directing attention in augmented realitydmprojects.lmc.gatech.edu/~lsand6/lsandthesis...despite the...
TRANSCRIPT
DIRECTING ATTENTION IN AUGMENTED REALITY
A Dissertation Presented to
The Academic Faculty
by
Logan Sand
In Partial Fulfillment of the Requirements for the Degree
Masters of Science in Digital Media in the Ivan Allen College of Literature Media and Communications
Georgia Institute of Technology May 2017
COPYRIGHT © 2017 BY LOGAN SAND
DIRECTING ATTENTION IN AUGMENTED REALITY
Approved by:
Dr. Michael Nitsche, Advisor School of Literature Media and Communications Georgia Institute of Technology
Dr. Jay Bolter School of Literature Media and Communications Georgia Institute of Technology
Dr. Janet Murray School of Literature Media and Communications Georgia Institute of Technology
Date Approved: April 24, 2017
To my mother, for all her support. I couldn’t have done this without her, emotionally.
To Jessica Burns, one of my closest friends. I couldn’t have done this without her, literally.
iii
TABLE OF CONTENTS
LIST OF FIGURES iv
SUMMARY vii
CHAPTER 1. Introduction 1
1.1 Existing Augmented Reality Products 2
1.2 Initial Ideation and Motivation 6
CHAPTER 2. Establishing Research 9
2.1 Graphic Novels and Guiding Attention 9
2.2 Vuforia as a Platform 17
CHAPTER 3. Project Design and Development 21
3.1 Beginning the Process 21
3.1.1 Proving the Concept 23
3.1.2 Creation and Implementation of Markers 24
3.1.3 Types Of Augmentations 26
3.2 Pilgrimage as a Concept 37
3.3 Constructing the Visual Framework 38
3.4 Designing the Augmentations 44
3.5 Incorporating the Augmentations 46
CHAPTER 4. Implementation 49
4.1 Implementing the Augmentations 49
4.2 Testing the Comic 56
CHAPTER 5. Conclusion 59
5.1 The Next Steps 63
REFERENCES 68
iv
LIST OF FIGURES
Figure 1a – Nintendo 3DS AR card 1
Figure 1b – Vuforia Base Trigger 1
Figure 2 – XCOM The Board Game 3
Figure 3a – PulZar Card 4
Figure 3b – PulZar Game 4
Figure 4 – Crayola Color Alive 5
Figure 5 – Marvel AR Application 7
Figure 6 – Stan Lee, How To Draw Comics The Marvel Way 10
Figure 7 – Eisner’s Comic Panel Flow 11
Figure 8 – The Enigma of Amigara Fault 13
Figure 9 – Astonishing X-Men #1 14
Figure 10 – Eisner’s variation in panel edges 15
Figure 11 – Nintendo 3DS AR use 17
Figure 12a – Regular Vuforia Image 18
Figure 12b – Vuforia Image With Tracking 18
Figure 13 – Marvel’s Avengers AR App 19
Figure 14 – Pokemon Go 20
Figure 15a – Transformers Original 23
Figure 15b – Transformers Augmented 23
Figure 16a – Istochnik, Page one 25
Figure 16b – Istochnik, Page two 25
Figure 17 – Bioshock Lighting Example 27
Figure 18 – Flower Color Example 29
Figure 19 – Dead Space Nursery 30
Figure 20 – Dead Space Sun 31
Figure 21a – Infamous: Second Son Paper Trail: Van 33
Figure 21b – InFamous: Second Son Paper Trail: Phone 33
Figure 22 – The Lost Woods 34
Figure 23 – WiiU Gamepad being used as scanner 36
Figure 24a – Star Trek Klingon 39
v
Figure 24b – Stargate SG-1 Spirits 39
Figure 25a – White Poodle Moth 40
Figure 25b – Red Poodle Moth 40
Figure 26 – Character Designs In-Comic 40
Figure 27a – Original Panel Normal 41
Figure 27b – Original Panel Augmented 41
Figure 28a – Sketch Layer 42
Figure 28b – Ink Layer 42
Figure 28c – Color Layer 42
Figure 28d – Text Layer 42
Figure 29a – Club Panel Normal 43
Figure 29b – Club Panel Augmented 43
Figure 30a – Page Seventeen Normal 44
Figure 30b – Page Seventeen Augmented 44
Figure 31a – Comic Image 46
Figure 31b – Vuforia Recognition Points 46
Figure 32 – Zero Star Vuforia Grade 46
Figure 33 – Comic Scene within Unity 48
Figure 34a – Lid to the Crate 51
Figure 34b – Crate of Weapons 51
Figure 35a – Key Card Start 52
Figure 35b – Key Card Success 52
Figure 36 – Keypad On Initialization 54
Figure 37 – Keypad, showing incorrect combination 55
Figure 38 – Keypad, showing correct combination 55
Figure 39a – Istochnik, Page Two 60
Figure 39b – Pilgrimage, Page Seven 60
Figure 40a – Showing art style of early pages 62
Figure 40b – Showing art style of late pages 62
Figure 41 – Scott Pilgrim Vs. The World 64
Figure 42 – Pest Caravan by Jessica Burns 65
vi
Figure 43a – Artace in the comic 66
Figure 43b – Artace fully shaded 66
Figure 44 – Amber Concept Sketches 67
vii
SUMMARY
Despite the failure of Google Glass to become the next hot item, augmented reality is quickly
becoming a booming market of design, both for technical industry purposes as well as home and
public entertainment. Many game companies, both video and tabletop, have begun to research and
develop for this hybrid medium. However, their designs are almost immediately flawed because it
becomes apparent that they are only just tiptoeing outside of their comfort zones in the creation of
the artifacts. The companies that are used to designing purely digital games use the tangible
creation as just a tool for image recognition without focusing too much on its design aside from
pure utilitarianism, and the companies that are used to designing tabletop games use the augmented
reality as minimal enhancement that’s usually optional for gameplay. The aim of my project is to
address the flux that’s created in any hybrid media when the user’s attention switches between the
digital and the tangible artifacts. I plan to extrapolate from Scott McCloud’s concept of closure
between graphic novel panels to create dynamic interest in an Augmented Reality setting, with the
attention shift between panels serving as a parallel to the shift between digital and physical
mediums.
1
CHAPTER 1. INTRODUCTION
As virtual reality (VR) becomes a booming market with the Oculus Touch, HTC Vive,
and Playstation VR to name a few, there is little public mention of its predecessor Augmented
Reality (AR). Where VR focuses on immersing the interactor entirely in a digital space and
experience, AR bridges the gap and places digital artifacts in the real world through the use of a
camera. Since it is both physical and digital most companies who make Augmented Reality
experiences tend to stay close to what they are used to developing, and the other medium
becomes almost incidental.
This is not necessarily a problem, but rather a lack of usefulness. If time and resources
are going to go into making part of a product, then shouldn’t that part of the product also have
value? So much augmented reality is done with blocky black and white images or clearly labeled
high contrast ones, and they do not mesh into the augmentation itself. They are used as markers,
and then once the device has recognized them they are obscured.
Fig. 1a-1b: Nintendo 3DS AR card (left) and Vuforia Demonstration Trigger (right)
The biggest reason for this is that images like these, with high contrast and clean lines,
are most easily recognized by the augmented reality programs. Vuforia, a commonly used AR
program, even incorporates an image grader that ranks targets on a scale from zero to five of how
2
easily the program can recognize the target image. Since these games and experiences are
developed for the mass public, it makes sense that easily recognized images would be used to
avoid any problems in target recognition. These images are lacking any greater context, and
merely serve as a marker for the program to work correctly.
The goal of this project was to design an augmented experience that would be cohesive
between the digital and tangible artifacts, making both halves critical to the full experience.
Graphic novels, or comic books, provided a framework for the execution, since attention is
already divided into segments, though within the same medium. By using the concept of
paneling in comic books as outlined by Scott McCloud and Will Eisner, an augmented graphic
novel could be created that would utilize the companion app in the storytelling as opposed to
providing incidental information to the user. It would allow the panels to dynamically change
and shift to show changes in the story world, building suspense and creating the illusion of
change, even though the comic has been pre-printed and is a static object. The interaction and
agency, or the illusion thereof, comes from the digital application. Because this project is
designing for interaction within a digital space, video games can be used as reference points to
how interaction is done in a digital space even though the app itself is not a traditional video
game. Many of the same principles of interaction apply, and can be used as a basis for guiding
player attention both on the screen and between the screen and the book.
1.1 Existing Augmented Reality Products
The inequality between digital and tangible design in modern Augmented Reality
gaming creates a rift in player attention, because the game designers are designing for
dominant and secondary dynamics. The user’s attention is meant to be focused on one artifact
over the other, with the secondary artifact acting as either a gimmick or a tool. For example, the
3
board games Munchkin (Steve Jackson Games, 2001) and XCOM: The Board Game (Fantasy
Flight Games, 2014) both have companion apps, but neither are necessary for gameplay. They
serve to assist with rules and scorekeeping, and XCOM’s diegetic design of the application helps
with immersion, but they are ultimately extraneous to gameplay.
Fig. 2: The board game uses the companion app as a way to integrate a digital interface for
rules and scorekeeping, rather than augmenting the game itself.
On the opposing side, digital game developers generally use their tangible artifacts as
markers to place digital objects in set points in the real world. PulZar (PlayStation Vita, 2012)
and the Nintendo 3DS’ Augmented Reality cards (Nintendo, 2011) are both prime examples of
the tangible artifact being a tool. While Nintendo’s AR Cards do have more detailed designs than
the QR-Code styled markers that PulZar uses, they are still utilitarian in their nature. The cards
are made up of straight lines and sharp angles, with very high contrast between the darks and
lights of the colors used.
