IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 20071647
A Survey of 3DTV Displays:
Techniques and Technologies
Philip Benzie, Member, IEEE, John Watson, Senior Member, IEEE, Phil Surman, Ismo Rakkolainen, Klaus Hopf,
Hakan Urey, Member, IEEE, Ventseslav Sainov, and Christoph von Kopylow
Abstract—The display is the last component in a chain of ac-
tivity from image acquisition, compression, coding transmission
and reproduction of 3-D images through to the display itself.
There are various schemes for 3-D display taxonomy; the basic
categories adopted for this paper are: holography where the image
is produced by wavefront reconstruction, volumetric where the
image is produced within a volume of space and multiple image
displays where two or more images are seen across the viewing
field. In an ideal world a stereoscopic display would produce
images in real time that exhibit all the characteristics of the
original scene. This would require the wavefront to be reproduced
accurately, but currently this can only be achieved using holo-
graphic techniques. Volumetric displays provide both vertical and
horizontal parallax so that several viewers can see 3-D images that
exhibit no accommodation/convergence rivalry. Multiple image
displays fall within three fundamental types: holoform in which
a large number of views give smooth motion parallax and hence
a hologram-like appearance, multiview where a series of discrete
views are presented across viewing field and binocular where only
two views are presented in regions that may occupy fixed positions
or follow viewers’ eye positions by employing head tracking.
Holography enables 3-D scenes to be encoded into an interference
pattern, however, this places constraints on the display resolu-
tion necessary to reconstruct a scene. Although holography may
ultimately offer the solution for 3DTV, the problem of capturing
naturally lit scenes will first have to be solved and holography
is unlikely to provide a short-term solution due to limitations in
current enabling technologies. Liquid crystal, digital micromirror,
optically addressed liquid crystal and acoustooptic spatial light
modulators (SLMs) have been employed as suitable spatial light
modulation devices in holography. Liquid crystal SLMs are gener-
ally favored owing to the commercial availability of high fill factor,
high resolution addressable devices. Volumetric displays provide
Manuscript received March 8, 2007; revised June 15, 2007. This work was
supported in part by EC within FP6 under Grant 511568 with the acronym
3DTV. This paper was recommended by Guest Editor L. Onural.
P. Benzie and J. Watson are with the School of Engineering at the Univerity
of Aberdeen, Aberdeen AB24 3UE, U.K. (e-mail: firstname.lastname@example.org).
P. Surman is with the Imaging and Displays Research Group (IDRG) at De
Montfort University, Leicester LE1 9BH, U.K.
I. Rakkolainen is with Tampere University of Technology, Tampere 33101,
Finland and also with Fogscreen Inc., Helsinki 00180, Finland.
K. Hopf is with the Fraunhofer-Institute for Telecommunications—Heinrich-
Hertz-Institut (HHI), Berlin 10587, Germany.
H. Urey is with the Optical Microsystems Laboratory, Department of Elec-
trical Engineering, Koc University, Sariyer 34450, Istanbul.
V. Sainov is with the Central Laboratory of Optical Storage and Processing
of Information, Bulgarian Academy of Sciences (CLOSPI-BAS), P.O. Box 95
Bremen D-28559, Germany.
Digital Object Identifier 10.1109/TCSVT.2007.905377
both vertical and horizontal parallax and several viewers are able
to see a 3-D image that exhibits no accommodation/convergence
rivalry. However, the principal disadvantages of these displays
are: the images are generally transparent, the hardware tends
to be complex and non-Lambertian intensity distribution cannot
be displayed. Multiple image displays take many forms and it is
likely that one or more of these will provide the solution(s) for the
first generation of 3DTV displays.
Index Terms—Holography, TV, 3-D displays.
outlined in other papers in this issue, there is a long chain of
activity from image acquisition, compression, coding transmis-
The concept of a 3-D display has a long and varied history
stretching back to the 3-D stereo photographs made in the late
19th century through 3-D movies in the 1950’s, holography
in the 1960’s and 1970’s and the 3-D computer graphics and
virtual reality of today. The need for 3-D displays and vision
grows in importance by the day, as does the number of applica-
tions such as scientific visualization and measurement, medical
imaging, telepresence, gaming, as well as movies and televi-
sion itself. Many different methods of 3-D displays have mani-
fest themselves over the last few decades, but none has yet been
able to capture the mass market. Much of development in 3-D
imaging and displays of the latter end of the 20th century was
spurred on by the invention of holography, and this was the
catalyst which led to some of the significant advances in au-
tostereoscopic and volumetric methods, whereas, advances in
techniques of virtual reality have helped to drive the computer
and optics industries to produce better head mounted displays
and other 3-D displays.
Many approaches have been outlined from simple stereo
with anaglyph glasses through to full parallax holography.
What technology is applied in a given circumstance will largely
depend on the application itself. For example, it maybe that a
full parallax, full color, interactive holographic display would
be used in air traffic control but that an autostereo-display is
more appropriate for low level computer aided drawing (CAD)
applications. What is clear is that no single approach is likely
to dominate and it will be the application which will determine
HE DISPLAY is the last, but not the least, significant as-
pect in the development of 3DTV and vision. As has been
1051-8215/$25.00 © 2007 IEEE
1648IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 2007
which technology is adopted. The form that such displays
would take is one aspect which needs considerable thought and
is a major factor in consumer acceptance. Will the consumer
want to see “Star Wars” images in a central table or will a flat
panel in the corner of a room suffice?
II. 3-D DISPLAYS AND ASSOCIATED PERCEPTION FACTORS
The technologiescurrentlybeingpursued for 3-Ddisplay can
be broadly divided into the following categories, although there
are various other methods of classification used and the termi-
nology is not always clear.
• Autostereoscopic displays (Sections III–V).
• Volumetric displays (Section VII).
• Holographic displays (Section VIII).
Strictly speaking, the term “autostereoscopic” could be ap-
plied to all of the first three categories above, since it describes
those displays which create a stereoscopic image without the
need for any form of special glasses or other user-mounted de-
vice. However, it is usual to restrict the term to cover displays
such as binocular (single user), multiview (multiple discrete
stereoscopic views) and holoform systems where only multiple
2-D images across the field of view are considered. Autostereo-
systems are limited by the number of viewers and eye or head
tracking is usually needed. In holographic displays the image
is formed by wave-front reconstruction, and includes both real
encoded into an interference pattern, this requires a high resolu-
tion recording medium and replay medium which place severe
constraints on the display technology employed. Nonetheless,
holography can be deployed in reduced parallax (e.g., stereo-
holography or lenticular) systems, which relax some of the con-
straints. Volumetric displays form the image by projection onto
a volume, or use discrete locations of luminescence within a
volume, without the use of light interference (e.g., rotating or
vibrating light emitting diode matrix). These displays tend to
have limited resolution. Head mounted displays such as those
using, for example, liquid-crystal-on-silicon (LCOS) devices or
retinal scanning devices (RSD) are disadvantaged by the re-
quirement to wear devices and in certain synthetic vision appli-
cations suffer from binocular rivalry . These display devices
be the focus of this paper . The more traditional stereo-tech-
nologies, all require the use of viewing aids and will not be cov-
ered in detail here.
In order to make a highly believable 3-D display, there are a
number of important physiological cues that must be presented
to the eye-brain system. The main requirement is to create the
illusion of depth or distance by using a series of cues such as
commodation , . If an autostereoscopic based system is to
be considered, conflicting cues can lead to discomfort and fa-
tigue. There are also a number of psychological depth cues such
interposition, retinal image size, texture gradient, and color. In
the case of a mismatch between convergence and accommoda-
tion eyestrain in stereoscopic virtual displays may occur. The
larger the depth of the stereoscopic virtual image, the larger the
mismatch and the resulting visual strain.
Binocular asymmetries, the difference between the left and
the widespread use of stereo image based displays. The fac-
tors that determine viewing comfort most strongly are vertical
disparity and crosstalk between the left and right images. Even
small amounts of each can cause noticeable discomfort .
There are many solutions offered to reduce the impact of
human factors on the acceptance of 3-D images, such as eye-
tracking, adaptive optical systems , or physically moving the
source-location in scanned beam displays, and controlling the
amount disparity and keeping it within a “comfortably view-
able” range , . For the presentation of a completely real-
istic image, the angles subtended by objects in the image should
be the same as in real life so that the image is orthoscopic,
but this is difficult to achieve for small displays . For op-
timum viewing of TV, the viewing distance should be about
three times the picture height . Another related phenom-
enon is the “puppet theatre” effect, which manifests itself by
objects in the image occupying a smaller region on the retina
than they would in a natural scene . This is intrinsically
no more disturbing than viewing the miniaturized 2-D images
that we currently seem to find acceptable. It can be reduced by
tance  and results in zero disparity. The brain, however, can
adapt to and correct for a wide range of ocular distortions and
the question arises as to whether these effects are as important
2-D image, without this really bothering us . The improve-
ment afforded by 3-D over monoscopic images has been widely
reported. Pastoor, for example, states that the overall psycho-
logical impact of a 3-D screen is equal to flat images twice their
In addition to human factors, many features of the display it-
self influence 3-D perception. If a pair of stereo images is of
differing quality, the subjective quality of the fused pair is de-
termined by the better image . If the high-frequency spatial
components are not, the image-pair can be fused readily. These
effects have important implications for stereo image coding.
