Originally posted on Medium
Like this essay? Follow me on Twitter for more!
It’s kind of miraculous to be able to say that Even the Ocean has been out for a year, when at times it felt like the game would never be out, after a 3.5 year development. These past years, I’ve wanted to write about Even the Ocean’s (hereon “ETO”) music. The game is available on Steam and itch.io.
Quick Facts on the Soundtrack
It has a run time of about 4 hours. The vast majority of the music was started and finished from 2013–2014, with a few straggling tracks into 2015/2016. Timbre-wise, the game mainly used the Carbon 2 software synth[1], and a few soundfonts that I used in Anodyne. The Carbon 2 synth is a subtractive synth with a few interesting filters, three oscillators, with a wavetable option for a few of them. It has an interesting built-in chorus and pitch shifting option as well as some interesting LFOs, making it a very flexible and simple synth. I haven’t used it since Even the Ocean, though.
Stylistically[2], the game is all looping music, about 25% ambient music, 30% more straightforward melodic music, then 45% “composed with the Ambient-melodic approach” (and varies in terms of how present melodies are.)
Location-wise, the music breaks down to:
“Mission Area Music” — places like Riverton — where Aliph goes for missions, talks to NPCs, fixes a power plant. This constitutes about 66% of the music.
“City Music”. These play in the city areas such as Whiteforge, during cutscenes, or walking around the city. This is about 25% of the music.
Everything else — peritextual music (title screen, etc), overworld, some ambient pieces.
In his book Understanding Video Game Music (2016), Tim Summers has an outline on “How to Hear a Video Game,” which offers ways of breaking down how we can analyze and listen to a game’s music. It does a great job of breaking down how we would otherwise intuitively critique or comment on a song, and organizing it into a outline format, covering things such as the historical context of the composer, the genre influences, the uses of rhythm, leitmotif, and more.
So I thought, game composers always know that the game influences how the music should sound. And we have this explicit way of analyzing a song. Is there a more explicit way to describe how a game influences how the music is composed?
When writing music, organizational forces (hereon “forces”) help the composer figure out what sort of music they should make, by helping to reduce the gigantic possibility space for a musical composition. At the simplest, this is something like, “a green grassy area, during good weather, should have a song in a major key or a song that feels positive.” That “green grassy area” is a spatial force — limits on the song driven by the type of digital space — whether it be hypertext, 2D or 3D — and its visual qualities. There are also gameplay forces, which include the game’s genre, style of gameplay, which will influence the music, as the music may need to enhance aspects of gameplay (a classic example being racing or shoot-em-up arcade games, which feature very energetic scores.) . And there are narrative forces, which could make a green grassy area have a sad song, if some character just died.
I call these forces because they are a nice visual metaphor, for how a composer’s mind is pushed towards certain sorts of compositional solutions. Rather than “constraints,” which conjure an image of being locked in a box. Although at times it seems like some composers are constrained by using only orchestral instruments…
It’s okay to ignore or go against a force — it’s more of a guideline than anything.
Sometimes when working on a large game, these implicit forces are not enough to satisfactorily realize a composition, and so the composer needs to add some self-imposed forces, which are further constraints on style that make it easier to know how to start one out of maybe over a hundred songs. A personal style helps here, but also specific rules for a game can help.
From Even the Ocean, categories 2 (city music) and 3 (everything else) were less constrained by self-imposed forces — for these I mostly wrote what would fit the scene, without worrying as much if it needed to obey a organizational rule. There’s stuff worth talking about those, but for now, I’ll write about Category 1, or Mission Area Music and its organizational forces.
Mission Area Music
At the highest level of abstraction, ETO’s gameplay can be broken down into two categories, mission and city areas. You receive a mission from the city, watch some cutscenes, maybe walk around the city. Then you go to a mission area, where you work through some problems with NPCs, get through a Power Plant and do your repairs.
A mission area can be further broken down into nature, power plant, and (power plant) core repair areas. For each mission area, each of these three subcategories has at least one song, occasionally with additional songs (such as wind ambiences for interiors in Boreas Heights, or a song for the library in Magdal Woods).
Because this structure of gameplay appears about 10 times, I decided to give myself some self-imposed organization forces (also “guidelines”) to help with composing. This would work by letting me know the style of music to use for a certain part of the game, which kickstarts the composition process.
An example of a set of songs for a mission area is Riverton, Riverton Power Plant, Riverton Plant Core. You can look through the OST and find the same three sorts of songs for each of the game’s 10 or so areas.
Implicit Organizational Forces
Again, these are things that the narrative, gameplay and visuals of ETO pushed the music towards.
Nature: Mostly outdoors, calm, not a lot of action. Speaking to NPCs. Walking around natural geological forms.
Power Plant: Platforming challenges. Indoors, mechanical, closed-feeling. Aliph trying to get to the center of the Power Plant.
Core Repair: The end of a level. More puzzle-y challenges in a single, nonlinear space. Repairing the Power Plant.