4
Fig. 3a: The PulZar game trigger card before the game begins.
Fig. 3b: The PulZar game in progress, with the cards obscured by the app.
The Crayola Color Alive coloring book and companion app is the closest ideal for this
project. In this system, the user has a coloring page and colors it in as they would normally.
Then, once they have finished coloring the page they can scan them with an iPad and see a 3D
recreation of what they have colored. From a technical standpoint all of this has been done
5
before, where the line-art of the coloring pages acts as an image tracker for the recognition
program. Then the image of the sheet is projected onto a 3D model of whatever the page
represents. This can be done in programs like Vuforia with little problem. However it provides a
more even distribution of attention between the physical and the digital, in the time has to be
taken to color in the page in order for the model to look correct. The pages can be used blank, of
course, but the model will just be paper-white then which is not the intended goal. When the user
colors in the lines they create pencil strokes and striations that then transfer onto the three-
dimensional model in the companion app, so they can immediately see the impact that they have
on the space.
Fig. 4: This is a screenshot from the Crayola Color Alive app, showing the colored model.
6
Initial Ideation and Motivation
Development for this comic experience began during Dr. Jay Bolter’s Augmented Reality
class at the Georgia Institute of Technology, and began to look critically at augmented reality as
a whole. The first realization was that companies tended to focus on one half or the other when it
came to Augmented Reality, and either the tangible real-world suffered for it, or the application
became unnecessary and therefore unused. Most AR applications relied on some form of QR
code for image recognition, and then the app hid the tag entirely once the experience began. To
me, this seemed less about merging physical and digital, and more about lacking a creative way
to anchor their digital game objects in a physical space. The first key requirement of Augmented
Reality is that “It combines real and virtual content” (Billinghurst, p77, 2015). The idea of
Augmented Reality means to make reality more than it already is, rather than to obscure what
exists in favor of the purely digital. This concept seems to be one that most current developers
are struggling with, because either the Augmented Reality is gimmicky to focus on the tangible
artifact, or the tangible artifact has been reduced as far as it can to a purely utilitarian form. This
means that the interactor still is only focusing on either the tangible or the digital artifact, with
the other serving as a means to an end.
Clarity and ease of recognition is not the only reason that digital game developers lean
towards black and white artifacts with sharp straight lines. The developers want the player to
focus on the digital, and the card itself is just a means of recognition and nothing else. This is
most obvious in the PulZar game, where the cards are simplified QR Codes that the Playstation
Vita recognizes. More often than not with this sort of game, the AR Card is hidden or obscured
once the camera recognizes it. This forces the player to focus solely on the digital artifact,
without being consciously aware of the process that is going on that allows them to see digital
7
objects projected into the real world. They pay no attention to the tangible marker, because it has
been effectively removed from their field of view.
Games are not the only medium that has tried to augment their features, comic books
have also tried to implement digital artifacts within their pages. Both DC Comics and Marvel
Comics developed AR companion smart-device apps for their comics, but neither one seemed to
fully embrace the digital aspect of the medium. The app released by DC Comics merely overlaid
a text-free version of the comic over the page, and had voice actors reading the text out loud.
Marvel Comics’ app, on the other hand, eschewed the comic entirely and when the AR panels
were scanned it brought up things like character biographies or concept artwork. Both of these
are entirely incidental, and not at all required for the full reading of the comic books. These apps
feel similar to more traditional second-screen apps that television shows use, providing bonus
behind-the-scenes information to be viewed during commercial breaks. It requires its own level
of attention be paid to it, and exists separate of the ongoing story like a behind-the-scenes section
on a DVD menu.
The Marvel AR app, which showed extra artwork but did not change the story.
But what if the app was involved in the story progression? What if the app aided in
telling the story, as opposed to just adding in unnecessary information or reading the book to
8
you? This was part of the main goals of this project; to develop an augmented reality experience
that would split the user’s attention between the digital app experience, and the physical artifact
that went with it. Graphic novels were chosen as the real-world artifact because the comic panels
themselves could be used as image-recognition markers, which would inherently make the
markers more visually interesting than a QR code.
Graphic novels naturally break down information into timed sections, called gutters.
Gutters are “...That space between the panels” where “...The comic’s reader takes two separate
images and transforms them into a single idea” (McCloud, p66, 1993). This is merely to say that
not every single thing is drawn out in a comic, otherwise it would be a flipbook. But the idea
behind gutters is a powerful one, and one that is exploitable for the purposes of this project.
While the reader may not be consciously aware of it, the gutters are a place where their attention
goes from one point in time to another, within the diegetic space. Each frame is a moment or
series of moments in time, depending on the size of the frame (McCloud, p95, 1993). Larger
frames have more content, and therefore are not just a single snapshot like a smaller frame would
be. The gutters dictates this spacing, since it divides the pages into panels. Since different gutters
have different connotations according to Will Eisner, it stands to reason that there could be a
designated gutter style for augmentation that would act as a trigger without having a specific
marker inside the panel itself (Eisner, p44, 2008). This way, the augmented panels would still be
clearly marked, but in an unobtrusive way that would not impact the artwork and storytelling
within the panels themselves.
9
CHAPTER 2. 2 ESTABLISHING RESEARCH
Graphic Novels and Guiding Attention
Stan Lee and Joe Buscema speak at length about how comic panels must be set up and
organized in their book Stan Lee’s How To Draw Comics: The Marvel Way. The reader’s eye
needs to flow naturally from panel to panel, and so the action needs to flow in a way that is
smooth, but also helps lead the reader from panel to panel. The way a character faces, or the way
a piece of machinery in the background is laid out, even the direction a character’s eyes are
pointing impacts the direction of the reader’s attention. In the context of this project these same
guiding principles can be used to push the reader from the page, to the app, and back without
having harsh signposts dictating what they have to be doing and how much time must be spent
on it.
10
Fig. 6: Page 111, Stan Lee How to Draw Comics the Marvel Way
In the above panel, it shows a sketched-out page of panels and demonstrates the concept
of implicit lines. Depth helps a lot to create this, and is the main method of controlling the
reading path. Panel one has Captain America in the foreground with Captain Britain behind him
on the reader’s right. He’s on the same depth plane as Red Skull in the next panel, who is aiming
a gun at Nick Fury. The line created by Skull’s leg and Fury’s shoulder leads the eye to panel
three, where the gun is mimicking this line. Fury’s ankle and elbow are outside the panel itself,
which leads the eye to panel four. This method of directing attention is commonly used by artists
because the mind wants complete images and by leaving parts of a whole outside of a panel it
11
guides the eye to where the omitted object is; in this case the next panel of the page. In an
augmented reality context, these concepts can be used to guide the reader back into the tangible
object from their device, rather than having the augmentation on a timer and disappearing after a
set length of time. It flows more organically, and lets the reader move at their own pace rather
than the pace dictated by the writer, taking their time where they feel like it is necessary to.
Below, Eisner has illustrated the most basic flow of attention through the panels. In western
comics, the last panel on the page should be in the bottom right corner, and the first page should
be in the top left. There are ways of working around this, such as having a full-page borderless
panel with smaller panels within, but this diagram is the standard basic layout.
Fig. 7: Showing the flow of a six-panel page. (Eisner, p42, 2008)
The fewer panels on a page, the easier this idea of controlling attention is to accomplish.
If the entire page is one panel, then the reader automatically looks at it in the correct order
12
because there is only one panel to look at. Three horizontal panels stacked vertically are also
almost a guarantee, as each panel will be read, then the reader progresses to the next one down.
With dynamic setup it becomes more difficult to ensure correct reading.
Scott McCloud, in his book Understanding Comics defines a panel as a segment of time.
The longer the panel, the more time is being consumed in it because “In learning to read comics
we all learned to perceive time spatially, for in the world of comics, time and space are one and
the same” (McCloud, p100, 1993). There are limits to this time dilation, namely the width of the
comic page itself. A panel is a moment, though not necessarily a still image. Too much action in
one panel and the reader becomes disoriented and confused, so the panels are broken up by
‘gutters’, or the blank space between panels. The placement of these gutters is crucial because
“We [comic artists] assume as readers that we will know what order to read panels in, but the
business of arranging those panels is actually quite complex. So complex, in fact, that even
seasoned pros will sometimes blow it” (McCloud, p86, 1993). The artist cannot assume that
every reader will think in the exact same way that they do, so the line of progression must be
made clear. While panels can be more than just rectangles, to be sure, the core structure of
comics sticks to a fairly grid-like composition. As shown above in the page from The Enigma of
Amigara Fault, diagonal gutters are used to create tension and unease, especially when they are
only slightly tilted. This page was chosen to show this technique because the panels begin
vertical, then slowly become more slanted as the page progresses. Artists of horror comics and
manga sometimes subtly tilt their gutters in this fashion because it will make the reader
uncomfortable even though they cannot precisely say why.
13
Fig. 8: The Enigma of Amigara Fault shows the flow of attentive direction. The
characters are facing the linear progression, with diagonal gutters dividing the last two
rows to increase tension. Ito, The Enigma of Amigara Fault
14
Fig. 9: A panel with no gutters serving as the background
This page from Astonishing X-Men #1 by Joss Whedon is an example of directing
attention with gutters, or lack thereof, while also forcing the reader into a specific mindset. The
two panels at the top are using McCloud’s ideas of small panels being a short segment of time,
but their width keeps them large enough to be interesting. The center panel breaks the gutters and
is zoomed out, giving it a sense of expanse and isolation to the reader. By removing the gutter
15
lines which are used to denote sections of time, it makes the reader see the large panel occurring
almost simultaneously with the three smaller block panels.
