Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02350687 2000-12-04
PCT/NZ99/00072
' ' :~ccGived 07 September 2000
~.,
-1-
A Hi4h Definition Volumetric Display System
(Courcelle)
Field of the Invention
The present invention relates to three-dimensional display systems. More
particularly,
although not exclusively, the present invention relates to volumetric three-
dimensional
display devices suitable for use in applications which include medical
imaging, computer
l0 aided design and molecular modelling. Such uses require high definition,
high quality three-
dimensional images as well as image quality which is substantially invariant
with viewing
direction.
The methods and apparatus discussed herein permit large numbers of voxels to
be
I S illuminated during each image refresh period. This enables image objects
and scenes to be
depicted at high resolutions and therefore represents a considerable advance
over existing
volumetric display systems. Further, the system described herein minimises
variations of
image quality with viewing direction.
20 Background to the Invention
The present invention relates to volumetric three-dimensional image
generation. In such
systems, image objects and scenes are depicted within a physical three-
dimensional
volume (display space).
At the present time, there are a relatively large number of ways in which
volumetric three-
dimensional display devices may be implemented. A fundamental difference in
the various
implementations resides in the method by which the physical volume (display
space) is
generated. Some systems use the rotational motion of a planar, helical or
other surface
~nnErup~D sHE~
IPE.A/btJ
CA 02350687 2000-12-04
PCT/NZ99/00072
' ' Received 07 September 2000
-2-
geometry. Other systems use a reciprocating planar surface. Further, static
volume
systems exist (or have been proposed) in which no motion is required to
generate the
display space. Examples of this latter type correspond to a display space
comprising a
volume of solid or gaseous material, or alternatively a three-dimensional
matrix of
optoelectronic devices where each of the optoelectronic devices is capable of
generating
an image voxel within the display space.
Images may be generated (i.e. voxels illuminated) in various ways. A number of
systems
use a beam addressing approach (beam addressed displays) using laser or
electron beams.
o In systems which use motion for display space generation, a suitably
directed and
modulated laser beam illuminates voxels at the points at which the beam is
made to
intersect with a semi-opaque surface. Similarly an electron beam, at its point
of
intersection with a phosphor coated surface, will stimulate an optical
emission and can be
used to generate a voxel. In the case of a static volume system, voxels may
for example
i 5 be generated by a two step excitation process at the point of intersection
of two non-
visible beams. Other systems use optoelectronic devices for voxel generation.
To date
displays of this type have employed a two dimensional planar matrix of
optoelectronic
devices located upon a rigid surface which sweeps out the display space.
Alternatively, in
the case of static volume displays, the optoelectronic array may take the form
of a three
2o dimensional matrix. Various other techniques also exist, however these tend
to be
variations on the above.
All of the above types of display generally implement or adapt known
technology and, in
applications requiring high image resolution, can exhibit inherent display
limitations which
25 are discussed in detail below.
Generally volumetric display technologies permit images to be viewed through a
wide, and
sometimes practically unrestricted, range of viewing angles. However, many of
the existing
AMENDED SHEET
~p~~ms
CA 02350687 2000-12-04
PCT/NZ99/00072
' received 07 September 2000
-3-
display implementations produce images having a perceived quality which is not
invariant
with viewing position. That is - images are faithfully reproduced only when
viewed from
certain locations. This therefore restricts the viewing angle. In general
image quality varies
as a consequence of viewing direction and image location. This therefore
restricts viewing
freedom despite the viewing freedom which is apparently offered by the display
technique.
This is a consequence of the optical characteristics of the image space and
cannot be
compensated for by the graphics engine.
It is considered that the present invention represents a significant departure
from both
conventional scanned beam devices (for example see International Application
Nos.
PCT/NZ93/00083 and PCT/NZ96/00028) and matrix devices (for example see US
patent
No. 4,160,973).
Systems, in which the image space is generated by the rotational motion of a
surface such
I5 as those described in the abovementioned documents suffer from a several
intrinsic
problems. Whilst these may not be a significant disadvantage in a number of
applications,
their impact increases with the demand for greater image quality, clarity
and definition. Such problems ultimately limit the system performance and may
set an
upper limit to the size of the display space.
One problem is caused by refraction within the body of the rotating surface.
For example,
in a beam addressed system, a surface layer /such as a phosphor coating)
interacts with
the beams) thus permitting the generation of visible voxels. This layer must
be fabricated
on a rigid body. Similarly, in a matrix system employing rotational motion,
the
optoelectronic components must be bonded to a rigid body. The rigid body
should be
transparent and will have finite thickness. Such characteristics will produce
refraction
effects when the body is viewed at an oblique angle whereby light emitted from
the layer
or matrix passes through the transparent material from which the rigid body is
formed. In
AMENDED SHEET
~Rf-'_;~ ~~: ° )
CA 02350687 2000-12-04
PCT/NZ99/00072
' ' Received 07 September 2000
-4-
some cases, total internal reflection in the rigid body will cause voxels to
be distorted
and/or attenuated. The effect of these aberrations is that when the rigid body
is viewed
edge on (or at angles close to this position), optical distortions may be
visible resulting
from the presence of the rigid body and/or occlusion caused by the
devices/material which
s generate the voxels.
For example, in the case of a system using a planar rigid body, the rigid
body, produces
images which, when viewed in line with the axis of rotation, are distorted by
a region or
band of strong attenuation. Although the impact of these effects is
ameliorated by
0 binocular vision (where the axis of rotation is vertical) they still give
rise to regions in
which the image is diminished in intensity.
