Game Scoring - Towards A Broader Theory PDF
Game Scoring - Towards A Broader Theory PDF
Scholarship@Western
4-16-2015 12:00 AM
Supervisor
Jay Hodgson
The University of Western Ontario
Recommended Citation
Enns, Mack, "Game Scoring: Towards a Broader Theory" (2015). Electronic Thesis and Dissertation
Repository. 2852.
https://ir.lib.uwo.ca/etd/2852
This Dissertation/Thesis is brought to you for free and open access by Scholarship@Western. It has been accepted
for inclusion in Electronic Thesis and Dissertation Repository by an authorized administrator of
Scholarship@Western. For more information, please contact wlswadmin@uwo.ca.
GAME SCORING: TOWARDS A BROADER THEORY
by
Mack Enns
“Game scoring,” that is, the act of composing music for and through gaming, is
distinct from other types of scoring. To begin with, unlike other scoring activities, game
scorer’s choices are thus first-and-foremost limited by available gaming technology, and
the “programmability” of their musical ideas given that technology, at any given
historical moment. Moreover, game scores are unique in that they must allow for an
unprecedented level of musical flexibility, given the high degree of user interactivity the
video game medium enables and encourages. As such, game scoring necessarily
constitutes an at least partially aleatoric compositional activity, the final score being
demonstrates this through case studies of the Nintendo Entertainment System sound
hardware configuration, and game scores, including the canonic score for Super Mario
Bros. (1985).
media.
ii
Acknowledgements
I wish to give many thanks to my advisor, Jay Hodgson, for his insight,
encouragement, patience and guidance in writing this thesis. I wish to thank my second
reader, Norma Coates, for her valuable insight and wisdom in this endeavour. Thanks to
my mom, my dad, my brother and my sister for listening to me rant, and for all their love.
Thank you to my partner, Jessie Wong, for supporting me in countless ways, and for
putting up with all the NES music. Finally, thanks to Jordan Pearson for introducing me
iii
Table of Contents
Abstract ii
Acknowledgements iii
Table of Contents iv
Chapter 1 1
Introduction 1
Chapter 2 28
Introduction 28
Configuration 31
iv
Channel Overview: Noise Channel 42
Chords 50
Chapter Summary 58
Chapter 3 59
Introduction 59
Conclusion 84
Chapter 4 86
Summary 86
Future Directions 89
v
Significance 90
Implications 92
Conclusion 94
Bibliography 96
vi
List of Tables
Table 6: Famicom Sound History Series: Mario the Music (Partial Track List) 82
vii
List of Figures
viii
1
Chapter 1
Introduction
scoring,” that is, composing music for and through gaming. As with film scoring, game
scoring supports, complements and elucidates the visual aspects of a broader narrative
medium, namely, the medium of video games. However, any further resemblance
between game scoring and film scoring is illusory, despite a glut of research which treats
A host of technical and aesthetic priorities, values, obstacles and concerns faces
the game scorer, which simply do not influence composition for other media. Most
their compositions. Game scores are only ever realized — they only ever exist as
something other than imperceptible digital bits — through gameplay, after all. So, too,
must game scorers compose for particular sound hardware configurations. Their
compositional activity is structured in its entirety, then, by the very limited set of
aesthetic possibilities each console’s peculiar Audio Processing Unit (APU) affords. If it
1
Work that approaches game scoring via film studies tools and concepts includes, but is not limited to:
Arrasvuori (2006); Boyd (2003); Chan (2007); Collins (2007d; 2008c); Hoover (2010); Jørgensen (2004;
2006; 2007a; 2007b; 2008b; 2009; 2010); Kamp (2014); Mera (2009a); Munday (2007); van Elferen
(2011); Whalen (2004; 2007); Wilhelmsson and Wallén (2010); Wood et al. (2009); and Zehnder and
Lipscomb (2004; 2006), among others.
2
cannot be programmed into a console’s APU, it simply cannot exist as part of a game
between music composition and game scoring is only possible if one insists on
be.3 To see game scoring as no less musical an artistic activity than, say, composing a
symphony, analysts need only adjust their analytic prejudices to include a host of modern
activities that may indeed appear unmusical on first glance, such as software
this thesis that game scores are somehow aesthetically equal to the ostensibly “great”
works of the Western Art Music canon, even if I prefer Koji Kondo to Beethoven most
days. My argument is that game scoring comprises a unique compositional mode which
is structured entirely by gaming technology and, thus, which remains always inherently
2
Work which relates game scoring to software programming more often comes from the computer science
discipline, such as: Aav (2005); Alves and Roque (2011); Baccigalupo (2003); Berndt (2009; 2011); Berndt
and Hartmann (2007; 2008); Berndt et al. (2006); Boer (2003); Borchers and Mulhauser (1998); Childs IV
(2007); Collins et al. (2011); Cunningham et al. (2011); Droumeva (2011); Ekman and Lankoski (2009);
Farnell (2007; 2011); Fay (2004); Friberg and Gärdenfors (2004); Grimshaw (2007; 2010); Grimshaw and
Schott (2007; 2008); Grimshaw et al. (2008); Havryliv and Vergara-Richards (2006); Hoffert (2007);
Holtar et al. (2013); Hug (2011); Huiberts (2010; 2011); Jørgensen (2004; 2006; 2007a; 2007b; 2008a;
2008b; 2009; 2010); Kayali (2008a; 2008b); Kayali et al. (2008); Knight (1987); Lendino (1998);
Lieberman (2006); Liljedahl (2011); Marks (2009); Mullan (2010); Murphy and Neff (2010); Nacke et al.
(2010); Pichlmair and Kayali (2007); Reiter (2010); Roux-Girard (2009; 2010); Sanger (2003); Sanders
(2010); Stockberger (2003); Tinwell et al. (2010); Toprac and Abdel-Meguid (2010); Villareal III (2009):
Weske (2002); Whitmore (2004); Wilde (2004); Wilhelmsson and Wallén (2010); and Wooller et al.
(2005), among others.
3
Since scoring is a form of music composition, I treat game scoring as music composition in this thesis.
Here, however, I differentiate my usage of “composition” with narrower uses of the term which do not
allow for activities involved in game scoring. For example, see: Demuth (1951); Dunstan (1909); Eisler
(1951); Ouseley (1886); and Perle (1963), among others.
3
video game music are plentiful, of course, but these tend to consider game scoring as an
offshoot of film scoring, and they seldom address the technical structure undergirding
game scoring and the “aleatoric” nature of the musical works that structure enables.5
4
Ludo-musicology is a relatively new field of research which focuses exclusively on video game music,
which differs from music composed for the non-performative visual arts or non-interactive film, for
instance. For more on this see: Ludomusicology.org.
5
Ludo-musicological studies of video game music include, but are certainly not limited to: Belinkie
(1999); Berndt (2009; 2011); Berndt and Hartmann (2007; 2008); Berndt et al. (2006); Brame (2011);
Bridgett (2010); Chan (2007); Childs IV (2007); Collins (2005; 2006; 2007a; 2007b; 2007c; 2007d; 2008a;
2009); d’Escrivan (2007); DeCastro (2007); Deutsch (2003); Farnell (2007); Fay (2004); Furlong (2004);
Gibbons (2009); Herber (2008); Hermans (2013); Hoffert (2007); Kaae (2008); Knight (1987); Lendino
(1998); Lieberman (2006); Marks (2009); Mera (2009a; 2009b); Sanger (2003); Schmidt (1989); Sweeney
(2011); Van Geelen (2008); Whitmore (2004); and Wooller et al. (2005).
6
For examples of musicological studies of game scoring, see: Allouche et al. (2007); Arrasvuori (2006);
Arsenault (2008); Baxa (2008); Belinkie (1999); Bridgett (2008); Carlsson (2008); Cassidy (2009); Chan
(2007); Collins (2005; 2006; 2007a; 2007b; 2007c; 2007d; 2008a; 2008b; 2008c; 2009; 2013); Collins et
al. (2007); Crathorne (2010); d’ Escrivan (2007); Deutsch (2003); Fritsch and Strötgen (2012); Fritsch
(2013); Furlong (2004); Gibbons (2009; 2011); Guerraz (2008); Herber (2008); Hermans (2013); Jørgensen
(2008b); Kaae (2008); Kamp (2009; 2014); Kärjä (2008); Mera (2009a; 2009b); Miller (2007; 2008a;
2008b; 2009); Munday (2007); Pichlmair and Kayali (2007); Reale (2011); Schütze (2008); Shultz (2008);
Smith (2004); Summers (2011); Svec (2008); Sweeney (2011); Tessler (2008); van Geelen (2008); Western
(2011); Whalen (2004; 2007); Wood, Harper, and Doughty (2009); and Youngdahl (2010), among others.
7
For more work on game scoring from media studies, see: Jones (2008); Thornham (2013); Waggoner
(2009); Wolf and Perron (2013); Crawford, Gosling and Light (2013); Domsch (2013); Egenfeldt-Nielsen,
Smith and Tosca (2013); Ensslin (2011); Gamboa (2012); Garrelts (2005); Huntemann and Aslinger
(2012); Juul (2011); Ruggill and McAllister (2011); Whalen and Taylor (2008); Paul (2012); and Newman
(2012); among others.
4
sociologists,8 and a host of researchers working in related disciplines,9 but these scholars
tend to consider game scoring vis-à-vis film scoring, rather than on its own terms, and
they very often neglect the actual compositional procedure of game scoring per se.10 As
such, I offer the following case study — specifically, of Richard Vreeland’s game score
for the 2012 Xbox Live Arcade “puzzle-platform” game FEZ — as a means of
elucidating and concretizing the theoretical terrain I intend to cover in this thesis. This
theoretical terrain is drawn largely from work by Zach Whalen (2004) and Karen Collins
(2013), both of whom pose salient questions about the interactive nature of game scoring.
The interactive element of videogames requires its own analysis […] Cognitive
theories of perception and questions of immersion versus engagement as a means
of understanding ‘flow’ or pleasurability in games allows for a richer
understanding of the complex communication involved in videogame music.11
8
For examples of sociological research on game scoring, see: Kohler (2005); Crawford, Gosling and Light
(2013); Miller (2007; 2008a; 2008b; 2009); Schmidt (1989); Smith (2004); Svec (2008); and Demers
(2006), among others.
9
For example, much research on video game music comes from a relatively new field called interactive
sound studies. For more on this see: Collins (2005; 2006; 2007a; 2007b; 2007c; 2007d; 2008a; 2008b;
2008c; 2009; 2013); Jones (2008); Thornham (2011); Waggoner (2009); Wolf and Perron (2013);
Crawford, Gosling and Light (2013); Domsch (2013); Egenfeldt-Nielsen, Smith and Tosca (2013); Ensslin
(2011); Gamboa (2012); Garrelts (2005); Huntemann and Aslinger (2012); Juul (2011); Ruggill and
McAllister (2011); Whalen and Taylor (2008); Paul (2012); and Newman (2012), among others. In
addition, a host of scholars have employed concepts from the field of psychology to analyze video game
music: Cassidy (2009); Jørgensen (2008b); Kamp (2014); Nacke and Grimshaw (2010); Nacke et al.
(2010); Sanders (2010); Tan (2010); Whalen (2004); and Zehnder and Lipscomb (2004; 2006), among
others.
10
For example, analysts, casual listeners and even gamers often confuse “video game soundtracks” with
actual game scores. Video game soundtracks are officially-released and licensed as recordings which
contain music from video games. The music contained on a video game soundtrack, however, differs from
game scores since it is offered as a fixed sequence of audio information. I explore this distinction in-depth
in Chapter 3 of this thesis.
11
Mihaly Csikszentmihalyi (1990) defines “flow” as a state of complete involvement or immersion in an
activity. To achieve a flow state, there must be a balance between the challenge of the task and the skill of
the performer. If the task is too easy or too difficult, flow cannot occur. Game developers desire the
cultivation of “flow” in the gamer because it is a pleasurable state.
5
Collins (2013) suggests that new media, such as video games, provide instances
of interactive sound which are unique for their diffused sources of composition. In
interactive sound design, she argues, not only the composer has a hand in the
compositional process, but also the designer, programmer, and even the gamer. Like
Whalen (2004), Collins (2013: 5) argues it is the very interactivity of gameplay which
produces immersion, given the dialectic of “feedback and control mechanisms” in the
gameplay experience. Collins’ (2013) argument is useful as a basis for a study of video
game music composition and will thus provide a primary model for the ludo-
Vreeland undertook to compose his celebrated score for FEZ, a puzzle-platformer video
game released for Xbox Live Arcade in 2012. All of what follows, including screenshots
of, and technical information about the music system for FEZ, is drawn from a
Vreeland himself at the 2012 Game Audio Network Guild Summit.12 This is by no
means an exhaustive examination, nor is it intended to be. The point of what follows is
12
Screenshots of actual gameplay are created by the author.
6
simply to examine some aspects of game scoring, and to demonstrate that they resemble
“music composition.”
which includes the game’s creator and designer Phil Fish, and its programmer Renaud
Bédard. The latter were responsible for most of the development of the game. Fish
determined the creative vision for the project, while Bédard made that vision a reality
through programming. It was only until after the game’s visuals were designed and
programmed that Vreeland was invited to compose and produce the game’s celebrated
score, in fact. Therefore his task was to musically elucidate a pre-conceived visual world
with its own spatial limitations, mechanics, aesthetics and logic, and to provide a score to
spaces known as “Rooms.” At the outset of the game, Gomez, the game’s “protagonist,”
protagonist in the classic 8- and 16-bit Super Mario Bros. series, Gomez has impressive
jumping abilities, which serve as the main element of gameplay in a world composed of
encounters a mysterious being known as the Hexahedron, who grants him a “magical fez
hat” that allows him to perceive a third dimension, which rotates the gamer’s perspective
at will. As Gomez experiments with his new ability, the Hexahedron unexpectedly
13
Non-Euclidean space is space which cannot be measured by Euclidean geometry, which is the study of
flat space. Non-Euclidean geometries introduce fundamental changes to our concept of space, as in FEZ.
7
fractures and explodes, causing the game to glitch, freeze and reboot, complete with
BIOS screen. Gomez awakens in his room with his ability to perceive and manipulate a
third dimension intact, and is charged with the task of recovering the scattered fragments
Even after Gomez acquires the ability to perceive the third dimension, gameplay
in FEZ remains largely two-dimensional. Depth, or the Z-axis, is only visible to the
player in the rotation of perspectives, and is not a factor in the actual obstacles and
chasms which Gomez must traverse. The player must manipulate these perspectives to
explore the world of FEZ and collect thirty-two cubes in the form of “cube bits,” “whole
which a rotation of perspective moves it to the other side of the screen, even though he
has not moved at all. Figures 1 and 2 are gameplay screenshots from FEZ, which
Fig. 2: Gomez atop the same tree as Fig. 1 with the perspective rotated once, ninety
degrees clockwise. He is now able to ascend the tower.
9
Players of FEZ must conceive of space in a different way than usual in order to
navigate its world. One of the first people who was allowed to explore this world,
silently, was Rich Vreeland. Not only did Vreeland have to conceive of space differently
in this initial run-through of FEZ, he was also forced to conceive of composition in a new
way. The incorporation of the music system into Fezzer, the game’s programming
system developed by Renaud Bédard, allowed for this new compositional approach.