Pretty simple stuff — still leaves room open for a lot of approaches. And now for some preliminary self-imposed forces, which are just reasoning about what would really make sense for the game:
Self-Imposed Forces (non-narrative)
Nature: Less intense gameplay implies music will be paid more attention. Should attempt to sit between catchy melodic song, like Anodyne’s Fields, to fit the presence of some humans, and an ambient piece, where very few if any melodic fragments exist, to fit the more outdoors setting.
Power Plant: Keep it upbeat when appropriate (or traditionally structured), focus on driving an immersive, slightly upbeat and driving feeling. But still not too in-your-face, because this is ETO and it’s meant to be a calm and cohesive experience.
Core Repair: Take place at the end of power plant levels, where you repair things in the power plant’s core, often progressing the story when finishing. A bit more focused, ‘realistic.’
Self-imposed Forces (Narrative)
Beyond the above needs, there was a narrative motivation — I wanted the process of moving through mission areas to represent different states of Aliph’s mind. This helped further narrow down how to write the music.
Nature: Nondiegetic (music not ‘coming from’ anything in the game’s setting like a radio or river) Aliph getting acquainted with new places and people, trying to calm her nerves before the work. Mostly calm music, but can shift into different feelings/moods depending on the themes of the area’s plot. Possibly escapist (doesn’t it always feel like that, for at least a moment, when we get off a plane or train?)
Power Plant: Nondiegetic. Aliph beginning her work, trying to get into a state of flow, trying to escape from the implications of her work. Should strongly contrast with Core Repair music.
Core Repair: Diegetic. I wanted this to have the player be able to hear, maybe feel, the sort of ambience, temperature that Aliph is feeling. To emphasize the stark silence that accompanies important decisions in real life. Aliph is a worker caught in a failed solution towards stopping environmental disaster. In silence or the buzz of A/C, people in power in the USA government sign their bills which affect millions the world over. So, I wondered, what does the silence surrounding Aliph’s decision sound like?
One of the major decisions was how to use ambience in the songs. Could you make all those categories ambient, like the Core Repair songs? Sure. Ambience-only is a tricky thing to work with — I think it can be harder to pull off unless you’re doing photorealism. Too much silence or ambience can easily come off as strange or surreal (which may be good! But not for ETO…) Generally, I like melodies and melodic fragments, which led me to deciding on…
The Ambient-Melodic approach.
Recall I wanted the Nature areas to sit between Anodyne’s melodies and more ambient sounds. So in order to make the Nature area music, I came up with an approach for music with ETO’s guidelines. Ambient-Melodic is not really a genre of music, but a thought process for composing. In 2014 I was thinking along the rough categories outlined above, but I didn’t really see it as an ‘approach’ until thinking about it now.
The idea is:
Know what ambient-melodic is (Generally) not appropriate for: character themes, cutscenes, places where you need to think about melody and that stuff in order to make a place memorable in a musical way. (Or, it could be appropriate! For example Humus’s theme is quite strange.)
Focus on timbre as the most important element of the song. (Of course picking a reasonable key, mode, whatever.) Designing an appropriate synth patch is the most important element. Everything else can follow.
Add in melodic fragments as necessary, first as sounds that might be more textural/ambient. Chord progressions set to rhythms (i.e., influence from traditional pop music) can help carve out a feel, and you can always dial back from obvious chords to melodic fragments drawing from those chords or a subtle bassline.
If the song seems to pull into a more melodic feel, and it fits the area, then go with it. But always check with concept or in-game art!
Likewise, if lots of melody isn’t working, then try to stay more in the ambient area.
The more ambient of these songs is The Great Cliff, Dreamdram Canyon, Boreas Heights, Restview Beach, Magdal Woods. These don’t really have rhythm to them, though they are in a set key signature, they are more about the collaging of melodic ideas and sounds.
Slight more melodic are Oscar Basin, Rainy Great Falls.
And much more melodic are Riverton, Karst Pass (easily my favorite song from the OST), Sunny Great Falls (though this song was transplanted into Great Falls, from a scrapped story idea in Magdal Woods.)
I think my first attempt at this approach was the song I made for an IndieStatik (a defunct game website) kickstarter compilation, called “In A Secluded Northern Place”. It’s melodic as far as songs go, but there are a lot of ambient textures popping up and fading in here and there — attempts at taking compositional approaches for sound effects but using them as instruments. Sleepy Glade was another attempt, but it was more of a melodic song with an ambient intro.
The thing I love about writing music for my games is that the music rarely falls into pre-existing music genres unless I want it to. So each song comes from a genre dictated by the game and the digital space the song takes place in. Inevitably I draw in outside influences and trends, but a lot of it comes from the game. This is why games are an ever-expanding frontier for music, like innovative music for films or plays. Their music can be hard to meaningfully categorize into a recent trend, and instead they prefer to be analyzed. If my game is about the Taiwanese diaspora, then the music is meant to enhance that narrative story and atmosphere.