Flexing of the rigid body may often exacerbate the abovementioned problem. The
rigid
body must therefore be capable of resisting any flexing produced by its
motion. Also, as
I S the size of the display space is increased, it may be necessary to
increase the thickness of
the rigid body in order to avoid flexing. However, as the thickness is
increased, the optical
distortions described above become more severe.
A further problem arises along the axis of rotation. In order to permit the
rigid body to be
20 as thin as possible (while at the same time ensuring mechanical stability)
it is often
necessary to provide additional support in the form of a centre shaft. A rigid
body having
a helical geometry provides greater mechanical strength than a planar
configuration.
However, in both cases, in order to view images in line with, and on the
opposite side of,
the axis of rotation to the position of the observer, it is necessary to look
through the
25 rotational axis and hence view light which has passed through both the
rigid body and
centre shaft (if presentl. Again this leads to optical distortions in the
image. Even in the
case of systems which employ a planar rigid body and no physical centre shaft,
the rigid
AMENDED SHEET
IPEA/AU
CA 02350687 2000-12-04
PCTlNZ99/00072
' ::~ceived 07 September 2000
-5-
body generates an artificial centre shaft with a diameter equal to one or two
times the
thickness of the rigid body.
Matrix systems employing rotational motion can have even greater associated
problems
whereby the optoelectronic devices add to the thickness of the rigid body and
may occlude
light emitted from other elements located at a greater depth within the
display space.
Further, the high density of electrical connections needed to activate each
device can
contribute to distortion or occlusion of images.
l0 Directed beam devices activate voxels sequentially or by the inclusion of
more beams,
directed beam systems are able to demonstrate a degree of parallelism (the
degree of
parallelism being no greater and often less than the total number of beam
scanning
devicesl. Since the generation of each image voxel occupies a finite time, for
each beam
directed system there is an upper limit on the number of voxels it is able to
generate during
15 each image refresh. Often the image refresh period equals the period of
motion of the rigid
body. It is not possible to reduce the period of motion of the rigid body
below that required
in order to produce substantially flicker-free images. Therefore, it is not
possible to extend
the time available for each image refresh beyond that required for tolerable
image flicker.
Thus, the only way of increasing the number of voxels illuminated during each
refresh
2o period (and therefore generate images containing more detail) is by using
more beams and
beam deflection devices.
When multiple beams and beam scanning devices are employed, registration
between the
beams becomes a major issue. Reliance may therefore have to be placed upon
highly
25 accurate mechanical alignment of the beam sources, beam scanning devices
and driving
electronics. The alignment accuracy required increases as the maximum
attainable voxel
density is increased. In systems using charged beams, the beams) may not
propagate in a
rectilinear manner. Manual or automated electronic alignment methods may
therefore be
~wtECv~~fw~ SHEET
i~a~:~~
CA 02350687 2000-12-04
PCT/NZ99/00072
' Received 07 September 2000
-6-
required such as those mentioned in International Application No.
PCT/NZ93/00083. In the
case of systems using laser beams, electronic alignment techniques are
unlikely to be
easily implemented.
In summary, in order to increase image detail (by enabling larger numbers of
voxels to be
illuminated at a higher density), it is necessary to increase the degree of
parallelism in
voxel activation. In the case of scanned beam devices, this necessitates the
provision of
multiple beams and beam scanners able to position voxels with great accuracy.
As the
voxel density is increased the required placement accuracy is greater and so
therefore is
0 the degree to which the beam scanning devices must be mechanically and
electrically
aligned.
The result is complex calibration procedures throughout the operational life
of the display
system. In certain cases electronic alignment techniques may be used, however
this may
I S cause further complication and additional computational overheads.
While the techniques proposed in the abovementioned International Applications
overcome
some of these difficulties, aberrations and calibration difficulties
nevertheless become
significant when the need for higher image definition arises.
Considering systems which do not use directed beams and motion for display
space
generation, and consist of a static three-dimensional matrix of optoelectronic
(or other)
elements. Each member of the matrix is able to give rise to a single voxel.
Consider a cube
containing an array of optical devices leg: LEDs) spaced regularly throughout.
These voxel
generation centres are activated in turn or simultaneously to produce the
desired image.
The optical sources are stationary and can therefore illuminate only a single
location in
space. A prime disadvantage of such a device is that a large number of voxel
generation
centres are required. For example, if the cube has sides of length 20cm, and
intervoxel
AMEN(~En SHEF'r
,. .
CA 02350687 2000-12-04
PCT/NZ99/00072
' Received 07 September 2000
spacing of 0.1 mm, then 8 million voxel generation centres would be needed.
This leads to
a corresponding large number of connections resulting in associated
obstruction and
aberration of the light as it traverses the display space. Furthermore, the
addressing of
such a large number of connections (data pathways) becomes infeasible as the
dimensions
of the image space are increased.
The number of connections and voxel generation centres can be greatly reduced
by display
space generation through the motion of a rigid body. It should also be noted
that the
number of external connections may be reduced byadopting a row and column type
IO addressing technique ("multiplexing"). For example, an array of n by n
optical elements
may be addressed by Zn connections. Elements within this array are addressed
sequentially. By groupingvoxel generation centres, it ~s possible to adjust
the degree to
which elements within the display space are addressed in both serial and
parallel form.