Eisner speaks at length regarding this topic in Comics and Sequential Art, outlining
various ways that comic borders can direct attention and create specific reactions. If the frame is
tight to a subject, then the attention is very direct. The reader is looking at the subject, and little
else. If a frame is large and distant, then the reader is made to feel isolated and alone, even if the
frame is crowded. Likewise, the actual border impacts the reader as well.
Fig. 10: “The ‘panel’ here is actually the doorway. It tells the reader visually that
the actor is confined in a small area within a wider one-- the building.” (Eisner, p49, 2008)
A rough, cloud-like border “...Defines the picture as being a thought or memory...”,
whereas “The absence of a panel outline is designed to convey unlimited space” (Eisner, p49,
2008). Two panels with diagonal parallel lines give a sense of dynamic tension, often used with
fighting characters facing each other. The panel can just be one wide panel stretching all the way
across with the characters on either side, but the line through the middle is symbolic of the break
in communication.
16
For this project, both McCloud and Eisner are invaluable resources given their experience
with the visual medium of sequential art. Their texts are more commonly used in reference of
structure and composition, but they have gems of attention guidance especially in regards to
panel flow. McCloud speaks in terms of time and the reader, which is important in regards to
splitting attention between two devices. A series of smaller panels take up less time than a few
larger panels, but does that equate whatever augmented reality experience that correlates with
that page is the real question. Eisner’s book talks a lot more about creating emotions and thought
processes, which are crucial for attention. The biggest part of both authors is flow. There needs
to be suggested lines within the panels and between the panels for them to flow properly. This is
the crux of reader attention in comic books. If the flow is broken or reversed, like a character in a
movie crossing across the camera line, then the reader becomes confused and it breaks the
experience for them. The flow of the story determines the flow of the actions, and the flow of the
actions determines the flow of the panels. These decisions need to be planned out in parallel to
each other, in order for the experience to be fully realized with maximum efficacy. The project
would have to be thoroughly planned from the start, otherwise it would fall into the same trap
that Marvel and DC did, of just having a comic book that happens to have a few AR artifacts,
rather than designing a cohesive experience that used both mediums effectively. The
augmentation had to flow with the comic rather than against it, which is why the concept of
gutters was so important from the start. By using the gutters as described by Eisner and McCloud
augmentations could be inserted into the comic, splitting the reader’s attention between page and
screen, without breaking the flow of the comic’s pacing. By keeping and maintaining the flow of
the comic, the reader can translate from the book, to the device, and back again without thinking
consciously about it.
17
Vuforia as a Platform
As previously stated, Augmented Reality is something that people have been working
with for years, and many people use it in different forms. One that is widely used is image
recognition, where a camera on a smart-device sees a picture and uses a program like Vuforia,
first released in 2013 by Qualcomm, which is made to create and grade image triggers, to project
a digital artifact onto the image. This is how apps like Nintendo’s 3DS AR game works.
Fig. 11: The Nintendo AR cards, and what they look like on a 3DS screen.
Vuforia prides themselves in being “The world’s most widely deployed AR platform”
with over 325,000 registered users and over 35,000 applications made. It is capable of
recognizing simple and complex 3D objects, as well as QR code style markers and custom user-
defined images (vuforia.com, 2017). This project uses the user-defined images, with the specific
comic panels themselves acting as triggers. It can be developed on Android and iOS devices, as
well as Microsoft HoloLens and several other phone-based VR headsets like the Gear, the Knox,
18
and the View-Master. It can be used with several development kits, the decision was made to use
it with Unity because it was well-documented and stable for use.
These programs do not recognize the image itself, instead there is an algorithm written
that automatically sets key points on images based on contrast, and when the camera detects the
matching points it generates the digital image based on the corresponding points. This means that
AR images within the same app have to be significantly different, otherwise you risk triggering
the wrong image and breaking the experience.
Fig. 12a-b: The regular image (left) and the Vuforia recognition points (right)
This method of AR intervention is used for experiences that are meant to be modular and
taken with the user, but it is not the only AR methodology. Another popular augmentation is
overlay based on camera position. In this style, the app gets the camera’s orientation and then
renders images in relation to it. This can be used anywhere without a physical target, because the
camera is only being used to create a background.
19
Fig. 13: The Marvel Avengers AR App released shortly after the first movie.
This does not require any image recognition or geolocation, because it is not tethering any digital
artifacts to real-world targets. It is just a game running with an invisible background, so the
camera overlays whatever it sees giving the illusion of the action taking place in the real world.
Images are not required for anchoring digital objects in space, however. With the AR
technique of geofencing, the app can access the internal global positioning system (GPS) and
uses it to find markers placed at geographic coordinates. This is how apps like Pokemon Go
work. Niantic links geographic coordinates to a gym, poke-stop, or pokemon, and then whenever
the player’s phone enters the geofence it triggers the corresponding event.
20
Fig. 14: Pokemon Go, showing the gyms and poke-stop locations.
Geofencing is the least reliable of the three methods of augmentation, because it relies on GPS
accuracy and margin of error. With the anchor being entirely digital, there is also the risk of user
error where the player thinks they are within the fence, but their phone doesn’t recognize it. This
is in direct opposition to the camera positioning method, where the application doesn’t need to
recognize any tangible artifact to work. Image recognition seems to be the comfortable middle
ground between the two, where the app needs to recognize a real world boundary of some sort,
but everything is consistent between people. It is the same app, and the same trigger images, so it
drastically removes the chances for user error. This is why image recognition was an ideal choice
for an augmented comic, because the experience had to be uniform between readers. Just as a
traditional comic book has the same panels on the same pages from reader to reader, so should
the augmented comic book. By having the augmentations be a part of the plot, it avoids the
failing that DC’s AR app had of not adding anything to the experience. The DC Comics app read
the exact same words out loud that were on the page, and did not actually change anything. The
21
Marvel Comics AR app brought new content to light, but unfortunately that content was aimed at
a narrow niche like new comic readers. A short biography of who Tony Stark is won’t be
interesting to someone who has been reading comics for decades, for example. By having the AR
experience tie in with the story, it puts all the readers on a level field.
CHAPTER 3. Project Design and Development
Beginning The Process
Pilgrimage was the third iteration of the idea of an augmented comic book, with the first
two serving as stepping stones towards the final product. Initially a proof of concept was made
using a few pages from the IDW comic Transformers: More Than Meets the Eye because it had a
vibrant color scheme and clear line work. The artwork was pre-made and ready to go, and only
needed to be put into a Vuforia library and augmented. This was almost immediately successful,
allowing the project to move on to step two; which was to create original artwork and augment it.
Adobe Photoshop was used to create these test panels, which were then augmented in the same
way as the Transformers panels were. A round of testing was done with two test pages at this
step, and the information gathered was used to create the final product of Pilgrimage.
It is important to plan the comic, including the augmented sections, ahead of time to
make sure the two parts are cohesive and make sense in tandem rather than just having two
separate experiences that are forced together. So, the first step was to find out how to control and
direct attention in a digital setting. Research done in the field of video games was helpful during
this step, because user interfaces focus so strongly on guiding user attention. While the intention
was never to make a traditional video game, the principles can be paralleled with any digital
interactive experience. The problem with a good deal of the existing research was that it was
22
almost always heavily slanted in the direction of either video games causing or solving attention
problems in younger children. There has been a great deal of research and documentation
regarding attention in interactive mediums, but a majority of it revolves around Attention Deficit
Hyperactivity Disorder (ADHD). Those in support of video games have papers stating that
games cut down on ADHD and other problems, where the opposing side has equally valid-
seeming data saying that video games exacerbate ADHD. When some of these papers are
establishing their background, it is shown that their bias is impacting their research because they
are using surveys with “...Questions [that] reflect the negative impact of video games on social
functioning and relationships including excessive video game use, neglecting work and social
life, anticipation, lack of control, and salience”. (Chan and Rabinowitz, 2008) Chan and
Rabinowitz state outright that they are using questions that reflect the negative impact of video
games, rather than a mix of positive and negative, or better yet neutral questions.
The first thing to do is set a marker. Image recognition style augmented reality needs
markers to work properly, otherwise it boils down to just another companion app. These markers
will also serve as a way to break the reader’s attention in the comic, forcing them to transfer to
their phone or tablet device. The Attention Blink method served as a groundwork for designing
how to mark augmentation, because according to their study participants were able to better
discern information that only had to be identified, rather than categorized. (Johnson and Proctor,
p207, 2004) It was also shown that in a sequence, “...Although participants are able to identify
and report a single target… Report of the second target often fails”. (Johnson and Proctor, p208,
2004). This reasoning is why the decision was made to have only one augmented panel per page,
because any more than that and a hierarchy is formed. One augmentation becomes the primary
one, and the rest become secondary and tertiary with the risk of being glossed over entirely by
23
the reader. It also served as a structure for relying on a trigger like a block of color, rather than a
more complex icon within the panel itself, since it was postulated in the Attention Blink study
that less complicated triggers are more easily and more quickly recognizable than complex ones.
Proving The Concept
To test whether comic panels could even work as Vuforia triggers, premade comic panels
were taken from IDW’s series Transformers: More Than Meets The Eye and put them into a
Vuforia database, then created a basic scene in Unity using the tutorials provided on Vuforia’s
forums. These panels worked well, due to the high contrast colors and thick, clean lines.