Moreover, Fezzer was no less dynamic than the world it was used to create: as he
composed, Vreeland was invited to propose ideas for the music system, which Bédard
would then implement into his programming. The music system they eventually
developed took the form of various tools and techniques integrated into Fezzer itself.
Thus, the production and composition of the score for FEZ was inextricably linked with
the development of gameplay and design. The three tools and techniques from Fezzer
which I will explain now are (i) the sequence context menu; (ii) the scripts browser; and
(iii) the main composition sequencer. These are names which I have given to these tools
and techniques, and not what Vreeland or Bédard may have called them.
10
Fig. 3: The sequence context menu in Fezzer. One of the appearing and disappearing
blocks in one of the Music Rooms is right-clicked, prompting the context menu.
(Vreeland [2012])
Fezzer allows the user to physically explore every aspect of FEZ as an omniscient
observer. Manipulation of perspectives is not necessary here as the user can already view
any area in three dimensions. Right-clicking on any element in the game world, such as
the block in Figure 3, prompts the sequence context menu. The sequence context menu is
a tool specifically for the assignment of sounds to physical elements within the game. It
is used for either sound effects or music given the scenario, or both as in the instance of
Figure 3. The “Sequence…” button allows the user to load a sample, piece of music or
sound effect into the menu. In the above example, Vreeland has loaded “3x_03” and
“3x_04” into the context menu as usable sound elements. These refer to the bright and
11
bit-crushed synth arrays that coincide with the appearance and disappearance of bright
Vreeland. The former had to adapt the gameplay to the rhythm of the music composed
by the latter. The blocks thus not only appear with the synth arrays, but appear on beat
with the level’s score. Vreeland said that this process involved thinking about music in
terms of proximity rather than order, or in terms of spatiality rather than temporality.
in Vreeland’s (2012) composition process, as he could now think about “which notes do I
want to happen near other notes so that they sound pleasing.” The word “near” in
Vreeland’s quote does not denote a nearness in time, but in (spatial) proximity between
elements in the world of FEZ. Bédard’s music system allows for a spatial conception of
music through its incorporation into the programming software itself. Right-clicking an
element and assigning it a sound may seem like a simple task, but it also prompts a new
Fig. 4: The scripts browser window in Fezzer. This example is again from one of the
Music Rooms, but the window may be pulled up in any area, as with the sequence
context menu. (Vreeland [2012])
Unlike the sequence context menu, the scripts browser window in Fezzer affects
an entire room rather than just any one single element. Scripts are programs which are
written for a specific run-time environment that can read and execute tasks in an
automated fashion. In other words, scripts are sets of tasks that can be performed by
programs that can interpret them, hence Fezzer deals with its own specific type of scripts.
The general nature of this definition points towards the wealth of possibilities with
scripts, as they can perform almost any function so long as the host program can interpret
them.
which lists each script’s “Id,” “Name,” “Trigger,” “Condition” and “Action.” The “Id”
13
of a script is simply an identifying number, while the “Name” column serves largely the
same function. In Figure 4 it is safe to assume that Vreeland left the “Id” and “Name”
that which sets the execution of a script in motion, but its implementation becomes more
complicated in specific cases. Figure 4 shows a scripts browser window with scripts for
Gomez ascends higher in the Music Rooms in FEZ, as in Figure 5, different musical
elements are added to and subtracted from the mix. Each trigger therefore indicates an
or lower than the specified altitude. For example, script four has “Volume[5], GoHigher”
as its trigger value, so any time Gomez goes higher than an altitude of “5,” the script is
triggered. The numbers which denote altitude are arbitrarily assigned to invisible blocks
in the Music Room, which are positioned by Bédard. The “Condition” column of the
scripts browser allows for any other conditions to be entered, such as time of day or
perhaps the amount of cube bits Gomez has acquired. In Figure 4 no extra conditions are
necessary, so the column remains unused. Finally, the “Action” column refers to what
It may be helpful here to reiterate the flexibility of scripts, and note that they are
governed by their own scripting language. “Volume[x]” thus refers to altitude in the
trigger field, rather than the volume of a sound, for example. The “Action” column uses
2]([Name], [Number of bars],” where “Target type” is the type of element being acted
14
upon, “Action” is the action to be taken, “Target type 2” is the sub-type of element being
acted upon, “Name” is the name of that element and “Number of bars” is simply the
length of the element. Script four, for instance, performs the unmute function on the loop
Fig. 5: Gomez ascends the first Music Room by jumping to bright red blocks as they
appear.
If Gomez ascends higher than the altitude marked by an invisible block as “5,” a
new musical element will therefore enter the mix, and it will remain there unless Gomez
descends lower than the marked altitude. When this happens — that is, when Gomez
descends below the designated altitude — gamers hear the opposite effect: the loop is
muted again. In this sense, Vreeland’s “composition” for the game is actually interactive,
what Vreeland calls “Music Gameplay.” Progress in the Music Rooms is signified by the
15
soundtrack, which rewards players with more elements of the song as they approach the
summit.
16
Fig. 6: The main composition sequencer window. This example is from work on one of
the Puzzle Rooms, which uses the song “Cycle.” (Vreeland [2012])
17
The main composition sequencer window is used mostly to determine the timing
logic of the elements in one of Vreeland’s “songs.” Like the scripts browser, the main
composition sequencer can make changes that affect an entire level, but also, like the
sequence context menu, it can be used to tweak single musical elements. The song name
can be entered or re-entered at the top of the window. The “Overlay Loops” list box
displays all the loops that can be in a level’s song, which can be added, removed and
Although it is not readily evident in Figure 6, I assume that you may select and
manipulate more than one loop at a time for faster workflow. Vreeland’s naming style
for his loops can be seen in the above example, and takes the form of “[Song Name] ^
is in the file name, so there is no guesswork necessary to determine which loop is which.
It is notable that the “Musical Element” field does not adhere to any specific type of
musical aspect, but instead serves solely to help programmers identify the loop. In some
The “Selected Loop Properties” area serves most of the functionality of the main
composition sequencer window. The “Loop Filename” is visible at the top, with a
browse button beside the text field. The “Trigger between after every…” area has two
text fields, with scroll arrow buttons, where a range of bars may be entered. In Figure 6
the song “Cycle” is split into many overlay loops, which play in the Puzzle Rooms
according to the settings entered here. The “Trigger” section, for instance, denotes where
the selected loop will play, within a given range if desired. This makes the actual song
18
heard during gameplay slightly unpredictable, or aleatoric, as loops may come and go
Below, the “Fractional time” checkbox allows for irregular time signatures to be
used in the deployment of loops. The “…and loop between…” section includes another
pair of text fields with scroll arrow buttons. These can be set to a range of the amount of
times the selected loop will play — another instance of aleatoric composition involving
chance operations. The length of the selected loop may be entered in the “The loop is…”
style incorporates the length of the loop in bars, so it is likely that he never uses the
“Detect” button. The “Delay first trigger by…” field can be set to denote the number of
bars after which the loop is played the first time. In this case, loops may be staggered in
The “One-at-a-time” checkbox is oddly placed, as its setting applies to the entire
song “system,” instead of just the selected loop. This setting works in conjunction with
the “Custom Ordering” text field below it, and allows the user to restrict the song to play
only one loop at any given time, while the “Custom Ordering” field dictates the order of
those loops. Alternatively, “Random One-at-a-Time Ordering” precludes the need for a
custom order, as it plays loops one at a time at random. The “Mute,” “Solo” and
“Preview” buttons are used to preview the song or selected loop within the main
composition sequencer window. Finally, the time of day checkboxes “Day,” “Night,”
“Dawn” and “Dusk” may be checked to specify when the selected loop may play
The “Base Properties” section of the composition sequencer allows for song-wide
changes to be made to tempo and time signature. As with many settings in this window,
these are musical elements which would normally be set in the compositional stages of
writing music. In non-interactive sources of music such as records, the tempo and time
signature are ordinarily set early in composition because they can dramatically change the
form of the song. This follows a more traditional approach to composition because it is
built upon the notion of a song’s “essence” which can be represented as notation (or sheet
inextricably linked to it. Settings such as tempo and time signature must therefore remain
malleable even late into the composition process. Alternatively, perhaps a better way to
express this difference would be simply to say that the composition process must remain
extended and “open,” right until the video game itself is complete.
The bottom section of the main composition sequencer actually deals with sound
effects, as Vreeland wanted the eight cube bits that make up a full cube to have
corresponding sounds that make up a full musical scale. The “Assemble Chord” drop-
down menu allows the user to choose the chord to be assembled, while each drop-down
menu in the “Shard Notes” area allows the user to choose a note for each cube bit to play.
20
As should now be evident, “game scoring,” that is, composing music for video
scorer’s choices are thus limited by available gaming technology, and the
“programmability” of their musical ideas given that technology. Moreover, game scores
are unique in that they must allow for an unprecedented level of musical flexibility, given
the high degree of user interactivity the gaming medium enables and encourages. As
activity, the final score being determined as much through chance gameplay as traditional
composition.14 Despite these unique attributes, though, and despite the recent
proliferation of game scoring in general, very little is written by scholars about game
analyze, in what follows I begin the process of considering the musical process of game
14
Aleatoric music is music in which some element of the composition is left to chance, and/or some degree
of freedom is afforded its performer. For more on this, see Rubin (2005) and Antokoletz (2014).
15
This is not to say the scholars do not write about music for video games. Rather, I argue that they treat it
as a cousin of film scoring, primarily, when game scoring requires consideration as a completely unique
compositional activity. See, for instance: Belinkie (1999); Berndt (2009; 2011); Berndt and Hartmann
(2007; 2008); Berndt et al. (2006); Brame (2011); Bridgett (2010); Chan (2007); Childs IV (2007); Collins
(2005; 2006; 2007a; 2007b; 2007c; 2007d; 2008a; 2009); d’Escrivan (2007); DeCastro (2007); Deutsch
(2003); Farnell (2007); Fay (2004); Furlong (2004); Gibbons (2009); Herber (2008); Hermans (2013);
Hoffert (2007); Kaae (2008); Knight (1987); Lendino (1998); Lieberman (2006); Marks (2009); Mera
(2009a; 2009b); Sanger (2003); Schmidt (1989); Sweeney (2011); Van Geelen (2008); Whitmore (2004);
and Wooller et al. (2005), among others.
21
acknowledges the “performative” and “interactive” aspects of video games, which create
the unique challenges and concerns that game scorers, and analysts of game scoring, must
address through their work.17 Indeed, as Collins (2013: 1) explains, “[video] games,
mobile phones, and other modern digital media alter the traditional relationships between
creator and consumer, audience and performance when the audience takes a participatory
In fact, game scorers arguably collaborate with players to produce the final score
for a game. In Super Mario Bros. (1985), for instance, the thematic content of each
a level. If players “beat” a level, the score triggers the “Flagpole Fanfare” theme, a
triumphant ascending melody (see Figure 7 below). Players who fail to conquer that
same level, on the other hand, hear the “Death Sound,” a comedic descending riff instead
(see Figure 8 below).19 Thus, whether or not the “Flagpole Fanfare” theme ever sounds,
16
Enns (2014).
17
For more on this see http://www.Ludomusicology.org.
18
Though this project will indeed place a significant focus on the gamer, it will not be a reception-based
discussion of the cultural ramifications of interactive media. Instead I offer a broader analytic of game
scoring from a purely musicological perspective. A review of important work on interactive media from
cultural studies would have to include: Collins (2005; 2006; 2007a; 2007b; 2007c; 2007d; 2008a; 2008b;
2008c; 2009; 2013); Jones (2008); Thornham (2011); Waggoner (2009); Wolf and Perron (2013);
Crawford, Gosling and Light (2013); Domsch (2013); Egenfeldt-Nielsen, Smith and Tosca (2013); Ensslin
(2011); Gamboa (2012); Garrelts (2005); Huntemann and Aslinger (2012); Juul (2011); Ruggill and
McAllister (2011); Whalen and Taylor (2008); Paul (2012); and Newman (2012), among others.
19
“Flagpole Fanfare” may be played at http://www.youtube.com/watch?v=3BsBXp6VkvU. The “Death
Sound” is playable at http://www.youtube.com/watch?v=M6KOEMJKdEI.
22
and how often, is up to the gamer, even while it is the game scorer who determines what
Fig. 7: A piano transcription of the “Flagpole Fanfare” theme heard when a player
successfully completes a level in Super Mario Bros. The ascending melody has a
triumphant or victorious thematic content. (“Flagpole Fanfare”)
24
Fig. 8: A piano transcription of the “Death Sound” which plays when a player loses a life
and fails to “beat” a level. The descending riff evokes disappointment, but is notably
shorter than the “Flagpole Fanfare” theme and adds a comedic element through
syncopated percussion. The “Death Sound” in SMB is archetypal for its effectiveness in
encouraging the player to attempt the level again, even after “dying” multiple times.
(“Death Sound”)
An obvious research question thus arises: how exactly is game scoring distinct
from scoring for other media? Before I can explain how I will answer this question, I
will first have to quickly consider scoring for non-interactive media. While film seems to
be a suitable medium for this comparison, to better elucidate game scoring’s more subtle
peculiarities I choose to analyze a medium that analysts, casual listeners, and even
gamers, chronically confuse with actual game scores: video game soundtracks. Video
game soundtracks are officially-released and licensed as recordings which contain music
from video games. The music contained on a video game soundtrack, however, differs
from game scores in that it is fixed and subject only to playback and equalization listens.
which are composed through gameplay (as I argue in Chapter 3 of this thesis).
25
Video game soundtracks are not open to aesthetic permutations given distinct
gameplay experiences, in other words. They have the same formal contours every time
they are played. Video game soundtracks are, in other words, ontologically “closed.”20
In effect, they are “idealized” versions of game scores, and in many cases impossible to
depends entirely on, and remains completely responsive to, gameplay and, thus, no single
This thesis includes four chapters in total. In Chapter 1, I have offered a critical
orientation to the subject of my thesis, namely, game scoring. This required providing a
broad survey of ludo-musicology and its primary subject (ie., game scores). I performed
a case study of the game scoring process for FEZ, in order to concretize the theoretical
terrain I examine in Chapters 2 and 3. I should note that a major aim of this research
project is to develop a working methodology for the study of game scores. In other
words, the methodology of this thesis will constitute an outcome, rather than just basis of
1. Examine the musical ability of the gaming technology (ie., the sound
hardware configuration) used to produce the game score.
20
I use the term “ontological” here as Martin Heidegger uses it, who notes that “ontological inquiries in
philosophy are concerned with [being]” (qtd. in Munday [2009]: “Ontology”). Thus, an ontological inquiry
into popular music scoring would analyze where such a process exists. Such an inquiry would reveal that
the process only exists and terminates in the production of the record. It is ontologically closed because
scoring does not continue in the playback of said record.
26
In Chapter 2, I argue that game scoring is distinct from other scoring, because it is
structured entirely by gaming technology. I demonstrate this through a case study of the
NES APU, wherein I survey this gaming technology’s musical possibilities and
limitations. I choose this as the subject of my case study because it is simply instructive.
