oscillated near the 1.3 Million Ringgit
mark. This is how “Volumetric Display Using LED” (codenamed
“Volex”) was born.
VOLEX
As we dug deeper into the project, the Project Team had to
look into a number of issues that need improvement in some
way to ensure that the resulting Prototypes meet everyone’s
expectations.
LED homogeneity
There are two ways to generate voxels in Swept-Surface
displays: either the light source is mounted on the moving
surface, or a beam is sent to the surface to be reflected in
synchrony with the expected object to be visualized. Volex
uses LEDs as the source of emission. Knowing the inherent
architecture of LEDs, this means that we must be very
careful when choosing a model, for many LEDs have a much
brighter area near the diode (usually at the center of the
LED). While some of that effect may not be discernible by
the human eye, the Project Team ran multitude of tests with
different types, shapes, luminosity and colors of LED until
we found the most appropriate model to fit the Project’s
needs.
LED
positioning
As LEDs travel the predetermined trajectory in space, they
will be emitting light to induce the optical illusion of an
object floating in mid-air. When a voxel is activated, it is
optimal if the resulting light emission can expand in a
spherical field without being blocked by any object. This
way, we ensure that viewers from all possible directions can
see that particular activated voxel at any given time. The
problem here is that, due to architectural constraints of
working with LEDs, it is not possible to totally remove all
the obstacles to each voxel. Consequently, the challenge
revolves around minimizing this ‘blockage” as much as
possible, to create a more satisfactory viewing experience.
Figure
8.
LED Emission blocked by a non-activated component
LED
dissipation angularity
Currently, there is no significant demand for the existence
of LED that can radiate light in a full spherical field
(i.e., in all directions). As such, LED manufacturers do not
currently produce such component. This is a significant
drawback when trying to develop a Volumetric Display using
LEDs. In optimal conditions, the emission elements would
emit light in all directions. Sadly, due to the nature of
components currently available, LED’s viewing angles
generally oscillate anywhere between 30° and 270°, although
the brightness of the voxel will be affected as the angle
increases even in wide-angle LEDs. This is why we must
consider resorting to generating a single voxel with
multiple LEDs: this is called a cluster. There are a number
of intricacies associated with this alternative solution,
but these will eventually become obsolete once spherical
LEDs start being developed.
Vibratory
concerns
Rotational speed is critical for a successful display. If
the speed is too slow, then voxels will be refreshed less
frequently, and this will result in a ‘flickery’ display.
The effects of extended exposure to flicker have not been
thoroughly studied, but we know from regular Television sets
that, once the refresh rate is sufficiently high, human eyes
will be ‘fooled’ into believing that there is no flicker at
all. It is thus vital to ensure that this minimum refresh
rate is observed at all times. This means maintaining a high
speed, which results in an increment of the vibrations
caused by small unbalances in the rotating disc. Since these
unbalances are exponentially magnified as speed increases,
it becomes both an operational and a safety concern to
ensure that the unit is as balanced as possible.
Voxel generation
As we have discussed above, there are a number of factors
that need attention for ensuring a functional hardware
architecture for Volumetric Displays. The software aspects
of the whole system are as complicated as the hardware. This
can be a very tricky part of the whole system: voxels may be
pre-arranged in a matrix with a discrete number of objects,
or we may latch the LEDs in such a way that their state can
be ON for a longer duration of time. It is generally simpler
to have a pre-set voxel resolution, and let the prospective
objects to be displayed adapt to the limited resolution, but
this will of course result in a loss of quality and realism.
By mapping all possible voxels and ‘slicing’ the –likely–
cylindrical volume swept by our emitting elements, we can
easily create objects within the set of possible voxels. All
we have to do is tell the LEDs when to be ON. Prototype 1 of
Volex has successfully achieved this. There is an obvious
problem trying to print 2D images of 3D objects, but a
Display Version of Volex should soon become available.
Data
transfer
Swept-Volume Displays involve high-speed motion. The data to
be displayed can either be stored on the moving unit, or can
be segregated to ensure a more flexible method to display
data. If we choose to have all the necessary components ‘on
board’, our display will need to be disassembled and we will
have to manually program a new dataset on the
microcontroller/microprocessor each time we want to have new
information added to the display. Instead, as we move
towards the final prototypes, our data will be streamed from
an external unit using one of the many available forms of
wireless data communications. This way, we will ensure that
our display can be more interactive and users may be able to
eventually connect them to a PC for data manipulation as the
technology matures.
CONCLUSIONS
There is a multitude of factors that must be envisaged when
developing Volumetric Displays: from the complexity of
hardware constraints to the intricacies of the software
system. Realism should at all times remain the main priority
when fine-tuning such a device, while other minor concerns
(vibration, noise, etc.) are also critical and must not be
overseen. Current technologies, while achieving a high
degree of realism, are overseeing the need for Consumer’s
support in order to successfully bring Volumetric Displays
to the masses. This is the gap the Volex is going to bridge.
For more information, visit
www.ucsi.edu.my/research/volex.html
REFERENCES
1.
The
American Heritage Dictionary of the English Language, Fourth
Edition, Boston: Houghton Mifflin, 2000.
2.
Wikipedia,
The Free Encyclopedia, September 2006, http://en.wikipedia.org/wiki/Stereoscopic
3.
Wikipedia,
The Free Encyclopedia, September 2006, http://en.wikipedia.org/wiki/Parallax
4.
Laser F/X
International, August 2006, http://www.laserfx.com/Backstage.LaserFX.com/Archives/Archives5.html
5.
Actuality
Systems, August 2006, http://www.actuality-systems.com/
6.
Design
Interactions, RCA, October
2006, http://www.interaction.rca.ac.uk/alumni/0002/jez/HTML/crd1/img/volume_img/one_unit.jpg
Andres M.
Trianon is the Project Leader for the VOLEX research and
Manager of UCSI Centre for Research Excellence. He can be
contacted at andres.m.trianon@gmail.com
|