aHuman
Interface Technology Laboratory
University of Washington
Box 352-142
Seattle, WA 98195
USA
BT Laboratories
Martlesham Heath
Ipswich, IP5 3RE
United Kingdom
grof@hitl.washington.edu , {jerry.bowskill, nick.dyer,
jason.morphett}@bt-sys.bt.co.uk
ABSTRACT
Two dimensional windows based interfaces may not
be appropriate for wearable computers. In this paper we draw on
established virtual reality techniques to design and evaluate
several alternate methods for information presentation in a wearable
environment. We find simple body-spatialised displays provide
benefits over traditional head-stabilised displays. Users found
the body-stabilised displays easier to use, more enjoyable and
more intuitive, and were able to perform significantly better
on a search task. Spatial audio and visual cues further enhanced
performance.
KEYWORDS
Augmented Reality, Wearable Computing, 3D Interfaces,
Spatial Information Display
INTRODUCTION
One of the broad trends emerging in advanced human-computer
interaction is the increasing portability of computing power.
Wearable computers are the next generation of portable machines.
Worn on the body they provide constant access to computing and
communications resources. However, for wearable computing to be
widely adopted there are unique interface challenges that need
to be solved. One of the most important issues is how to present
and interact with information in a wearable environment. In this
paper we apply traditional virtual reality techniques to develop
and evaluate spatialised interfaces for wearables and present
results comparing user performance with these interface to a more
traditional wearable interface.
BACKGROUND
The field of wearable computing encompasses a very
wide range of devices. In general, a wearable computer may be
defined as a computer that is subsumed into the personal space
of the user, controlled by the wearer and has both operational
and interactional constancy, i.e. is always on and always accessible
[1]. Wearables are typically composed of a belt or back pack PC,
head mounted display (HMD), wireless communications hardware and
an input device such as touchpad or chording keyboard. This configuration
has been demonstrated in a number of real world applications including
aircraft mantainence [2], navigational assistance [3] and vehicle
mechanics [4]. In such applications wearables have dramatically
improved user performance, reducing task time by half in the case
of vehicle inspection [4].
There are unique challenges in designing interfaces
for wearable computers. Although most current wearable applications
use traditional two-dimensional GUI's, these interfaces have been
optimised for desktop use and are less than ideal for the wearable
platform, both because of the nature of the tasks wearables are
used for and the unique input and output devices wearables have.
For example, wearable input devices must be able to be used with
one hand, when out of view, and at an arbitrary orientation. Previous
researchers have used speech [2], one handed twiddlers [5], half
keyboards [6] and dials [4] in wearable interfaces. Less work
has been done on the graphical interface. Wearables predominantly
have small monoscopic head mounted displays with limited resolution
and a narrow field of view. We are interested in developing interface
metaphors that are ideally suited for head mounted displays on
a wearable computer. Wearables are currently most commonly used
for data access and display so we have initially focussed on the
problem of 2D information presentation using a monoscopic display.
In this paper we show how spatial display techniques can be used
to improve the display of 2D informaiton in a wearable computing
environment.
SPATIAL INFORMATION DISPLAY
Information presentation using a head mounted display has been well studied in the virtual reality arena. Since most wearable displays are see-through or see-around displays, augmented reality interfaces are most relevant to our work. In this setting information can be presented in a combination of three ways:
Head-stabilised - information is fixed to the users viewpoint and doesn't change as the user changes viewpoint orientation or position.
Body-stabilised - information is fixed relative to the users body position and varies as the user changes viewpoint orientation, but not as they change position.
World-stabilised - information is fixed to real world locations and varies as the user changes viewpoint orientation and position.