Progressive scanning in a TV camera, as opposed to interlaced,
improves perceived image quality. At a field rate of 50 Hz an
interlaced scan only appears to have a resolution of 1.2 times
that of a progressive scan with half the number of lines .
Crosstalk is arguably the most important factor concerning the
design of autostereoscopic 3DTV. Reduction in disparity can
reduce crosstalk, as can color selection . A useful means
of reducing the effects of crosstalk is afforded by the “kinder,
gentler approach” that aims at “just enough reality” . Image
flicker influences the method of image multiplexing, and hence
the display-type. The critical flicker frequency (CFF)—the fre-
quency at which flicker becomes just visible—increases with
luminance  (from 13 Hz at a luminance of 3.4
(lm/m sr) to 51 Hz at 343 nits). However, other factors also in-
fluence CFF, such as age, foveal or peripheral vision and the
luminance of screen surround . Thin-film translator liquid
crystal displays (TFT-LCDs) will exhibit flicker at half the field
frequency. Driving the display in such a way that the phase of
the flicker is inverted on alternate rows can eliminate this .
BENZIE et al.: A SURVEY OF 3DTV DISPLAYS: TECHNIQUES AND TECHNOLOGIES1649
Depth plane quantization is the separation of the image into dis-
crete planes, due to the effect of the horizontal pitch of the dis-
play pixels producing disparity quantization. The threshold for
this to become apparent is 0.8 min of arc . Disparity quan-
tization is doubled in simple lenticular 3-D displays where left
and right images are displayed on alternate pixel columns.
III. MULTI-IMAGE DISPLAYS
Multiple image displays, where two or more images are seen
across the width of the viewingfield, can take three basic forms.
In the first category, a large number of views are produced in
order to give the appearance of smooth motion parallax. As
these give a hologram-like appearance, they are referred to here
as “holoform.” Displays where a smaller number of discrete
views are presented across the viewing field are termed multi-
view. The simplest multiple image displays are binocular where
only a single stereo pair is displayed. Head tracked displays
come under the category of multiple-image.
Holoform displays aim to provide smooth motion parallax as
the viewer traverses the viewing field laterally. This is achieved
by producing a large number of closely spaced discrete views
that give an image with a hologram-like appearance. One early
approach , “the Stereoptiplexer,” used picture information
from rapidly moving cine film to produce a moving “aerial exit
pupil” from scanning mirrors. The exit pupil is effectively a
vertical narrow aperture that traverses the region between the
the “aerial exit pupil” with a physical slit. The earliest work on
moved rapidly in front of a CRT screen. This method requires
an extremely fast frame rates to achieve display of video. A ver-
sion of this display using a 400-element ferroelectric array with
no moving parts was later developed .
Integralimaging can be consideredas a variationof holoform
displays. McCormick et al.  adapted the concept to produce
an autostereoscopic display with full motion parallax. Integral
imaging uses an array of small lenses that are either spherical
or cylindrical to produce the familiar stereoscopic photographs
where a lenticular sheet of vertically aligned cylindrical lenses
provide pictures with a horizontal parallax only 3-D image. A
variation of this locates a lens array behind a transmission-type
display and inserts a polymer-dispersed liquid crystal between
the lens array and the collimated backlight . This feature
enables 3-D/2-D conversion.
Quasi-holographic 3-D display systems such as holographic
stereograms, are an alternative approach to display. With holo-
graphic stereograms, multiple views across the viewing field
are produced using diffractive optics. St Hilaire  considers
the effect of the image appearing to “jump” between adjacent
is under-sampled. An optimum of 20 views per interocular dis-
tance to achieve the appearance of smooth motion parallax was
The Hungarian company Holografika has a 3-D display that
they call quasi-holographic . QinetiQ, advertise an “Au-
tostereo 3-D Display Wall” on their website . The display
consists of an array of 40 projectors that project images on to a
screen that diffuses light in the vertical direction only. The ef-
fective width of the array is increased by employing two side
mirrors that produce virtual extensions to the actual array. The
image is produced both in front and behind the screen.
B. Multiview Displays
In multiviewdisplays,a series ofdiscrete views are presented
across the viewing field. One eye will lie in a region where one
perspective is seen, and the other eye in a position where the
adjacent perspective is seen. The number of views is too small
for continuous motion parallax but strategies such as merging
one image into the adjacent image and limiting the disparity in
order to keep the apparent image content close to the plane of
lenses, either in the form of Fresnel lenses or sheets of cylin-
drical lenses, parallax with “point” light sources: holographic
viewing zone formation and “Cambridge” type displays with
A method, simple in principle but cumbersome to implement
, uses an arrangement of projection lenses, a Fresnel field
lens and vertical diffuser to produce a series of viewing zones
(4, 7, 13, or 21) depending on configuration across the viewing
Vertical lenticular screens can be used to direct light from
columns of pixels on an LCD into viewing zones across the
viewing field. A liquid crystal layer lies in the focal plane of
the lenses, and the lens pitch is slightly less than the horizontal
pitch of the pixels in order to place viewing zones at the chosen
optimum distance from the screen. In this case, three columns
plays with four zones are described in  and NHK devel-
oped a systemin 1990.Simple multiviewdisplays withthis
construction suffer vertical banding on the image known as the
“picket fence” effect. Secondly, when a viewer’s eye traverses
the region between two viewing zones, the image appears to
by the simple expedient of slanting the lenticular sheet in re-
lation to the LCD . The latest embodiment of the Philips
display presents nine images across the viewing field . The
slanted lenticular screen has a pitch that is 1.5 times the pixel
pitch of the LCD and is slanted at an angle of
resolution is reduced by a factor of three in both the horizontal
and vertical directions. This display is switchable between 3-D
and higher resolution 3-D. This is achieved with the use of an
active system where a liquid crystal material is in contact with
Although multiview displays are limited in the quality of the
stereo effect, the size of the usable viewing region and the re-
stricted depth of image field, the actual appearance of these
displays is remarkably good, bearing in mind their simplicity.
Other similar displays are marketed by Stereographics  and
Sanyo . The Sanyo display is four-view device that uti-
lizes a parallax barrier (a pin hole array) and a 40-in LCD with
768 pixels. “Point” light sources behind an LCD are
used by a Korean group to direct images to the viewing zones
1650IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 2007
. A simple 32
light sources are obtained from a collimated blue laser beam
passing through a microlens array. Each light source lies be-
hind a 6
6 array of pixels, therefore providing both horizontal
and vertical parallax. The paper acknowledges that the display
suffers from the loss of resolution—this is particularly severe as
both parallaxes are available.
Holographic optical elements (HOE) are another method
used to form viewing zones .
Moore etal.  optedfor temporal multiplexingwhere a se-
ries of images is presented in sequence. Although a sufficiently
fast device wasnotavailableto performthisfunction,theopera-
mission LCD rather than CRT system. Instead of a view being
displayed on an LCD, it is projected on to the Fresnel lens by a
lens located where the illumination sources were. A ferroelec-
tric liquid crystal shutter replaces the illumination sources. This
shutter is in the Fourier transform plane of the projection lens,
and its real image forms the exit pupil. However, this is not seen
on the image perceived by the viewer. Various versions of the
display provide up to 16 views, and color. Color is obtained by
using a Tektronix sequential liquid crystal shutter. Another ver-
sion  uses a 50-in concave spherical mirror to overcome the
ages are derived from three primary-color CRTs whose outputs
are combined by dichroic mirrors for maximum efficiency. A
further development provides two sets of viewing zones that are
derived from two sets of CRT subsystems. This version is par-
ticularly suitable for two-player arcade. Recently, the method
was revised to incorporate a fast ferroelectric display, with an
array of LEDs as the light source .
32 pixel display is described, where the
IV. MULTIPLE IMAGE: BINOCULAR, FIXED VIEWING ZONES
a single pair of viewing zones is produced. These can be of
five basic types; the viewing zones can be formed by lentic-
jector methods produce exit pupils that are real images of the
projector lenses, with the image formed on a double lenticular
screen or a special reflecting screen. Parallax displays use an
opaque mask, an array of orthogonally polarized image multi-
plexing elements or a series of line illumination sources behind
A simple lenticular screen display may consist of an LCD
with a screen infront of it.The screenconsists ofa seriesof ver-
tically aligned cylindrical lenses with a pitch slightly less than
double the horizontal LCD pitch. This allows for the viewer
being at a finite distance from the screen. As the majority of
LCDs have the RGB subpixels in the vertical stripe configura-
tion, the LCD has to be operated in the portrait mode in order
to avoid coloration and distortion of the colors in the image.