That is, the NES APU is an ideal subject because it has relatively simple musical
abilities. This is true of all my case studies, such as the choice of levels in Chapter 3, for
instance. With these findings, I am able to discuss specific compositional strategies NES
game scorers developed in response to the technology of the NES APU. This reveals
how NES game scoring is structured by NES technology, and, more importantly, how
case study of a canonic aleatoric composition, namely, John Cage’s TV Köln (1958), in
order to survey the aleatoric tradition, which developed long before the release of the first
video game. Here I outline the basic tenets of aleatoric music, such as the loss of
sounds which analysts would normally consider “extramusical sound effects.” I use this
27
crucial as its “music.” Game scorers must recognize, integrate, and in many cases
“compose” sound effects as they work, and game scores are marked by many instances of
such sounds.
the “performance” of aleatoric music in video games. This will involve a case study of
“World 6-2,” one “level” of Super Mario Bros. for NES. I analyze game scores produced
by my own gameplay, and I expect to find many “chance operations” and opportunities
for “performer freedom” in my scores for this level. Finally, I elucidate these and game
scoring’s aleatoric nature through a comparison of my game scores for World 6-2 to the
future directions for continuing my research. I expect the results of this research to prove
that game scoring is an unique compositional activity which is distinct from other
scoring. I conclude by considering the significance of the completed research to the field
Chapter 2
Introduction
“Game scoring,” that is, composing music for video games, is distinct from other
game music exists within, and as a part of, a much larger medium, namely, the medium
of video games.21 As such, a host of priorities, values and concerns peculiar to that
medium inhere in the game scoring process, which do not inhere in other compositional
activities.
allocation during game development, for instance. This “ocularcentrism” has had
21
This thesis focuses exclusively on game scoring. For more on gaming culture per se consult: Schott and
Horrell (2000); Carr (2005); Jansz, Avis, and Vosmeer (2010); Kontour (2012); Condis (2014); Chen
(2014); Williams, Hendricks, and Winkler (2006); Kirkpatrick (2012); Schleiner (2001); Bryce and Rutter
(2002); Morris (2004); Natale (2002); Steinkuehler (2005); Holbert and Wilensky (2010); Taylor (2003);
Maguire et al. (2002); Schleiner (1999); Kücklich (2005); Wirman (2014); Deuze, Martin, and Allen
(2007); Gros (2007); Örnebring (2007); Kennedy (2002); Sotamaa (2003); Sich (2006); Nieborg (2005);
Salen (2007); Daniels and Lalone (2012); Cover (2006); Lin (2008); Murray (2006); Corneliussen and
Rettberg (2008); Kingma (1996); and Jakobsson (2011), among many others.
29
perhaps the most profound influence on game scoring.22 Music is routinely subordinated
to graphics when games are produced, even if players have cited music as a crucial facet
of the gaming experience since the advent of home gaming consoles.23 The
way: musical ideas must be “programmable,” as it were, even as the hardware and
software resources earmarked for musical programming are chronically scarce. Koji
Kondo (2010: n.p.), who composed the scores for the epochal Super Mario and Legend of
Due to the differing capabilities of game systems, the way I make music has
changed. The Famicom could only produce 3 tones and didn’t have a large
variety of sounds, so I had to do a lot of scheming. There wasn’t a lot of memory,
either, so I had songs where I couldn’t fit everything in, and I made songs with a
limited number of sounds. When the Super Famicom came along, it had 8 tracks
to work with.24
Kondo suggests that different gaming consoles present different possibilities for
scoring, and that composers must adjust their scoring strategies accordingly. In other
words, each console provides a particular set of rules and limitations which
fundamentally structures the game scorer’s compositional ideation and practice in unique
ways. If there is no means to program a musical idea, the game scorer must consider
other options. This has ever been the case. Indeed, technology is likely to structure and
22
For more on ocularcentrism and acoustic space see, for instance, Sterne (1997) and Hodgson (2007).
23
Game scores — such as the soundtracks for, to name some better known examples, Super Mario Bros.
(Koji Kondo, 1985), Ice Climber (Akito Nakatsuka, 1985), The Legend of Zelda (Koji Kondo, 1986),
Metroid (Hirokazu Tanaka, 1986), Mega Man (Manami Matsumae, 1987), and Mega Man II (Takashi
Tateishi, 1988) — have been celebrated on their own merits since their release. For more on this, see:
Murphy (2012); Campbell (2013); and Hannigan (2014), among many others.
The Nintendo Family Computer, or “Famicom,” is a video game console released in Japan in 1983. Its
24
North American counterpart, the Nintendo Entertainment System, or “NES,” was released in 1985.
30
restrain the game scorer’s compositional process for as long as such technology is
Even now I compose with the amount of memory in mind. So I can’t say the
process is entirely without limitations. On Mario Galaxy, for example, I didn’t
use a live orchestra, I made the music to match up with the game, so by
synchronizing with the on-screen action the songs changed interactively. For the
boss battles, you power up and become stronger when you take damage, right? At
that point, the orchestra grows fuller, the chorus comes in… that’s game music for
you.26
restraints on the compositional process — do game scorers navigate when they work? To
answer this question, I will provide a detailed case study of game scoring for the
Nintendo Entertainment System (NES), surveying how that technology structures the
compositional process for NES games in particular. This requires examination of that
survey well known moments when game scorers have, to borrow Kondo’s term,
“schemed” within the NES’ sound hardware configuration to produce their celebrated
game scores, consciously compromising and adjusting certain aesthetic concepts to better
suit the NES hardware. I focus in particular on the way game scorers have “schemed”, or
navigated, the crucial first step of the scoring process, namely, so-called “orchestration”
25
This restraint is further exacerbated by the ocularcentrism inherent in gaming culture, as resources are
allocated to visual ends.
26
The citation refers to Super Mario Galaxy, a 3D platform game developed and released by Nintendo in
2007, whose score Kondo composed with Mahito Yokota.
31
(i.e., selection of musical instruments or timbres for different musical ends). In fact,
The following case study considers the Audio Processing Unit (APU) of the
By considering the musical nature of each of the NES APU’s five available channels for
scoring, I will elucidate game scoring’s general technological structure. That is, by
demonstrating that all aesthetic possibilities in game scoring for the NES are in the first
instance determined by the NES’ sound hardware configuration, that composers are free
only insofar as they may assemble and superimpose only those musical terms that the
NES APU can generate, I will concomitantly demonstrate that, in general, all
programmed, they simply cannot exist; and whether or not a musical idea can exist in
32
game scoring is determined in toto by the sound hardware configuration used to actualize
it.
processing unit implemented in the NES’ Central Processing Unit (CPU).27 The APU is
comprised of five discreet channels: two pulse wave generators (PWC), a triangle wave
generator (TWC), a noise generator (NC), and a delta modulation channel (DMC) which
triggers low-resolution (i.e., shorter bit-depth) audio samples. According to the official
“timers,” which are “clocked” by an overarching “word-clock” count from the CPU.28
Timers are responsible for clocking the actual waveform generators in each sound
channel, and they provide modulation (i.e., sound processing) parameters for each
available channel in the APU.29 The main difference between the APU’s waveform
channels, and its DMC, is that the former generate their own sounds in “realtime,” via
analogue monophonic synthesizers, while the latter stores, recalls and triggers digital
27
The CPU for the NTSC (North America and Japan) Famicom and NES was the Ricoh 2A03, or RP2A03,
and for the PAL (Europe and Australia) NES, the Ricoh 2A07, or RP2A07. Further technological
specifications will always be for the NTSC NES, except where noted. Purely technical information is taken
from the official NES development Wiki.
28
This term refers to frequency, where “clock speed” would indicate the frequency at which a CPU is
running, for instance. “Clocking” another processing unit such as a timer would then refer to providing
information at regular intervals, at some fraction of the frequency which the CPU is running at.
29
According to Truax (1999: “MODULATION”): “Whenever a parameter of a sound or audio signal […]
is varied systematically, the signal is said to be modulated.”
33
audio samples from memory. I now turn my attention to considering each individual
channel in the APU in greater detail, in turn, in the section immediately following.
The NES APU contains two identical pulse wave channels. These channels have
a “bright” and “sharp” timbre, which is to say, they oscillate frequencies falling in the
such, composers tend to use these channels to convey the primary melodies of their game
scores (Schartmann [2013]). Moreover, since they have two identical PWCs at their
disposal, scorers will often orchestrate their melodies as a unison, shared between both
PWCs. When composers see fit to use other channels to convey their melodies, they
maximum volume, and three different “volume envelope shapes,” are available in the
NES APU.31 These envelope shapes include: (i) constant; (ii) linear decreasing; and (iii)
looping linear decreasing, or, “sawtooth” (see Figure 9 below). However, game scorers
30
For example, see Koji Kondo’s “Dungeon Theme” for The Legend of Zelda (1986), which uses the
triangle wave for its main melody to frightening effect.
31
An “envelope shape” is the shape generated by a graph of one parameter of sound, such as volume,
versus time. A sound’s volume envelope is related to its “ADSR.” As White (1987: “ADSR”) notes:
“[ADSR is short for] ‘Attack decay sustain release,’ time constants associated with signals generated by
electronic music synthesizers. The attack time is the time it takes the signal level to rise from zero to its
maximum value. The decay time is the time required for the level to fall to the sustain value, and the
sustain time is the time it remains at this value. The release time is the time it takes for the level to fall to
zero after the sustain time is elapsed […] The ADSR actually defines the envelope of the generated signal.”
34
seldom use a “constant” envelope shape per se ((i) in Figure 9 below). This envelope
shape is deployed so that, later on, a more sophisticated envelope generator can be used
to modulate it, producing a more complex shape. The “linear decreasing” shape ((ii) in
Figure 9 below), on the other hand, is typically deployed to emulate the decay and release
of acoustic instruments, that is, to simulate the manner in which acoustic instruments fade
to silence.32 Finally, the “sawtooth” envelope shape ((iii) in Figure 1 below) is used to
produce a variety of results such as, to name a celebrated example, the electric guitar
timbre heard in Takashi Tateishi’s “Opening” theme for Mega Man II.33
Fig. 9: The three envelope shapes generated by the PWCs in the NES APU. These are (i)
“constant;” (ii) “linear decreasing;” and (iii) “sawtooth.”
Aside from these envelope shapes, the NES APU’s two PWCs have three distinct
Duty cycle is the fraction of time that a system is in an “active“ state. The duty
cycle of a square wave is 0.5, or 50%. Some music synthesizers, including square
32
This said, the volume envelope of acoustic instruments seldom decrease in a linear fashion. The NES is,
of course, incapable of emulating such sounds in a verité manner.
33
Tateishi is credited by the alias “Ogeretsu Kun” in the game’s credits, for some reason unbeknownst to
the author.
35
channels of 2A0334 and VRC6,35 can vary the duty cycle of their audio-frequency
oscillators to obtain a subtle effect on the tone colors (“Duty cycle”)
The “system” in this definition simply refers to a sound wave, while “active state” refers
to the state of a waveform above the horizontal axis. Changes in duty cycle alter the
timbre of any given sound. Game scorers have four variable duty cycles available to
them through the NES APU’s PWCs: 12.5%, 25%, 50%, and 75%. Figure 10 shows
34
The NES’ CPU, developed by Ricoh, also referred to as the “RP2A03.”
35
The VRC6 (Virtual Rom Controller, revision 6) is a memory management controller developed by
Konami primarily for Castlevania III: Dracula’s Curse (originally released as Akumajō Densetsu in Japan,
1989), released for the NES in 1990. Memory management controllers comprise many kinds of special
chips designed by video game developers and implemented in NES and Famicom game cartridges, to
extend the abilities of the stock NES and Famicom consoles. The Japanese Famicom, unlike the North
American NES, had the ability to generate extra sound channels with these chips. For example, Konami’s
VRC6 added the ability to generate two extra square waves and one sawtooth wave for the score of
Akumajō Densetsu. The scores for Castlevania III and Akumajō Densetsu are markedly different due to
different scoring structures provided by different technological configurations — which pertain to musical
capabilities in particular — while the games retain nearly identical visuals and gameplay.
36
Fig. 10: The duty cycles available to the NES APU pulse wave channels. The 75% duty
cycle is instead offered as an inverted 25% duty cycle to illustrate that it has a nearly
identical (in fact, indistinguishable to the human ear) timbral quality to a normal 25%
duty cycle. Thus, the NES, for musical purposes, only has three distinguishable duty
cycles for its two pulse wave channels. (“Duty Cycle”)
A lower duty cycle produces a thinner, “sharper” timbre, while a higher cycle
produces a fuller, “smoother” timbre. A 50% duty cycle thus produces the fullest and
smoothest sound available through the PWCs, a 12.5% duty cycle produces its thinnest
and sharpest sound, and a 25% duty cycle falls directly between these timbral extremes.36
The NES game scorer is not strictly limited to using only one or another duty cycle,
however. Composers can program the APU to produce variations in duty cycle at any
given moment, even when one of the PWCs is in the midst of oscillating a particular
36
A “pulse wave” with a 50% duty cycle is more commonly referred to as a “square wave,” since its active
and inactive states are of equal length.
37
exponentially. More often, though, as Schartmann (2013) explains, duty cycles are
The introduction to “Wood Man’s Theme” in Mega Man 2 (1989)…. begins with
a low-percentage duty cycle, only shifting to 50% — a much “rounder” sound —
when the theme begins in earnest. Thus the music’s introductory measures are
played by a different “instrument” than the principal theme.
“sweep unit.” The “sweep unit” increases or decreases a PWC’s “period,” that is, its rate
of oscillation, which in turn determines the frequency or “pitch” of the sound the PWC
produces (higher periods of oscillation produce higher frequencies, while lower periods
produce lower frequencies). Pitch-bending is most often used by NES game scorers to
create a “vibrato” effect, as can be heard in Koji Kondo’s “Flute” melody for The Legend
of Zelda (1986).37
The Triangle Wave Channel (TWC) has only a limited range of musical
capabilities. This channel is most often used by game scorers to generate low
frequencies, that is, frequencies below 450 Hz, typically to set bass parts in a score.
Though square waves are generally considered an ideal synthesized bass timbre, given
37
Sound effect or music? Do sound effects count as game scoring? I address these broader questions in
Chapter 3.
38
their constant amplitude and harmonic structure, triangle waves are better-suited for this
task than any other waveform available in the NES APU.38 A triangle wave generally
sounds much “smoother” and “rounder” than a pulse wave, for instance, because it alone
intense harmonic structure, and a longer period of decay, than do pulse waves, meaning
that the TWC alone generates fundamental frequencies below 450Hz without
concomitantly outputting loud harmonic content above about 7 kHz (See Figure 11
below).39
38
According to White (1987: “Overtones”): “[overtones] are tones produced by a musical instrument which
are higher in frequency than the fundamental […] All musical instruments produce complex sound
waveforms which repeat at their fundamental [or lowest] frequency.”
39
To be clear, the TWC produces fundamental frequencies up to, and beyond, the supersonic limit of
human hearing (20kHz). I am talking exclusively about harmonic content in this statement, that is, the
harmonics comprising a (fundamental) frequency’s overtone content.
39
Fig. 11: A comparison of the dynamic envelopes of a triangle wave and a square wave.
“T” is the period value. The left charts show the waveforms in terms of amplitude and
time, while the right charts show each waveform’s harmonics in terms of decibels and
frequency. Note that while each contain odd-ordered harmonics, those of the triangle
wave “roll off” much faster than those of the square wave. (“NDLs Vs. Linear Filters: An
Illustration”)
As noted, game scorers frequently use the TWC to produce a reliably “smooth”
and “round” bass line. However, the channel can also be used to generate frequencies
above 450 Hz, resulting in a timbre most closely resembling that of a flute (this flute
sound is featured prominently in the “Title” theme from The Legend of Zelda). The TWC
can also generate a sound like a tom-tom drum, when used to oscillate a rapidly
descending glissando. This sound is heard in various songs throughout the Mega Man II
40
This is not an exhaustive list of possible uses for the TWC, obviously. At this point, being interested in
primarily surveying the manner by which the NES APU structures game scoring for it, I am interested only
in surveying the TWC’s most common uses in relation to its technological capacities.