However, multiplexing does not reduce the total number of connections which
must be
I5 made to the voxel generation centres as a whole. Each must be directly
addressed and,
given the large number of voxel generation centres required, it is necessary
to have a large
number of connections within the display space. This can lead to image
occlusion and
distortion. It is noted that the connections to each optical element may take
a non
electrical form.
Further potential problems may be found with a 3-D array implementation if
constructed in
a known arrangement of the optical devices and connections. At certain viewing
angles,
the optical devices and related components will align. Thus, an obstruction
aberration may
be visible as image striation. This may be present in more than one viewing
direction,
depending on the geometrical layout of the obscuring components.
Examples of a matrix display employing rotational motion include that
described in US
Patent No. 4,160,973 (Berlin). The Berlin device uses a 2-D matrix of LEDs
which are
AMENDED SHEET
CA 02350687 2000-12-04
_ PCT/NZ99/00072
Received 07 September 2000
_g_
rotated through a volume creating a display space which is in the shape of
right cylinder.
While reducing the number of optical sources and corresponding connecting
wires, this
construction does not provide the desired optical homogeneity and isotropy.
The presence
of the rigid body upon which the LEDs are mounted may interfere at certain
angles with
the propagation of light. Such aberrations result not only from the rigid body
itself, but
also from the regular layout of the LED's on the planar array. For any viewing
direction,
there will always be a situation (in any complete revolution) where all of the
LEDs lie in a
plane parallel to a line joining one of the viewer's eyes with the centre of
the display
device. Further, on the rigid body there will exist a large number of
connections. The
layout of the connections may of necessity follow a regular pattern. When the
components forming the array align, an obscuration aberration may be visible
in the form
of image striation. Refraction and possible total internal reflection effects
will cause further
image distortions (as described above in the case of systems using rotational
motion).
These distortions are implicit in systems using the rotational motion of a
rigid body and
~5 result directly from the difference between the refractive index of the
rigid body and that
of the empty display space which it sweeps out.
Also, the presence of a central mounting shaft will interfere with optical
emission.
Although this may not be a significant disadvantage in many applications, it
would be
2o desirable to have a completely optically homogenous isotropic image space
which includes
a uniform central core.
A further limitation of known volumetric displays relates to the passage of
voxel data into
the display space and hence to the appropriate voxel generation centres..
Consider a
25 control and graphics processing system which is responsible for taking
voxel data from a
host computer and passing it in the appropriate form and at the appropriate
time to the
display. The operator should, in the ideal case, be able to treat the display
space as a 3-D a
Cartesian space (using whichever form of coordinate system is appropriate).
That is - the
AMEhIDED SHEET'
rPE~~~r~
CA 02350687 2000-12-04
PCT/:~IZ99/00072
Received 07 September 2000
-9-
display space should be able to represent data points (voxels) on a regular
grid-like spacing.
Thus, there should be a direct mapping of coordinate values into image space
voxels. The
display space should therefore exhibit the properties of homogeneity and
isotropy with
respect to voxel placement. Neglecting the discrete nature of the voxel
generation
devices, the number and spatial distribution of voxels which comprise an image
should be
invariant with its position.
It is possible that these conditions may never be fully realised. However, the
present
invention provides a solution which is expected to maximise these display
space metrics. A
to further benefit offered by the present invention is an increase in the
number of voxels
which may be illuminated within each image refresh. This is best described by
examining
the limitations to various current systems.
Voxel data originating from the host computer and suitably processed by the
graphics
t5 processing system passes to the display hardware via a number of data
pathways. It can
therefore be considered that the data pathways act as the interface between
the graphics
processing hardware and the voxel activation devices. Considering for example,
a swept
volume-directed beam system employing rotational motion. In such a
configuration, each
beam deflection system is responsible for sequential voxel activation. Several
beams may
20 be required in order to allow voxels to be generated throughout the display
space (see
International Application No. PCT/NZ93/00083 for example). The degree of
parallelism
exhibited by such a display is therefore no greater than (and perhaps less
than) the number
of beam deflection mechanisms. Increasing the number of beam deflection
mechanisms
will increase the degree of parallelism. However, the degree of parallelism is
inherently
25 limited in such devices. Since, as noted above, increasing the number of
beams/beam
deflection mechanisms leads to calibration and registration problems.
AMENDED SHEET
IPE~/Al.~
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
io -
In systems employing a 3-D matrix of static elements, the degree of
parallelism may be
reduced byadopting a row and column addressing technique. The number of voxel
generators is unacceptably large and is coupled to the dimensions of the
display space.
Similarly, the extent of the connectivity also increases.
In the case of a display employing the rotational motion of a planar matrix
element, such as
that described in US Patent No. 4,160,973 (Berlin), the problems in data
transmission
reside in writing sufficient data to the rotational planar LED array per image
refresh period.
The Berlin device permits parallel data transfer from the rotating display
electronics to the
to LED array. However, the rotating display electronics memory is updated via
a serial optical
link. Therefore, it would appear that an inherent upper limit to the rate of
data transfer wilt
limit the number of voxels which may be modified during each image refresh
period. The
Berlin system allows for incremental updating of the image (i.e. a percentage
of the total
number of available voxels). However, it is likely that high definition images
involving gross
~ 5 movement and covering large regions of the image space would exceed the
data transfer
capabilities of such a data transfer link. That is - while the graphics
processing system may
exhibit any required degree of parallelism, and while the 2-D optoelectronic
array may be
configured to exhibit parallelism, the data transfer link is inherently serial
and will ultimately
limit the throughput of information into the display system.