Left and right images are displayed on alternate columns (par-
allel to the lenticular lenses) . An LCD can be operated in
the normal landscape mode in a display utilizing a chequered
mask and orthogonally aligned lenticular screens . The left
and right images are multiplexed on alternate pixel rows. The
display consists of five layers, in order from the back; back-
light panel, chequered pattern mask, vertically aligned lentic-
ular sheet, horizontally aligned lenticular sheet and the LCD.
The horizontal pitch of the mask and the pitch of the verti-
cally aligned lenticular sheet are about double the LCD hori-
zontal pitch. These perform the same function, but in a different
manner, to the lenticular screen in the previous method. Rows
of the chequered pattern are directed to the appropriate rows of
pixels by the horizontally aligned lenticular sheet. A more re-
cent lenticular display, is capable of 2-D/3-D switching .
Microlenses activated on incident of polarized light can be used
to either divert the images to the appropriate eyes in the 3-D
mode, or enable light to pass through without deviation to allow
the display to be used in the normal 2-D mode .
In projection displays, only a single pair of viewing regions
is formed, thereby restricting the viewing region even more. A
producing diamond-shaped regions, the right image is directed
to the right side of the viewing field, and the left image to the
left. The distance from the screen over which 3-D can be seen
is considerably increased. Also, when both eyes are in the right
side of the field, a right image is seen by both of them, and a left
image by both eyes when they are in the left side of the field.
A. Parallax Barrier
The first of the parallax methods uses thin vertical illumi-
nation lines behind the LCD in order to direct the light to the
appropriate viewing regions . The lines are produced on a
diffusing screen by a lenticular sheet that is mounted behind it.
The primary light sources are a series of vertical lamps that are
located behind slit apertures and are focused by the lenticular
sheet. An earlier display used masks to producea simpleand ef-
fective means of supplying 3-D to a single viewer . In later
versions parallax barriers, both behind and in front of an LCD,
are used to present images with virtually no Moiré fringing.
These barriers consist of masks that have vertical apertures in
Polarization can be used to provide what is effectively a ver-
tical slit mask. In one system, an image-multiplexing screen
consists of an array of vertical strips of dichroic polarizing ma-
terial of alternating orientation; a Fresnel lens produces exit
pupils from a pair of illumination sources that have polarizers
in front of them to select the odd and even columns of pixels
. Difficulties were encountered with the manufacture of the
multiplexing screen. The optics are very basic as the pupils are
produced with a Fresnel lens. Also, the multiplexing barrier
would probably have been more simply obtained by using an
off-the-shelf micropolarizer array .
Sharp Corporation and Sharp Laboratories Europe, Ltd.
(SLE) jointly developed a TFT 3-D LCD, which can switch
between 2-D and 3-D display modes . The LL-151-3-D
monitor is capable of displaying dynamic 3-D images. Using
a parallax barrier, light from the LCD is split so that different
patterns reach the viewer’s left and right eyes. With the DTI
(Dimension Technologies Inc.) display , viewing zone for-
mation is accomplished with a special illumination pattern and
optics behind the LCD screen which make alternate columns
of pixels visible to the left and right eyes of a viewer sitting
BENZIE et al.: A SURVEY OF 3DTV DISPLAYS: TECHNIQUES AND TECHNOLOGIES1651
directly in front of the display and in certain regions to either
side. This displays left and right images of stereo pairs on
alternate columns of pixels on the LCD. The left image appears
on the odd numbered columns and the right image appears on
the even numbered columns. For example, for an LCD with
1024 columns and 768 rows, each complete stereoscopic image
consists of 512 columns and 768 rows. Both halves of a stereo
eyes. This is accomplished with a special illumination plate
located behind the LCD. Using light from compact, intense
light sources, the illumination plate optically generates a lattice
of very thin, very bright, uniformly spaced vertical light lines,
in this case 512. The lines are precisely spaced with respect
to the pixel columns of the LCD. Because of the parallax
inherent in our binocular vision, the left eye sees all of these
lines through the odd columns of the LCD, while the right eye
sees them through the even columns. The left eye sees only the
left eye portion of the stereo pair, while the right eye sees only
the right eye portion. This enables the observer to perceive the
image in 3-D.
V. MULTIVIEWER HEAD TRACKED DISPLAYS
Surman et al.  developed a display with multiviewer ca-
presenting a stereo image pair simultaneously, can form the
basis of a 3-D display. Regions are formed in the viewing field
where a left, or a right image only, are seen across the complete
area of the screen. The positions of these exit pupils follow the
and enable a high degree of freedom of movement. The exit
pupils are produced by novel optics, located behind the LCD
that replace the conventional backlight. The display operates
in a similar manner to anaglyph (red/blue glasses) or polarized
glasses, but without the need for glasses. The core technology
is based on a spatial multiplexing screen and an illumination
system with steering optics. The steering optics incorporates a
large number of white LEDs whose outputs are determined by
viewers’ head positions. Motion parallax cannot be displayed,
tage in a television system as it enables the simplest capture and
transmission, and places the smallest demands on the display in
terms of the amount of information that has to be displayed.
VI. HEAD TRACKING SYSTEMS
Real-time position and orientation tracking of viewers is an
important component for many 3-D displays , . For
some displays or applications, only orientation or position can
be tracked. This imposes many limitations but also simplifies
the task significantly. There already exist small, cheap and
accurate inertial sensors which can be attached to HMDs.
The accuracies of static position and orientation and dynamic
movement measurements are important. Also features like
sample rate, number of targets tracked, range of tracking, la-
tency, update rate, registration, and space requirements may be
important. It is highly important that the viewer is tracked and
the scene gets updated very fast. Many head tracking methods
are available including electromechanical, electromagnetic,
acoustic tracking, inertial tracking and optical tracking. No
single tracking method is universally applicable. There are
dozens of methods for position and orientation tracking for
various purposes and they are well documented e.g., in virtual
reality textbooks. Hybrids of all the methods can naturally be
used. They improve the accuracy but usually require more cal-
culation. State-of-the-art head trackers deploy passive (optical)
markers and active (optical, acoustic, magnetic) emitters and
receivers as well as inertial system components such as gyros,
gravimeters and accelerometers. Some advanced systems even
combine different components, e.g., optical and inertial subsys-
tems, in order to make the tracker more robust against changes
in the environment or in the case of occlusions. Generally,
these systems are intrusive since they require the user(s) to
be tethered to the measurement equipment, or at least to wear
some parts of the equipment.
The earliest head tracking displays used infrared light re-
flected from the viewer’s head. In a Fresnel lens single viewer
display , the head detector moves with the lens. Illumina-
tion of each side of the viewer’s head by two wavelength bands
of infrared light has also been applied . A further display
uses a large format convex lens to illuminate the left side of the
viewer’s head with IR in the 830–870-nm range, and the right
side of the head in the 930–970-nm band. The outputs of a pair
of cameras with matching filters are used to control the illumi-
nation from a monochrome 2-D display directly without the use
of any additional processing. This method is used in other dis-
plays by the same researchers , . Retinal reflection is
used insome head trackers, for instance using“Red-eye” reflec-
tion . Identifying the time during which the viewer blinks
and measuring the size of the pupil has been considered . A
to extract interest points in the images . Those points repre-
tion,template matchingand detection of synchronously moving
pairs. Near to real-time (120Hz) frame rates were achievedbya
high-precision single-person 3-D video head tracker , .
VII. VOLUMETRIC DISPLAYS
a defined volume of space. As volumetric displays create an
image in which each point of light has a real point of origin
in space, the images may be observed from a wide range of
viewpoints and angles. Additionally, the eye can focus at a real
point withinthe imagegiving a sense of ocularaccommodation.
Such displays are usually more suited for computer graphics
natural image capture. However, their most important drawback
with regard to TV displays is that they invariably suffer from
image transparency where parts of an image that are normally
occluded are seen through the foreground object. Another dif-
ficulty is the inability to display surfaces with a nonLambertian
niques have been proposed but few successful large scale de-
vices are available, although a number of smaller scale displays
have been developed. Volumetric displays can be of two basic
types: virtual image where the voxels are formed by a moving
or deformable lens or mirror, and real image where the voxels
are on a moving screen or are produced on static regions.