40
The TWC can generate higher frequencies than the PWC generates, because of its
special timer. That said, we shall see that these frequencies are typically used to produce
a “glitch” effect. The highest frequency that the NES PWC can generate is
approximately 12.4 kHz.41 Humans are only capable of hearing frequencies from 20 Hz
to 20 kHz, and even then most hearing humans older than eighteen years of age do not
hear very well above 16 kHz. This means that the PWCs are not capable of servicing the
upper expanses of human hearing (ie., 12-20 kHz). The TWC, on the other hand, can
generate supersonic frequencies (ie., frequencies over 20 kHz), because its timer is
silencing the triangle channel, without sacrificing valuable CPU cycles from the 2A0343
for a “silence” request.44 When game scorers experimented with this technique, however,
they found that an oddly percussive sonic artifact — ie., a “popping” noise — sounded
whenever the TWC returned to oscillating in the audible range. The supersonic
frequency is initially generated when scorers write a timer value of zero which, according
to the programming equation ftri = fCPU/(32*(t + 1)) (where “ftri” is the resultant frequency
41
Determined by the equation fpulse = fCPU/(16*(t+1)) where “fpulse” is the resultant frequency of the pulse
wave, “fCPU” is the base frequency of the CPU (1.78977267 MHz for a North American NES) and “t” is the
timer value.
42
“Clocking” simply refers to the process of providing information at regular intervals, in this case at some
fraction of the frequency which the NES CPU is running at. The TWC’s maximum frequency on a North
American NES is actually 55.9 kHz. This value is determined by the equation ftri = fCPU/(32*(t + 1)) where
“ftri” is the resultant frequency of the triangle wave, “fCPU” is the base frequency of the CPU and “t” is the
timer value.
43
The 2A03, also referred to as the “RP2A03,” is the name of the NES’ CPU developed by Ricoh.
44
Perhaps more than any other technique I examine, this technique encapsulates game scoring. To
compose a tacit section for a particular instrument, the scorer must actually compose frequencies above the
human audible threshold, that is, supersonic frequencies.
41
of the triangle wave, “fCPU” is the base frequency of the CPU and “t” is the timer value),
generates the channel’s highest available frequency. This frequency is so high, however,
that the mixer receives an irregular and abrupt sequencer value, for which it cannot
compensate. This results in “artefacting,” that is, audible distortion, most closely
resembling a “popping” noise. This “popping” can be heard in the score for Mega Man
The TWC has a rhythmic advantage over the other APU channels, because of the
accuracy of its timer. Developers of the NES APU felt it necessary to imbue only the
TWC with the clocking accuracy required to achieve “pinpoint” rhythmic precision, that
is, developers deemed it necessary to devote a crucial portion of only the TWC’s CPU-
load to achieving rhythmic rather than textural precision. This additional feature thus
technologically structures — it provides the only technical means for achieving — the
As noted, the NES APU’s two PWCs are more likely to be used to set tracks with
sustained pitches and upper-midrange frequency content, which is to say, for setting
melody and rhythm section parts. The bass section in NES game scores, however, are
often very repetitive and melodically simple, and require rhythmic accuracy over and
above anything else. This said, I should quickly note that this does not mean that
composers can only use the TWC to provide bass support for upper-register melodies.
Some composers have even gone so far as to use a TWC bass line for the primary
melody. This inversion of compositional convention can most notably be heard in the
“Underworld” theme from Kondo’s score for Super Mario Bros. (1985).
42
The NES APU Noise Generator Channel (NGC) oscillates “noise,” that is, sound
featuring an irregular or “random” waveform. Actually, the NGC outputs two different
kinds of noise: “white noise” and “periodic noise.”45 Most commonly, however, the
NGC is set to “white noise mode” and used to set the percussive elements of a game
score. In fact, scores most often use the NGC to orchestrate the components of a typical
“trap” drum set: kick drum, snare drum, hi-hat, et cetera. The sound of a snare drum ((i)
in Figure 12 below), for instance, emerges when scorers set the NGC to “white noise”
mode, and shape its dynamic envelope so it features a rapid “attack” and gradual “decay”
and “release” contours. An open hi-hat ((ii) in Figure 12 below), on the other hand,
emerges when scorers filter the NGC’s “white noise” through an envelope featuring
gradual “attack” and “release” phases, while the sound of a closed hi-hat ((iii) in Figure
12 below) emerges from exactly the same envelope contour, but with a rapid “release”
45
According to Kaernbach (2000: pgh. 1): “Noise is a sound with an irregular, random waveform. Unlike a
musical or speech sound, it contains a lot of different frequencies. It is called “white noise” if all audible
sound frequencies are represented with the same strength. This designation is in analogy to vision: white
light contains all visible frequencies of light.” Similarly, periodic noise is a sound with many different
frequencies, though these frequencies eventually repeat, unlike with white noise. The NES NGC is not
capable of producing “true” white noise, as its frequency pattern does repeat after 32767 steps. This
pattern, however, is too long for the human ear to notice its regularity, and it ends up sounding like white
noise anyway. The NGC generates periodic noise through a frequency pattern either 93 or 31 steps long,
depending on where it is in the 32767-step sequence when it is triggered.
43
Fig. 12: The three dynamic envelopes mentioned in the paragraph above. They are (i) a
snare drum, (ii) an open hi-hat, and (iii) a closed hi-hat.
The NGC’s “periodic noise” mode is less frequently evoked by game scorers, and
it is seldom used to orchestrate the percussion elements of a score because of its generally
harsher, more metallic texture. Periodic noise is a sound with an irregular waveform
sound than white noise, for instance, and which may even sound melodic. In fact, scorers
44
occasionally even use the mode to set a score’s melody. NES game scorers simply prefer
the NGC’s “white noise” mode for setting percussive elements because it has less pitched
content and, thus, sounds more like acoustic percussion instruments. This said, braver
NES game scorers have occasionally experimented with using the NGC’s “periodic”
mode to produce unprecedented effects. A good example can be heard in the “Quick
Man” theme from Mega Man II. In this case, the NGC switches rapidly back and forth
between “white noise” and “periodic noise” modes, to produce a complex rhythmic
pattern. The music thus underscores and heightens the intensity of the gameplay, even as
The Delta Modulation Channel (DMC) is unique amongst all the channels
comprising the NES APU, in that it triggers rather than oscillates, that is, it is used to
sequence audio samples stored in its memory. Though the NES sequencing capacities are
primitive by modern standards, the very fact that the NES featured a DMC for triggering
samples when the NES was first released, in 1985, was extraordinary for any home video
game console of its time.46 Moreover, the DMC expanded the breadth of sounds a game
46
For more on the history of sound technology in gaming, see Belinkie (1999) and Grimshaw (2010).
45
Through the technology of the DMC, NES game scorers buttress their
compositions using any (sampled) sounds they want, albeit in compromised “quality.”47
Game scorers most often use the DMC to produce sound effects, vocal “bites” and
percussion.48 Koji Kondo, for instance, used the DMC extensively in his score for Super
Mario Bros. 3 (1988 in Japan; 1990 in North America), mainly to sample the percussion
instruments (i.e., bongos, timpani and tom-toms) heard in the eight different “World
Map” themes players access sequentially as they progress through the game. This
Now that I have surveyed some of the capacities of each of the five channels
which comprise the NES APU, whose sonic capacities in turn comprise the entire sonic
palette available to NES game scorers, I turn my attention to exploring some of the more
conventional compositional uses game scorers devise for these channels. In so doing, I
turn, so-called: (i) “2-channel Echo;” (ii) “1-channel Echo;” (iii) “arpeggios;” (iv)
“triangle kick drums;” and (v) “melodic samples.” After I have done this, I will briefly
47
I would contend that 1-bit samples also have their own aesthetic, so “compromised quality” is used here
only to denote the radical simplification of audio information performed by the DMC, and the fact that a 1-
bit sample features a significantly lower “figurative” resolution than, say, a 24-bit sample.
48
A vocal “bite” is simply a brief sample of a vocal phrase, whether musical or not.
46
demonstrate how game scorers draw all of these disparate musical terms together, into a
To be clear, many of the game scoring techniques detailed below would ramify in
At present, we are accustomed to the ability to apply effects such as “echo” or “reverb”
with the click of a button in a modern Digital Audio Workstation (DAW) such as Apple’s
Logic Pro. At least in the case of the NES APU, game scorers do not have such a luxury.
Furthermore, in the case of all game scoring, the ability to generate such effects depends
This phenomenon is not entirely new, though, as echo effects appear in many
programmatic orchestral works. Baroque scores from the 17th and 18th Centuries, such as
J.S. Bach’s French Suites, often contained echoes as imitative devices, whereby a
musical motto was played by an orchestra and then repeated immediately afterwards
much more quietly. Similarly, game scorers do not simply sample and replicate echoes
but, rather, create echo effects through musical means. Sound effects, in game scoring,
then, are musical devices, even if traditional musicological analyses would dismiss them
One of the most common game scoring techniques for the NES APU is so-called
game scores because so few channels are available to orchestrate it.49 In fact, the only
channels composers can use to create a two-channel echo are the PWCs, since the
lengthening delay rates. And the same is true, of course, for reverberation effects.50
both PWCs, each successive iteration played in alternating channels at successively lower
amplitudes and ever lengthening delay rates. The “African Mines” theme, composed by
Hiroshige Tonomura for Ducktales (1989), provides a clear example of this effect in
game scoring, in this case deployed to evoke the reverberant acoustics of an underground
mine.51 Neil Baldwin likewise uses the effect in his “Puzzle Room” theme for Magician
(1990), the effect here evoking the cramped, dark and confusing landscapes of each
puzzle level. As its name suggests, however, two-channel echo requires a significant
amount of the NES APU’s available resources, monopolizing two of only five available
channels. Game scorers thus developed a technique for creating echo and reverberation
effect using only one voice very early on in the NES’ development. I explain this next.
49
Echo is a reflection of sound which arrives at the listener at least 25 milliseconds after the direct sound.
50
According to Izhaki (2008, qtd. in Hodgson [2010: 171]), “reverb” is: “the collective name given to the
sound created by bounced reflections from room boundaries … In modern times, we use reverb emulators,
either hardware or software plug-ins, to simulate this natural phenomenon.”
51
Developed by Capcom, the company which also develops the Mega Man series. Key personnel from that
series were tasked with developing Ducktales for the NES and the Game Boy, and handheld video game
console developed and released by Nintendo in 1989. The Game Boy version utilizes the Game Boy sound
hardware, similar to the NES APU, in order to create the same effect with two pulse waves.
48
resources, NES game scorers create “echo-like” sounds using only one channel. Geoff
Follin, for instance, achieves this effect by combining dramatic dynamic leaps with pitch
bends (downward glissandi). The effect can be heard clearly in Follin’s score for
Wolverine (1991), a licensed action video game based on the Marvel Comics superhero
of the same name. Follin’s theme for “Level 1” includes descending melodic accents
from one of the PWCs. Follin adjusts the volume envelope of each accent to decay
rapidly at first, but release slowly, and he deploys downward pitch-bends to emphasize
the lengthened release time. This results in an eerie pulse wave, accompanied by what
Tim Follin, Geoff Follin’s brother, provides another good example of single-
channel echo, and the highly individuated nature of its compositional production in game
scoring. In this case, the effect appears in Follin’s score for the feature film-licensed
game, Indiana Jones and the Last Crusade (1991, Taito).52 Instead of using repeating
and diminishing dynamic leaps combined with pitch bends to create the effect, however,
as his brother did in the score for Wolverine, Tim Follin here exploited the
pitches have less subjective loudness than higher pitches.53 In the “Tank – Cutscene”
theme, for instance, Follins sets repeating downward pitch bends in one PWC to create a
single-channel echo. Each pitch of the theme thus sounds as though its volume swells,
even though each pitch features a slow, constant fade throughout. This, combined with
the downward pitch bends, creates a reverberant effect using only a single PWC.
Neil Baldwin developed yet another method for producing single-channel echo in
his score for Hero Quest, an unreleased NES game developed in 1991. While the above
apply echo to a pattern of changing eighth notes for the “Final Track” theme of Hero
Quest (1991). This required a different technique. Baldwin thus decided to compose
52
The “Follin Bros.” have collaborated on game scores and are collectively known as such in gaming
culture. They are celebrated by chip music enthusiasts for developing techniques such as single-channel
echo.
53
The ear is not equally sensitive to all frequencies, particularly in the higher and lower ranges. In 1933,
Fletcher and Munson charted the response to frequencies across the entire audio range, as a set of curves
showing the sound levels of tones perceived as equally loud. These curves are called “equal loudness
contours” or “Fletcher-Munson curves.” According to White (1987: “Fletcher-Munson Effect”): “The most
sensitive range of human hearing is between 3kHz and 4kHz; the sensitivity falls off rapidly at lower
frequencies and somewhat more slowly at higher frequencies. In other words, sounds must be more
powerful at lower and higher frequencies than 3 to 4kHz in order to be heard at the lowest audible levels.”
50
quieter duplicates of each note of the melody, delayed by roughly an eighth note. This
produces a much larger sense of space in the “Final Track” than elsewhere in the score,
as percussion.
Table 2: A few NES tracks which prominently feature single channel echo.
pitches are articulated sequentially rather than simultaneously. Arpeggios are extremely
important in NES game scoring but difficult to produce, again because of the limited
number of channels available in the NES APU. Triadic chords are simply impossible to
produce via the NES APU, in fact, because only two of its channels, namely, the PWCS,
are capable of producing identical timbres simultaneously. Game scorers thus face yet
51
another technical dilemma: triadic chords are fundamental to the music they compose, yet
they lack the technical means to produce them. To solve this dilemma, NES game
scorers produce triadic block chords psychoacoustically, as it were. That is, rather than
sounding all three pitches of a triadic chord simultaneously, game scorers arpeggiate the
pitches so quickly that the human ear is incapable of distinguishing one waveform from
another. Consequently, the ear “sums” the component waveforms into a single triadic
unity.
by historians with Commodore 64 game scores and its MOS Technology Sound Interface
Device (SID).54 However, facing similar technical limitations, NES game scorers —
particularly those working in Europe, where the SID remains popular — have adopted
“psychoacoustic” arpeggiated chords. The technique can be heard in, among other
scores, Silver Surfer (Tim and Geoff Follin, 1990), Magician (Neil Baldwin, 1990), Skate
or Die 2 (Ron Hubbard, 1990), Solstice (Tim Follin, 1990), Darkman (Jonathan Dunn,
1991), M.C. Kids (Charles Deenan, 1992), and Asterix (Alberto González, 1993).
54
For a history of video game music up until 1999, see Belinkie (1999).
52
Table 3: NES tracks which feature arpeggio and psychoacoustic block chords
prominently.