Fig. 15a-b: The original panel (top) and the augmented change (bottom)
While these worked correctly, this step showed a very important flaw about mobile
development that was thankfully discovered early in the process. Originally the intention was to
create full 2D animations for the augmentations, and they had been created for this testing step
using the premade panels and animating the saturation levels. The video textures worked well in
24
Unity while they were tested, but caused errors during their upload to the iOS device The error
was that mobile devices do not support animated textures from Unity, so it was a relief to learn
that early on with five smaller animations rather than later on in the process with twenty-five
panels in the pipeline. Any animations would have to be done in Unity by animating objects,
rather than importing video files and applying them as textures.
Creation and Implementation of Markers
Over the course of this project contact was established with a professional artist, Jessica
Burns (https://www.artstation.com/artist/burnsjess), about visual ideas and how to keep a full
comic looking cohesive. Ms. Burns earned a B.S. degree from East Tennessee State University,
then went into the professional field with work for Digital Extremes, Play Likely Mobile, and
Van Ryder Games on her resume. She also has extensive experience with sequential art, most
recently with “Pest Caravan” (2016) hosted on Tapastic.com. With this experience she was able
to give useful critique throughout the process of this project, both on the side of the artwork as
well as the interactions.
Originally the plan was to keep the comic’s gutters an average width compared to most
mainstream comics, and use a bright red border around any panels that are augmented. When
speaking to Ms. Burns about this, she warned strongly against it because it would be disjointing
for the reader. The goal was not to break attention, but rather to nudge it in a different direction
and keep the reader inside the flow of the story world. She suggested that the gutter lines be
made wider, and to create a secondary border of a cooler color around augmented panels. What
Ms. Burns explained was that warmer colors appear further forward in the composition than
cooler colors, and since black is the coldest color in the spectrum any color you compare it to
will appear to come forward. The AR panels were then bordered in a bright green, half as wide
25
as the black gutters themselves. This way the augmented panels could be marked and identified,
but it was not as harsh of a break as the red outlines would have been.
Fig. 16a-b: Two pages from the testing comic, Istochnik
During testing, using the same process as the Transformers panels, these two pages from
the initial comic worked well for Vuforia. Having a mostly grayscale palette to choose from
meant that any colors used would be emphasized, and the green would be unmistakable because
it was the only vivid color being used. Vuforia works best with non-gradiated colors and clearly
defined edges, so it made each panel easily recognizable for the program. The problem for this
comic came when writing the script for it and after a few pages realized that it was writing for a
regular a comic and forcing augmented reality into it, rather than designing a comic to be used
with augmented reality. Unfortunately this meant that the original concept had to be scrapped in
favor of one that would lend itself more readily to the process. By shifting the tone from
26
obscured exploration to action-adventure, the AR elements fit much more cohesively with the
story and it diegetically makes more sense. This was the main flaw in the Marvel and DC
augmented reality apps, because there was no diegetic action within the digital space. The
information was ultimately extraneous to the story, and not necessary for the full enjoyment of
the comic book.
Types of Augmentations
The next step was to design the augmentations themselves, with each one different from
the last but also unified within the story world. Using video games as the basis for the
augmentations was advantageous, in that it laid the groundwork for what is successful and what
is not. Looking at how games control the player, trends began to emerge in methods for
manipulation.
1. Light and Color
Whether it is being used to illuminate writing on a wall, or having the lights in a room
flicker before going out, light is an effective way to keep attention focused on the device rather
than the book. These are listed together because it is difficult to convincingly change an object’s
color without changing the light shining on the object, or being emitted by the object, in some
way. Light and color are two sides of the same coin, since one cannot exist without the other.
Light illuminates, which allows the eye to experience the color that the light is reflected from.
An example of light guiding attention is on a stage, with Francis Reid saying that stage
lighting is specifically engineered to control the audience. In his book Stage Lighting Handbook,
he postulates that “The obvious technique is to light only the selected area of the stage while the
rest is blacked out. However, it is also possible to make a subtle but positive selection of vision
27
by balancing the selected area to a brighter level than the rest of the stage. It is surprising how
even the smallest light re-balance can help to concentrate audience attention on the appropriate
action area”. (Reid, p6, 2001) This information can be used directly in augmented reality, by
changing the lighting in a panel it immediately directs the audience from looking at the panel as
a whole, to focusing them on the key element of that panel. In a more digital setting, the
Bioshock games developed by Irrational Games rely heavily on darkness and light to direct
attention. One scene in particular had a power-up on a pedestal beneath a spotlight, with enemies
hiding in the darkness around the player. They’re there and immediately dangerous, but almost
no players see them because they are so focused on the power-up.
Fig. 17: A high-contrast area in Bioshock, showing where you need to go while obscuring
enemies and traps. (PS3)
Color serves as a gradient rather than a scale, because there are shades and combinations
of different colors whereas light is a sliding scale of off to on. According to Elliot and Maier,
“Those with normal color vision experience a vast and rich chromatic palette, with estimates
28
reaching up to 2.3 million discernible colors that may be seen together in an “almost infinite”
number of possible combinations” so it must be executed and handled differently. (Elliot and
Maier, 2014) Colors create depth within a flat panel, changing how the eye perceives the scene
even though it is all just ink on a page. “Warm or dark colors tend to advance while cool or light
colors tend to recede” which was echoed by Jessica Burns during her critiques. (Fraser and
Banks, p70, 2004) The emotional impact of certain colors has been studied for centuries, going
back to Goethe in 1810 describing the idea that “...Plus colors, namely, yellow, red-yellow, and
yellow-red, were thought to induce positive feelings such as lively, aspiring, and warm, whereas
minus colors, namely blue, red-blue, and blue-red, were said to induce negative feelings such as
restless, anxious, and cold”. (Elliot and Maier, 2014) Flower by ThatGameCompany uses color
very intelligently, especially since their game is already so vibrant. They constantly use high-
contrast colors to make sure the players know which directions to go, and which flowers are
important. Light is also used in this, but color plays the biggest part in the equation. With such
high saturation, the game had to focus on complementary colors: Red flowers on green grass,
blue flowers on dried yellow grass, white flowers on black grass. Always in small patches, never
overwhelming, just enough to get the player where they need to go.
29
Fig. 18: A wide shot that the camera zooms to after blooming a small circle. (Flower, PS4)
2. Movement
According to the paper “Attention-Based Motion Perception”, the human brain can detect
low-level motion independent of conscious thought, meaning that motion inherently garners
attention subconsciously. In the context of this project, it could be a ship moving past the
window, or a character moving within a panel to open a door. These small animations, between
ten and fifteen seconds, would be enough to keep the viewer’s attention diverted, without making
them feel like they’re watching a cutscene. This also works in reverse, with no motion in an area
where there should be movement. It builds suspense, and forces the user to focus to try and find
what they feel should be there.
An example of movement as direction is in the EA game Dead Space 2. Survival horror
games are excellent at directing player attention, because they build suspense and create tension
within the player by definition. In one scene, the player goes through a nursery area that is quiet
and still. But there is a room at the edge of the field of view with shadows moving. This guides
30
the player into the next area, but also creates tension. There are no enemies in this room, but it
feels like there should be because the shadows are moving on the wall.
Fig. 19: The Nursery in Dead Space 2, combining spinning lights and eerie music. (PS3)
Another instance of this technique, from the same game, comes slightly later in the same
area. There is a school gymnasium, much like the gymnasiums most Americans see in primary
school, where a massive horde of zombies is fought. All the doors are locked, so the player is
forced to go up on the stage. The walkway is narrow, and there are set pieces all around. After
the big fight, there is no movement and very little light. Everything is still, and the player is on
edge. At the end of the first row, a large and smiling sun prop drops down from the ceiling.
There is nothing scary about it, but after such a long period of no motion or movement, the
tension has built to such a level that even the smiling sun makes the player jump.
31
Fig. 20: This scene is dark and ominous before the sun drops down, the sudden motion
scaring the player. (PS3)
In comic books motion is implied, because they are not animated. By augmenting the
panels animations become possible, bringing the reader more closely in line with the artist’s
original concept. It was concluded by Cavanagh that “once a feature is sufficiently visible to be
tracked, then the sensations of motion for the feature are based on low-level signals”, so in the
context of AR this would mean that once the augmentation is superimposed onto the comic, the
reader would then see the movement without having to consciously think about watching it.
(Cavanagh, p1565, 1992) The reader would be able to see the movement as a whole within the
panel, not just as a moving object.
3. Interaction
Video games use interaction to engage the player and hold their attention for potentially
32
hours, so interactivity would be a key aspect of this project. Interaction serves as a way to draw
attention, while making the viewer feel involved in the story world. There is no true agency in
the medium of a graphic novel, given that the book is pre-printed, but the illusion of false agency
can be created by using the page, rather than a specific panel, as a boundary. Attentiveness
within game worlds has been studied previously, resulting that “Eye-tracking data shows that in
the first person shooter game, players paid attention only to the center of the screen, where the
cross of their gun was located. Most of the time, they shifted their eye-sight only to read the
information that appeared on the corner of the screen”. (Yan and El-Nasr, p6 , 2006). This
example of a first-person shooter, in this case Halo 2, shows that interactors focus on where the
interaction is, looking away only when new stimuli is added to their field of view.
Since this is an augmented novel, as opposed to a fully digital game, the interactive
elements will be relatively limited, with the goal time of taking under one minute to complete.
Examples of this could be scrolling through text on a computer screen, or entering a key code
that the protagonist discovered a few pages ago. Limited in scope to avoid taking too much
attention away from the physical artifact, but enough to pull and hold attention on the digital.
This is the largest holder of attention, and so must be used sparingly as opposed to light or
movement, which can be done quickly and retain the same efficacy. Augmented Reality would
lend itself easily to making a large grandiose puzzle that would take hours to solve, but that
would defeat the purpose of this project. The goal is to split attention roughly equally between
the analog and digital artifacts, not focus on one over the other.