1652IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 2007
A. Virtual Image Methods
One of the earliest virtual image methods , , uses a
mirror of varying focal length to produce a series of images
at different apparent distances. The variable curvature of the
mirror allows smaller movement than would be required from
a moving flat surface giving the same effect. The mirror con-
sists of a thin silvered Mylar film that is attached to the front
of a loudspeaker which is driven at between twenty and several
hundred Hertz. A variation on this uses a stretchable membrane
mirror ,  of 1.2 m diameter that can be varied over a
large range of F-numbers. The image can be produced both in
front of, and behind the plane of the mirror. Lenses can also
be used to produce a similar effect. In the “xyzscope”  a
rotating lens is used to effectively vary the distance between
the object and the lens centre. A combination of variable focus
lenses and integral photography has been proposed by Yana-
gasiwa et al. . However, the lens array described would be
difficult to make due to the varying radius of curvature needed
quirements of the human visual system for viewing objects in
depth. The accommodative cue is generated in hardware using
wavefront shaping deformable membrane mirrors. In addition,
software cues, such as blurring, relative size, and occlusion, are
tested for their ability to trigger appropriate accommodative re-
sponses in 3-D displays.
B. Real Image Methods—Moving Displays
A solid image can be produced in a volume of space by
displaying “slices” of the image on a moving screen (swept
volume). If, for example, a sphere has to be displayed, this can
be achieved by displaying a series of circles of varying size on
to a moving surface. Most attempts to create volumetric 3-D
images are based on swept volume techniques, because they
can be implemented in the near term with today’s hardware
Favalora et al.  have developed a system with a 100 mil-
lion voxel resolution sufficient for video display. This is ob-
tained by presenting 200 radially-disposed slices consisting of
768 pixel images. These are provided from a modified
projector that can supply hundreds of colors. Images are pro-
at 30 Hz. This is being developed by Actuality Systems under
the trade name Perspecta™ . It creates an image within a
desktop 20-inch dome. It has limited immersion, no entry into
volume, and is only suitable for small objects. However, it be
made interactive by external cameras, which track the user’s
hands on the dome.
Another rotating screen system, by Holoverse , ,
, with 8 bit color depth is based on Texas Instruments’
DMD (Digital Micromirror) Technology. They claim to
eliminate the wobble and sway of competing systems, while
increasing the resolution, color depth, and update rate. Their
“HoloDeck Volumetric Imager” has an effective cross-sectional
resolution of 1024
768360 in evenly distributed rotational
planes. Further examples of rotating screen devices are the
FELIX 3-D-display, which utilizes a helical screen and RGB
lasers . The 3-D image appears to be inside the cylinder,
where an outer wheel with vertical slits revolves clockwise
at a fast rate, while an inner wheel moving counterclockwise
at a slower speed lined vertically with LEDs projects thin
slices of a face. The system also requires a 360 digital camera
surrounding the object, and the data to be sent to the cylindrical
tube. The rapid succession of image slices seen through the
slits produces the illusion that the viewer can see the person’s
entire face at once, in 3-D.
C. Real Image Methods—Static Displays
Static displays are those where voxels are produced on sta-
tionary regions in the image space. Some of these methods have
the potential to overcome the problem of image transparency
such as that where UV light is piped, via fiber optic guides, to
individual voxels in a medium composed of a fluorescent dye
. The fluorescing region is not completely transparent as in
virtual image or moving screen methods. The device described
consists of a stack of transparent spacers, with layers of fluores-
cent dye between. The device was small at 11
FELIX” volumetric display is based on fluorescence excitation
crystal, which describes the complete volume of the display, is
dopedwithions ofrareearthelements.These ions are excitedin
two steps by two intersecting infrared laser beams and so start
to shine. This process is called two-step up-conversion and is
based on the absorption of two IR-photons by a rare earth-ion
which then emits a visible photon. The medium in which the
picture appears has to be solid, transparent and capable of being
doped. These demands are fulfilled by crystals, glasses and, in
future, even by plastic.
Dolgoff  describes a simple two-plane method which
uses a partially reflecting mirror to combine the real foreground
image behind it with the reflected background image. The
foreground image is brighter in order for it to appear opaque.
This type of display is not suitable for video, but does provide
a simple and inexpensive display that can be very effective
in applications such as advertising. An attempt to overcome
the transparency problem was made by Son et al. . They
proposed the use of an SLM in front of a translucent volumetric
image in order to block light in directions that would normally
be occluded. However, it seems that the speed and spatial
resolution of the SLM will be insufficient for this method to be
FogScreen have patented a 2-D projection screen which uti-
lizes a bank of “fog-like” particles . The projected image
appears to float in space and encourages the audience to in-
teract with it. It is possible to project different images on both
sides without interfering each other because screen transmits
light more than it reflects. It can also be made interactive by
integrating 2-D tracking. Although only 2-D it is a volumetric
display in the sense that the floating image is formed within a
volumeof free space.It is difficultto add more projection layers
to create depth, as the projection between the layers would need
some space and the screens shine through to the next layer.
However, it can be extended into a pseudo-3-D screen by using
dual-sided rendering, head-tracking rendering and stereoscopic
BENZIE et al.: A SURVEY OF 3DTV DISPLAYS: TECHNIQUES AND TECHNOLOGIES1653
VIII. HOLOGRAPHIC DISPLAYS
While holography is well established amongst the scientific
eral public. Many have seen the 3-D holographic images found
on credit cards but are unfamiliar with the high quality images
that holography can produce; even fewer people have witnessed
dynamic holograms or holographic movies. Color holography
produces 3-D images of startling reality, with depth cues and
parallax that are difficult to achieve by other techniques. The
main advantages of a holographic system are the recording of
the true 3-D wavefront of a scene and the retention of motion
parallax. However, this is offset against the need for a very
high bandwidth and the difficulty of obtaining natural shading.
The display of captured and transmitted 3DTV video signals
by holographic means is highly desirable . However, a full
parallax, large area, interactive, moving, color holographic dis-
play, which is thought by manyto be the ultimate goal of 3DTV,
requires incremental and parallel development in many essen-
tial areas of technology before it can be brought to fruition. For
example, a large display of say 100 mm diagonal will need dra-
niques toenable aspatial light modulator(SLM)tobe manufac-
Law” continues to apply then it could be ten years before a dis-
of SLMs requires advances in interconnection technology and
software required to drive them. Color displays require devel-
opment of compact, safe lasers or LEDs with sufficient coher-
ence and power. Synthesis of high-resolution computer gener-
ated holograms (CGH) with high-spatial frequency content for
3-D object presentation is most crucial for creation of true 3-D
perception of real-world scenes with randomly distributed dif-
fuse objects. The task of CGH synthesis is particularly chal-
lenging in the case of dynamic displays which require fast ap-
proaches for simulation of the underlying physical phenomena
for real-time representation . Compression within holog-
raphy has been considered in various publications . Given
the present state of modern computer power and capacity, such
an approach will rely on approximation techniques and corre-
spondingly fast algorithms for its implementation.
Historically, holographic movies and television have been
limited by their reliance on high-resolution emulsions for
recording and reconstruction. More recently, developments
in electronic sensor technologies and SLMs have moved the
technology into the digital domain where transmission of
holographic data is possible. The main drawback is the reso-
lution of the existing dynamic SLMs such as LCD or DMD
is not enough for reconstruction of the 3-D images with high
quality. Only acoustooptic modulator (AOM) SLMs can pro-
vide such a resolution in the Bragg diffraction mode but they
are 1-D structures whichentails omitting verticalparallax. Prior
to the development of high resolution charge coupled devices
(CCDs) or CMOS optical sensors, holographic recording media
have included: silver halide emulsions, photochromic, bacteri-
is required that can resolve high spatial frequency interference
patterns in the order of 1000 linepair/mm (for large angle off
axis recording). Much of the work of reflection holography
may be applied to the dynamic case of holographic television
–. A number of challenges need to be overcome before
holographic television is ready for the mass market. However,
the prospect of full wavefront dynamic reconstruction of 3-D
images is tantalizingly close and must be aspired to.
An alternative to conventional holographic recording and re-
play is optical scanning holography (OSH) whereby 3-D infor-
mation is recorded in a single 2-D scan. OSH uses heterodyning
techniques, accomplished by the use of an AOM , . A
novel approach is to directly acquire horizontal parallax holo-
grams from a scene using OSH and later reconstruct on these on
a SLM . Reduced parallax holographic systems are another
alternative, i.e., systems which sacrifice one plane of parallax
in order to reduce information content or increase brightness
. It has often been said that someone viewing a hologram
for the first time only notices the presence of vertical parallax
when “jumping up and down with excitement”! It is true that
in the case of 3DTV or movies, the viewer will normally be
seated and unaware of vertical parallax. However, in an oper-
combination of a lenticular sheet display and LCD represents a
good cross-over of technologies with the autostereoscopic ge-
ometry. The 3-D content reduction for real-time visualization
of 3-D moving objects could be realized by hybrid holograms
with synthesized mosaic structures. The hybrid hologram rep-
resents a combined structure of high spatial frequency analogue
posea problemtoa holographicdisplay whentheoriginalscene
is illuminated by lasers. The natural shading from the sun or
room lighting would be lost in reconstruction.