Each of these presentation methods require increasingly
complex head tracking technologies, as shown in table 1.0. The
registration requirements also become more difficult progressing
from head to world stabilised images; no registration is required
for head stabilised images, while complex calibration techniques
are required to achieve good world stabilisation [7].
| Information Presentation | Tracking Required |
| head stabilised | None |
| body stabilised | Orientation |
| world stabilised | Position and Orientation |
Body and World stabilised information display is attractive for
a number of reasons. As Reichlen[8] demonstrates, a body-stabilised
information space can overcome the resolution limitations of head
mounted displays. In his work a user wears a head mounted display
while seated on a rotatable chair. By tracking head orientation
the user experiences a hemispherical information surround - in
effect a "hundred million pixel display". World-stabilised
information allows annotating the real world with context dependent
data and creating information enriched environments [9]. This
increases the intuitiveness of real world tasks. For example,
Rekimoto uses world-stabilised virtual tags to label parts of
real world objects [10] while researchers at the University of
North Carolina register virtual fetal ultrasound views on the
womb[11]. In general spatial information displays enable humans
to use their innate spatial abilities to retrieve and localise
information. They also allow other cues such as spatialised audio,
virtual annotations and stereopsis to aid performance.
In a wearable setting, spatial information display can be used
to overcome the resolution and field of view limitations of the
HMD and provide information overlay on the surrounding environment.
This is important because the information presented on a wearable
is often intimately linked to the users real world location and
task. Despite these advantages, most wearables only use head-stabilised
information display. A notable exception to this is the work of
Feiner et. al. [3] who have developed a wearable campus navigation
aid that displays world-stabilised virtual labels on surrounding
buildings. Although not in a wearable environment, Feiner et.
al. [12] have also demonstrated 2D head-, body- and world- stabilised
windows in an augmented reality environment. This extended their
previous work which the combined a head-stabilised virtual display
with a laptop screen to overcome the size limitations of the screen
[13].
To date there have been no usability studies showing the usefulness
of spatialised information display on a wearable computer. In
this paper we provide an empirical comparison between information
display types and compare user task performance on the same task
with different display styles. We focus on comparing body-stabilised
to head-stabilised information presentation and also examine how
audio-visual spatial cues can be added to body-stabilised spaces
to further improve performance.
A WEARABLE INFORMATION SPACE
In our work we have chosen to begin with the simplest
form of body-stabilised display; one which uses one degree of
orientational freedom to give the user the impression they are
surrounded by a virtual cylinder of information. Figure 1.0 contrasts
this with the traditional head stabilised wearable information
presentation.
Head-Stabilised Body-Stabilised
A one degree of freedom spatial display has a number of advantages:
A head mounted display allows only the portion of the information
space in it's field of view to be seen. Thus, there are two ways
the data can be viewed in a cylindrical body-stabilised space;
by rotating the information space about the users head, or tracking
the users head as they look about the space. The first requires
no additional hardware and can be done by mapping mouse, switch
or voice input to direction and angle of rotation, while the second
requires only a simple one degree of freedom tracker. The minimal
hardware requirements make cylindrical spatial information displays
particularly attractive. In this paper we compare both interactions
methods to each other and to information presented in a head stabilised
manner. The cylindrical display space also allows us to use audio
and visual spatial cues to aid performance which we describe in
the later half of the paper.
This research was conducted on a custom built 586
wearable PC with 20mb of RAM running Windows '95. A hand held
Logitech wireless radio trackball with three buttons was used
as the primary input device. The display was a pair of Virtual
i-O iglasses! converted into a monoscopic display by the removal
of the left eyepiece. The Virtual i-O head mounted display can
either be used in see-through or occluded mode, has a resolution
of 262 by 200 pixels and a 30 degree field of view. The iglasses!
also have a sourceless inertial and gyroscopic three degree of
freedom orientation tracker. Figure 2.0 shows a user wearing the
display and wearable computer.
For the purpose of our user studies, a simulated
body-stabilised information space was created by texture mapping
images to polygons placed in a cylinder around the users viewpoint.
Head-stabilised information was shown as a stack of texture maps
attached to the user's view point. The Direct3D graphics library
was used and the interface was deliberately kept simple because
the wearable has no graphics acceleration hardware. With eight
sample images the simulation ran at over 15 frames a second.
SPATIAL DISPLAY EXPERIMENTS
In the following sections we describe two experiments with our wearable interface. The first examines display performance effects and the second the benefit of spatial cues.
Expt 1: Display Performance Effects
In the first experiment we compare how easily users can find information from eight pages of data displayed in the following conditions:
A) Multiple head-stabilised pages: All the pages are stacked on top of one another so that only the top most page is visible. When the user holds the right trackball button down and rolls the trackball in the positive or negative Y direction they scroll forwards and backwards at a constant rate through the stack of pages. The pages are attached to the users viewpoint so changing head orientation has no effect on page shown.