The Playstation 4 title InFamous: Second Son has a downloadable content mission called
Paper Trail that forces this interaction between two screens in order to advance the storyline.
After playing for a period in the game, the player must log onto a series of websites and solve
33
puzzles and crack codes, like piecing together shredded paper or hacking into a website. In one
instance, the player must call a number that tells them what to do. But they can only get that
number from finding a business card, turning it over, and logging into a website with the
password found there. While a useful example, this style of multifaceted interaction is much
more complex than what is necessary within this project, because it divides the time too
greatly. Each section of Paper Trail only takes a few minutes, but the sections nest together and
it can take almost an hour to get back to playing the game itself.
Fig. 21a-b: This shows website artifacts that directly impact games on a console. Infamous:
Second Son Paper Trail hacking storyline. (PS4)
For the graphic novel being created, interactions are limited to only a few instances
because they are so much more involved on the reader’s part than seeing a color change or
watching something move on a screen. The user has to physically make a change to the digital
space which, as stated, focuses their attention on the interaction and little else. The interactions
are not be incorporated into the panels for this reason, rather they are styled like pop-up windows
to obscure the panel itself. If an augmentation is interactive then that is the focus, and making it
smaller to fit within the panel itself would only be a hindrance.
34
4. Sound
Sound as a stimulus pairs with movement, so having sound as a stand alone augmentation
must be done with care to avoid breaking the illusion. Movement without sound can be
acceptable, but sound without movement is rarely seen. The glaring exceptions to this is when
either there is a speaker or device or the goal is to create tension, then the sound is being created
by something out of the field of view. Hearing has a separate set of rules from vision, which
must be recognized in order to use it effectively. Unlike vision, “...It is impossible for subjects to
process more than a few simultaneously presented auditory stimuli”. (Johnson and Proctor, p96,
2003) In the context of a graphic novel this information is useful, because “...The dichotic
listening paradigm shows that people can direct attention to just one source of stimulation when
the source is well defined by a basic characteristic such as location”. (Johnson and Proctor, P101,
2003) Meaning that as long as the source of a sound is identified, or at the very least implied in
the setting of the scene, it will hold attention more effectively than if it were a non-diegetic
sound like background music.
Fig. 22: The Lost Woods, Ocarina of Time (N64)
35
A classic example of this comes from the Nintendo64 game The Legend of Zelda:
Ocarina of Time. In one of the early stages there is a maze of rooms, each with four doors. While
the music is playing, it is noticeably louder near one of the doors to notify the player that it is the
correct one. Coming back to Bioshock, the player is often directed by the sounds of yelling,
gunfire, or in some cases a player piano within the worldspace. Sound is often used as direction,
whether with its presence or the lack thereof.
Sound within a comic book begins to toe the line between comic and animatic, because it
engages more senses than just the eyes. To restate the information from Attention, sound can be
used as an anchoring device within the worldspace, provided that it is given context. In a comic,
aside from the characters themselves the anchor could be a speaker on the wall, a phone, or an
alarm clock. This bridges the gap between digital and physical because the app is not providing
any new visual information, and the user is able to look at the comic page while hearing the
audio playing.
The goal of this project is shifting attention between two artifacts, one that is digital and
one that can be physically manipulated. One of the most prominent utilizations of this type of
attention shifting is the Wii U from Nintendo, with games like “Legend of Zelda: Wind Waker”,
“Sonic Boom”, and “Batman: Arkham City Armored Edition” all forcing the player to look
between the main screen and the smaller Wii U tablet. However, in each of these instances the
player is either explicitly told what to do and where, or left entirely to their own devices. In
“Wind Waker”, the tablet is used purely as an inventory system, mostly ignored by the player. In
“Sonic Boom” the only function of the tablet beyond a basic controller is a secret object detector,
not necessary for gameplay and can be forgotten almost immediately after it is revealed to the
player. In every instance, the Wii U’s games seem to ignore the gamepad almost entirely with the
36
exception of multiplayer versus play. Even then it only uses the gamepad as a sort of ‘black box’,
obscuring information between players to create new dynamics in games that involve one player
chasing another.
Fig. 23 Scanning in-game artifacts. Batman: Arkham City: Armored Edition (WiiU)
The Wii U touched on several augmented reality chances but failed to go in depth with
any of them within a game. One possible reason that Marvel and DC made their apps non-
essential to the enjoyment of the comics was that not everyone had access to a smartphone,
especially the younger members of their audiences, and they did not want to risk alienating any
of their audience in the pursuit of new technology. That argument cannot be used by Nintendo
concerning the WiiU, because the console needed the gamepad to function. The gamepad was
the controller, and a necessary part of the setup. The technology was there, and was used in a few
games, like the previously mentioned “Batman: Arkham City Armored Edition” and “Sonic
Boom”, but even then they were not core gameplay elements. “Sonic Boom” used it as a way to
find inconsequential money caches, and “Batman: Arkham City Armored Edition” rarely
37
required the use of the gamepad as the scanner. These are both augmentations, but they were
digital augmentations on an already digital environment, so it came up short within the
experiences.
Pilgrimage as a Concept
After it was made clear that the original comic idea would not be usable, the next steps
had to be planned out much more carefully. For Istochnik, which was the working title of the
first iteration, the idea for the comic happened and started the art began immediately, writing the
story as pages were finished. But this method quickly proved to be problematic, so for the second
comic it was clear that the full script would have to be written out with the augmentations
outlined before any art began. It started small, with a sentence or two describing the story, and
then a paragraph outlining the story in broad details. After the paragraph had been written the
full script could begin, page by page with the panel layouts, dialogue, character actions, and most
importantly the augmentations. What had happened with Istochnik was the page first, and then it
had to be figured out where AR could be inserted into it which led to an incohesive end result.
Reflecting on the original process it became clear that artwork had to become secondary to the
experience, with the focus being on writing a story that could use the augmentations within it
rather than have them shoehorned in as an afterthought.
The story began to form, influenced by tropes within media currently dominating the
field and developed in line with familiar storytelling genres. Establishing the setting as an alien
spaceship afforded for many freedoms as far as augmenting went, because advanced technology
would allow for certain things that modern-day Earth does not have; especially in the realm of
lighting. The hardest part about writing was getting started because whatever was written then
would be what became the final product, and being locked into something was unnerving.
38
The second draft of the story, the one which ultimately became Pilgrimage, revolved
around an alien race looking for new worlds to colonize. The protagonist, named Artace, was a
naive low-level worker who had little knowledge of anything beyond his specific job within the
ship. This allowed for organic story flow to take place, with information being revealed to Artace
at the same pace that the reader got it. Ultimately Artace and two of his close friends, Qaalen and
Eijis, discover a treacherous plot by their newly militaristic government, led by Galra, and have
to make the decision whether or not they want to accept it or risk everything to change it.
Constructing The Visual Framework
After the script was finished the artwork had to begin almost immediately, since it would
be take the longest amount of time. There were many concept sketches, but they had all been for
Istochnik which had a faceless human protagonist in a futuristic suit. Pilgrimage needed to have
at least a partially non-human cast, because it would allow for more creative freedom as far as
characterization went. Inspiration initially came from classic science-fiction television shows,
and patterns in the alien designs began to emerge. A large percentage of aliens were humanoid
with one or two distinctions, and usually around the face. Klingons in Star Trek had forehead
ridges, and the Spirits in Stargate were humanoid except for gill-like appendages where their
noses should be.
39
Fig. 24a-b: A Star Trek Klingon (Left) and the Stargate SG-1 Spirits (Right)
For the character designs it needed to be nonhuman, but anthropomorphic. Recognizable facial
features make emotion easier to convey, so a humanoid shape is preferable when designing a
character. The decision was made to abstract it one step and use insects as a stylistic base as
opposed to familiar animals like dogs or cats, which are often anthropomorphized into humanoid
characters. With this in mind, the process of looking at various insect species began, scrolling
through image galleries on Google and Pinterest. After a few days of looking and thinking the
Venezuelan Poodle Moth presented itself as the base insect to work from. It lent itself well for
variation because the species itself can be a variety of colors, and since it has hairs, in an
anthropomorphic state that could translate to hairstyles, allowing for easy differentiation between
characters.
40
Fig. 25a-b: Two colorings of the Venezuelan Poodle Moth
Fig. 26: The in-comic character variations based off of the Poodle Moth
The environment design of the world space heavily referenced science fiction tropes like
sliding doors, keypads, with the walls and ceilings very angular and streamlined. The futuristic
environments also allowed for augmentations without stretching the disbelief too far.
41
Fig. 27a-b: The original panel (left) and the augmentation for that panel (right)
These augmentations are meant to work within the story world not against it, so the panels had to
be set up in a way to show an empty state where the augmentation would go. In this example
case, it is the blank computer screen that lights up when the image is recognized by the
application.
The creation of the pages was the most time consuming process of making the whole
comic. Each page had to go through several steps, and in order to keep a cohesive visual
appearance from page to page the same process had to be followed. This effectively means that
each page had to be drawn two different times, before being colored in. The augmented reality
programming could not begin until the pages were finished because the pages were the markers,
so this gave the project a very strict set of directions to follow. Thankfully the script was readily
available to work from, so this section of work was only translating written descriptions to a
visible depiction of the words.