A further weakness associated with a technology employing the rotational
motion of a 2-D
optoelectronic matrix such as that described in the previous paragraph resides
in the
varying workload of the optoelectronic components. The length of the path
(track) swept
out by each of the optoelectronic devices is proportional to their radial
distance from the
axis of rotation. Given that each of these devices is responsible for voxel
generation along
the circumference of a circle at intervals (tracks) which are constant
throughout the display
space, it follows that the workload of each device is in direct proportion to
its distance
from the axis of rotation. Therefore, a device which is closer to the axis of
rotation is
AMENDED SHEEP
~~'r ::, ro,a 1
CA 02350687 2000-12-04
PCT/NZ99/00072
rieLeived 07 September 2000
-ii
responsible for the generation of fewer voxels per rotation (image refresh
periodl than one
wi~ich lies at a greater radial distance from the axis of rotation. The
devices at the greatest
distance from the axis of rotation will have the greatest potential workload
and those
closest to the rotational axis the smallest. Thus, while the bandwidth of the
interconnection between the multiplexing electronics and the optoelectronic
components
must be the same for all elements in the array, the total amount of data which
may
potentially be passed to each element during each image refresh period is will
vary with the
radial distance of each element from the axis of rotation. A further problem
is the
increasing voxel density at small radii. The optical emission devices are of
finite size and
l0 unacceptable densities can occur close to the axis of rotation.
In order to permit grey scale and colour mixing it is necessary to pass to
each element an
analogue signal the magnitude of which will determine the intensity of an
illuminated voxel.
The data processed by the graphics processing hardware is generally in digital
form. As
15 the number of data pathways out of the graphics processing hardware is
increased in order
to facilitate parallel voxel generation within the display system, it follows
that there is
usually a need for a greater number of digital to analogue converters (DAC'sl.
Normally
there would be at least one such DAC per data pathway. For example, in the
beam
addressed system described above the number of DAC's would normally correspond
to
20 three times the number of beams able to address the display simultaneously.
The need for
three DAC's per data pathway reflects the general need in the case of a beam
addressed
system to provide two DAC's responsible for beam deflection and one DAC for
grey scale.
Large numbers of DAC's considerably increase the cost, size and power
consumption of
the graphics processing hardware. There is a further need to pass a large
volume of data
25 across a rotating intaface. Serial methods described in the content of the
Berlin device are
not considered capable of effecting the anticipated transfer rates.
AMEND5ED SHEET
~~~~'11.~
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
-12-
It is an object of the present invention to provide a high definition three-
dimensional
volumetric display which overcomes a number of the abovementioned
disadvantages by
providing a display with an image volume which exhibits little or no obscuring
aberration
effects, which has data transfer characteristics which allows serial and
parallel data
multiplexing in order to match the control system to the display electronics,
allows a
reduction in the number of DAC's or which at least provides the public with a
useful
choice.
Disclosure of the Invention
to
In one aspect the present invention a three-dimensional volumetric display
comprising:
a solid, substantially optically transparent support structure adapted to be
rotated
around an axis of rotation by a drive means; and
I5 an array of optical sources dispersed throughout said support structure in
such a
manner that the optical sources do not lie in any single plane, the optical
sources being
electrically connected to a connection means which is adapted to receive
control signals
from display circuitry to drive the optical sources according to the control
signals wherein,
in use, the drive means and the optical sources may be driven in such a manner
that a
20 three dimensional image may be generated within the support structure.
The optical sources arm preferat~ly dispersed substantially uniformly
throughout the display
structure. The support structure preferably occupies substantially the entire
display
volume within which an image may be generated. The optical sources may be
dispersed
25 substantially throughout the display structure.
Preferably, as the display structure rotates, the optical sources trace out
circular tracks
around the axis of rotation wherein the optical sources are dispersed at
locations on the
AMENDED SHEF_T
,:..
CA 02350687 2000-12-04
PCT/NZ99/00072
received u7 September 2000
circular tracks in such a way that the optical sources spatial density is
substantially
uniform throughout the display structure.
The array of optical sources may be calibrated by observing the illumination
of each of the
s activation sources in such a way and at such a time so that their relative
locations on their
corresponding tracks may be determined.
In a preferred embodiment, the optical sources are located on planes
perpendicular to the
axis of rotation of the support structure, whereby for any single plane, the
optical sources
o are located relative to one another on corresponding tracks so that the
optical sources form
a spiral when viewed along the axis of rotation, the spiral traversing a path
outwards from
the axis of rotation to an outer edge of the support structure.
Preferably the planes are arranged one on top of the other and sequententially
displaced
~5 around the axis of rotation in relation to one another so that the optical
sources lie on the
surface of a helix.
In an alternative embodiment the number of optical sources on a track may vary
as a
20 function of the displacement from the axis of rotation, alternatively the
optical sources on
a single track may be selected to effect the emission of selected colours.
The support structure may be in the shape of right cylinder, a sphere or other
three-
dimensional shape suitably adapted to the viewing application for which it is
intended.
The support structure may be embedded in a stationary secondary support
structure, the
secondary support structure having substantially the same refractive index as
the support
structure, the arrangement being such that there is as small a gap as possible
between the
AMENDEr'~ ~NE~...
Ip~,:, ~;~,
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
- 14 -
support and secondary structure while allowing relative movement between the
support
and secondary structure whereby the construction of the secondary support
structure is
such that the visual isotropy of the generated three-dimensional image is
enhanced.