B) Cylindrical with trackball control: Pages are spaced equally about the surface of a body-stabilised cylinder with the user at the centre. Holding the right trackball button down and rolling the trackball in the positive or negative X direction rotates the cylinder clockwise or counter clockwise at a constant rate about the users head. The head tracker is not used so head rotation has no effect on viewpoint.
C) Cylindrical with head tracking:
Pages are displayed on the surface of a cylindrical space as above.
When the right trackball button is held down, the yaw angle of
the user head motion is measured by the head tracker and used
to set the camera viewpoint rotation about the vertical axis.
Head motions in other directions (pitch and roll) have no effect
on the viewpoint.
Condition A simulates how data is displayed in most
current wearable applications. In all conditions the pages are
exactly the same size. A snap-to function was used in the cylindrical
conditions (B and C) so that when the user released the right
button the view would snap to the page taking up most of the visual
field. This was to ensure that when the user stopped manipulating
the cylindrical space they could only see one page, just as in
the head-stabilised condition.
Experimental Task
Since many wearable applications involve data display
and retrieval, the subject's task was to find which page contained
a certain target icon. Eight pages of unique graphical icons were
used with five icons shown on each page. A head-stabilised target
icon was shown in the upper right hand corner of the users display
and was visible at all times. Figure 4.0 shows a sample page with
target icon.
In all conditions the trackball buttons were used to start and
stop user interaction. When the user released the right trackball
button and clicked the middle button the target icon was compared
to those on the currently visible page. If there was a match a
new target icon was automatically displayed, otherwise the current
icon remained. A set of eight target icons were used for each
condition and time measurements were taken between button presses
to measure search time.
Twelve subjects took part in the experiment, seven males and five
females aged between 19 and 35. Some of the subjects had experience
of virtual environments, but none had used a wearable computer
before. Subjects were given a standardised test of spatial ability
[14] and all had normal natural or corrected eyesight and normal
hearing. Subjects were also all right-eye dominant, the eye covered
by the monoscopic display. To begin the experiment they were given
several minutes training with each condition until they felt comfortable
with the interface. While training, a calibration routine measured
the amount of time it took each subject to view all the pages
under each condition. A Latin squares design was then used; all
subjects experienced all three conditions, but in a different
order to minimise order effects. Three sets of eight target icons
and corresponding pages were created and each subject used the
same set of icons. Although the image sets were different, the
target icons occurred in the same order in all image sets.
For each display condition subjects were given a total of three
minutes to complete two tasks. First they were to find each of
the eight targets as quickly as possible and time to complete
each search was measured. After completing the search they were
to use the remainder of the time to remember the order of the
pages in the information space. Following each condition subjects
were tested on the workload of the task using a the NASA standardised
workload assessment battery which asked questions about mental,
physical, temporal and emotional effort during the task [15].
They were also tested on their recall of the information space
by giving them a set of paper images of the pages they had just
seen and asking them to place the pages in either a stack or cylinder
corresponding to the space they had just experienced. After completing
all three trials users were given a post experiment and asked
to rate each condition according to ease of use and understanding
of where the information was. The complete post experiment questionnaire
is shown in appendix A.
Results
Performance
The three display conditions had different amounts
of inherent system delays, for example condition C required polling
the head tracker. A simple normalisation technique was used to
produce performance values that could be compared across conditions.
The average time it took each subject to view the entire information
space under each display condition was measured during the training
and pre-experiment calibration period. After the experiment, the
subject's performance time for each condition was then divided
by this normalising factor. In calculating the average search
time, the first result from each trial was also discarded to ensure
that subjects were always starting their search from the same
point; the location of the first target object.
Users performed significantly faster in the two body-stabilised
display conditions (B and C). Table 2.0 shows the resulting original
average search times, calibration values and normalised values
for each condition. Although subjects performed quicker in the
head stabilised condition (condition A), there were less inherent
delays in this condition, shown by the calibration times. Thus
the normalised values give a better indication of relative performance.