42
Fig. 28a-d: Rough lines (Top left) Inks (Top Right) Color (Bottom Left) and Text (Bottom
Right)
The more pages that had been made, the more care had to be taken about maintaining color
accuracy because every time a new character or location was added, the colors had to remain
consistent across the rest of the comic. This meant that every time Artace or Galra were drawn,
the colors would be exactly the same and the coloration would be exactly the same between page
43
four and page twenty. After all of the pages were finished, the artwork for the augmentations had
to be made. This took longer for some panels than it did others, because some panels had more to
change. For example, the augmentation on page five was the addition of nightclub style lighting
and adjusting of the color curves. This only took a few minutes with the gradient tool and Curves
adjustment layer. On the other hand, pages that were an entirely different panel, like on page
seventeen, took just as long as making one of the regular panels.
Fig. 29a-b: Original panel for page five (Left) and augmentation for it (Right).
44
Fig. 30a-b: Original panel for page seventeen (Left) and augmentation for it (Right).
Not all of the panels had visual augmentations but rather had audial augmentations, and
this presented a fresh challenge. Making the music from scratch would have been time
consuming and difficult, since production level tools are not readily available. As an audio
resource a royalty free sound library (FreeSFX.co.uk) was used in this project, so professional
quality sounds were available. This audio was used in several augmentations in tandem with
image changes to create key points by combining attention methods. Some of the audio did have
to made myself, but those were the specific plot-related audio clips. With creative post-
processing, sound clips for Artace, Galra, and a newscaster were created that all sounded
different and unique.
The creation of this comic, as with any comic, had to follow a series of steps. First the
story had to be written, and then the artwork had to be made. The artwork had several steps as
well, beginning with finding references, then the concept art, then sketches, inks, and colors for
every page. After that the dialogue was added, and this would be the step that most comic books
45
are completed, but since this comic has augmented reality functionality the final step in this
process was the augmentation. The visual augmentations followed the same steps as one of the
regular panels, and the audial augmentations only had to be found and prepared to be
implemented. The final step was to link the augmentations to the images.
Designing the Augmentations
With the comic and its augmentations made, the final step in the process was to create a
Vuforia database that could be used with Unity. There are many programs that support Vuforia
use, including Georgia Tech’s own Argon program. Each option had its own set of positives and
negatives, but ultimately Unity was chosen because it had thorough documentation for Vuforia
use, had public online forums for the discussion of problems and troubleshooting, and allowed
for offline use. Unity also afforded the ability to create single stand-alone apps that had their own
individual icon, as opposed to Argon which required entering the application and then loading a
bookmarked webpage that contained the program.
The database in Vuforia was where the artwork for Pilgrimage hit its first stumbling
block. While uploading the marker panels, most of them were ranked four and five star quality
but a few were ranked either one or zero star. What this means in the context of Vuforia was that
some of the images would be much more difficult to recognize as triggers than others. These
recognition points are created at the intersection of lines, where the markings are clear and can
be recognized consistently and repeatedly over the lifetime of the application. Lower ratings
mean that the image could potentially take longer to recognize, not be recognized at all, or flicker
while it’s being shown onscreen.
46
Fig. 31a-b: The original image (left) and the Vuforia target recognized points (right) of a
five-star image trigger.
Fig. 32: A Vuforia point map of a zero star image trigger.
This hurdle taught that Vuforia did not create recognition points in areas with gradients,
or soft lines. Upon noticing this pattern, going back and changing a few key points in the failing
images brought them up from zero to one star, making them usable. According to the Vuforia
47
website, for basic image recognition one star will suffice. Their grading system also works on
three-dimensional shapes, and three stars are the minimum recommended amount for any object
recognition that is not just a flat plane. So the decision was made to at least try and see if they
would be recognized at one-star quality, fully aware that if they were not these trigger panels
would have to be redone. Thankfully this was not the case, because even one-star graded images
were easily recognizable by the app in this context, meaning that the next phase of the project
could begin.
Incorporating the Augmentations
This phase was where the project moved from Photoshop to Unity, which was potentially
the greatest limitation going into the project. While Unity and Unreal are both established game
engines, Unreal Engine Four (UE4) is used widely in game development for games like the
“Gears of War” series and “Kingdom Hearts III” where Unity was used for games like “A Night
in the Woods” and “Super Mario Run”. In addition, Unreal Engine’s Blueprint coding workspace
meant that code never had to be typed out longhand, and the flow of data inside the scene was
illustrated, making it easier and faster to figure out where the problem in the code was. Unity
uses either C# or Java to program, with each section of code existing within its own text file
rather than a shared space within the engine. The code cannot be edited inside of Unity, but
instead must be opened in an external program, altered, and then returned to Unity.
Despite this, Vuforia was the limiting factor and only Unity had an official plugin for it.
There was a third-party plugin for Unreal Engine Four (UE4), but the risk of using a potentially
unstable third-party product was too great to take. In their developer portal Vuforia has produced
a plugin explicitly for use with Unity, which contains premade objects for an AR camera and
image recognition target objects. This accelerated production, removing the steps of creating a
48
virtual scene that would be necessary if using Argon. On their site they made a very clear set of
step by step instructions on how to set up the scene, and how to link digital artifacts to the trigger
images. The image triggers were a premade object within the Vuforia package, so all that had to
be done for them was put them in the scene. Since the device camera was controlling the scene
camera, it did not matter where the triggers were placed. They were ultimately put in a grid
pattern, simply to keep track of which pages corresponded with what images without having to
reference back to the full page files.
Fig. 33: Unity project scene with the twenty-six trigger images
To link an artifact to the trigger, it has to be a child object of that image trigger. Then the
Vuforia package code creates that object in relation to the image when it is recognized by the
device camera. To test the recognition, cube primitives were created as child objects of every
image, centered on the trigger. This was the fastest and easiest way to see if any panels would
fail recognition and need to be redone. Thankfully this was not the case, and all of the image
49
triggers functioned as they were supposed to. During this step, small deviations began appearing
in the functionality of Vuforia’s image recognition from a screen versus from a paper copy. The
camera tries to adjust for brightness, and since digital screens produce their own brightness the
light contrast is much higher than reading from a paper, and in certain panels that had high
brightness contrast the camera struggled to read the image. The first panel consistently took
much longer to register than any other panel when testing between the desktop and an iPad, but
once it was printed out on paper it registered much faster. Since the final product was always
meant to be a tangible book this was acceptable, but it would hinder others sharing the comic as
a digital file and viewing it from their computers. Even if the file were shared, it would not
function optimally until it was printed out onto paper.
As a game engine, Unity is easily used for mobile development since it does not require
complex rendering or ray casting, something most phones cannot handle. Vuforia recognizes
this, and has produced a toolkit specifically to be used in the development of AR apps within
Unity. This expedited several processes in setup that would have been arduous in other
programs, like establishing an AR Camera and identifying the individual image triggers. Testing
with basic primitives, Vuforia was able to recognize each image trigger and anchor the cubes
correctly, which meant that the real augmentations could be connected.
CHAPTER 4. Implementation
Implementing Augmentations
The basic augmentations were done first, which were the image to image panel changes.
An object plane had to be created as a child object to the trigger panel, and then textured with the
augmented differences. Starting with the image to image transitions allowed for the conventions
50
of Unity to be picked up slowly, so that one thing could be worked on at a time in order to not be
overwhelmed by learning a new user interface, new coding language, and new naming
conventions. By the time these panels had been completed there was a routine, and the next step
up in difficulty could be taken.
The animated panels were the logical procession from image-to-image transitions, but
thankfully Unity has an animation editor that is similar, although simplified, to Maya’s; a
program made by Autodesk. It creates the animation file (.anim) as well as the controller
automatically, and it combines the keyframe creator and the curve editor into the same window
to streamline the process of animating and refining. Since the animations were only image planes
the changes were not as complex as a full character model, limited to translation and rotation.
Keyframe style animation is the same across platforms, so there was no need to learn an entirely
new animation toolkit.
Adding sound to panels was the first major stumbling block. In the Vuforia
documentation, it said to create an audio source parented to the image trigger to create sounds.
While this worked, technically, it played the sounds as soon as the scene started rather than on
image recognition, which meant that as soon as the app began the sound began playing. It was
playing successfully though, so it was only a matter of starting and stopping the audio. A
separate script had to be created within Unity to hold the audio file, rather than linking it to the
image trigger itself. Upon testing, the audio played exactly as intended on recognition, but only
played once. With the shorter clips, like the alarm clock and electricity crackling in pages eight
and twenty-four respectively, the sound only lasted a few seconds before cutting off. Unity
included a ‘Loop’ function in the audio object, and after turning that on the sound played for as
long as the reader wanted to stay on that panel. The decision was made not to loop the three
51
dialogue boxes because it would not diegetically make sense for that audio to be looped. If the
reader wants to hear it again they can re-scan the image and hear it, but they do not automatically
loop.
The last group of augmentations to create was the interactive ones. Some of these
required more complexity than others because the interactions were more in-depth. The first
hurdle that needed to be passed was figuring out how to touch and drag. In Unity, the tag
OnMouseDown() is also used for touch on devices so the code could be tested on the computer
rather than having to upload the file to the designated iPod every time there was a new iteration.
The first interaction that was coded was the augmentation on page fourteen of the comic; moving
the lid of a crate out of the way. There was no spatial recognition or variables to be coded, but it
was moving an image plane out of the way. In the original comic, Istochnik, the object was
outright destroyed on touch, but for this panel there needed a build up to the reveal. By dragging
the lid off it slows the reveal compared to the lid simply vanishing from existence.