A. Underlying Technologies and Techniques
As suggested above, the development of a full scale interac-
tive holographic display requires the parallel development of
many underlying technologies. While the spatial resolution of
constantly improving. SLMs modulate a coherent light source
based on input control parameters. The two main types of SLM
are optically and electrically addressable. With an optically
addressable SLM a light source is used to write the prescribed
spatial pattern onto the SLM. Electrically addressed SLMs can
convert electrical signals into an interference pattern. Here,
each pixel may be individually addressed. The main qualities
required from a SLM in digital holography are that they are
fast, have high transmittance and a good optical efficiency .
Existing SLM technologies include LCDs, AO displays and
digital micromirror devices. Liquid crystal SLMs are elec-
trically addressed devices with a driving mechanism similar
to that used in commercially available LCTVs. LC displays
can be used to write a dynamic interference pattern. Displays
such as metal–insulator–metal (MIM) and twisted-nematic
(TN) have been used in electronic holography displays .
Liquid-crystal-on-silicon (LCOS) devices use a combination
of a liquid crystal and a mirror to perform optical modula-
tion. The main advantage of LCOS technology is the high
1654IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 2007
fill factor of up to 93%. An optically addressed liquid crystal
device (OALCD) converts an intensity pattern into a phase
or amplitude modulation. The backside (writing side) of the
display is illuminated by the intensity pattern which creates a
corresponding 2-D-modulation of the refractive index in the
liquid crystal film on the other side of the display (reading
side). The incoherent light is converted into a voltage applied to
the liquid crystal layer. Therefore, the magnitude of refractive
index change is dependent on the incident intensity.
Digitalmirror devices form anothergroupofsystemsfor
spatial light modulation. DMDs consist of an array of tiltable
micro-mirrors mounted on hinges over a CMOS static RAM
chip. Pixel counts of more than 1280
with pixel pitches less than 16 m. The mirrors are addressable
via binary data sent to the SRAM, which produces an elec-
trostatic charge distribution, causing the individual mirrors to
tilt either ON or OFF. DMDs have been utilized in an array
to enable holographic video display . One advantage of
DMDs overLCDs is that the incidentlight is reflected with high
efficiency, while the liquid crystal systems always suffer from
a certain amount of light absorption, even if they are changing
the phase only. Additionally, because of the lower absorption,
mirror based devices can be used with higher light intensities
without running into thermal problems. DMDs are used in a
wide-range of applications such as optical switching, spec-
troscopy, scanners, confocal microscopes, hologram memory
devices, telecommunication applications, biomedical imaging,
interferometry and bar-code readers.
Polymer dispersed liquid crystals (PDLC) and their holo-
graphically formed counterpart (HPDLC) are composite ma-
terials with LC nano-scale droplets dispersed in photopolymer
film. They can provide low cost, high resolution, high contrast,
with good switch-on/switch-off times, but this is at the cost
of relatively high control voltage and limited viewing angles
–. AO SLMs have been used for some time in 3-D
of travelling acoustic waves and a coherent light source within
a medium, to modulate the properties of the transmitted optical
wavefront. The AO medium may consist of a piezo-electric
transducer bonded to a suitable crystal such as fused silica. On
application of a radio frequency (RF) signal to the medium,
the acoustic wave acts like a “phase grating” travelling through
the crystal at the acoustic velocity of the material and with
an acoustic wave-length dependant on the frequency of the
RF signal. The incident laser beam is then diffracted by this
grating. Acoustooptic SLMs generate a 1-D modulation and
require a scan mechanism. The scanning optics in AOM system
requires synchronization to the fringe data stream and optical
processing to for the reconstruction.
The major disadvantage of SLMs compared to conventional
emulsion-holograms is the lower space-bandwidth product
(SBP) of these devices. SBP is defined as the product of the
device dimension and the pixel frequency. With a resolution
of 3000 line-pair/mm and a size of 20 cm
hologram has a SBP in the region of 1000, whereas current
SLMs approach 100. This low SBP has a strong influence
on the resolution limit. The smaller the pixel pitch and the
larger the wavelength, the smaller is the achievable distance
1024 can be obtained
between the object and SLM during the reconstruction for the
same object. The reduction of pixel size in combination with
an increasing number of pixels leads to a decreasing speckle
size in the reconstructed image. Another advantage of a shorter
distance between SLM and object is the reduced intensity loss
and thus better reconstruction of the object. SLM technologies
are constantly evolving with reduced pixel pitch and increased
The development of lasers to meet the specific requirements
of holographic display systems requires some thought. A color
holographic display using an LCOS SLM, would require the se-
lectionof three wavelengths,to providechrominance and wave-
length matching of the characteristics of the SLM and human
eye. Typical wavelengths for this would be blue 476 nm, green
532 and red 648 nm . The spatial coherence of lasers used,
must be considered to provide even illumination of the SLM
(Hologram) and temporal coherence must be considered with
regards to speckle. The power output of such lasers and their
compactness for a television system must also be calculated, to
provide adequate illumination. High power lasers would be re-
quired ( 300 mW, visible), along with careful thought on the
ocular hazards (for example, suitable beam expansion, fail safe
mechanisms etc.). The two main methods for color holography
are color mixing  and time-division multiplexing illumi-
nation . In color mixing, three lasers are combined using
beam splitters to form a color hologram. In time-division multi-
plexing, the lasers are rapidly alternated using an aperturing ar-
rangement or pulsed lasers. Recently, a color electroholography
system using three colored reference sources that required no
additional optics (e.g., shutters, etc.) was demonstrated .
Color displays require development of compact, safe lasers or
LED’s with sufficient coherence and power.
IX. FUTURE OF 3-D DISPLAYS
Clearly no display method is without its problems or limita-
tions. The development paths which have to be followed before
a full 3-D display can be realized are very complex. Given the
current state-of-the-art, nonholographic displays, such as volu-
metric or autostereo, are in a more advanced state of develop-
ment and it is felt that they are more likely to reach the market
place in a shorter time frame. Many approaches have been out-
lined, from simple stereo through to full parallax holography.
It is not clear which particular technology will dominate future
3-D displays. However, it is thought that it will be the applica-
tion that drives the technology development. Those stereo tech-
nologies, which require glasses or other viewing aids, are un-
likely to be accepted by the general consumer.
an autostereoscopic walk-around true 3-D display. In general,
there are numerous bright ideas, but the field is still in its in-
fancy. Economics dictates that there must be enough market de-
mand to provide and develop products for the future. This is
not the case for most volumetric displays. Many volumetric dis-
plays already have some special applications which may have a
profitable market. However, it is difficult to see any of them re-
Although there is much activity surrounding head-tracked
3-D displays, there are few methods that have the potential to
BENZIE et al.: A SURVEY OF 3DTV DISPLAYS: TECHNIQUES AND TECHNOLOGIES1655
be developed into a multiple viewer display suitable for televi-
sion. Any display that uses a Fresnel lens has a limited area over
which exit pupils can be formed and are unlikely to form the
basisof amultiuser display.Methodutilizing alenticular screen
will suffer likewise. The geometry of the parallax methods does
not allow any of them to serve more than one viewer. A number
of successful HMDs have been developed for simulator and
gaming applications. There is a huge amount of interest from
military, industrial, and consumer markets in having realistic
3-D displays and rendering tools. However, the lack of accom-
modation depth cues in stereo display leads to binocular rivalry.
Some projection methods have the potential for multiple viewer
operation, but these would require extremely large housings.
A holographic display will inevitably rely on digital holog-
raphy and spatial light modulation techniques. The quality of
ability of highresolution SLMsand CCDs. Thecapture of holo-
graphic images using CCDs will provide the high resolution
data required to drive the SLM. Alternatively, computer gen-
erated holograms may be used to drive the holographic recon-
struction. Research is required into finding suitable SLMs for
reconstruction and techniques to improve on the limited resolu-
tion provided by these sensors.
Liquid crystal displays are well suited to dynamic digital
holography since they are electronically addressable. HPDLC
displays offer high resolution compared to the conventional
LCOS displays. LCOS displays are commercially available,
but their limited spatial resolution and pixel count produces a
small imaging area. LCDs have formed the basis of most im-
plementations of holographic TV. The knowledge gained from
optimizing LCOS displays for holographic reconstruction, will
be invaluable in the progress of the technology and creating
larger displays. Learning how to maximize the limited reso-
lution will produce techniques that optimize the efficiency of
later displays and help to reduce bandwidth. Further, research is
required into methods of producing realistic color images from
multiplex SLMs. MEMs devices can be used as spatial mod-
ulators for the holographic display, with commercial devices
available such as the DMD. DMDs provide pixel counts and
pixel pitches of a similar order to current LCOS technology,
however, each pixel acts in a binary intensity mode that reduces
the quality of reconstruction. The main advantage of DMDs
is that light is reflected with a high efficiency compared to
LCOS technologies. It is thought that until and unless other
technologies are developed holography will be reliant on LCOS
technology for the display. Further it is expected that the SBP
of these displays will increase, thus improving the imaging area
and quality of reconstructions.
mentum, it is felt that the advances currently being made in
autostereo displays suggest that a multiviewer, high resolution,
bright display could be achieved earlier than a holographic one.