A one factor ANOVA on the normalised result found these to be
significantly different across conditions (F = 4.88, df = (2,24),
p=0.016, Fcrit = 3.40). If the system delays in each condition
had been the same then subjects would have performed one and a
half times faster with body-stabilised information spaces that
with a traditional head-stabilised interface. There was no difference
in performance between body-stabilised conditions and no significant
correlation between spatial ability and search times or normalised
values.
| Results | |||
| Avge. Search Time | 6.07 | 8.33 | 8.06 |
| Avge. Calibration Time | 7.18 | 14.31 | 13.4 |
| Normalised Result | 0.91 | 0.59 | 0.68 |
Subjects found it easier to recall the information space in the
head-tracked condition than in the other conditions. Only two
out of the twelve page layouts for this condition were incorrect,
as opposed to five each for the non-head tracked spatial display
and the head-stabilised display. Subjects commented that it was
easy to remember the page ordering in the head tracked condition
because the pages always stayed in a fixed position with respect
to their body. In some cases they used real world objects which
could be seen through the display to help them remember the location
of the virtual pages.
Subjects felt the head tracked condition involved significantly
more physical work than the other display conditions. In response
to the question, "How much physical activity was involved?",
on a scale of 1 to 9 (1=low, 9=high) subjects gave the average
scores shown in table 5.0. As expected subjects found the head
tracked condition more physically demanding; a one factor ANOVA
finds a highly significant difference between conditions, [F =
8.70, df = (2,33), p < 0.0001, Fcrit = 3.28]. There was no
significant difference across any of the other workload survey
questions.
| Physical Work |
Users felt the spatialised conditions more enjoyable, intuitive and easier to find the target objects with. Table 6.0 summarises the average values given for the first four post experiment survey questions. In these questions subjects were asked to score conditions on how easy it was to find the target, to remember where all the information was, how enjoyable it was, and how intuitive the interface was. They were asked to score the answers on Likert scales with anchors of 1 at the negative end and 7 at the positive end. The complete survey is shown in appendix A. There was a significant difference across conditions for responses to questions on how easy is was to find the target, how enjoyable it was, and how intuitive it was. Comparing between body stabilised conditions, a one-tailed ttest finds that subjects felt it easier to find target with the trackball control than head tracking (t = 1.83, df = 16, p=0.043, tcrit = 1.75). There were no other significant differences between spatialised conditions.
| Conditions | ANOVA | |||
| Questions | A | B | C | P value |
| Find Target* | 3.6 | 5 | 4.3 | p < 0.01 |
| Remember | 3.8 | 4.7 | 4.6 | p = 0.32 |
| Enjoyable* | 3.3 | 4.8 | 4.4 | p<0.05 |
| Intuitive* | 4.3 | 4.9 | 5.5 | p<0.025 |
Subjects were also asked to rank each condition in order for the
same questions, where the best condition was ranked first and
the worst last. Subjects rankings were significantly different
across conditions for how easy it was to find the target, whether
they understood where all the information was, and how intuitive
the interface was. In all cases subjects ranked the spatialised
conditions (B and C) better than the head-stabilised condition
(A). Table 7.0 summaries the average rankings, the associated
Kruskal-Wallis scores (K values) and significance levels.
| Rank
Questions | |||||
| Easiest* | 2.75 | 1.5 | 1.75 | 10.5 | p<0.01 |
| Liked Best | 2.50 | 1.75 | 1.75 | 4.5 | p<0.20 |
| Understanding* | 2.97 | 1.75 | 1.33 | 16.2 | p<0.001 |
| Intuitive* | 2.67 | 2.08 | 1.25 | 12.2 | p<0.005 |
As with the first set of questions,
subjects found the head-tracked condition most intuitive. They
also felt the body-stabilized conditions gave a better understanding
of where all the information was in the display space.
In this first experiment we explored user performance
on a search task within head-stabilised and body-stabilised information
displays. Users performed better in the body-stabilised conditions
(taking system delays into account), and also perceived that it
was easier to find the target information in such conditions.
This performance improvement happened because users had a better
understanding of where pages were and also could see upcoming
pages as they rotated their viewpoint.
Although there was no performance difference between
body-stabilised conditions, many users commented that they found
the head-tracked condition more intuitive and natural to use.