The optical sources may be light emitting diodes, optoelectronics materials
stimulated by
non-visible radiation, devices which may have a two state (on/off) operating
state.
Preferably the array of optical sources are connected to display electronics
in such a way
that they may be activated by means of parallel data which is multiplexed,
where the
l0 degree of serial/parallel mixing is controllable and dependent upon
matching the display to
control electronics.
Preferably the operation of the display electronics should be sufficiently
parallel so as to
ensure that the serial data links do not restrict the system bandwidth.
IS
Preferably, the optical sources are grouped into regions which are each driven
in parallel
where those regions as a whole are driven serially, each region having a
number of high
workload and low workload in order that each data pathway has substantially
the same
potential data transfer capacity, wherein the workload is proportional to the
radius of the
20 track on which the optical sources) is located.
Preferably the rotating components of the volumetic display are coupled to
stationary
components by means of a multichannel optical data link.
25 The support structure may incorporate a reciprocating core incorporating a
reduced number
of optical sources the reciprocating motion when coupled with rotational
motion producing
a plurality of virtual tracks.
AMENDED SHEET
IPEq/Atl
CA 02350687 2000-12-04
, PCT/NZ99/00072
15 Received 29 September 2000
According to a further aspect of the invention there is provided a display
including:
a support structure supporting a plurality of optical sources dispersed
throughout
said support structure in such a manner that the optical sources do not lie in
any
single plane;
rotation means for rotating the support structure about the axis of rotation;
translation means for translating the support structure;
drive circuitry for activating the optical sources; and
control means for actuating the rotation means, translation means and drive
means in such a manner that optical sources may be activated at locations on a
plurality
of tracks through which the optical sources pass so as to generate image
elements at
desired locations on a plurality of such tracks.
Brief Description of the Drawings
The present invention will now be described by way of example and with
reference to the
drawings in which:
Ficture 1 illustrates a simplified dispersed optical source display having
optical
sources located on planar spiral patterns;
Figure 2 illustrates a top view of the dispersed optical source display of
Figure
1;
Figure 3 illustrates a simplified schematic perspective view of a single plane
of
a dispersed optical source display;
Fi4ure 4 illustrates a top view of the display device shown in figure 4;
Figure 5 illustrates detail of a calibration method appropriate for a display
device of figure 4;
Fi4ure 6 illustrates a schematic of an N x M planar emitter array; and
AME(J~D SN~GT
IREAlfiU
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
- 16
Figure 7 illustrates a representation of the dispersal of optical emitters
through a
display volume.
s The present invention may be classified as a moving matrix type volumetric
display. The
closest prior art of which the applicant is aware is US patent No. 4,160,973
(Berlinl.
As noted above, ideally a three-dimensional volumetric display will produce
images which
faithfully reproduce the object concerned without the display system producing
optical
to artefacts resulting from the system itself.
This requires that the voxel refractive index be the same or similar as the
refractive index
of the volume in which the image is to be generated. The expression "the same"
is
understood to mean that the light emitted from the image encounters
substantially identical
15 optical conditions when traversing the display device. Therefore, in the
limiting case, the
physical apparatus which produces the light emissions (by means of illuminated
optical
sources) within the image generation volume is completely transparent and does
not
interfere with optical emissions produced by voxels within the swept volume.
2o Therefore, the ideal image generation volume is a transparent solid with a
constant
refractive index throughout its volume. This contrasts with swept volume
scanned beam
devices (such as an electron beam impinging on a phosphor screenl. Here, the
glass
structure supporting the phosphor has a different refractive index from the
rest introduces
refractive aberrations when the screen is at an oblique angle to the viewer.
Such
25 aberrations are caused both by refraction distorting light from voxels as
well as a total
internal reflection within the glass.
AMENDED S~iEET
IFrq~~l I
CA 02350687 2000-12-04
PCT/NZ99/00072
I~eccived 07 September 2000
-t7_
To overcome these problems, a planar matrix display could be embedded in a
support
structure having the same refractive index as the matrix display. This would
avoid
refraction aberrations. However, obscuration aberrations would still be a
problem because
of the highly ordered arrangement of optical emission devices and their
associated
electrical connections.
A solution to this is to disperse the emission devices through the display
volume.
Figure 1 illustrates a simplified example of a display device in which the
optical sources are
o dispersed. Referring to figures 1 and 2, a cylindrical transparent support
structure 20
contains planar arrays 15a, 15b 15c of optical sources 21 a, 21 b etc. For
clarity, only
three planes are shown spaced widely apart. The planar array construction is
not intended
to limit the scope of the invention and merely represents an implementation
which is for
convenience in manufacture and construction. Accordingly, the display can be
generalised
I5 to a rotating volume containing distributed optical sources. The dispersion
may be
described as random insofar as randomness is interpreted to exclude
statistical clustering.
To illustrate the principle of dispersion of the optical emitters, Figure 7
illustrates a diagram
of emitter layout in a cylindrical display system. The x and y axes correspond
to a section
through the cylinder with the x axis being the radial direction and the y axis
the height.
2o These axes are different from those shown in the other figures. The z axis
corresponds to
the track length which is proportional to the radius measured from the axis of
rotation. The
plane L corresponds to the track length as a function of radius and it can
clearly be seen
that to maintain emitter density toward the axis, and given the finite size of
the emitters,
there will be congestion problems near the centre. This will be discussed
further below.
25 Referring again to Figure 7, the matrix array in Solomon would be
represented by the plane
of emitter locations T (coplanar with the x and y axes) and the present
invention emitter
locations as, for example, the artside of the display, as being along the
track path indicated
by the letter S.