The authors would like to thank everyone within the 3DTV
consortium (EC within FP6 under Grant 511568) who has pro-
vided help and contribution.
 N. Navab, “Developing killer apps for Industrial Augmented Reality,”
IEEE Comput. Graph. Appl., vol. 9, no. 3, pp. 16–20, Mar. 2004.
 R. Patterson, M. Winterbottom, and B. Pierce, “Perceptual issues in
the use of head-mounted visual displays,” Human Factors, vol. 48, pp.
 W. A. IJsselstein, P. J. H. Seuntiens, and L. M. J. Meesters,
“State-of-the-art in human factors and quality issues of stereo-
scopic broadcast television,” ATTEST Proj. Deliverable 1, European
Commision Project IST-2001-34396, 2002 [Online]. Available:
 L. Meesters, W. A. Ijsselsteijn, and P. J. H. Seuntiens, “A survey of
perceptual evaluations and requirements of 3-D tv,” IEEE Trans. Cir-
cuits Syst. Video Technol., vol. 14, no. 3, pp. 381–391, Mar. 2004.
 F. L. Kooi and A. Toet, “Visual Comfort of binocular displays,” in Dis-
plays.New York: Elsevier, 2004.
 N. L. Silverman, B. T. Schowengerdt,J. P. Kelly, andE. J.Seibel, “En-
gineering a retinal scanning laser display with integrated accommoda-
tive depth cues,” Soc. Inf. Displ. Techn. Dig., no. 2, pp. 1538–1541,
 S. Pastoor, “Human factors of 3DTV: An overview of current research
at Heinrich-Hertz-Institut Berlin,” IEE Colloq. Stereosc. Telev., 1992,
Digest No. 1992/173, p 11/3.
 J. Konrad, “Enhancement of viewer comfort in stereoscopic viewing:
Parallax adjustment,” in Proc. SPIE, Stereosc. Displ. Appl. X, 1999,
vol. 3639, pp. 179–189.
 S. Pastoor, “3-D-television: A survey of recent research results on
subjective requirements,” in Signal Process.: Image Commun..
York: Elsevier, 1991, p. 24.
 C. W. Smith and A. A. Dumbreck, “3-D TV: The practical require-
ments,” Televis.: J. Roy. Televis. Soc., pp. 12–22, Jan./Feb. 1988.
 H. Yamanoue, “The relation between size distortion and shooting con-
ditions for stereoscopic images,” in Proc. Asia Displays’95, 1995, pp.
 H. Yamanoue, M. Nagayama, M. Bitou, and J. Tanada, “Orthostereo-
scopic conditions for 3D HDTV,” in Proc. SPIE, Stereoscop. Displ.
Appl. IX, 1998, vol. 3295, pp. 111–112.
 E. Mulkens and J. W. Roberts, “Effects of display geometry and pixel
structure on stereo display usability,” in Proc. SPIE, Stereoscop. Displ.
Virt. Real. Syst. VIII, 2001, vol. 4297, pp. 276–289.
 S. Pastoor, “3-D-television: A survey of recent research results on
subjective requirements,” in Signal Process.: Image Commun..
York: Elsevier, 1991, p. 23.
 D. Marr, Vision.New York: W. H. Freeman, 1982, pp. 147–148.
 P. M. Schweiwiller, A. A. Dumbreck, and A. D. Chapman, “Videotape
recording of 3D television pictures,” in Proc. SPIE Stereoscop. Displ.
Appl., 1990, vol. 1256, p. 239.
scopic views,” in Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst., 1994,
vol. 2177, pp. 92–96.
 M. Siegel, “Perceptions of crosstalk and the possibilities of a zoneless
autostereoscopic display,” in Proc. SPIE, Stereoscop. Displ. Virt. Real.
Syst. VIII, 2001, vol. 4297, pp. 34–41.
 R. Kingslake, Optical System Design.
 C. A. Maxwell, “Flicker science and the consumer,” Inf. Displ., pp.
7–10, 1992, 11/92.
 A. G. Knapp, “Television applications,” SID Residential Summer
 T. Bardsley, “The design and evaluation of an autostereoscopic com-
puter graphics display,” Ph.D. dissertation, De Montfort Univ., Le-
icester, U.K., 1995, p 144.
 R. Collender, “3-D television, movies and computer graphics without
glasses,”IEEE Trans. Consum. Electron., vol. CE-32, no. 1, pp. 56–61,
 H. B. Tilton, “Large-crt holoform display,” in Proc. IEEE Int. Displ.
Research Conf., 1985, pp. 145–146.
 I. Sexton, “Parallax barrier display systems,” in Proc. IEE Colloq.
Stereoscop. Televis., 1992, pp. 5/1–5/5.
 M. McCormick, N. Davies, andE. G. Chowanietz,“Restrictedparallax
images for 3D t.v.,” in Proc. IEE Colloq. Stereoscop. Telev., 1992, pp.
 J. H. Park et al., “Depth-and viewing-angle-enhanced 3-D/2-D con-
vertible display based on integral imaging,” in Proc. 11th Int. Displ.
Workshops, Niigata, Japan, 2004, pp. 1455–1458.
New York: Academic, 1983,
1656IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 2007
 P. St. Hilaire, T. Honda, Ed., “Modulation transfer function of holo-
graphic stereograms,” in Proc. SPIE, Int. Conf. Appl. Opt. Hologr.,
1995, vol. 2577, pp. 41–49.
 S. Pastoor, “Human factors of 3DTV: An overview of current research
at Heinrich-Hertz-Institut Berlin,” in Proc. IEE Colloq. Stereoscop.
Telev., 1992, p. 11/3.
 T. Balogh, “Method and apparatus for producing 3-D picture,” U.S.
Patent 5801761, Sep. 1, 1998.
 C. Cameron, “Collaborative autostereo 3-D display system,” presented
at the Proc. Joint Meeting 3-D Technol., San Jose, CA, Apr. 25, 2005
[Online]. Available: http://www.sid.org/chapters/uki/qinetiq.pdf
 A. R. L. Travis, “Autostereoscopic 3-D display,” Appl. Opt., vol. 29,
pp. 4341–4342, 1990.
 S. Hines, “Autostereoscopic video display with motion parallax,” in
Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst. IV, 1997, vol. 3012, pp.
without glasses,” in Proc. SID, 1990, vol. 31, no. 3, pp. 263–266.
Displ. Appl. X, 1999, vol. 3639, pp. 84–91.
 S. T. de Zwart, W. L. IJzerman, T. Dekker, and W. A. M. Wolter, “A
20-in. switchable auto-stereoscopic 2D/3D display,” in Proc. 11th Int.
Displ. Workshops, Niigata, Japan, 2004, pp. 1459–1460.
 L. Lipton, “Method and Apparatus for Maximizing the Viewing
Zone of a Lenticular Stereogram,” U.S. Patent No. 65190888, Feb.
 S.-S. Kim, K. H. Sohn, V. Savaljev, E. F. Pen, J. Y. Son, and J.-H.
Chun, “The optical design and analysis for super multiview 3-D
imaging system,” in Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst.
VIII, 2001, vol. 4297, pp. 222–226.
 S. Takahashi, T. Toda, and F. Iwata, “Full color 3D-video system using
grating image,”in Proc.SPIEPractical Hologr.X,1996,vol.2652,pp.
 J. R. Moore, N. A. Dodgson, A. R. L. Travis, and S. R. Lang, “Time-
Displ. Virtual Real. Syst. IV, 1996, vol. 2653, pp. 10–19.
 N. A. Dodgson, J. R. Moore, S. R. Lang, G. Martin, and P. Canepa, “A
50 time-multiplexed autostereoscopic display,” in Proc. SPIE, Stere-
oscop. Displ. Virt. Real. Syst. VII, 2000, vol. 3957, pp. 177–183.
 C. M. G. Lee and A. R. L. Travis, “Flat-panel autostereoscopic view-
sequential 3D display backlight,” in Proc. 11th Int. Displ. Workshops,
Niigata, Japan, 2004, pp. 1479–1482.
 T. Bardsley, “The design and evaluation of an autostereoscopic com-
puter graphics display,” Ph.D. dissertation, Sch. Comp. Sci., De Mont-
fort Univ., Leicester, U.K., 1995.
 H. Morishima,H. Nose, N.Yaniguchi,K. Inoguchi,and S.Matsumura,
“Rear cross lenticular 3D display without eyeglasses,” in Proc. SPIE,
Stereoscop. Displ. Appl. IX, 1998, vol. 3295, pp. 193–201.