Head tracking was particularly valuable for recall because users
could remember where the information was located relative to their
bodies. Several users commented on how the head tracking allowed
them to associate pages with real world objects. However users
found it more physically demanding and some commented on the social
acceptability of using head motions in public spaces.
An advantage of using spatialised displays is that additional audio and visual spatial cues can be presented to aid performance [16], [17], [18]. In this second experiment we examined how spatial cues could affect performance in a head-tracked body-stabilised display. In addition to the head tracker body-stabilised condition (condition A) described in experiment 1, three other conditions were tested:
B) Spatialised audio: A head tracked cylindrical body-stabilised display with a three dimensional spatialised audio cue played at the location of the target page. The audio cue consisted of a sample of white noise. The equal frequency distribution of white noise makes it easy to localise [19].
C) Visual cues: A head tracked cylindrical body-stabilised display with head-stabilised arrows overlaid on the users field of view to show them which way they should turn their heads. When the target is closer in the clockwise direction, the right arrow is shown, and the left when it is closer in the counter clockwise direction. A square between the arrows changes colour when the user is looking directly at the target. Figure 5.0 shows the visual cues.
D) Visual cues and spatialised audio: A head tracked cylindrical body-stabilised display with the addition of both head-stabilised arrows and spatialised audio cues described above.
Head tracking was used in all four conditions, but
the same spatial cues could have been applied to a non-head tracked
body-stabilised display such as condition B in the last experiment.
Real time audio spatialisation was performed entirely on the wearable's CPU, causing a significant drop in graphics performance. To remove frame rate effects, spatialised audio was played in all conditions, but the head mount speakers were disconnected for the non-audio conditions. This ensured a constant frame rate across all conditions.
The same set of subjects used in the first experiment
was used in the second and the task was the same, although new
sets of target icons and pages were used. Once again, after the
entire experiment they answered the same survey questions listed
in appendix A, modified to have four conditions for each question.
Subjects were not asked to recreate the information space as in
the first experiment.
Results
Adding spatial cues significantly aided performance. Figure 6.0 shows the average search times are 35 percent faster with each of the audio and visual cue conditions than with no additional cues. A one-factor ANOVA finds a significant difference between conditions (F = 8.05, df = (3,44), p < 0.0001, Fcrit = 2.81), but there was no difference in performance between spatial cues (conditions B, C and D), (F = 0.03, df = (2,44), p =0.96, Fcrit = 3.28). Subjects took the same amount of time regardless of the spatial cue used.
There was a significant correlation between spatial ability and
performance when using audio-only cues (R = -0.61, df = 12, p<0.05).
Subjects with higher spatial ability completed the task in less
time. However there was no correlation between spatial ability
and performance on any of the other conditions. This may be because
a spatial audio cue requires the user to form an accurate mental
model of the information space relative to the audio source. A
visual annotation provides a more immediate cue which requires
fewer mental spatial manipulations.
Users felt that spatial cues made significantly easier to find
the target, and made the interface more intuitive to use. Table
8.0 summarises the average scores from the first four questions
of the post experiment survey and the corresponding one factor
ANOVA values. For the question on how easy it was to find the
target, there was also a significant difference between spatial
cue conditions, (F = 4.13, df = (2,33), p = 0.025, Fcrit = 32.8).
Subjects felt it easier to find the target when there were spatial
audio cues, even though there was no difference in performance
between cue conditions.
| Cue Conditions | ANOVA
P value | ||||
| Questions | A | B | C | D | |
| Find Target* | 3.3 | 5.4 | 4.5 | 5.8 | p<0.001 |
| Remember | 4.0 | 3.3 | 3.1 | 3.2 | p = 0.22 |
| Enjoyable | 3.8 | 4.8 | 4.6 | 4.9 | p = 0.16 |
| Intuitive* | 4.1 | 5.3 | 4.5 | 5.4 | p < 0.05 |
Subjects also ranked the cue conditions in order for the same
questions. Rankings were only significantly different across conditions
for the question of how easy it was to find the target. In all
cases subjects ranked the spatialised audio conditions better
than the other conditions, but not significantly so. Table 8.0
summarises the average rankings and the associated Kruskal-Wallis
K scores and significance levels.