/;MEhIDED SHEET
IP~AIAU
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
_ Is _
In the embodiment shown in figure 1 , the optical sources are arranged in a
path (lying in a
plane) traversing a spiral from the axis of rotation of the display, to an
outer edge. Each of
the planes, upon which the optical sources are located, are rotationally
displaced with
respect to one another (around the axis of rotation of the support structurel.
Therefore,
there is no preferred alignment orientation when the display volume is viewed
from any
direction.
A suitable dispersal regime may be where optical emitters are located at
different locations
o along a track. While being constrained on a circular track, the emitter
locations can be
dispersed so that the emitter density is uniform over the whole volume of the
display
system. In this way there would be no preferred direction which would result
in
obscuration of one emitter by another. In the specific case of emitters being
set out in
spirals on adjacent planes, if the planes are sequentially displaced, the
resulting emitter
I S dispersion pattern will reduce (but not eliminate) obscuration. The
preferred method is
where the location of the emitters on the tracks (ie; their rest locations) is
dispersed in
such a manner that the emitters are spread evenly though the display.
Colour or gray scale capability may be included in the present implementation
by locating
20 (for example) three emitters on a track and timing their emission so that
appropriate
colours are produced.
The alternative, and preferred version, is that shown in figure 3. In this
embodiment, the
optical sources are dispersed on the planes ~rvhich make up the support
structure. The
25 planes are sandwiched one on top of the other. Also, the optical sources
are a uniform
refractive index volume thus reducing the effect of distributing the
wire/connection array to
avoid localised high density groups of related components/structure.
AMENDED SNC~
IPEA/AI)
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
- 19
There may be 'tensing' effects visible in the image due to the refractive
index of the
support structure when compared to air. However, under certain circumstances
this may
enhance the view of the image by magnification of image detail and a reduction
in image
interference caused by real world objects located on the opposite side of the
display space.
This would be most apparent in a spherical display space. In cases where the
airldisplay
relative refractive index produces unacceptable distortion, it may be possible
to embed the
rotating display support inside a body of similar, if not identical,
refractive index. For
example, a rotating spherical display volume device may be embedded in a
static cubic
block. The rotating and static components would be optically coupled so that
there is little
io or no refractive gradient or interface between the two. This construction
may be suitable
for cases Where the application requires a specifically shaped display unit.
At the centre,
the voxel density will be high as the paths traversed by the optical sources
will be shorter
than those towards the edge. The voxel density is therefore ideally reduced by
decreasing
the number of optical sources in the core region. This congestion will be
discussed further
I S below.
Referring again to figure 3, a single plane 24 is shown having five optical
sources 25a-e.
Of course, this number is illustrative only and a practical display will
incorporate a large
number of optical sources. As the plane is rotated around the z axis (the z
axis being
2o perpendicular to the x and y axesl, each of the optical sources will trace
out concentric
circular paths or tracks 26. Optical sources are activated in synchronisation
with the
rotational motion of the support structure and voxels may therefore be
positioned in the
display space as required. The optical sources need to be illuminated at
precisely known
positions in space in order to produce an image. In the present case, the
positions of the
25 optical sources must be calibrated with reference to the rotational
position of the support
structure. In the case of the optical emitters located in a known dispersal
regime on the
tracks, calibration may be performed using the known locations of the optical
sources in
each plane. Alternatively, all optical sources may be simultaneously
momentarily activated
AMEfVpED SHEEP
~ c .=,q,A ~ ~
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
-zo
and normalised or rotated onto a plane passing through the axis of rotation of
the support
structure (or any other reference plane). Such a procedure is schematically
shown in figure
whereby a random array of optical sources 25a-a are advanced or delayed onto
corresponding locations 25a'-e'on the Y axis. This may be done by using a
video camera
imaging the display from above and using a reference optical source located on
the
reference plane for other known location) and delaying the activation of other
optical
sources in relation to this source. The unknown optical sources delay or
advances their
activation so that they lie on the Y axis. Other calibration methods may be
feasible and it
is envisaged that the implementation of the synchronisation step would be
within the
I o scope of one skilled in the art.
i ne cpucal sources may be thought of as light emitting means powered and
activated by
electrical connections using wires. As a consequence of the dispersion of the
optical
sources within the display volume, the wires will also be dispersed and thus
the display
IS device will not exhibit any preferential obscuring aberrations. It is
feasible that electrical
connections may be effected by means of thin, virtually invisible, metallic
fibres encased in
a uniform refractive index support medium.
Referring to the data flow constraints, if we consider a planar array of N x M
optical
20 sources, the total number of connections is generally 2NM (assuming two
connections per
optical source or NM + 1 for a common groundl. These connections can be
addressed in
parallel as they are completely independent. This would be effective if the
data supply
mechanism was able to carry NM data pathways at any one time. In this model,
the
connection between each graphics path line channel and each optical source
would take
25 the form of a one to one mapping. However, with the inclusion of grey scale
capability or
a display having a practical resolution, each optical source would require the
application of
an analogue voltage level and for systems having even a moderate voxel
activation
capacity, the size of the driving circuitry would be very large. For example
for a display 30
AMENDED SHEET
IPEA/AU
CA 02350687 2000-12-04
PCT/NZ99/00072
' ~> ~~-~~~ ~ed 07 September 2000
aw.w v
-21
cm high and 1 5 cm in radius, at 1 mm element separation there would be 45,000
elements. That is - 45,000 pathways without serial links. For this reason, it
is highly
desirable to incorporate a partially sequential data transfer method. To this
end, the highly
parallel array may be interconnected in a series of planes. This can be
thought of as sub
dividing the N x M display into a sequence of individually parallel connection
arrays but
with the subdivided sections being multiplexed. This would result in the
number of
interconnections being significantly reduced. The effect of such a reduction
is the inclusion
of serial activation of the elements within each 2D segment. However, it can
be seen that
the degree of parallelism can be adjusted to match the control system. This
adjustability
to in the design of the volumetric display allows optimisation of the
particular display to
achieve a required level of definition, voxel spacing and subsequent image
quality coupled
with the characteristics of the control system. It is considered that the
inclusion of sucn
adaptable multiplexing represents a highly desirable attribute of the present
novel system.