 J. Harrold, D. J. Wilkes, and G. J. Woodgate, “Switchable 2D/3D dis-
play -solid phase liquid crystal microlens array,” in Proc. 11th Int.
Displ. Workshops, Niigata, Japan, 2004, pp. 1495–1496.
 G. J. Woodgate, “Lens array structure,” U.S. Patent 7215475, 2007.
 D. Trayner and E. Orr, “Autostereoscopic display using holographic
optical elements,” in Proc. SPIE, Stereoscop. Displ. Appl. VII, 1996,
vol. 2653, pp. 65–74.
 J. Eichenlaub, “A lightweight, compact 2D/3D autostereoscopic lcd
backlight for games, monitor and notebook applications,” in Proc.
SPIE, Stereoscop. Displ. Virt. Real. Syst. IV, 1997, vol. 3012, pp.
 J. Hamasaki, “Aberration theories of lenticular and related screens,” in
Proc. Int. Workshop Stereoscop. 3-D Imaging, Santorini, Greece, Sep.
1995, pp. 14–31.
 J. Y. Son, V. V. Smirnov, Y. S. Chun, and S. S. Kim, “Non-glasses
type stereoscopic display system based on polarization,” in Proc. SPIE
Stereoscop. Displ. Virt. Real. Syst. VI, 1999, vol. 3639, pp. 132–136.
 S. Faris, “Multi-Mode Stereoscopic Imaging System,” U.S. Patent
6195205 B1, Feb. 27, 2001.
 LL-151-3-D 15” XGA LCD Monitor: 3-D Displ. Monitor Without
the Need For Special Glasses, Sharp, 2005, Company Brochure.
Sharp Systems of America [Online]. Available: http://www.sharpsys-
 C. Kaufmann, A. Liu, and D. Burris, J. D. Westwood, Ed. et al., “DTI
autostereographic display: Initial evaluation,” Medicine Meets Virtual
Reality 2000 2000, IOS.
 P. Surman, I. Sexton, R. Bates, W. K. Lee, and K. C. Yow, “3-D televi-
sion: Developing a multimodal multiviewer TV system of the future,”
J. Soc. Inf. Displ. 12, vol. 153, pp. 153–159, 2004.
 W. Barfield and T. Caudell, Eds., Fundamentals of Wearable Com-
puters and Augmented Reality.
 J. P. Rolland, L. D. Davis, and Y. Baillot, “A survey of tracking tech-
puters and Augmented Reality, W. Barfieldand and T. Caudell, Eds.
Mahwah, NJ: Lawrence Erlbaum, 2001, pp. 67–112.
 A. Schwartz, “Head tracking stereoscopic display,” in Proc. IEEE Int.
Displ. Res. Conf., 1985, pp. 141–144.
 Y. Nishida, T. Hattori, S. Omori, J. Suzuki, K. Katayama, and S.
Sakuma, “Simplification of infrared illumination of stereoscopic
liquid crystal TV,” in Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst.
II, 1995, vol. 2409, pp. 96–100.
 T. Hattori, “Stereoscopic display employing head-position tracking
using large format lenses,” in Proc. SPIE, Stereoscop. Displ. Appl. IV,
1993, vol. 1915, pp. 2–5.
 Y. Nishida, T. Hattori, S. Sakuma, K. Katayama, S. Omori, and T.
Fukuyo, “Autostereoscopic liquid crystal display II (practical applica-
tion),” in Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst., 1994, vol.
2177, pp. 150–155.
 C. H. Morimoto, D. Koons, A. Amir, and M. Flickner, “Pupil detection
pp. 331–335, 2000.
 H. Imai, T. Susumu, and M. Imai, “Viewing point detection system
using specific image processing for eye-position tracking autostereo-
scopic display,” in Proc. SPIE, Stereoscop. Displ. Appl. X, 1999, vol.
3639, pp. 92–98.
 H. Heidrich, A. Schwerdtner, A. Glatte, and H. Mix, “Eye position
detection system,” in Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst.
VII, 2000, vol. 3957, pp. 192–197.
 S.-W. Shih, Y.-T. Wu, and J. Liu, “A calibration-free gaze tracking
technique,” in Proc. ICPR 2000, 2000, vol. 4, pp. 201–204.
 K. Talmi and J. Liu, “Eye and gaze tracking for visually controlled
interactive stereoscopic displays,” Signal Process.: Image Commun.,
vol. 14, pp. 799–810, 1999.
 A. C. Traub, “Stereoscopic display using rapid varifocal mirror oscil-
lations,” Appl. Opt., vol. 6, no. 6, pp. 1085–1087, Jun. 1967.
 S. Mckay, S. Mason, L. S. Mair, P. Waddell, and S. M. Fraser, “Stereo-
scopic display using a 1.3 diameter stretchable membrane mirror,” in
Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst. VI, 1999, vol. 3639, pp.
 S. Mckay, G. M. Mair, S. Mason, and K. Revie, “Membrane-mirror-
based autostereoscopic display for teleoperation and telepresence ap-
vol. 3957, pp. 198–207.
 J. Fajans, “Xyzscope—A new option in 3-d display technology,” in
Proc. SPIE, Vis. Date Interpret., 1992, vol. 1668, pp. 25–26.
 N. Yanagisawa, K. Kyung-Tae, J.-Y. Son, M. Tatsuya, and O.
Takatoshi, “A focus distance controlled 3DTV,” Proc. SPIE, vol. 3012,
pp. 256–261, 1997.
 B. T. Schowengerdt and E. J. Seibel, “True 3-D scanned voxel displays
using single and multiple light sources,” J. Soc. Inf. Displ., vol. 14, no.
2, pp. 135–143, 2006.
 G. E. Favalora, R. K. Dorval, D. M. Hall, M. Giovinco, and J. Napoli,
“Volumetric 3-D display system with rasterization hardware,” in Proc.
SPIE, Stereoscop. Displ. Virt. Real. Syst. VIII, 2001, vol. 4297, pp.
 Perspecta, Actuality Systems, Inc., Burlington, MA 2004, .
 D. J. Solomon, “Visual Special Effects Displ. Device,” U.S. 6404409,
Jun. 11, 2002.
 D. Bahr, K. Langhans, M. Gerken, C. Vogt, D. Bezecny, and D.
Homann, “Felix: A volumetric 3D laser display,” in Proc. SPIE
Projection Displ. II, San Jose, CA, 1996, vol. 2650, pp. 265–273.
 D. L. Macfarlane and G. R. Schultz, “A voxel based spatial display,”
in Proc. SPIE Stereoscop. Displ. Virt. Real. Syst., 1994, vol. 2177, pp.
 G. Dolgoff, “Real-Depth imaging: A new imaging technology with in-
expensive direct-view (no glasses) video and other applications,” in
Proc. SPIE, Stereoscop. Displ. Virt. Real. Syst. IV, 1997, vol. 3012,
control method for solid objects 3-D display,” in Proc. SPIE, Stereo-
scop. Displ. Virt. Real. Syst. IV, 1997, vol. 3012, pp. 62–272.
 Palovuori, “Method and apparatus for forming a projection screen or
projection volume,” U.S. Patent 6819487 B2, Nov. 18, 2004.
Mahwah, NJ: Lawrence Erlbaum,
BENZIE et al.: A SURVEY OF 3DTV DISPLAYS: TECHNIQUES AND TECHNOLOGIES1657
 S. DiVerdi, I. Rakkolainen, T. Höllerer, and A. Olwal, “A novel walk-
through3-Ddisplay,”in Proc.SPIE,Stereoscop.Displ. Virt.Real.Syst.
XIII, 2007, vol. 6055, pp. 1–10.
 L. H. Enloe, J. A. Murphy, and C. B. Rubinstein, “Hologram trans-
mission via television,” Bell Syst. Tech. J., vol. 45, no. 2, pp. 225–339,
 G. E. Moore, “Cramming more components onto integrated circuits,”
Electron., vol. 38, no. 8, Apr. 19, 1965.
 L. Ahrenberg, P. Benzie, M. Magnor, and J. Watson, “Computer gen-
erated holography using parallel commodity graphics hardware,” Opt.
Exp., vol. 14, no. 17, pp. 7636–7641, 2007.
 H. Yoshikawa and J. Tami, “Holographic image compression by
motion picture coding,” in Proc. SPIE, Practical Hologr. X, 1996,
 T. Kubota, “Recording of high quality color holograms,” Appl. Opt.,
vol. 25, pp. 4141–4145, 1986.
 J. L. Walker and S. A. Benton, “In-situ swelling for holographic color
control,” in Proc. SPIE, Practical Hologr. III, S. A. Benton, Ed., 1989,
 P. Hubel and L. Solymar, “Color-reflection holography: Theory and
experiment,” Appl. Opt., vol. 30, no. 29, pp. 4190–4203, 1991.