As noted, percussion in NES game scores is typically set using the NC and the
DMC. Game scorers wanting to give their kick drums extra “punch,” for instance, pair
noise from the NC with audio samples of a kick drum triggered in the DMC. However,
the latter may require simply too much storage memory to be a feasible component of a
game score, so composers must resort to other means to produce the effect. One such
means involves oscillating a triangle wave bass part in the TWC, but rapidly bending its
pitches downwards, that is, by orchestrating extremely fast downward glissandi in a bass
line produced by the TWC. These glissandi, when filtered through an extremely short
sustain and release envelope, produce a sound more like a kick drum than a low
frequency triangle wave. Like the “psychoacoustic block chord” technique elucidated in
the section immediately above, this “kick drum” technique is also especially popular in
53
European game scores for the NES, and can be heard in scores for Hero Quest, Silver
Table 4: NES tracks which prominently feature the TWC kick drum technique.
NES game scorers use the DMC mostly to trigger audio samples of vocals,
percussion and sound effects. It is simpler to use the DMC to produce these sounds,
because they rely less on pitch than timbre and timing to produce their musical effects.
Using the DMC melodically is much more complicated than using the NES APU’s other
four channels, because the DMC has specific limitations with regard to “re-pitching”
samples:
55
This technique is most commonly used for kick drums, to be sure. However, it is possible to use it to
create other drum sounds, such as a woodblock, by incorporating the same technique at higher frequencies.
54
Because of these limitations, each of the highest DMC pitches in an NES game
score requires their own samples. Lowering the pitch of these samples causes them to
sound at a slower rate, producing distortion and significant “tonal artefacting” (i.e.,
detuning). A full chromatic scale is difficult to obtain through re-pitching, because the
fifteen other set pitches are not arranged chromatically, even when starting with a tonic of
C4. Retro Game Audio (2012: “NES Audio: Sunsoft Bass and Melodic Samples”)
outlines the resultant pitches with a starting sample at a pitch of C in the fourth octave:
C4 G3 E3 C3 A2 G2 F2 D2
C2 B1 A1 G1 F1 E1 D1 C1
Yet even these notes are not exactly correct, as some notes output sharp or flat. The
56
Retro Game Audio (2012: “NES Audio: Sunsoft Bass and Melodic Samples”).
55
The “period hex values” correspond to information read by the DMC memory
reader as waveform period lengths. In the table above, these periods are converted to
frequencies in Hz and finally notes with deviation given in cents, where one cent is 1/100
of a semitone. NES game scorers may play their desired samples at only these
frequencies.
The games created by Sunsoft during the latter half of the NES’ shelf life
contained some of the few scores to use DMC samples melodically. Batman: Return of
the Joker (1991), Hebereke (1991), Journey to Silius (1990), Gremlins 2 (1990),
Gimmick! (1992) and Super Spy Hunter (1992) all use DPCM samples primarily to add
definition to the bass sections of their scores. In these scores, a real electric bass guitar
was sampled by Sunsoft composer Naoki Kodaka, and played back through DPCM.
These samples accompany a triangle wave bass, and the combination of these two
channels results in a sound with the smooth and round timbre of a triangle wave but
In order to accomplish this, however, a chromatic scale was necessary for the
sampled bass section. Kodaka responded to the strange pitch table of the NES DMC by
A# B C C# D
From these five notes he could then use the re-pitch function of the DMC to lower these
samples, resulting in something close to a full chromatic scale. Since these samples
mimic the notes of the TWC, the discrepancies in pitch and timing — as stated, lowering
pitch is not an exact process and causes slower playback — created by the DMC are less
noticeable to the listener. In fact, these discrepancies are more likely to be interpreted as
musically interesting, rather than inaccurate, if they are noticed by the listener.
Exact calculation of the resultant amplitude of all the channels in the NES is
almost impossible, due to the APU’s non-linear mixing scheme. Each channel contains
audio signal. These DACs are implemented in such a way that produces non-linear
interactions between channels. For example, a high value in the DMC output unit will
reduce the volume of the TWC and NC, while a high TWC output has no effect on the
Koji Kondo used this peculiarity of the NES APU to his advantage in the score he
produced for Super Mario Bros. (1985). Kondo re-set the dynamic output of the DMC
simply to limit the TWC’s dynamic output. In fact, Kondo did not even use the DMC to
example is not meant to be taken as a unique case, though, as the non-linear nature of the
NES APU mixer affects every decision the NES game scorer makes. Since the
manipulation of one channel will likely affect another’s output, the entire arrangement
changes with, say, a duty cycle change on the PWC. NES game scorers must know not
only how their orchestrations will affect the output of a single channel, but also how that
57
According to White (1987: “Limiter”): A “limiter” is “a special type of compressor which prevents the
signal from exceeding a certain preset level.” White (1987: “Compressor”) also notes that a “compressor”
is an “audio device which reduces the dynamic range of a signal.”
58
Chapter Summary
As a case study of the NES APU, and the way it structures game scoring for the
NES, this chapter is by no means exhaustive. It is not meant to be. Neither is it meant to
stand alone as a work dedicated to increasing knowledge about NES hardware. It should
I hope I have demonstrated in this chapter that, when they compose, game scorers
think musically in relation to a particular sound configuration hardware, which limits the
possibilities they can imagine for composition in very particular ways. Indeed, musical
which game scorers compose. This is not to say that musical ideation and creativity in
configurations. In this chapter I have sought to elucidate this structure using the NES
APU as a case study, paying particularly close attention to some of the better known
musical possibilities that game scorers have gleaned — or, to borrow Kondo’s phrase,
“schemed” — from that particular sound configuration hardware. In the next chapter, I
scrutinize the aesthetic product of this process, namely, game scores per se, especially as
Chapter 3
Introduction
“Game scoring,” that is, composing music for video games, is distinct from other
given the nature of the gaming medium. Most crucially, video games are interactive,
meaning that the gamer actively performs the gaming experience, though their actions are
Video games, mobile phones, and other modern digital media alter the traditional
relationships between creator and consumer, audience and performance when the
audience takes a participatory role in instigating sound events. When using such
media audiences may, through their actions, be responsible for evoking sounds,
selecting them, altering or shaping them, or creating new sounds, thus playing an
active role in the composition of their own soundscapes. This active role of the
audience raises interesting questions about the ways in which we theorize sound
in media. What does it mean to interact with sound? Who is the audience, and
who is the creator of such co-creative, interactive sonic constructions?
examine these issues strictly with regards to game scoring, that is, from a strictly ludo-
60
technology, and game scores are programmed by composers, designers and programmers,
but a game score is only ever finally realized through gameplay. The gamer is a crucial
partner in the game scoring process, serving as its primary performer, insofar as their
gameplay realizes and, in so doing, dictates the final compositional structure of a game
score.
The duration of any game score, for instance, is always determined by the
duration of a given gameplay session. I may play Super Mario Bros. (1985) for three
hours, the result of which would be a long, fragmented score punctuated by gameplay
sounds of failure and/or triumph. Alternatively, I may instead play the game for only
thirty seconds, the result of which is a very different piece.59 Koji Kondo is primarily
responsible for composing (in a traditional sense) the various fragments which I have
heard in both these instances, but the order in which these fragments sounded would
depend entirely on my gameplay choices. Thus the harmonic and melodic design of the
58
Like ludo-musicology, interactive media studies is a burgeoning field at present. Work from this
discipline with a focus on game audio includes, but is certainly not limited to: Allouche et al. (2007);
Berndt et al. (2006); Berndt and Hartmann (2008); Berndt (2009); Berndt (2011); Borchers and Mulhauser
(1998); Bridgett (2008); Fay (2004); Fish (2003); Guerraz (2008); Herber (2008); Hoffert (2007);
Jørgensen (2006; 2008); Kaae (2008); Rayman (2014); Collins (2005; 2006; 2007a; 2007b; 2007c; 2007d;
2008a; 2008b; 2008c; 2009; 2013); Jones (2008); Thornham (2011); Waggoner (2009); Wolf and Perron
(2013); Crawford, Gosling and Light (2013); Domsch (2013); Egenfeldt-Nielsen, Smith and Tosca (2013);
Ensslin (2011); Gamboa (2012); Garrelts (2005); Huntemann and Aslinger (2012); Juul (2011); Ruggill and
McAllister (2011); Whalen and Taylor (2008); Paul (2012); and Newman (2012), among others.
59
These examples may be taken to further extremes. If I leave any game running in any state for an entire
year, the score I have helped compose is one year long, and so on.
61
Given the role of gameplay in game scoring, I argue that it should be understood
compositional activity in which one or more musical elements are left to chance, as well
options.61 In this case, the chance operation involved in the music’s composition is a roll
of the dice, and the element of composition chance determines is the order in which the
pre-composed sections of the piece are performed.62 Similarly, any composition with
composition. That is, I argue that a game score involves both chance operations and a
the “performer” of a game score is not a musical performer but a “ludal” one, that is, a
60
Work which relates video game music to aleatoric composition includes: Collins (2009); Lieberman
(2006); Philips (2014); Rayman (2014); Summers (2011); Paterson et al. (2011); Young (2012); Paul
(2013); Custodis (2013); Pannerden et al. (2011); Lerner (2014); d’Escrivan (2007); Bullerjahn (2010);
Hermans (2013); and Mitchell (2014).
61
The word “aleatory” or “aleatoric” is derived from the Latin word alea, which actually means “dice.”
62
For example, see: Der Allezeit Fertige Menuetten- und Polonaisencomponist (German for “The Ever-
Ready Minuet and Polonaise Composer”) (1757) composed by Johann Philip Kirnberger; Einfall Einin
Doppelten Contrapunct in der Octave von sechs Tacten zu Machen ohne die Regeln Davon zu Wissen
(German for “A method for making six bars of double counterpoint at the octave without knowing the
rules”) (1758) composed by C.P.E. Bach; and Table pour Composer des Minuets et des Trios à la Infinie;
avec deux dez à Jouer (French for “A table for composing minuets and trios to infinity, by playing with two
dice”) (1780).
63
For example, Ornette Coleman’s Free Jazz (1961) was the first album-length improvisation, making it at
least partially aleatoric.
62
haptics inform the gamer’s gameplay choices, so game scores are subject to improvised
operations and “performer freedom” embedded within game scores.64 This will require
study of actual gameplay. Given the limited space I have here, however, and given the
breadth of scoring possibilities available through even the simplest of video games, I
shall only examine one part, or “level,” of a game, namely, the 22nd level of Super Mario
Bros. I do this to elucidate how gameplay patterns determine the final game score a
gamer hears during each gameplay. I shall compare the game scores I produce for Super
Mario Bros. to a non-aleatoric musical source, which is the game’s “official soundtrack.”
Video game soundtracks differ from game scores because they are fixed, “idealized”
game scores, divorced from gameplay itself. Before I do any of this, though, I briefly
survey the aleatoric tradition in Western Art Music, to provide an historical and analytical
context for game scoring. I do this by scrutinizing John Cage’s TV Köln (1958) in detail.
While composers such as Henry Cowell, John Cage, and Christian Wolff took the
first decisive steps towards the creation of an aleatoric compositional style in the early
1950s, there exist many examples of aleatoric music which pre-date the twentieth
64
Roig-Francoli (2008) coined the term “aleatoriality” to designate the degree of chance in a composition.
63
There has always been some degree of rhythmic, harmonic, and formal freedom
for both the composer and the performer throughout earlier centuries, as
manifested in the use of rubato or ad libitum indications, fermatas and grand
pauses, improvisation in cadenza-like passages, realization of figured bass, and
even the elimination of the barline resulting in metric freedom, as in keyboard
fantasias of C.P.E. Bach.
In the early part of the twentieth century, composers such as Arnold Schoenberg, for
instance, made more radical attempts to “free the performer” from the rigid structures of
The use of tone clusters and the elimination of the barline in works of Ives and
Cowell resulted in both harmonic and rhythmic indeterminacy, while the use of
Sprechstimme permitted a degree of pitch and harmonic indeterminacy in works
of Schoenberg and others.
founded on the idea of total composer control over all aspects of the performance of a
composition. From the early twentieth century on, an ever-increasing gap formed
between the tenets of integral serialism and the principles of aleatoric composition.
These two compositional styles reached their most extreme polarization in the early
1950s, when the American composer John Cage began to compose pieces in which the
The American composer John Cage was the first to fully incorporate chance
operations in his compositions from the early 1950s. Cage was influenced by the
ragas and talas, which he took from Carnatic and Hindustani music traditions.
Moreover, his philosophy of music derived from his interest in Zen Buddhism. Cage
We need first of all a music in which not only sounds are just sounds but in which
people are just people, not subject, that is, to laws established by any one of them,
even if he is “the composer” or the “conductor.” […] The situation relates to
individuals differently, because attention isn’t focused in one direction. Freedom
of movement is basic to both this art and this society.
Thus Cage advocated both for a relinquishment of control on the part of the composer
and a degree of indeterminacy for the performer. In this way, both roles, as we would
normally conceptualize them, are problematized by Cage’s music from this time.
indeterminacy in the realm of performance. Figure 13 shows the score for this piece.
The score for TV Köln incorporates both traditional Western notation and a notation
system of Cage’s own invention. The piece is divided into four sections, which Cage
explains are of equal duration, though he does not specify their absolute durations.65
65
In effect, the piece can be of any duration, similar to a game score, as noted above.
65
Fig. 13: The score for TV Köln. This was accompanied by a page of instructions which
explain Cage’s peculiar notation system. (“TV-Köln 1960”)
The instruction sheet accompanying the score provides a legend for four of the
[The] lines [of each section], unlike the staves of traditional piano music, do not
separate bass and treble clef, or right hand and left hand, but distinguish the type
of sound played or — more accurately — where that sound is produced: on the
keys (K), inside the piano (I), on the piano’s exterior surface (O), or somewhere
other than the piano (A).
Strangely, Cage did not provide an explanation for the letter “P,” which appears on the
fifth line of the third section. Finally, Cage instructs that a note’s proximity to a line may
TV Köln’s score is unique in that it does not instruct performers to play particular
sounds, but indicates the way that sounds should be produced. As Morgan (1991: 317)
notes, what “is indicated are essentially actions, not musical events.” Cage creates a
general musical structure for the performer to follow, though he leaves considerable room
for that performer’s agency. The score for TV Köln allows for “freedom of movement,”
as Cage puts it, meaning that many of its compositional parameters are left to the agency
of the performer. For example, the absolute duration of any musical event in the piece is
which corresponds to pressing a piano key, to “A” which corresponds to almost any
action, to “P” which could signify anything. Finally, the difference between stemmed
themselves, however they see fit. Yet, as Morgan (1991: 318) suggests, despite all these
has a definite “shape,” because the events that constitute it have been arranged in
a given temporal relationship with one another. This shape even has, however
inadvertently, a certain traditional quality. The piece opens with a single isolated
event; continues […] with a relatively quick succession of events […] and closes
with two more isolated events.
through various means, all of which granted the performer greater agency, and the score
such as the “auxiliary” sounds he indicates in TV Köln — is salient to the current analysis
of game scoring, particularly with respect to “sound effects.” Cage’s score “musicalizes”
analyses. Similarly, as the next section will demonstrate, game scoring includes a
Before I investigate examples of actual game scores, I should clarify the role of
sound effects in game scores, and the way I intend to conceptualize sound effects in this
thesis. One might equate video game sound effects to extraneous sounds heard in the
orchestra pit of a ballet recital, for instance, such as the squeaks of chairs or the coughs of
performers. This analogy is unsuitable for video game sound effects, however. In the
example of sounds heard in the orchestra pit, the composer of the piece does not take
these extraneous noises into account. They are considered “extramusical,” as it were.