Fig. 34a-b: The lid (left) and the crate that it obscures (right)
52
The next interaction required spatial recognition, because it was a key-card unlocking a
cell. The card had to move over the reader in order to unlock it, and the OnOverlap() or
OnIntersect() tags felt clumsy and imprecise. Instead the key card’s x and z coordinates were
used to create a logic statement of “if (transform.position.x >= “boundaryLeft” &&
transform.position.x<= “boundaryRight” && transform.position.z >= “boundaryBottom” &&
transform.position.z <= “boundaryTop”). This if statement would only be true if the key card’s
center met these requirements. The coordinates from the image of the card reader were then
established, using the dark square as the boundaries.
Fig. 35a-b: The card reader as it begins (left), and after it recognizes the key card (right).
This presented a problem, however, because the coordinates of the card after recognition
did not match the world space coordinates in the editor scene, and would not trigger the change.
The code worked correctly and was then tested by putting the transform.location statement on
either side of the equation and it triggered immediately. The proper set of coordinates to use had
to be determined, and then implemented. So in the key card’s drag script, two small but useful
lines of code were added at the end: Debug.log(transform.position.x); and
53
Debug.log(transform.position.z);. This would print out the coordinates that were needed for
recognition. So the program was initialized and the card was dragged to the correct boundaries,
giving the printed values to create a proper if statement. After these values were entered the card
scanner worked immediately and work could begin on the keypad puzzle panels.
Two different panels both use keypads, so a lot of the code was able to be reused between
them. The biggest change was the name of the global variables, and the root script that held
them. Also the second keypad was a decoding game, but the first one was right or wrong since
the code is provided for the reader to get them used to the interaction, so the second one changes
the position code to yellow if the number is in the combination, but in the wrong place. Invisible
geometry was used to create trigger volumes for each button on the keypad, then created a string
to hold the code and an integer to hold the number of numbers in the code. The number keys
only work if the string is less than the maximum code value, at which point the user can only use
the escape and enter keys. Escape clears the numbers regardless of their value, but Enter checks
to see if the string is correct. If the string is correct then Enter destroys the keypad volumes as
well as the first keypad to display the success screen, and if not it resets the string and values.
This was the longest designed interaction, and so it only occurs twice within the comic. The first
time was used as a teaching moment since the code is provided and the player can see what is
correct and what is incorrect, and then the second time is the actual puzzle interaction.
The code for this part was tedious and complex, because each key was almost the same
except for changing the numerical value, and it created a lot of scripts because each key needed
its own to house its own individual number. The keypads were created quickly, the biggest
problem was how to indicate the correct values. 3D primitive objects were used to give the pad
depth, and ultimately using spheres as color indicators seemed the most successful because they
54
were reminiscent of the LED bulbs in an actual indicator pad. The recognition code for this was a
layered if statement, where it checked the number of characters in the combination string, and if
the button pressed was correct it turned green, and if it was incorrect it turned red. The second
keypad followed the same conventions, with the addition of an Else If step that checked if the
button pressed was anywhere in the combination. If it was in the combination but in the wrong
place, the sphere turned yellow. The color changes were handled using the code
“GetComponent<Renderer> ().material.color = new Color (R, G, B);”. Each step in the If
statement during Void update() sets the code, and the Void Start() phase sets the base color as
blue. This is also the color that the spheres reset to when Esc or Ent are pressed if the
combination is incorrect, indicating that the values have not been entered.
55
Fig. 36: The code keypad upon initialization
Fig. 37: The code keypad showing the incorrect combination. Yellow indicates the right
number but wrong space, red indicates the wrong number, and green indicates the right
number in the right space.
Fig. 38: The success screen, which is the right combination after pressing enter.
56
Implementing the augmentations was successful in terms of the technology because they
all worked correctly, and the augmentations matched the initial script so the implementation was
also successful in terms of the diegetic story world because they fit in with the plot rather than
being extraneous. While creating the artwork was the most time consuming part, writing and
implementing the code took more thought, because variables had to be set and called, and
animations had to run. Numbers had to be dealt with precisely in order for changes to happen
correctly, and with the interactions on the number pads permutations had to be calculated to
show the correct sequence of colors on the lights. Once the application was working there was
only one step left, to test the comic with people who had not seen it, and make changes
accordingly.
Testing The Comic
Once the augmentations were working and the application had been installed onto a
device, this prototype was tested with a group of participants unfamiliar with the project in order
to get feedback. They were handed the application and the completed comic on paper, and left
them to their own devices. They were not told that the green bordered panels were the
augmented ones, because there was an explanation on the inside cover on how to use the app. In
comics, traditionally the inside cover is used to hold the credits, publisher information, copyright
information, and a brief synopsis of the plot so far. With Pilgrimage, this was the first issue so
space was available to insert a few sentences on how to access the augmentation. In most cases
the readers assumed that the green panels were the augmented ones right away even without
reading the first page, but three of the eight people in the group took several pages to figure that
out so in the final draft of the comic the explanation of the augmented function of the comic was
57
put inside of the same green border to set the connection immediately without having to use the
mobile device.
During the initial testing, it was noticed that their fascination with the augmented panels
kept them on the panels through several loops of the animation. While this was unexpected, it
actually evened out the time that they spent reading the comic and looking at the augmentations.
While testing the Vuforia triggers there was concern that the loops were too short and the
attention would be focused too heavily on the comic with an uneven ratio, but this was from the
perspective of someone who had written the comic and knew what the triggers were. Since
others who were reading the comic knew nothing about it and were experiencing it for the first
time, they were engrossed in the augmentations and went through several loops. While these
changes were being planned out consultation began on how to transition back from digital to
tangible, but because of the nature of comic panels this seemed unnecessary. Comic panels are a
section of time, but the reader can look at them for as long as they want. Some people read
comics for the story and don’t pay much attention to the pictures, but others may pay more
attention to the pictures than the plot. By limiting how long each augmentation lasts it would
artificially shorten the panel, making the reader have to re-scan the panel. With this in mind the
decision was made not to have the three audio dialogue panels loop because during testing, the
participants commented that the audio began to bleed together and was blurring the ending and
the beginning together. Since the dialogue is diegetic as part of a conversation, as opposed to a
beeping alarm or thumping club music, it would make sense for it not to loop.
What was also noticed was that the augmentations did not evenly split attention, which
was mentioned when outlining the four means of holding attention. It was a surprise, however,
when it was audio that held the least amount of attention. The dialogue held attention but the
58
looped audio did not hold attention for more than a couple seconds, which is less than the looped
animations. The interactive keypad panels held attention much longer because it was a puzzle to
be solved, and the participants visibly reacted to their success. The first keypad is supplied with a
combination on the same page, setting the expectations and controls for when it reoccurs later in
the book. The participants reacted to this with nods, and general words of understanding. They
saw the four spheres, and saw the four numbers on the page, and assumed that they went
together. Later in the comic they have to figure out the combination themselves, which took an
average of several minutes and garnered frustration from the participants, but a bigger reaction
upon success. None of the testers would move on from that page until they had discovered the
combination, and even though the only reward is a green light they still reacted physically to
their success. There are many more comic panels than there are augmented panels, so the
attention levels balance out. Overall the people that got to try it did spend more time reading the
comic than looking at the augmentations, but not by too wide of a margin.
After observing the test group, a few small changes had to be made to the app but the
design criteria for the project were satisfied. A full length single issue comic book was made, and
a digital application was made to go with it. The app augmented designated panels of the comic
book, with panel gutters serving as a transition marker as opposed to an icon inside the panel like
Marvel did with theirs. Using the methods taken from research into attention manipulation as
well as digital interaction a wide variety of augmentations were able to be made that held
attention, but did not feel repetitive according to the test participants. Attention was split
approximately equally, on average between the participants.
59
CHAPTER 5. Conclusion
At the beginning of this project, there were two concrete goals in mind. To create an
augmented reality experience that would split the interactor’s attention evenly between the
digital and tangible mediums, and to create AR markers that were more visually interesting than
a typical QR code. Graphic novels were the ideal medium to combine artwork and augmented
reality because the panels serve as a way to divide time and attention into segments for the
reader. The artwork itself would be accomplished without relying on an outside artist. Video
games were studied as a means of directing attention to an interactor within a digital space,
concepts that were transferable to the AR app that was developed. A basic proof of concept was
done first using premade comics to see if comic panels were even usable as AR image trackers.
This test was successful, and progress continued to making test pages of original artwork to see
if they would be as effective. The first idea worked thematically as a graphic novel, but did not
lend itself to augmentation so it had to be scrapped in favor of a new story that would be more
cohesive with augmentation. A script was also written the second time and completed before the
artwork even began, rather than starting in Photoshop and writing the story page by page as they
were completed. Because of this the artwork could be planned out much more efficiently, and
ended up with a product that fulfills the initial premise of the project successfully.
60
Fig. 39a-b: The first comic Istochnik (left) and the final comic, Pilgrimage (right).
The risk of using Unity for the first time was higher than would be preferable, but
thankfully as a program it is well documented and has a downloadable plugin for Vuforia, which
is a commonly used image recognition program. Because of this the documentation was
indispensable and allowed for troubleshooting problems as they arose. The largest perceived
difference between Unity and other game engines was the syntax of the specific language, and
now that the syntax has been learned if another similar project had to be made, the coding could
be completed in probably half the time.
Vuforia, on the other hand, was fascinating to work with and is so open ended that it
allows for almost endless variation. The project shows that more than just QR Codes can be used
for image recognition, full artwork can be as well. With a few exceptions, meaning the panels
61
that failed the Vuforia grader, the comic panels functioned naturally as image triggers without
having to consciously think about that during their creation. In fact, while testing Vuforia with
comic pages at the very beginning using premade comic panels from Transformers: More Than
Meets The Eye, all of those panels rated five stars in the Vuforia grader. This theme of
artistically minded triggers over purely utilitarian triggers can be expanded upon into educational
materials, street art, or just larger scale graphic novels. The pages of Pilgrimage made for this
project amount to one issue in a series, which average between twenty-five and thirty pages not
including advertisements. This comic hits in the middle ground at twenty-six pages, and has a
self-contained story that does not require any outside information to be enjoyed.