The above multiplexing technique can be readily adapted to the dispersed
optical element
15 display.
As noted above, the size of a track swept out by an optoelectronic element is
proportional
to the radial distance, R, from the axis of rotation. As a consequence for
elements
approaching the axis of rotation i.e. for elements with decreasing values of
R, there is less
20 space to disperse these elements within the display space. Further, the
potential number
of voxels which each element may create during each revolution of the display
space is
directly proportional to the radial distance, R, of the element from the axis
of rotation.
Accordingly, as the axis of rotation is approached, there is less space for
element dispersal
25 and a decrease in the potential workload for each element. This is where it
can be seen
that for small values of R and finite optical element size, there will be
congestion. The lack
of space for dispersal may cause optical problems (e.g. image distortion)
should the
refractive index of each element fail to properly match that of the display
space material.
AMENDED $I-I!EET
iPEA/AU
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
-22
.,
This would be particularly so if an observer views an image component located
on the
opposite side of the axis of rotation to that of the observer. This problem
may be
minimised by reducing the density (i.e. number) of optoelectronic elements for
smaller
values of R.
Simply reducing the density of elements will affect the homogeneity of the
display space.
To circumvent this effect, a translational (or reciprocating) motion may be
superimposed
upon the rotational motion of the display space towards its centre. Therefore,
that the
display space material may consist of an outer rotational component and an
inner
cylindrical core.
These two materials are in close proximity and are matched in terms of the
refractive
index. Both rotate the outer core at speeds above the flicker fusion
frequency. The inner
cylindrical core also rotates, however its speed of rotation may differ from
that of the outer
l5 material.
Further, the inner cylindrical core has a vibrational motion (parallel to the
axis of rotation)
superimposed upon it. The rotational speed of the inner core will generally be
greater than
that of the outer display space material. The amplitude of the vibrational
motion permits
20 each optoelectronic element to pass through a number of track positions in
the Y direction.
For instance, should the amplitude of the vibrational motion be sufficient to
allow each
optoelectronic element to pass through two track positions in the Y direction,
then the
number of elements required in the Y direction may be halved. This introduces
the concept
of virtual tracks. A virtual track may be defined as a track within the
display space which
25 does not contain its own optoelectronic element from a different position
within the display
space for voxel generation within the virtual track.
AMENDED SHEET
IPEA/AU
CA 02350687 2000-12-04
PCT/NZ99/00072
ri Y~_ei ved 07 September 2000
_ 2; _
If, for example, an inner core radius of 5 cm is considered then an outermost
element upon
this inner core would in the case of no vibrational motion sweep out a
distance of
approximately 30 cm per display space revolution, and given a voxel spacing of
1 mm,
would be responsible for generating 300 potential voxels. If the density of
elements is
halved by the imposition of vibrational motion it would be necessary to have a
vibrational
motion of at least 300 times the rotational frequency of the display space.
The vibrational
amplitude being in the order of 3-4 mm for an inter-voxel spacing in the Y
direction of
approximately 1 mm. It is noted that by increasing the rotational frequency of
the inner
core it is possible to reduce the vibrational frequency.
The vibrational motion could be effected by various means such as by
electromechanical
stimulation or could be derived from the difference in rotational speed
between the outer
display space and inner core. Increasing the amplitude of motion would permit
a further
reduction in the number of optoelectronic elements required in the Y
direction, i.e. parallel
15 to the axis of rotation.
The display space material as a whole may also be flexible and highly
controllable thus
permitting the vibrational motion to extend from the outside to the inside at
an ever
increasing amplitude. This avoids the discontinuity caused by the outer core
and the inner
20 vibrating core envisaged above.
As noted above, the present invention provides a method and apparatus for
parallel voxel
generation and permits a large number of voxels to be generated during each
image refresh
or display space revolution. An important aspect of the implementation upon
coupling the
25 voxel data from the graphics engine or control system into the display
space for passage to
the electro-optical elements. Should this link simply take the form of a
serial optical
connection, then it is likely, given the large number of elements which are
contained in the
display space, that only a fraction of them could be addressed/updated during
each display
4AAEntp~n gHEET
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
-24
space revolution or image refresh. That is - a serial optical link would limit
the bandwidth
of the connection between the graphics engine and the display space electro-
optical
element multiplexing system. In order to maintain a high bandwidth connection
and permit
a very large fraction of the elements to be updated (if not potentially all)
during each
display space rotation, the present invention may use a parallel optical
connection between
the static outside world and the rotating/vibrating display space.
In a preferred embodiment, this connection takes the form of two concentric
cylinders.