68
The instruments which make “music” are not the same objects which produce these
not only an inevitable but a complementary musical device. After all, a game score may
only be realized through the act of gameplay, a state which contains many more sounds
than those programmed as “music” per se. Game scorers must be highly aware of these
sounds, to the extent that their scoring process is tailored significantly to accommodate
them. Moreover, they are “scored” precisely as, say, a triangle-wave bass line would be.
Whereas the ballet orchestra composer only tolerates incidental sounds as extramusical
accidents, the game scorer thus integrates sound effects into the compositional process.
In short, the game scorer composes sound effects, and thus they must be fully aware of
the parameters under which sound effects “intrude” upon their scores. Koji Kondo
The thing I consider to be most important is making the game more fun. There
are three things I keep in mind. First of all, each game has a unique rhythm or
tempo, so I try to capture that and compose music that fits the game’s rhythm.
Second of all, the balance. For games, it isn’t just the music, one also has to
consider sound effects, the balance of the volume, the balance between left and
right channels, and make sure the sound effects [are] more prominent. Third,
putting in variations in the music to fit with the interactivity of the game. For
example, speeding up the tempo when time is running out or changing the music
that plays depending on the player’s location.
Kondo considers sound effects every time he scores a game. They must be louder
than the background music he composes, he notes. Even if Kondo does not consider
sound effects “music,” then, their presence heavily informs his scoring choices. In fact, I
would argue that the musical consequences of video game sound effects — that is, how
69
game scorers accommodate these phenomena in the composition of their own “music” —
When a player directs Mario to collect a coin in Super Mario Bros. (SMB), for
instance, a satisfying “ching!” effect sounds. This sound effect is of primary importance
to the “fun” of the game, as it serves as an “aural reward” each time players direct Mario
to collect a coin. In addition, this sound effect directly affects the orchestration of the
background music in SMB. It is created by the two pulse wave channels of the NES, so
any melody which utilizes either or both of those channels is attenuated each time Mario
collects a coin. For a very brief period, as this sound effect plays, the background music
contains only a bass line, provided by the TWC, and a drum loop, provided by the NC.
“Castle” themes change each time Mario grabs a coin. This effect was used more
prominently for the U.S. version of Super Mario Bros. 2, where the pulse channels are
muted each time the player pauses the game, leaving only the triangle and noise channels
Game scorers work with, and alongside, sound effects to produce a coherent aural
world for games. Sound effects must match the music's aesthetic, and vice versa. The
analytic categories of “sound effects” and “music” are only useful, then, in the
categorization of sounds in a game score, but neither is more central or crucial to a game
score. One does not belong to the world of the game score, while the other belongs to the
world of gameplay, for instance. A game score is a product of gameplay, after all. To
demonstrate this, I now turn my attention to the game score for “World 6-2” of SMB.
70
This case study examines the game score for the 22nd level of SMB, namely,
“World 6-2.” I use SMB as the setting for this game score because of its simplicity and
relatively recognizable structure. In addition, SMB is an NES game, so Chapter 2’s case
study of the NES sound hardware configuration will inform this analysis. I shall examine
each section of the level, starting with a visual overview of each section. Figure 14 is an
Fig. 14: “World 6-2” of Super Mario Bros. During gameplay, the gamer may only see a
small section of this level. The pipe labelled “A” leads to an “underworld” area, while
the “B” pipe takes Mario back to the “overworld.” The “C” pipe leads to an
“underwater” area, while the pipe labelled “D” takes Mario back to the overworld. Point
“E” is a secret “beanstalk” which Mario may climb up to a “bonus” area, while “F” is
where Mario may fall back to the overworld. Finally, pipe “G” leads to another
underworld area, which has the “H” pipe as its exit back to the overworld. (“Super Mario
Bros. Maps”)
72
In SMB, players must navigate Mario through “worlds” or levels by using his
fireballs, depending on whether he has a “Fire Flower” “power-up.” Mario may also
Enemies such as “Piranha Plants” intermittently block his way down pipes, while
with green shells, and “Buzzy Beetles” (turtle-like creatures with harder, darker shells),
An almost infinite number of game score variations may result from any given
“playthrough” of World 6-2 of SMB. Musically, the beginning of the level is always the
same: Mario starts outside the castle from the previous level, and the first notes of the
“Overworld” theme play. From here, the game score for the level may conclude at any
point in which Mario “dies,” an action which triggers the “Death Sound,” a comedic
descending riff punctuated by what sounds like bongos from the NC. A number of
actions result in Mario’s death: running into any enemy while in “small Mario” form,
falling down “pits,” or allowing the time limit to expire, which, in World 6-2, is 400
“seconds.”66 A game score can be cut short by any of these occurrences, or it may be
extended indefinitely if at any time players press the “Start” button, which pauses the
If players choose to progress through the level normally, the game score will
usually be accented with the “Boing!” sound effect produced when Mario jumps, an
66
One “second” in Super Mario Bros. is equal to 0.4 seconds in real life, so the actual time limit for World
6-2 is 160 seconds.
73
action crucial to navigating the game world. This sound effect plays through one of the
pulse channels in such a way that it does not normally affect the melody of the song, as
with the “Coin” sound effect. The latter, as noted before, uses both pulse channels of the
NES APU, so the main melody of the “Overworld” theme, also requiring both PWCs, is
In the overworld of World 6-2, there are only three blocks which contain coins.
Two of them are invisible and yield one coin each, and they are indicated by the dotted
square outlines in Figure 14 on p.71 of this thesis. The last of these is a “Ten-Coin
Block,” indicated by a coin inside a brick block just past the first pipe, accessible only by
finding and jumping on the invisible coin block beneath it, and avoiding or killing the
Koopa Troopa below. A Ten-Coin Block may yield anywhere from one to ten coins,
depending on how fast the player can repeatedly make Mario jump into it. Hitting this
block triggers both a “thud” sound effect, presumably meant to signify Mario’s head or
hand hitting the block, and of course the sound effect for collecting a coin. If the player
repeatedly hits the block — as they are encouraged to by the “Ching!” sound — this can
significantly change the game score for the level. In that case, the melody of the
“Overworld” theme is attenuated, and its bass line is similarly interrupted by the
numerous “thuds” of Mario hitting the block, which require the TWC’s resources. The
sound effect of Mario’s jumps will play at this time as well, so in this instance the score
is pervaded by sound effects. Whether or not these sound effects ever sound in Kondo’s
score for SMB is determined completely by each gameplay, as is the tempo and number
of sound effects.
74
In this same area, a Koopa Troopa poses a threat below the invisible coin block.
If Mario jumps on this Koopa Troopa, the “Stomp” sound effect plays, and the Koopa
hides in his shell. The player may then navigate Mario towards the shell, which makes
him kick it and send it sliding across the ground, bouncing off any obstacles and killing
any enemies in its path, as well as Mario. This has varying musical consequences. If I
make Mario kick the shell between the first two pipes, for instance, the shell will rapidly
bounce back and forth, triggering a “Thud!” sound — a sound louder than the “thud”
heard when Mario hits a block — each time it hits a pipe, thereby adding an offbeat
Other sound effects are triggered when Mario obtains a “power-up” like a “Super
these items. Appropriately, the “power-up sound” is an ascending melody, while the
“damage sound” is composed of the same notes in the reverse order. Power-ups carry
over from level to level in SMB, and since World 6-2 is a later level in the game, Mario
may already have a Super Mushroom or a Fire Flower, effectively changing the
possibilities of such sound effects even occurring. Furthermore, Mario’s Fire ability,
obtained through collecting the Fire Flower power-up, has its own sound effect which
sounds each time Mario throws a fireball, and is a means of killing enemies from a
distance.
star in a brick block after the twentieth pipe in Figure 14. When Mario is invincible in
SMB, the background music switches entirely from whatever theme is playing to the
75
repetitive “danceable” melody. Players have considerable control over if and when this
piece plays. In fact, it is gameplay that dictates if and when Mario collects the Starman
power-up, and thus whether or not the “Invincibility” theme ever even sounds. Given the
Starman’s erratic movements, however, the player may wish to make Mario collect the
Starman and fail, in which case whichever theme is currently playing simply continues.
World 6-2 has four hidden areas, accessible to Mario via pipe or beanstalk, each
with their own background music. The entrance to the first of these, an underworld area
stocked with coins, is the pipe labelled “A” in Figure 14. Figure 15 below shows this
underworld area.
Fig. 15: The secret underworld area of World 6-2 in Super Mario Bros. When the player
navigates Mario down the pipe labelled “A” in Figure 2, he falls from the top of this
screen at the position labelled “A” here. Ten coins and one Ten-Coin Block await Mario
in this area. (“Super Mario Bros. Maps”)
76
When players navigate Mario to the underworld in World 6-2, the background
music changes entirely to the “Underworld” theme. This piece contains short,
punctuated, sporadic-sounding riffs on the pulse channel, each divided by large sections
of silence. At the end of each loop there is a longer, descending melody composed of
these quick notes, which resolves the tension created by the heavy use of silence. These
cramped underworld areas in Super Mario Bros. In the first underworld area of World 6-
2, the player may make Mario jump on the pipe and hit the Ten-Coin Block for up to ten
coins. They may also jump up, run over, and jump into the brick structure housing the
other ten coins. Again, whenever Mario grabs a coin, the pulse channels sound the
“Ching!” effect instead of the melody. Alternatively, players may choose to bypass the
coins altogether and make Mario head straight for the pipe, effectively cutting the
“Underworld” theme section of the score short. Finally, the player may remain in this
area for the remainder of the time limit, which would result in a game score concluding
with the “Underworld” theme and finally the “Death Sound.” Traversing the “B” pipe
leads Mario back to the pipe labelled “B” in Figure 14, which once again triggers the
If players navigate Mario down the pipe labelled “C” in Figure 14, he ends up in
triggers the background music to switch to the “Underwater” theme. The “Underwater”
theme, unlike the other themes in Super Mario Bros., is a slow waltz-like melody,
perhaps meant to evoke the motion of Mario gracefully swimming past obstacles and
77
enemies. Again, any occurrence of the “Coin” sound effect interrupts the melody line of
the “Underwater” theme, as it does with the other themes. Mario has a new sound effect,
similar to his “jump” sound effect but shorter, quieter, and lower in pitch, which plays
whenever he takes a stroke through the water, which he does whenever the player presses
the “A” button. In order to avoid being sucked down by the currents, the player must
rapidly press “A” to make Mario swim vigorously, whenever he is over the “pits.”
Naturally, this produces many instances of the “swim stroke” sound effect. If a player
“dies” here, then the game score of course ends with the “Underwater” theme and,
Fig. 16: The hidden underwater area in World 6-2 of Super Mario Bros. Threats to
Mario include squid-like creatures known as “Bloopers,” and more pits, which this time
have currents which pull him downward. Mario sinks downward from the point labelled
“C,” if he travels down the pipe labelled “C” in Figure 14. (“Super Mario Bros. Maps”)
The third and final type of hidden area accessible in World 6-2 of Super Mario
Bros. is the “bonus stage.” This area is only accessible if the player makes Mario find the
beanstalk hidden in a brick block just above the eleventh pipe of the overworld. If Mario
climbs the beanstalk at the point labelled “E” in Figure 14, he ends up at point “E” of the
Fig. 17: The “bonus stage” of World 6-2 of Super Mario Bros. In order to reach the
coins arranged in the sky, the player must make Mario jump onto a cloud platform, which
promptly begins moving horizontally to the right, all the way to the end, where it finally
leaves the screen. To exit the area, Mario simply has to fall back down to the overworld
at the point labelled “F,” which returns him to the point labelled “F” in Figure 14.
(“Super Mario Bros. Maps”)
Bonus stages in SMB have the same theme as the “Invincibility” theme. The only
difference is that while the “Invincibility” theme has a set duration — the duration being
the specific period of Mario’s invincibility — the “Bonus Stage” theme may continue for
as long as Mario is in the bonus stage, a duration which the player has control over.
There are many coins in bonus stages, so the pulse channels must often rapidly switch
between playing the melody of the “Bonus Stage” theme and playing the coin sound
effect, as players make Mario jump (triggering the “Jump” sound effect) up and down
from the cloud platform. The only way to “die” in a bonus stage is to run out of time,
80
which would result in the game score for this level ending with the “Bonus Stage” theme
progressing to the end of the level, marked by a flag, which, if Mario touches it, triggers
the “flag” sound effect and finally the “Course Clear Fanfare” theme, a triumphant
ascending melody. The “flag” sound effect is a slide from low to high frequencies on the
two PWCs. This sound effect increases in length, and its final pitch increases, the higher
the player makes Mario jump onto the flagpole. Furthermore, if the timer’s last digit is a
“1,” “3,” or “6,” then the respective amount of fireworks will appear over the castle,
which have their own “explosion” sound effect provided by the NC. In effect, the ending
of a level can contain many different arrangements of sounds, all of which depend upon
gameplay patterns.
Finally, if at any time during the level the timer reaches “100,” the “Hurry Up!”
jingle plays, and the song currently playing then increases to a faster tempo. Each theme
in Super Mario Bros. therefore has a “Hurry Up!” variation, which can occur at any
physical point in the level. For example, I may choose not to move Mario in World 6-2
until the timer has reached “100,” which would produce a fast version of the
“Overworld” theme. Similarly, if I am running out of time in the underwater area, the
“Underwater” waltz will play at a faster rate. Even the already fast tempo of the
“Invincibility” theme doubles when this happens, resulting in an especially frenetic piece.
81
Despite its ubiquitous and archetypical status, the complete game score for SMB
may not be found on any single officially-released music album. Famicom Sound
History Series: Mario the Music (2004) is one of a handful of releases which contain all
the background music pieces from SMB, though it contains none of the game’s sound
effects. Those wishing to own an album which contains these sound effects must seek
out Super Mario History 1985-2010 (2010), a booklet and soundtrack CD bundle which
is only included with the game Super Mario All-Stars Limited Edition (2010) for the
Nintendo Wii. Unlike actual game scores, officially-released video game soundtracks
The most important difference between game soundtracks and game scores, for
my purposes, is that while game scores are inherently interactive, game soundtracks are
dependent upon gameplay patterns, a game soundtrack exists as a fixed sequence of audio
information. One listens to game soundtracks, while one performs game scores. Table 6
is a partial track list for Famicom Sound History Series: Mario the Music.
82
Table 6: Tracks 4-11 of Famicom Sound History Series: Mario the Music, which
includes all of the background music from Super Mario Bros., as well as the “Death
Sound,” which, arguably, may be considered a sound effect.
similar manner. The “normal” version of the theme plays at its normal tempo, then, after
it has looped a few times, the “Warning” sound plays, and the so-called “Hurry Up!”
variation on the theme plays at its faster tempo. Each track thus includes sounds which
one might hear when playing the game, though the compositional structure of each track
never changes. While I may navigate Mario to die after 5 seconds of playing World 6-2
of SMB, resulting in a brief game score, the tracks contained on Famicom Sound History
Aurally, the most obvious difference between playing SMB and playing Famicom
Sound History is the lack of sound effects in the latter. Game scoring involves a
recognition and integration of sound effects into the musical process, and game scores
contain many instances of sound effects which are triggered as a direct result of gameplay
usual sound effects heard in-game, for a more “pure” presentation of the music. If a
game soundtrack offers a game’s sound effects at all, they are typically bunched together
under one “Sound Effects” track, or, as with Super Mario History, they are offered as
Table 7: Tracks 11-20 of Super Mario History 1985-2010, which includes most, but not
all of the sound effects in SMB.