Jessica Burns was sent a copy of the comic once it was finished, and we sat down in a
Google Hangouts session to talk about what worked and what didn’t. Her first comment was on
how much the artwork improved comparing the first pages to the later pages, which is
subjectively true. There had been little practice when the project began, and working for eight
hours a day on drawing improved the art drastically, so the end of the comic doesn’t quite match
the beginning under scrutiny.
62
Fig. 40a-b: Page four (left) had stiff lines, and no real character interaction. They were
drawn as poses rather than as characters. Page twenty-three (right) shows motion and
interaction while fighting, rather than just stiff posture.
Another major critique of hers was the lack of depth in the comic’s artwork. All of the
colors were flat, meaning that there was no lights or shading on the characters or the
environment. The comic should be shaded to some extent without question, but that is a time
consuming process in sequential art because the shading needs to be consistent across panels.
This is a similar concept to keeping colors consistent, but then each character has an additional
set of shades for each color on them so the decision was made not to shade the panels. It did not
impact the Vuforia recognition, which was the goal to begin with. For the purposes of this
project, the artwork did not need to be masterful, and would have taken two or three times longer
to make if the pages were to be fully shaded and detailed. Instead a less realistic art style was
63
used, though it was more complex than Istochnik because it was still effective. Mainstream
comic books are not made by one person, generally they are made by teams of workers who have
one designated task. They have writers, pencilers, inkers, and colorists, not to mention letterers
which were circumvented by using a font rather than hand lettering every dialogue box. In
hindsight, this was one of the bigger risks of this project. It was established that comics would
work as triggers based off of experimentation using Transformers comics, but by undertaking all
of the steps of production by one person it was reducing the work of roughly eight people on a
professional comic book to one person on an amateur one. A script was written, illustrated, and
augmented, all in just under three months; something that most people would have deemed not
possible, or not likely, for one person to accomplish. Not only that, but the visuals and
augmentations were completed almost exactly how the script established them without having to
sacrifice any pages to save time on the artwork or on the coding side.
The Next Steps
In the end, this project was a success in that a problem was identified, and then an
intervention for that problem was developed. This is not the end of the project, because there are
several things that could be done in the future to improve it. First of all, the amount of content
would almost have to be changed. There would not be one augmentation per page, rather one
every few pages, but the augmentations would be more meaningful. While they diegetically
work within the story, there were several pages of the comic that had augmentations simply
because it had been set as an arbitrary rule that every page had to have one augmentation. This
would only work on a larger scale though, so it would either have to be a fully made graphic
novel like “The Watchmen” or “Persepolis”, or a syndicated series like the publications that
Pilgrimage was modeled after. That way the same amount of augmentations would be provided
64
but they would be meaningful and, more importantly, better spread out. In this comic, each page
was augmented so it became expected and routine. But if the pages were augmented irregularly,
then it would always be a surprise when the next one came. This is not to say that the comic is
not successful, only that the augmentations would be more effective if they were less frequent. If
there were only one augmentation every two or three pages, it could be saved and used when it
would be more thematically appropriate. This would create focus on it, since it would be less
common than it is now.
In the next iteration of this comic the artwork would be done in higher quality, with full
lighting and shading on every panel. This goes back to the conversations and consultations with
Ms. Burns, in making the artwork more professional and finished. Comics that have lineart and
color but no shading need to be heavily stylized so the reader knows it is a conscious choice,
rather than an inability to do so.
Fig. 41: Bryan Lee O’Malley’s series “Scott Pilgrim vs The World”
65
Fig. 42: An example of Ms. Burns’ work, in her online comic Pest Caravan.
Scott Pilgrim is highly stylized as a comic, and it uses lots of motion lines to fill up the
otherwise empty space that would have been occupied by shading detail. It works because of that
exact reason, because it is not just line art. Pilgrimage has lineart and color, but no real level of
detail beyond the lineart. Shading in the comic would be time consuming, to be sure, but it would
make it much more presentable to a higher caliber audience.
66
Fig. 43a-b: An idea of how even just character shading can improve a comic.
In addition to this full character breakdown drawings would have to be done, so that
reference would be readily available as to what the characters looked like from the front, sides,
and back. While making the artwork for Pilgrimage, this was one of the largest time investments
because whenever Artace or Qaalen appeared in the comic at an unusual angle, it would require
finding reference poses as well as figuring out what their character would look like from that
angle.
67
Fig. 44: Pest Caravan character study for protagonist Amber, courtesy of Ms. Burns.
Even though these are rough sketches of the character, it gives just enough definition that
the artist can take them and work with them. These are similar to the augmentations that Marvel
had in their app, where if the user scanned a panel it would show character breakdowns like this
one for that page. In just this breakdown, Ms. Burns shows her character’s face from four
different angles, the scale plating, the larger armor plating, a range of emotions, and an enraged
form.
Almost all of the changes that would be made would be cosmetic, making the artwork
more professional and detailed. The exception would be the dispersal of the augmentations
throughout a larger body of work, rather than concentrating twenty-six augmentations within
68
twenty-six pages of comic book because it did start to feel unnecessary in a few panels according
to the preliminary testing. The rule was set that there should be one augmentation per page, and
that rule was adhered to for the sake of the process. Overall this project is successful when
compared to its initial intent, especially since it all works the way that was intended. The gutters
divided attention, and by stylizing the gutters of the comic attention was able to be guided
roughly evenly between the comic book itself and the application.
69
Works Cited
Banks, Adam and Tom Fraser. Designer's Color Manual: The Complete Guide to Color Theory
and Application. San Francisco: Chronicle Books, 2004.
Billinghust, Mark, Adrian Clark, and Gun Lee. “A Survey of Augmented Reality”. Functions and
Trends in Human-Computer Interaction, vol. 8, No. 2-3, pp. 73-272. 2015. https://files.t-
square.gatech.edu/access/content/group/gtc-a168-6c57-54b5-9980-762e61a1c021/gtc-
71de-5ab7-500a-b57f-75ac1ce94218/Readings/Billinghurst.pdf
Burns, Jessica. Pest Caravan. Tapastic. Aug. 31, 2016. https://tapastic.com/episode/444271
Buscema, Joe and Stan Lee. Stan Lee’s How To Draw Comics: The Marvel Way. Simon and
Schuster; New York City, 1978.
Cavanaugh, Patrick. "Attention-Based Motion Perception." Science, vol. 257, no. 5076, 1992,
pp. 1563, Advanced Technologies & Aerospace Database; Materials Science &
Engineering Database; ProQuest Central; ProQuest Environmental Science Collection;
Research Library,
http://prx.library.gatech.edu/login?url=http://search.proquest.com/docview/213542961?a
ccountid=11107.
Chan, Phillip A. and Terry Rabinowitz. “A Cross-Sectional Analysis of Video Games and
Attention Deficit Hyperactivity Disorder Symptoms in Adolescents”. Annals of General
Psychiatry, 5:16, 2006. Web. Accessed Sept. 5, 2016. http://annals-general-
psychiatry.biomedcentral.com/articles/10.1186/1744-859X-5-16
Clark, Nicholas. Flower. ThatGameCompany, 2009. Playstation 3.
Eisner, Will. Comics and Sequential Art: Principles and Practices from the Legendary
Cartoonist. W. W. Norton & Company; London, 2008.
Elliot, Andrew J. and Markus A. Maier. “Color Psychology: Effects of Perceiving Color on
Psychological Functioning in Humans”. Annual Review of Psychology, 2014. P95-p120.
Hill, Sefton. Batman: Arkham City Armored Edition. Rocksteady Studios and Warner Bros.,
2011. Wii U.
Ito, Junji. Gyo Ugomeku Bukimi. The Enigma of Amigara Fault. Big Comic Spirits; 2001. Viz
Media; 2003.
Johnson, Addie and Robert W. Proctor. Attention: Theory and Practice. Thousand Oaks: SAGE
Publications, 2003.
Levine, Ken. Bioshock. Irrational Games, 2007. Playstation 3.
McCloud, Scott. Reinventing Comics: How Imagination and Technology Are Revolutionizing an
Art Form. HarperCollins; New York City, 2000.
McCloud, Scott. Understanding Comics: The Invisible Art. HarperCollins; New York City,
2000.
70
Miyamoto, Shigeru. The Legend of Zelda: Ocarina of Time. Nintendo, 1998. Nintendo64.
Miyamoto, Shigeru. The Legend of Zelda: Wind Waker HD Edition. Nintendo, 2013. Wii U.
Pashler, Harold. Attention. Abington: Psychology Press; 1998.
Reid, Francis. The Stage Lighting Handbook. New York: Routelage, 2001.
Schofield, Glen and Matthias Worch. Dead Space 2. Visceral Games and Electronic Arts, 2011.
Playstation 3.
Senn, Chris. Sonic Boom: Rise of Lyric. Sega, 2014. Wii U.
Vargus, Nic. “Nintendo Wii U Review (2012)”. IGN, 2012. Web. Accessed Sept. 5, 2016.
http://www.ign.com/articles/2012/11/21/wii-u-review
Whedon, Joss. Astonishing X-Men. Issue 1. Marvel Comics; New York. 2004.
Yan, Su and Magy Seif El-Nasr. “Visual Attention in 3D Video Games”. Penn State University,
2006. Web. Accessed Aug. 25, 2016. http://dl.acm.org/citation.cfm?id=1178849