 P. M. Hubel, “Recent advances in color reflection holography,” in
Proc. SPIE, Practical Hologr. V, S. A. Benton, Ed., 1991, vol.
 T.-C. Poon, “Three-dimensional television system using optical scan-
ning holography and spatial light modulator,” J. Inf. Displ., vol. 3, no.
 T.-C. Poon, “Recent progress in optical scanning holography,” J.
Hologr. Speckle, vol. 1, pp. 6–25, 2004.
 T.-C. Poon, T. Akin, G. Indebetouw, and T. Kim, “Horizontal-par-
allax-only electronic holography,” Opt. Exp., vol. 13, pp. 2427–2432,
 K. Maeno, “Electro-holographic display using 15 mega pixels LCD,”
Proc. SPIE, vol. 2652, pp. 15–23, 1996.
 T. Kreis, P. Aswendt, and R. Höfling, “Hologram reconstruction
using a digital micromirror device,” Opt. Eng., vol. 40, pp. 926–933,
using liquid crystal television spatial light modulators,” J. Electron.
Imaging, vol. 2, no. 2, pp. 93–99, 1993.
 P. F. V. Kessel, L. J. Hornbeck, R. E. Meier, and M. R. Douglass, “A
MEMS-based projection display,”Proc. IEEE, vol. 86, pp. 1687–1704,
 M. Heubschman, B. Manjuluri, and H. Garner, “Digital micromirrors
enable holographic video display,” Laser Focus World, vol. 40, no. 5,
 R. T. Pogue, R. L. Sutherland, M. G. Schmitt, L. V. Natarajan, S. A.
Siweski, V. P. Tondiglia, and T. G. Bunning, “Electrically switchable
bragg gratings from liquid crystal/polymer composites,” Appl. Spec-
trosc., vol. 54, pp. 12A–28A, 2000.
 H. Xianyu, J. Qi, R. Cohn, and G. P. Crawford, “Total-internal reflec-
vol. 28, pp. 792–794, 2003.
 K. Beev, T. Angelov, A. Petrov, and S. Sainov, “Investigation of Bragg
gratings recorded in polymer-dispersed liquid crystals,” J. Optoelec-
tron. Adv. Mater., vol. 6, no. 3, pp. 799–803, 2004.
 P. St. Hilarie, S. A. Benton, M. Lucente, and P. M. Hubel, “Color im-
ages with the MIT holographic video display,” in Proc. SPIE Practical
Hologr. VI, 1992, vol. 1667, pp. 73–84.
 L. Onural, G. Bozdagi, and A. Atalar, “New resolution display device
for holographic three dimensional video: Principles and simulations,”
Opt. Eng., vol. 33, no. 3, pp. 835–844, 1994.
 M. S. Peercy and L. Hesselink, “Wavelength selection for true color
holography,” Appl. Opt., vol. 33, p. 6811, Oct. 1994.
 K. Sato, “Animated color 3-D image using kinoforms by liquid
crystal devices,” J. Inst. Telev. Eng. Jpn., vol. 48, pp. 1261–1266,
 T. Shimobaba and T. Ito, “A color holographic reconstruction system
by time division multiplexing with reference lights of laser,” Opt. Rev.,
vol. 10, pp. 339–341, 2003.
 T. Ito, “Color electroholography by three colored reference lights si-
multaneously incident upon one hologram panel,” Opt. Expr., vol. 12,
no. 18, 2004.
 D. J. Solomon, “Three-dimensional volumetric display system,” U.S.
Patent 4983031, Jan. 8, 1981.
 D. J. Solomon, “High contrast display device enclosure system using
Philip Benzie (M’07) received the B.Eng. degree
(hons.) in electrical and electronic engineering
and the Ph.D. degree from Aberdeen University,
Aberdeen, U.K., in 2001 and 2006, respectively. His
Ph.D. dissertation was on the application of finite
element analysis to holographic interferometry for
He is currently working at Aberdeen University
as a Research Fellow and Section Leader for holo-
graphic displays in the European Network of Excel-
lence wide 3DTV project. His current research inter-
John Watson (SM’05) received the Ph.D. degree for
the University of St. Andrews, U.K., in 1978.
He joined the U.K. Atomic Energy Authority at
the Dounreay Fast Reactor site in 1976 where he
worked on instrument development for nuclear plant
inspection. He joined Robert Gordon’s Institute of
Technology, Aberdeen University (AU), Aberdeen,
U.K., in 1981 as a Lecturer in Electronic and Elec-
trical Engineering before moving to Department
of Engineering, AU, in 1984. He was appointed
to Senior Lecturer, then Reader and to a Chair (Professorship) in Optical
Engineering in 2004. His main research interests are in holography, subsea
laser welding, laser induced spectral analysis, and optical image processing.
His research group at AU have developed two underwater holographic cameras
for the measurement and analysis of marine organisms and particles.
Prof. Watson is a Fellow of The Institution of Engineering and Technology, a
Fellowof The Instituteof Physics, and an elected member ofthe Administrative
Committee of OES.
Phil Surman received the B.Sc. degree in electrical
and electronic engineering from Heriot-Watt Univer-
sity, Edinburgh, U.K., in 1971 and the Ph.D. degree
on the subject of head-tracked stereoscopic displays
from DeMontfort University, Leicester, U.K., in
The multiuser 3-D display developed within the
EU funded ATTEST project is a continuation of his
Ph.D. work. A laser-based version of this display is
Ismo Rakkolainen receivedtheDr.Tech.from the Tampere UniversityofTech-
nology, Tampere, Finland, in 2002.
bara, in 2004–2006, a Senior Researcher at Tampere University of Technology,
23 journal and conference proceeding articles, 1 book chapter, 4 patents, and
several innovation awards. His primary research interests include 3-D display
technology, 2-D midair displays, virtual reality, interaction techniques, novel
user interfaces, and the impact of information technology on society.
Klaus Hopf received the Dipl.-Ing. degree in electrical engineering from the
Technical University of Berlin, Germany, in 1986. He has been with the Hein-
rich-Hertz-Institute since 1987 and performed work within government-funded
research projects on the development of novel technologies for image represen-
tation in high resolution (HDTV) and 3-D. He worked on the ATTEST project,
managing the development of a headtracking autostereoscopic 3-D display, tar-
geted towards 3DTV applications in the Internet domain. His present interests
focus on image representations using autostereoscopic displays and implemen-
tations of novel human-machine interfaces in augmented reality environments.
Currently he is involved in the participation of Fraunhofer HHI in the Network
1658IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 17, NO. 11, NOVEMBER 2007
Hakan Urey (M’03) received the B.S. degree from
1992, and the M.S. and Ph.D. degrees from Georgia
Institute of Technology, Atlanta, in 1996 and 1997,
all in electrical engineering.
He is currently an Assistant Professor of Electrical
Engineering at Koç University, Istanbul, Turkey. He
worked for Bilkent University, Ankara, and Georgia
Tech Research Institute, Atlanta, as a Graduate Re-
search Assistant, and Call/Recall Inc., San Diego, as
a coop student. He joined Microvision Inc., Seattle,
WA, in 1998 and he was the Principal System Engineer when he left and has
been consulting for the company since 2002. He has published more than 70
journal and conference papers, 6 edited books, 2 book chapters, and he has 15
issued and several pending patents.
His research interests are in scanning systems, microoptics, optical MEMS, op-
tical system design, and display and imaging systems a member of SPIE, OSA,
IEEE-LEOS, and the president of the Turkey chapter of IEEE-LEOS.
Ventseslav Sainov is the Director and Scientific Leader of the Holographic and
holography, interferometry, optical recording, and processing of information.
His expertise is in the fields of light sensitive materials for permanent and
dynamic holographic recording, optical and digital processing of interference
patterns, nondestructive testing, profilometry and 3-D micro/macro measure-
ments. He studied physics at the University of Hanover. Afterwards he worked
10 years in different laser companies developing CW lasers for metrology and
printing applications. During this time, he prepared his doctoral thesis “Tune-
able Microcrystal Laser for Absolute Distance Interferometry.” Since 2002, he
is with the Central Laboratory of Optical Storage and Processing of Informa-
tion, Bulgarian Academy of Sciences, Sofia, Bulgaria, as the Head of the Op-
tical Metrology Department.
Christoph von Kopylow received the Dipl.-Phys.
degree in physics at the University of Hanover,
Hanover, Germany, in 1992.
He worked for ten years in different laser com-
panies developing CW lasers for metrology and
printing applications. During this time he prepared
his doctoral thesis “Tuneable microcrystal laser for
absolute distance interferometry.” Since 2002, he
has been with Bremer Institute fuer Angewandte
Strahltechnik GmbH, Bremen, Germany, as Head of
the Optical Metrology Department.
Dr. Kopylow is working in the field of shape and deformation measurement
and nondestructive testing with coherent and noncoherent optical methods.