The sound effects of SMB are offered on Super Mario History unaccompanied.
Oddly, the game score for SMB is unable to even produce sound effects in this manner.
Background music is always playing during gameplay, which is the only state which may
produce such sound effects in-game. Not only is a game score capable of orchestrations
not possible on a game soundtrack, then, a game soundtrack may similarly contain
Conclusion
because of its dependence on chance operations and gameplay. In this chapter, I have
examined game scoring on its own terms, by analyzing actual game scores, resulting
from my own gameplay of Super Mario Bros. Just as John Cage wrote rules for how
sounds are to be produced in his scores, game scorers like Koji Kondo are cognizant of,
and in many cases determine, the parameters under which sounds and music are triggered
in gameplay. For example, during one playthrough of World 6-2, I had directed Mario to
the bonus stage with less than 100 seconds left on the timer. The “Bonus Stage” theme
was triggered by my entrance to the area, and it played at a faster tempo because of the
current state of the timer. As I proceeded to collect as many coins as possible, the
background music only played the frenetic TWC bass line and NC drum loop, both
PWCs’ resources exhausted by numerous instances of the “Coin” sound effect. The
creation of this strange soundscape is only possible through my own gameplay patterns.
In fact, game scores are only ever materially realized by gamers, through gameplay.
Game soundtracks, as I hope I have shown, are simply fixed sequences of audio
game scoring example is impossible to hear on any soundtrack for SMB, because game
soundtracks lack the interactivity involved in game scoring. Game scoring exists as part
85
of gaming, per se, a medium which is inherently interactive, so it follows that game
Chapter 4
Summary
“game scoring,” that is, composing music for and through gaming. As I hope I have
shown, game scoring involves a host of technical and aesthetic priorities, values,
obstacles and concerns which do not influence scoring for other media. Most
their compositions, as game scores are only ever realized through gameplay.
Accordingly, game scorers compose for particular sound hardware configurations, and
through a case study of a particular gaming sound hardware configuration: the NES APU.
Just as a game scorer would approach this sound hardware technology, I examined the
APU for its musical abilities. As Koji Kondo suggests, the NES sound hardware
configuration structures NES game scoring in its own peculiar way, which is different
from other gaming sound hardware configurations. For example, the NES APU offered
only five discrete sound channels, while the later SNES sound hardware had eight.
87
As it turns out, NES game scorers developed unique and innovative compositional
strategies in order to program musical ideas into the APU (which in many cases would be
impossible otherwise). After examining the NES APU’s musical possibilities and
this sound hardware configuration’s abilities and limitations. These examples were
meant to demonstrate both the technological structure of game scoring, and the broader
point that this compositional activity resembles software programming more than any
“unmusical” nature. Programming, coding and gaming are involved in game scoring, and
these activities are no less musical than writing notes on a staff. I forego the outdated
distinction between programming and composition, and instead suggest that game
gameplay (the duration of any game score, for instance, is always determined by the
unexpected benefit, Cage composed “sound effects” for this piece, which would normally
expect, integrate and even compose sound effects into their scores. I used this similarity
between game scoring and aleatoric composition to argue that the distinction between
“music” and “sound effects” is irrelevant to the game scoring analyst. Game scorers
program video game sound effects just as they program video game music, so I
I then performed an in-depth case study of World 6-2 of SMB to elucidate game
for my gameplay choices in World 6-2, I did not expect to find so many. Even in an
early game for the NES (an earlier home gaming technology by today’s standards), Koji
Kondo had to program numerous musical ideas for numerous gameplay states in SMB.
For example, when I direct Mario to collect the “Starman” power-up, the game score
switches entirely to the “Invincibility” theme. I outlined all the ways in which my
gameplay could affect the game score for World 6-2, in an effort to demonstrate that the
sountracks for SMB. This comparison highlighted both the “fixed” nature of video game
soundtracks, as well as the “unfixed” nature of game scores. Game scoring is aleatoric
because the gamer presents “chance” and a degree of “performer” freedom to the final
composition. As expected, the video game soundtracks for SMB were only “idealized”
renditions of game scores, and were even impossible to produce by gameplay itself. This
Future Directions
With this thesis, I have many directions in which to take my research. I find the
interaction between the game scorer (and the gamer67) and gaming technology most
configurations, just as the gamer develops ludal strategies in response to the world, rules
fact, game scoring may affect the latter in a peculiar way. Beyond greed, my motives for
directing Mario to repeatedly collect coin after coin — especially in the bonus stage —
stem from a “circular causal” relationship generated by the system of visuals, haptics, and
most importantly for my purposes, sound of Super Mario Bros. In this case, the “Coin”
sound effect exists as both a feedback and control mechanism, because it is both a reward
and a motive for my gameplay patterns (Whalen: 2010). I am, in effect, involved in a
Similarly, the act of game scoring, that is, scoring in a traditional sense, by a
technology. For example, Koji Kondo composed the score for Super Mario Bros. in a
music editor program he wrote himself in Family BASIC, a dialect of the BASIC
programming language68 that is used to program the Famicom. As such, his written
67
By now it should be clear that the gamer is, in fact, part game scorer.
68
Beginner’s All-purpose Symbolic Instruction Code, or “BASIC,” is a family of general-purpose, high-
level programming languages.
90
“score” for SMB exists as code in the Family BASIC language, which he could make
changes to, in order to effect various musical outcomes (one of which is the realtime re-
orchestration of APU channels in the event Mario collects a coin). Kondo’s feedback and
control mechanisms are the sounds he is able to produce from the NES APU through this
code, as they constitute his motive and reward. As such, this “closed signaling loop” is
indicative of a “cybernetic system,” which is any system which involves this kind of
“circular-causal” relationship.
Future directions for my research could involve discussing the “closed signaling
loop” involved in the cybernetic systems of gameplay and game scoring. This would
avenue.
Significance
media studies in kind, its most significant contribution is within the field of ludo-
yet to be done on game scoring per se.69 This thesis is relatively unique in that it
69
Examples of work on game scoring that currently exist include, but are certainly not limited to: Belinkie
(1999); Berndt (2009; 2011); Berndt and Hartmann (2007; 2008); Berndt et al. (2006); Brame (2011);
Bridgett (2010); Chan (2007); Childs IV (2007); Collins (2005; 2006; 2007a; 2007b; 2007c; 2007d; 2008a;
2009); d’Escrivan (2007); DeCastro (2007); Deutsch (2003); Farnell (2007); Fay (2004); Furlong (2004);
91
concerns itself with video game music from a “scoring” perspective. Indeed, recent
studies of interactive media have tended to overlook the act of game scoring itself,
fixating instead on the product of that process.70 Most fundamentally, then, this thesis is
though culturally significant, remains conspicuously absent from the lion’s share of
The ludo-musicological analysis of game scoring I have performed here, that is, a
approaches to film scores are unsuitable for studying game scoring. Despite attempts by
scholars to make terms and concepts from film studies fit in studies of game scores, such
analyze other media map clumsily onto the video game medium, and may even work to
Moreover, analysts, non-academic writers, and gamers alike often confuse video
game soundtracks with actual game scores. In this thesis I drew a technological
distinction between these two media: a video game soundtrack is a record, realized by
Gibbons (2009); Herber (2008); Hermans (2013); Hoffert (2007); Kaae (2008); Knight (1987); Lendino
(1998); Lieberman (2006); Marks (2009); Mera (2009a; 2009b); Sanger (2003); Schmidt (1989); Sweeney
(2011); Van Geelen (2008); Whitmore (2004); and Wooller et al. (2005).
70
Examples include, but are certainly not limited to: Jones (2008); Thornham (2011); Waggoner (2009);
Wolf and Perron (2013); Crawford, Gosling and Light (2013); Domsch (2013); Egenfeldt-Nielsen, Smith
and Tosca (2013); Ensslin (2011); Gamboa (2012); Garrelts (2005); Huntemann and Aslinger (2012); Juul
(2011); Ruggill and McAllister (2011); Whalen and Taylor (2008); Paul (2012); and Newman (2012),
among others.
71
Work that approaches game scoring via film scoring analysis models includes, but is not limited to:
Arrasvuori (2006); Boyd (2003); Chan (2007); Collins (2007d; 2008c); Hoover (2010); Jørgensen (2004;
2006; 2007a; 2007b; 2008b; 2009; 2010); Kamp (2014); Mera (2009a); Munday (2007); van Elferen
(2011); Whalen (2004; 2007); Wilhelmsson and Wallén (2010); Wood et al. (2009); and Zehnder and
Lipscomb (2004; 2006), among others.
92
game score is only realized through gameplay, and produced by gaming technology.
This thesis is an analytic model designed to examine game scoring vis-a-vis concepts
drawn directly from gaming rather than borrowed from other cultural forms (such as
Implications
The act of game scoring can be analyzed through a multitude of theoretical lenses.
and its result is a working methodology for analyzing game scoring via a ludo-
musicological research framework. More specifically, one of the outcomes of this thesis
was a methodology for analyzing any particular game score. While I chose to analyze the
NES APU for its relative simplicity, and my scores for SMB for my familiarity with its
“world,” my research methods may be applied to all gaming technologies and game
scores, regardless of their complexity. As I have shown, game scores must be analyzed
first-and-foremost in terms of their technology, and then through the gameplay used to
realize them. The methodology I developed in this thesis allows for such an analysis.
machines in their own right. Mobile gaming is growing more and more popular each
day, though very little research has been conducted on the process of “mobile game
93
scoring.” Mobile devices present yet another technology for the game scorer to
orchestrate, though they present very different aesthetic challenges than, say, the NES
APU. Most significantly, mobile video games are developed for numerous different
devices, so their game scores are structured by numerous different sound hardware
configurations, and realized by numerous different gameplay scenarios. Even the choice
of wearing headphones has its effects on the realization of a mobile game score. My
methodology allows for an analysis of this peculiar type of game score, too, since it
approaches game scoring from a broad perspective, and certainly includes scoring for
mobile games.
Sociological and psychological research on video game music may also benefit
from my analysis of game scoring. For instance, many studies from these fields on video
game music focus on the widespread use of electronic gambling machines.72 These
studies tend to analyze sounds and music used in these machines, as “feedback” and
“control” mechanisms which encourage playing (and spending). As noted, the nature of
these mechanisms is dependent on the nature of the game scoring process, which itself is
gambling machine, and thus structures the game scoring process in a unique way. Since
music is posited by these researchers as a crucial element in encouraging play, they may
72
For significant examples, see: Collins et al. (2011); Noseworthy (2009); and Dixon, Trigg and Griffiths
(2007), among others.
94
machines, and their games, are scored. This may reveal more aspects of game scores
which facilitate more explicit capitalistic intentions than do “regular” game scores.
Research from psychology, sociology, media studies and cultural studies alike
have begun to address the extremely “gendered” identity of the average gamer.73 While
the proliferation of mobile gaming may have increased the amount of female gamers in
recent years, the gender identity of gaming remains predominately masculine. Moreover,
this phenomenon was much more pronounced in earlier gaming eras. How has game
of how that music is created in the first place, thus allowing for a more thorough
examination.
Conclusion
To be sure, “video game music” differs from game scoring, in that the former is a
“thing,” while the latter remains an ongoing activity. While video game music may be
73
For example, see: Brown et al. (1997); Eden et al. (2010); Feng et al. (2007); Chess (2011); Ferguson,
Cruz and Rueda (2008); Greenberg et al. (2010); Jantzen and Jensen (1993); Homer et al. (2012); Behm-
Morawitz (2014); Dietz (1998); Cruea and Park (2012); Scharrer (2004); Miller and Summers (2007);
Perry (2011); Hamlen (2010); Williams et al. (2009); Ogletree and Drake (2007); Gailey (1993); Ivory
(2006); and Soukup (2007), among many others.
95
Bibliography
Jansz, Jeroen, and Lonneke Martens. 2005. “Gaming at a LAN Event: The Social Context
of Playing Video Games.” New Media & Society 7 (3): 333–55.
doi:10.1177/1461444805052280.
Jantzen, Gitte, and Jans F. Jensen. 1993. “Powerplay — Power, Violence and Gender in
Video Games.” AI & Society 7 (4): 368–85.
Jennifer, Martin. 2012. “Keeping Up with the Virtual Joneses: The Practices, Meanings,
and Consequences of Consumption in Second Life.” University of Western
Ontario. http://ir.lib.uwo.ca/etd/386.
Jones, Steven E. 2008. The Meaning of Video Games: Gaming and Textual Strategies.
Routledge.
Jørgensen, Kristine. 2004. “Sounds and Sources in Sacred.” In Proceedings from Cosign
2004. Split, Croatia.
———. 2006. “On the Functional Aspects of Computer Game Audio.” In Proceedings of
the Audio Mostly Conference. Interactive Institute/Sonic Studio, Pitea, Sweden.
———. 2007a. “On Transdiegetic Sounds in Computer Games.” Northern Lights: Film
& Media Studies Yearbook 5 (1). doi:10.1386/nl.5.1.105_1.
———. 2007b. “‘What Are These Grunts and Growls Over There?’ Computer Game
Audio and Player Action.” PhD Dissertation, Copenhagen University.
———. 2008a. “Audio and Gameplay: An Analysis of PvP Battlegrounds in World of
Warcraft.” Game Studies 8 (2). http://gamestudies.org/0802/articles/jorgensen.
———. 2008b. “Left in the Dark: Playing Computer Games with the Sound Turned Off.”
In From Pac-Man to Pop Music: Interactive Audio in Games and New Media,
edited by Karen Collins, 163–76. Aldershot, Hampshire and Burlington, VT:
Ashgate.
———. 2009. A Comprehensive Study of Sound in Computer Games: How Audio Affects
Player Action. Monograph. Lewiston, NY: Mellen.
———. 2010. “Time for New Terminology? Diegetic and Non-Diegetic Sounds in
Computer Games Revisited.” In Game Sound: Game Sound Technology and
Player Interaction: Concepts and Developments, edited by Mark Grimshaw, 78–
97. Hershey, PA: IGI Global.
Juul, Jesper. 2011. Half-Real: Video Games Between Real Rules and Fictional Worlds.
MIT Press. http://books.google.ca/books?id=mZP6AQAAQBAJ.
Kaae, Jesper. 2008. “Theoretical Approaches to Composing Dynamic Music for Video
Games.” In From Pac-Man to Pop Music: Interactive Audio in Games and New
Media, edited by Karen Collins, 75–92. Aldershot, Hampshire and Burlington,
VT: Ashgate.
Kaernbach, Christian. 2000. “Periodic Noise FAQ.” Christian-Albrechts-Universitat Zu
Kiel. http://www.uni-kiel.de/psychologie/emotion/team/kaernbach/pn/faq.htm.
Kamp, Michiel. 2009. “Ludic Music in Video Games.” Master’s Thesis, Utrecht
University.
———. 2014. “Musical Ecologies in Video Games.” Philosophy & Technology 27 (2):
235–49.
Kärjä, Antti-Ville. 2008. “Marketing Music through Computer Games: The Case of Poets
of the Fall and Max Payne 2.” In From Pac-Man to Pop Music: Interactive Audio
in Games and New Media, edited by Karen Collins, 27–44. Aldershot, Hampshire
and Burlington, VT: Ashgate.