One of these cylinders is stationary and the other co-rotates with the display
space. Upon
l0 the surface of the stationary cylinder is a 2D array of light-emitting
devices and on the
rotating cylinder is a 2D array of optical receptors each capable of
converting light signals
received from the emitters into eiectricai signals. Both cylinders are
positioned around the
axis of rotation of the display space. The principle of operation is best
considered by
imagining oneself to be a light receptor. The physical layout of the
transmitters would
I5 cause an observer (receptor) to see a series of light sources rising and
sinking in sequence.
During certain periods no light source would be visible to the observer.
During periods
where the light source is sufficiently strong, the controlling hardware would
modulate the
light source in order to permit transmission of information from the source to
the observer.
The physical arrangement of the light sources and receptors is such as to
ensure that a
20 maximum number of light sources are able to transmit at any time to a
maximum number
of receptors. The use of a plurality of transmitters and receptors permits
parallelism in data
transfer. If a single ring of transmitters and receptors is considered (this
ring being located
around the circumference of the cylinders) then the operation may be easily
understood.
By varying parameters such as the total number of transmitters and receptors
together
25 with their spacing and any collimating system placed around each receptor,
we can vary
the degree to which parallel data transfer may take place. By duplicating
these single rings
of transmitters and receptors along the length of the two cylinders we can
duplicate the
AMENDEn SHEET
~~'~ ~ .~8.; 1
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
-25
degree of parallelism. It is noted that these rings may be physically
displaced around the
circumference.
In order to effect proper data transfer it is necessary for the controlling
hardware to direct
(route) information to the appropriate transmitters and to have a clear
knowledge of which
receptor will contain the transmitted information. The system therefore
employs an
optoelectronic synchronisation link from which routing information can be
derived and
which further provides the control system with timing information which is
necessary in
order to generate images within the display space (see International
Application
l0 PCT/NZ93/000831. The optoelectronic link takes the form of a single
transmitter and single
receptor, the transmitter being located upon the cylinder containing the 2D
array of
optoelectronic receptors and the electro-optical receiver being located upon
the cylinder
containing the 2D array of electro-optical transmitters. The transmitting
device operates
continuously and provides the receptor with a single-pulse high revolution of
the display
I5 space which corresponds to a single rotation of the cylinder equipped with
the
optoelectronic receivers. This pulse is used to obtain timing information for
voxel output
into the display space and further routing information for the passage of
information
through the optoelectronic parallel data transfer link.
20 In order to tolerate a small degree of mechanical misalignment in the
positioning of the
electro-optical transmitter and receiver arraysit is desirable to have a
degree of duplication
in the information which is passed through the parallel data transfer link and
intelligent
hardware which co-rotates with the display space and which is capable of
performing
amalgamation of incoming data, error correction and which is multiplexing of
signals to the
25 voxel elements. It is further noted that in order to achieve maximum
bandwidth through
the parallel data transfer link it is highly desirable to minimise the extent
of the night times.
These are periods in which no transmitter is able to pass signals to any
particular receptor
(this being caused by the geometrical arrangements between the two arrays and
AMENDED SHEET
~PEAIAU
CA 02350687 2000-12-04
PCT/NZ99/00072
Received 07 September 2000
-2G
corresponds to no transmitter having risen sufficiently above a receptors
horizon
(collimated horizon)). As a transmitter moves above a receptors horizon not
only must it be
aware of which receptor it is to encounter, but it must also be aware of the
time at which
it has risen sufficiently in order that a signal of sufficient amplitude can
be received by the
receptor. In one particular embodiment each transmitter will judge from
information gained
from the once-per-revolution synchronisation signal at what point it should
commence
transmitting information to a particular receptor and at what point it should
stop. Between
these two periods the transmitter will transmit an integral number of voxel
descriptors. By
exercising caution in terms of the amplitude of signal required by a receptor
to receive the
l0 transmitted information without error. It can generally be assured that
error correction will
not be necessary. This however reduces the useable bandwidth of each
transmitter and
receiver. The degree of parallelism exhibited by the parallel data transfer
link enables
however a sufficiently high bandwidth for the passage of information to the
display space
hardware and will enable the possibility of a high percentage if not all of
the voxel element
I5 positions to be updated during the course of each image refresh period.
As noted above, the present invention is not limited to a cylindrical support
structure and
other geometries such as spherical, conical or other rotationally symmetrical
body are to be
considered within the scope of the invention.
It can be seen that the present invention provides for a substantially
optically isotropic and
homogenious image generation volume as well as the capability of connection
regime
which may be adjusted in terms of the degree of multiplexed serial/parallel
mixing so as to
match the control system.
Either the dispersed optical source regime or other arrangements may be
particularly
adaptable to manufacture as a large number of identical planar optical source
arrays can be
manufactured from which the (in the present case) cylindrical support
structure can be
AMEPJf7ED SHEET
CA 02350687 2000-12-04 pCT/NZ99/00072
a eceived 07 September 2000
_27_
~
built up by placing one layer upon the other. Each plane may be rotated
randomly to
produce the dispersed scattering of the optical sources through the display
volume. In the
case of a spiral optical source array, offsetting adjacent planar elements
would lead to a
random helical surface sweeping out an image volume.
Where in the foregoing description reference has been made to elements or
integers having
known equivalents, then such equivalents are included as if they were
individually set
forth.
Although the invention has been described by way of example and with reference
to
particular embodiments, it is to be understood that modifications~and/or
improvements may
be made without departing from the scope of the appended claims.
IS
25
PSPEC59050
AMENDED SHEET
IpEAIAU