|
| |||||
| Easiest* | 3.3 | 2.1 | 2.7 | 1.9 | ||
| Liked Best | 3.1 | 2.1 | 2.8 | 2.1 | ||
| Understanding | 2.5 | 2.2 | 2.7 | 2.6 | ||
| Intuitive | 3.3 | 1.9 | 2.7 | 2.2 | ||
As expected spatial cues significantly helped task
performance. More interestingly there was no difference in performance
despite the dissimilar nature of the cues. The audio cues gave
absolute target location information, telling the user which
direction they needed to rotate their head and by how much. In
contrast, the visual annotations only gave relative information
about target location, showing the user which way they needed
to rotate and when they had arrived at the target page. Audio
cues also rely on a different sense than the visual cues, and
some users commented on the extra visual overloading that the
visual cues caused. This may explain why users felt they performed
better in the conditions using audio cues (B and D). Several users
also mentioned that they found it difficult when both visual and
auditory cues were used together and often concentrated on only
one of the cues.
However, the addition of spatial cues didn't increase
the understanding that users had of the information space. Users
commented on how they attended to the spatial cues rather than
the pages when searching for targets, affecting knowledge of
the space.
CONCLUSIONS
In this paper we addressed the problem of wearable
information displays and have shown how even simple spatialised
displays provide benefits over traditional head-stabilised displays.
Users found the body-stabilised displays easier to use, more enjoyable
and more intuitive, and were able to perform significantly better
on a search task. The fact that there was no difference in results
between the head-tracked and non-head tracked methods of viewing
the display space imply that these benefits come from the spatialisation
of the information itself and can be achieved with no additional
head tracking hardware. However user preferences and information
recall results suggest that there may be types of spatial interfaces
and tasks where head tracking is desirable. This warrants further
investigation.
Adding spatial cues to the body-stabilised display
dramatically improved performance. Audio or visual spatial cues
both give the same performance benefit, although spatialised audio
caused a significant graphics performance decrease due to increased
CPU load, suggesting that visual cues may be more practical for
current wearable applications.
These results are only the first in a series of explorations
on wearable information displays. In the future we plan to implement
some real applications in a body-stabilised space and look at
long term use. We will also explore spaces with additional degrees
of freedom and different forms of visual and audio cues, including
adding absolute information to visual cues and varying sound sources
for audio cues. Finally, we will investigate how visual and audio
cues can be combined in a more intuitive manner, such as using
audio for coarse peripheral navigation and visual cues for selection
among objects in the field of view.
[1] Mann, S. Smart Clothing: The "Wearable Computer" and WearCam. Personal Technologies, Vol. 1, No. 1, March 1997, Springer-Verlag.
[2] Espisito, C. Wearable Computers: Field-Test Results and System Design Guidelines. In Proceedings Interact '97.
[3] Feiner, S., MacIntyre, B. Hˆllerer, T. A Touring Machine: Prototyping 3D Mobile Augmented Reality Systems for Exploring the Urban Environment. To appear in Proceedings of the International Symposium on Wearable Computers, Cambridge, MA, October 13-14, 1997.
[4] Bass, L., Kasabach, C., Martin, R., Siewiorek, D., Smailagic, A., Stivoric, J. The Design of a Wearable Computer. In Proceedings of CHI '97, Atlanta, Georgia. March 1997, pp.139-146. New York: ACM.
[5] Thad Starner, Personal Communication 1997.
[6] Matias, E., MacKenzie, S., Buxton, W. A Wearable Computer for Use in Microgravity Space and Other Non-Desktop Environments. Companion of the CHI '96 Conference on Human Factors in Computing Systems, pp. 69-70. New York: ACM.
[7] Azuma, R., Bishop, G. Improving Static and Dynamic Registration in an Optical See-Through HMD. Proceedings of SIGGRAPH '94 (Orlando, FL, 24-29 July 1994), Computer Graphics, Annual Conference Series, 1994, pp. 197-204
[8] Reichlen, B. SparcChair: One Hundred Million Pixel Display. In Proc. IEEE VRAIS '93. Seattle WA, September 18-22, 1993, pp. 300-307.
[9] Bowskill, J., Morphett, J., Downie, J. A Taxonomy for Enhanced Reality Systems. Submitted to the International Symposium on Wearable Computing 1997.