A lot of music I write gets lumped into ‘ambient’ or ‘soundtrack’, which may be useful from distinguishing it from rock or hip-hop, but ambient is not so much a genre as a descriptor of the lack of melody. Likewise, ‘soundtrack’ is also a barely descriptive term, merely indicating that the song is meant to be played alongside a particular set of moving (film, theatre) or traversable (games) images. Of course, then maybe all music should exist with labels corresponding to where we hear them! Instead of pop music, we’d have Whole Foods music, Panda Express music, Radio music, Concert music, Target music, or Fuddrucker’s Parking Lot music… perhaps just ‘retail’ or ‘commercial’ music, then.
Additional songs
It’s not a hard rule that each mission area in ETO had to have three songs. If you poke through the OST, you’ll see many of the areas have additional or extra songs. Riverton has a song for the indoors of a storeroom. There’s a song for a starfish in Restview. The final 3 areas have short ambient pieces before their ‘power plant’ song.
That is, even if you impose a structure on yourself, you can still embellish or ignore that structure when composing, if you think you can make an area’s music extra special.
Further Organizational Ideas
The Nature-Power Plant-Core Repair was the main way of figuring out the composition process. But I did use some ‘guiding leitmotifs’ to help determine melodies / melodic contour for the songs.
These were the “Earth, Air, and Sea” motifs. You can hear them here (they play in order.)
These aren’t driven home in recognizable ways like some other games, but were instead used as ways to help me think about how to compose melodies and melodic fragments. Sea is a mysterious sounding motif. Air is more of a trill, going back and forth between notes. Earth is a positive-sounding, yet short, melodic fragment. Here’s a probably incomplete list of where you can hear some of them.
“EARTH”
Restview: 0:15, Riverton Power Plant: 0:58 (liberally used), Earth Geome (opening), Karst Pass (1:05)
“AIR”
Dreamdram (0:12) , Dreamdram Plant , Magdal Woods (0:04 ), Magdal Plant (0:04), Air Geome (1:39)
Exception: Karst Pass (3:00 , though this technically is an “EARTH” area),
Exception: Great Cliff — motif is not present, but has delayed/blurry sounds mimicking Magdal and Dreamdram sounds.
“SEA”
On Organizational Forces
By the time we open a DAW to compose a piece, the possibility space for the song has been narrowed by organizational forces. I think this is good, though, and worth observing… seeing how certain ideas help us go from an infinity of music to an eventual .mp3 or .ogg file.
When the normal forces of a game are not enough, the idea of self-imposed organizational forces can help further narrow that possibility space until it feels easier to make a good piece of music. And at that point, a composer can either proceed by following what the forces point to, or by analyzing and possibly subverting them!
Alternatively, as a designer, you can think about these ideas and use them to better give audio direction to a composer!
I’ve used similar strategies for brainstorming for my other recent projects. For the 2016 Perfect OST, music to describe various spaces that were categorized into Physical, Digital, and Algorithm/System. Places like a dance club, or the desktop of a computer, or the notion of a “server room” containing the “cloud” or a sea of data.
For the soundtrack for my upcoming All Our Asias, I try to progress from more melancholic/darker/rougher ambiences, to brighter/hopeful melodies. This also accompanies the sorts of spaces and developments in the game’s story. By structuring the OST this way, it lets me more easily figure out how to start the sounds.
Do good things come from these constraints? I think so! My game Anodyne’s OST was helped by using mostly MegaDrive/Genesis/SNES soundfonts. Even earlier games of mine had PxTone to thank. Many memorable soundtracks rely on key sounds, instruments and synthesizers that give the games a characteristic sound, and it’s part of why there is a lot of nostalgia for OSTs like Chrono Trigger and Final Fantasy 6.
This is why I’m super into making my own patches for synths (I usually make 3–4+ for each song, sometimes re-using, etc.). My patches aren’t super complicated, but I think when you make a lot of patches you more easily can develop a style and signature sound for a game’s OST.
Without writing your own patches, naturally you tend to gravitate towards orchestral palettes such as strings, piano, woodwinds. And while these traditional instruments, under the hand of a skilled composer, can be used in an original fashion, it’s a lot easier to fall into existing musical tropes that result in a soundtrack that fails to help enhance the game.
Even though writing synth patches you may tend to fall into designing sounds that fit into the role of a mallet instrument, or a string or wind instrument, the timbre will be far different and more unique than if you multisample a violin or piano.
Further essays in this series, which may never come out given that I finished this essay in November 2017 but didn’t publish it till January (oops…), may include..
Specifics on the instrument patches for ETO and how those sounds progressed from Anodyne’s use of soundfonts
Analysizing the music theory structure of some songs, or breaking down some sound effects/synth patches.
Footnotes
[1] Carbon 2 is a Reaktor synth- I don’t know if you can get it anymore, or if it’s included with Reaktor 6 or whatever. I stopped using it years ago in favor of Ableton Live and Operator, but it’s worth checking out if you get a chance I guess!
[2] More specific breakdowns:
25% “Power Plant” Music (Melodic-ambient style tending towards the melodic, more upbeat.)
19% “Nature area” Music (Melodic-ambient tending towards ambient)
13% Power Plant Core music (Ambient)
13% Character Themes+ Melodic Cutscene Music (Mostly melodic)
13% Melodic city songs, or the Overworld map.
12% Ambient city songs/cutscene music.
5% Peritextual (Title, menu, credits)