104
Conference (I/ITSEC).
Marks, Aaron. 2009. The Complete Guide to Game Audio for Composers, Musicians,
Sound Designers, and Game Developers. Burlington, MA and Oxford: Focal
Press.
Marks, Aaron, and Jeannie Novak. 2009. Game Development Essentials: Game Audio
Development. Monograph. New York: Delmar.
Mera, Miguel. 2009a. “An Interview with James Hannigan.” Music, Sound and the
Moving Image 3 (1): 99–115.
———. 2009b. “Invention/Re-Invention.” Music, Sound and the Moving Image 3 (1): 1–
20.
Miller, Kiri. 2007. “Jacking the Dial: Radio, Race, and Place in Grand Theft Auto.”
Ethnomusicology 51 (3): 402–38.
———. 2008a. “Grove Street Grimm: Grand Theft Auto and Digital Folklore.” Journal
of American Folklore 121 (481): 255–85.
———. 2008b. “The Accidental Carjack: Ethnography, Gameworld Tourism, and Grand
Theft Auto.” Game Studies 8 (1). http://gamestudies.org/0801/articles/miller.
———. 2009. “Schizophonic Performance: Guitar Hero, Rock Band, and Virtual
Virtuosity.” Journal of the Society for American Music 3 (4): 395–429.
Miller, Monica K., and Alicia Summers. 2007. “Gender Differences in Video Game
Characters’ Roles, Appearances, and Attire as Portrayed in Video Game
Magazines.” Sex Roles 57: 733.
Mitchell, Helen R. 2014. “Fear and the Musical Avant-Garde in Games: Interviews with
Jason Graves, Garry Schyman, Paul Gorman and Michael Kamper.” Horror
Studies 5 (1): 127–44.
Morgan, R.P. 1991. Twentieth-Century Music: A History of Musical Style in Modern
Europe and America. Norton. http://books.google.ca/books?id=LIxrQgAACAAJ.
Morris, Sue. 2004. “Shoot First, Ask Questions Later: Ethnographic Research in an
Online Computer Gaming Community.” Media International Australia 110: 31–
41.
Mullan, Eoin. 2010. “Physical Modelling for Sound Synthesis.” In Game Sound
Technology and Player Interaction: Concepts and Developments, edited by Mark
Grimshaw, 340–60. Hershey, PA: IGI Global.
Munday, Rod. 2007. “Music in Video Games.” In Music, Sound and Multimedia: From
the Live to the Virtual, edited by Jamie Sexton, 51–67. Edinburgh: Edinburgh UP.
Munday, Roderick. 2009. “Glossary of Terms in Heidegger’s Being and Time.” Visual
Memory. http://www.visual-memory.co.uk/b_resources/b_and_t_glossary.html.
Murphy, Conor. 2012. “Orchestras That Play Video Game Music.” Big Fish Games.
http://www.bigfishgames.com/blog/orchestras-that-play-video-game-music/.
Murphy, David, and Flaithrí Neff. 2010. “Spatial Sound for Computer Games and Virtual
Reality.” In Game Sound Technology and Player Interaction: Concepts and
Developments, edited by Mark Grimshaw, 340–60. Hershey, PA: IGI Global.
Murray, Janet H. 2006. “Toward a Cultural Theory of Gaming: Digital Games and the
Co-Evolution of Media, Mind, and Culture.” Popular Communication 4 (3): 185–
202.
Nacke, Lennart, and Mark Grimshaw. 2010. “Player-Game Interaction Through Affective
Sound.” In Game Sound Technology and Player Interaction: Concepts and
106
Perle, George. 1963. Serial Composition and Atonality: An Introduction to the Music of
Schoenberg, Berg, and Webern. Berkeley, CA: University of California Press.
Perron, Bernard, ed. 2009. Horror Video Games: Essays on the Fusion of Fear and Play.
Jefferson, NC: McFarland.
Perry, Stephen. 2011. “Top MCS Downloads for 2010: Sexual Content, Video Game
Violence, and Gender Studies Grab Attention.” Mass Communication and Society
14 (4): 403.
Phillips, Winifred. 2014. A Composer’s Guide to Game Music. Cambridge, Mass. and
London: MIT Press.
Pichlmair, Martin, and Fares Kayali. 2007. “Levels of Sound: On the Principles of
Interactivity in Music Video Games.” Digital Games Research Association
Conference. Tokyo.
Rayman, Joshua. 2014. “Experimental Approaches to the Composition of Interactive
Video Game Music.” Master’s Thesis, University of East Anglia.
Reale, Steven. 2011. “Chaos in the Cosmos: The Play of Contradictions in the Music of
Katamari Damacy.” ACT 2 (1). http://www.oalib.com/paper/2082553.
Reiter, Ulrich. 2010. “Perceived Quality in Game Audio.” In Game Sound Technology
and Player Interaction: Concepts and Developments, edited by Mark Grimshaw,
153–74. Hershey, PA: IGI Global.
Roig-Francolí, Miguel A. 2008. Understanding Post-Tonal Music. Boston: McGraw-Hill.
Roux-Girard, Guillaume. 2009. “Plunged Alone into Darkness: Evolution and the Staging
of Fear in the Alone in the Dark Series.” In Horror Video Games: Essays on the
Fusion of Fear and Play, edited by Bernard Perron, 145–67. Jefferson, NC:
McFarland.
———. 2010. “Listening to Fear: A Study of Sound in Horror Computer Games.” In
Game Sound Technology and Player Interaction: Concepts and Developments,
edited by Mark Grimshaw, 192–212. Hershey, PA: IGI Global.
Rubin, Justin. 2005. “Aleatoric Music.” University of Minnesota Duluth.
http://www.d.umn.edu/~jrubin1/JHR%20Alea%201.htm.
Ruggill, J.E., and K.S. McAllister. 2011. Gaming Matters: Art, Science, Magic, and the
Computer Game Medium. University of Alabama Press.
http://books.google.ca/books?id=CVkfAnl01_EC.
Salen, Katie. 2007. “Toward an Ecology of Gaming.” The Ecology of Games: Connecting
Youth, Games, and Learning, 1–17.
Sanders, Timothy. 2010. “Time Perception, Immersion and Music in Videogames.” The
24th British Human Computer Interaction Conference. London.
Sanger, George. 2003. The Fat Man on Game Audio: Tasty Morsels of Sonic Goodness.
Monograph. Indianapolis: New Riders Games.
Scharrer, Erica. 2004. “Virtual Violence: Gender and Aggression in Video Game
Advertisements.” Mass Communication and Society 7 (4): 393–412.
Schartmann, Andrew. 2013. Maestro Mario: How Nintendo Transformed Videogame
Music into an Art. New York: Thought Catalog.
Schleiner, Anne-Marie. 1999. “Parasitic Interventions: Game Patches and Hacker Art.”
Disponibile Online. http://www.opensorcery.net/patchnew.html.
———. 2001. “Does Lara Croft Wear Fake Polygons? Gender and Gender-Role
Subversion in Computer Adventure Games.” Leonardo 34 (3): 221–26.
108
doi:10.1162/002409401750286976.
Schmidt, Brian L. 1989. “Designing Sound Tracks for Coin-Op Games, Or, Computer
Music for Under $65.00.” In Proceedings: 1989 International Computer Music
Conference, 273–76. San Francisco: Computer Music Association.
Schott, Gareth, and Darrin Hodgetts. 2006. “Health and Digital Gaming: The Benefits of
a Community of Practice.” Journal of Health Psychology 11 (2): 309–16.
doi:10.1177/1359105306061189.
Schott, Gareth R., and Kirsty R. Horrell. 2000. “Girl Gamers and Their Relationship with
the Gaming Culture.” Convergence: The International Journal of Research into
New Media Technologies 6 (4): 36–53. doi:10.1177/135485650000600404.
Schütze, Stephan. 2008. “The Creation of an Audio Environment as Part of a Computer
Game World: The Design for Jurassic Park - Operation Genesis on the Xbox"™
as a Broad Concept for Surround Installation Creation.” Organized Sound 8 (2):
171–80.
Shultz, Peter. 2008. “Music Theory in Music Games.” In From Pac-Man to Pop Music:
Interactive Audio in Games and New Media, edited by Karen Collins, 27–44.
Aldershot, Hampshire and Burlington, VT: Ashgate.
Sich, Christy. 2006. “From Game Studies to Bibliographic Gaming: Libraries Tap into
the Video Game Culture.” Bulletin of the American Society for Information
Science and Technology 32 (4).
Smith, Jacob. 2004. “I Can See Tomorrow in Your Dance: A Study of Dance Dance
Revolution and Music Video Games.” Journal of Popular Music Studies 16 (1):
58–84.
Sotamaa, Olli. 2003. “Computer Game Modding, Intermediality and Participatory
Culture.” New Media, 1–5.
Soukup, Charles. 2007. “Mastering the Game: Gender and the Entelechial Motivational
System of Video Games.” Women’s Studies in Communication 30 (2): 157.
Steinkuehler, Constance A. 2005. “The New Third Place: Massively Multiplayer Online
Gaming in American Youth Culture.” Tidskrift Journal of Research in Teacher
Education 3 (3): 17–32.
Sterne, Jonathan. 1997. “Sounds Like the Mall of America: Programmed Music and the
Architectonics of Commercial Space.” Ethnomusicology 41 (1): 22-50.
Stockberger, Axel. 2003. “The Game Environment from an Auditive Perspective.” In
Level Up: Proceedings from DiGRA 2003.
Summers, Tim. 2011. “Playing the Tune: Video Game Music, Gamers, and Genre.” ACT
2 (1). http://www.oalib.com/paper/2085371.
“Super Mario Bros. Maps.” Ian-Albert.com.
http://ian-albert.com/games/super_mario_bros_maps/.
Svec, Henry. 2008. “Becoming Machinic Virtuosos: Guitar Hero, Rez, and Multitudinous
Aesthetics.” Loading... 2 (2).
http://journals.sfu.ca/loading/index.php/loading/article/view/30.
Sweeney, Mark. 2011. “Analysing Crysis.” Master’s Thesis, University of Oxford.
Tan, Siu-Lan. 2010. “Effects of Built-in Audio Versus Unrelated Background Music on
Performance in an Adventure Role-Playing Game.” International Journal of
Gaming and Computer-Mediated Simulations 2 (3).
Taylor, T.L. 2003. “Multiple Pleasures: Women and Online Gaming.” Convergence: The
109
Villarreal III, Armando. 2009. “The Effects of Music on Gameplay and Immersion.”
Master’s Thesis, MIT.
Vreeland, Richard. 2012. “Philosophy of Music Design in Games.” presented at the 2012
Game Audio Network Guild Summit.
Waggoner, Zach. 2009. My Avatar, My Self: Identity in Video Role-Playing Games.
McFarland.
“Weekly.” VGChartz. http://www.vgchartz.com/weekly/.
Weske, Jörg. 2002. “Digital Sound and Music in Computer Games.” 3DAudio.
http://3daudio.info/gamesound/index.html.
“We Spend 3 Billion Hours a Week as a Planet Playing Videogames.” TED.
http://www.ted.com/conversations/44/we_spend_3_billion_hours_a_wee.html.
Western, Gary. 2011. “Top Score: The Evolution and Significance of Video Game
Music.” CSU Monterey Bay.
http://garywestern.files.wordpress.com/2011/04/garywestern_topscore.pdf.
Whalen, Zach. 2004. “Play Along - An Approach to Videogame Music.” Play Along - An
Approach to Videogame Music 4 (1). http://www.gamestudies.org/0401/whalen/.
———. 2007. “Case Study: Film Music vs. Video-Game Music: The Case of Silent
Hill.” In Music, Sound and Multimedia: From the Live to the Virtual, edited by
Jamie Sexton, 68–81. Edinburgh: Edinburgh UP.
110
Whalen, Z., and L.N. Taylor. 2008. Playing the Past: History and Nostalgia in Video
Games. Vanderbilt University Press.
http://www.books.google.ca/books?id=SqMqmGWgo3wC.
White, Glenn D. 1987. “ADSR.” In The Audio Dictionary: 7.
———. 1987. “Limiter.” In The Audio Dictionary: 130.
———. 1987. “Fletcher-Munson Effect.” In The Audio Dictionary: 95-96.
———. 1987. “Overtones.” In The Audio Dictionary: 166.
Whitmore, Guy. 2004. “A DirectMusic Case Study for Russian Squares for Windows XP
Plus Pack.” In DirectX 9 Audio Exposed: Interactive Audio Development, edited
by Todd M. Fay, Scott Selfon, and Todor J. Fay. Plano, Texas: Wordware.
Wilde, Martin D. 2004. Audio Programming for Interactive Games. Monograph. Oxford
and Burlington, MA: Focal Press.
Wilhelmsson, Ulf, and Jacob Wallén. 2010. “A Combined Model for the Structuring of
Computer Game Audio.” In Game Sound Technology and Player Interaction:
Concepts and Developments, edited by Mark Grimshaw, 98–130. Hershey, PA:
IGI Global.
Williams, Dmitri, Nicole Martins, Mia Consalvo, and James D. Ivory. 2009. “The Virtual
Census: Representations of Gender, Race and Age in Video Games.” New Media
& Society 11 (5): 815–34.
Williams, J.P., S.Q. Hendricks, and W.K. Winkler. 2006. Gaming as Culture: Essays on
Reality, Identity and Experience in Fantasy Games. McFarland, Incorporated.
http://books.google.ca/books?id=i7UBWz6LBK4C.
Wirman, Hanna. 2014. “Playing by Doing and Players’ Localization of The Sims 2.”
Television & New Media 15 (1): 58–67. doi:10.1177/1527476413505001.
Wolf, M.J.P., and B. Perron. 2013. The Video Game Theory Reader. Routledge.
Wood, Simon, Graeme Harper, and Ruth Doughty. 2009. “Video Game Music - High
Scores: Making Sense of Music and Video Games.” In Sound and Music in Film
and Visual Media: An Overview, 129–48. New York and London: Continuum.
Wooller, Rene, Andrew R. Brown, Eduardo Miranda, Rodney Berry, and Joachim
Diederich. 2005. “A Framework for Comparison of Processes in Algorithmic
Music Systems.” In Generative Arts Practice, 109–24. Sydney: Creativity and
Cognition Studios Press.
Youngdahl, Erik. 2010. “Playing the Past: History and Nostalgia in Video Games.”
Undergraduate Thesis, Westleyan University.
Young, David M. 2012. “Adaptive Game Music: The Evolution and Future of Dynamic
Music Systems in Games.” Bachelor’s Thesis, Ohio University.
Zehnder, Sean M., and Scott Lipscomb. 2004. “Immersion in the Virtual Environment:
The Effect of a Musical Score on the Video Gaming Experience.” Journal of
Physiological Anthropology and Applied Human Sciences 23 (6): 337–43.
———. 2006. “The Role of Music in Video Games.” In Playing Video Games: Motives,
Responses and Consequences, 241–58. Mahwah, NJ and London: Lawrence
Erlbaum Associates.
111
VITA
EDUCATION
WORK EXPERIENCE
PUBLICATIONS
Enns, Mack. 2014. “Game Scoring: FEZ, Video Game Music and Interactive
Composition.” In Post Conference Proceedings, edited by R. Hepworth-Sawyer,
J.Hodgson, R. Toulson & J.L. Paterson. 2013 Innovation in Music Conference:
York, UK.