[10] Rekimoto, J., Nagao, K. The World through the Computer: Computer Augmented Interaction with Real World Environments. In Proceedings of User Interface Software and Technology '95. (UIST '95), November 1995, New York: ACM.
[11] Bajura, M., Fuchs, H., Ohbuchi, R. Merging Virtual Reality with the Real World: Seeing Ultrasound Imagery within the Patient. In Proceedings of SIGGRAPH '92. Computer Graphics 26, no. 2 (July 1992), pp. 203-210.
[12] Feiner, S., MacIntyre, B., Haupt, M., Solomon, E. Windows on the World, 2D Windows for 3D Augmented Reality. Proc. UIST '93 (ACM Symposium on User Interface Software and Technology), Atlanta GA, November 3-5, 1993, pp. 145-155.
[13] Feiner, S., Shamash, A. Hybird User Interfaces: Breeding Virtual Bigger Interfaces for Physically Smaller Computers. In Proc. UIST '91 (ACM Symposium on User Interface Software and Technology), Hilton Head SC, November 11-13, 1991, pp. 9-17.
[14] Cube Comparisons Test, In Standard Ability Tests, Educational Testing Service, Princeton, New Jersey, 1976.
[15] Workload Assessment Battery, NASA Johnson Space Centre, Houston, 1989.
[16] Cohen, M. Integrating Graphical and Audio Windows. Presence: Teleoperators and Virtual Environments, Vol. 1, pp468-481.
[17] Fadden, D., Braune, R., Wiedemann, J. Spatial Displays as a Means to Increase Pilot Situational Awareness. In Pictorial Communication in Virtual and Real Environments, edited by S. Ellis, Taylor and Francis, London, 1991, pp.172-181.
[18] Burraston, D., Hollier, M., Hawksford, M. Limitations of Dynamically Controlling the Listening Position in a 3D Ambisonic Environment. In Proceedings of the 102nd Convention of the Audio Engineering Society, March 22nd -25th, Munich, Germany, 1997.
[19] Alton-Everest, F. The Master handbook of
Acoustics, 3rd Edition, McGraw-Hill, New York,
1994.
APPENDIX A: POST EXPERIMENT SURVEY
After the two experiments subject were given the
following post experiment survey.
For experiment two the conditions were modified to:
A - head tracked information display
B - head tracked information display with audio cues
C - head tracked information display with visual cues
D - head tracked information display with audio and visual cues.
The questions were also modified to score answers for each of the four conditions.
You have just finished an experiment on wearable information displays with three conditions:
A - head stabilised information display
B - cylindrical information display with mouse input
C - cylindrical information display with head tracking
For each of these conditions please answer the following questions:
1) How easy was it to find the target?
1 2 3 4 5 6 7
1=not very easy 7=very easy
For the head stabilised condition (A):
For the cylindrical condition with mouse input (B):
For the head tracked condition (C):
2) How easy was it to remember where all the information was in the information space?
1 2 3 4 5 6 7
1=not very easy 7=very easy
For the head stabilised condition (A):
For the cylindrical condition with mouse input (B):
For the head tracked condition (C):
3) How enjoyable was this condition?
1 2 3 4 5 6 7
1=not very enjoyable 7=very enjoyable
For the head stabilised condition (A):
For the cylindrical condition with mouse input (B):
For the head tracked condition (C):
4) How intuitive was the interface to use?
1 2 3 4 5 6 7
1=not very intuitive 7=very intuitive
For the head stabilised condition (A):
For the cylindrical condition with mouse input (B):
For the head tracked condition (C):
PART B
For the following questions you will be asked to rank all the conditions in order on a scale of one to three and give a brief explanation for the ranking. The three conditions were:
A - head stabilised information display
B - cylindrical information display with mouse input
C - cylindrical information display with head tracking
1) Which condition was easiest to find target (1 = easiest, 3 = hardest)
A: B: C:
2) Which condition did you like the best (1 = most enjoyable, 3 = least enjoyable)
A: B: C:
3) Which condition did you feel like you had the most understanding of where all the information was (1 = most understanding, 3 = least understanding)
A: B: C:
4) Which condition had most intuitive interface (1 = most intuitive, 3 = least intuitive)
A: B: C: