Special Topics In Gameology is an in-depth look at a specific corner of the gaming world, in miniseries form. For this edition of the feature—Aural Histories—we’re interviewing game creators about the role of sound effects in video games. In this first installment, Anthony John Agnello talks to Limbo composer Martin Stig Andersen.
Video games are loud. It’s not just the classic digitized soundtracks, the thumping orchestral scores, or the reappropriated pop that drones over the action. Music is only one small organ in games’ vast body of sound. They are also littered with peculiar sound effects. The plop of Mega Man landing on the ground, the rustle when Nathan Drake tumbles behind cover in Uncharted; most things in games are noisy, but they’re all noisy for a reason. The Gameological Society discussed the process of making Limbo’s sound effects with Playdead’s Martin Stig Andersen, who explains how he determines which sounds are necessary in a fantasy world.
The Gameological Society: What is the role of sound in video games?
You can make sounds for all kinds of things, but in the end, what is most important is to communicate the game design.
Martin Stig Andersen: I guess the most important one is to communicate the game to the player. It’s often you can make sounds for all kinds of things in the game, but in the end, what is most important is to communicate the game design in some way. That’s what I tried to do in Limbo, to narrow the sounds down so they always have something to do with the gameplay. It informs the player and helps the player to solve the puzzles. As soon as you only play sounds that are important to the game, the player will actually start to listen to the sounds. Whereas if you have sounds for all kinds of arbitrary objects in the game, in the end, the player will just filter it out, because if there’s not enough relevant information in it, then it’s not important. I think that’s really one important thing.
Then of course, like in Limbo, creating the atmosphere is very important too. I guess it’s kind of a balance in a way. Definitely I would say the most important thing is to communicate the game and the gameplay.
Gameological: Something I loved playing Limbo is that one of the very first things you hear is the boy’s feet on the ground. I know that when you were making the sounds for the game you used recorded sounds off a lot of old analog recording equipment and then manipulated those recordings digitally. Walk me through how you created the sounds of the boy’s feet.
Andersen: I actually created the footstep sounds with the needle of a record player. There’s this small distorted sound when you move the needle on the record like [makes sound with voice *ksch ksch ksch ksch*] and also lift the pickup and let it fall down again, which gave a small pop. I also used a generic sound that’s playing. It’s playing all the time. No matter which material the boy is walking on, you will always hear this kind of noise component. So that’s one component.
The other component is more specific. That would be recordings of footsteps on grass, dirt, whatever. Then they would be mixed together. Then, initially, I was told to really economize the amount of source material, to keep it down to 30 or 50 megabytes. I ended up at 60, I think. So really from the beginning, I was trying to economize with the amount of materials.
Basically, I chopped everything up in grains. I had sounds separately for the heels and for the ball of the foot. The speed of the boy determines how fast the heel and toe sounds are played after each other. Faster will always be at the same time and when he walks, there will be more space between the heel and the ball or toe or whatever you call it.
Then, maybe as you noticed, the sound is quite loud, which is kind of a challenge. When you play games, you often go nuts listening to the foley sounds because it’s just sounds all the time, like a small machine. But at the same time if you just turn it down, it’s not like the boy gets softer. The environment just gets louder. We tried to establish the boy as being relatively loud in the environment in order to make the environment soft.
Gameological: These volume dynamics seem very important in Limbo. Like you said, it’s very different from other games like a Super Mario Bros. where the sounds are bombarding you all the time. What do you think the other differences are between designing a game like Limbo, which is far more subtle, and that kind of loud bombastic Super Mario Bros type?
I would never imagine a pure recording of a bird in the woods of Limbo. It would totally break the world for me.
Andersen: One of our original ideas was to create a game that didn’t sound like a video game. [Laughs] It sounds kind of like a stupid idea. Why do games sound like games? I guess it’s because of this generic setup. You attach sounds to different objects, and then they mix themselves in accordance to where they’re spaced in the geometry. You set up everything and when you play the game, the game mixes itself. I guess that’s kind of the traditional way of making games and why they end up sounding like games in a very objective way. You would never hear a film in the cinema with these kind of objective sounds. They always have more selected sounds. That was what we were aiming at. So it’s in accordance to different states, and it’s only a few selected sounds that you would actually hear. It’s more a subjective mix that you would often do in cinema. It’s not about what you would necessarily hear in real life. It’s more about pinning out a few objects that are important, and then you hear those sounds associated with those objects.
Gameological: You’ve said that your most hated sounds are things that sound fake. How do you make things sound real?
Andersen: Maybe it’s easier when you’re just avoiding things like voice. [Laughs] Limbo was special. You wouldn’t really like the boy with a voice. He has an anonymous face, and so we wanted an anonymous sound for him.
One thing that makes sounds in video games sound fake is the amount of repetition. In real life, you will never hear the same sound every time a foot hits the ground. It will never sound totally the same. As far as Limbo is concerned, it’s just cutting things into small grains, then sewing them together in the game, creating a kind of ambience. You would never hear a straight loop. It will always be cut into maybe 10 or 15 small separate grains and they would be stitched together randomly in a way and then shaped by different parameters and properties.
In that sense, Limbo was a bit easy—maybe easy isn’t the right word. The visuals were really trying to avoid things that were too clear or too easy to identify. You have these more anonymous or ambiguous graphics. This really allows you, the player, to interpret and project your own associations into the image. Like I said before, the face of the boy—he looks different depending on who’s watching him. It’s much more difficult to make a face in a game that’s totally detailed because it’s much more difficult to keep it alive. It’s easy to look fake. That’s probably the same with sound. You’re trying not to be too literal. It’s more about keeping things open to interpretation.
Like when you hear a bird in the game. I would never imagine a pure recording of a bird in the woods of Limbo. It would totally break the world for me. So those sounds are heavily distorted, and by doing that, you can hear less of the identity of the sound, so it sounds more generic and distorted somehow. This way, I can more easily repeat those sounds. It’s much harder to really recognize the origin because it’s distorted. As opposed to a high-fidelity sample of a bird—the third time you heard it, you would recognize all the tiny details in the sound, and then it would start to sound fake.
Gameological: Limbo still feels like a game with a story. As much as the boy is ambiguous, it still feels likes there is a point A and a point B. An arc. I think the sound is a big part of that. How do you go about creating the story using only sound?
Andersen: When I came into the project, it was quite late, so I had a general impression about the game in its entirety when I started working on it. In the opening of the game, you could believe it to be naturalistic. There’s no drone or melodic and musical things going on. But as you progress through the game, I introduce it really subtly along the way. Then I allow it to evolve into a more abstract, droney 8-bit soundscape in a way.
What that means for story is that even though there’s not really a story in the game, there was my own idea or interpretation of the game. There’s a small boy traveling through this violent world of Limbo. At some point in the game, this brutality has something beautiful to it. You have this kind of contrast, a kind of a schizophrenia. In the most terrible moments, I would add some kind of divine soundscape or something like that. My own interpretation of that was the boy being traumatized, becoming habituated to this violent situation. It even feels like he starts to like being in this kind of violent environment.
My favorite example is when he encounters the Gatling guns. If you added some traditional action music and atmosphere, it would just be too banal, too one-dimensional. I much more like the idea of adding these small, divine soundscapes. It creates something beautiful and also ironic in a way. I don’t know if you had the same sensation.
Gameological: I had very much the same sensation. The ending of the game, when you reverse polarity in the gravity section when the boy slams through an invisible plane of glass. It’s a terribly violent moment, but it is revelatory. It is that mix of almost humorous violence and revelation. That moment works. Were there any examples of things that you tried to implement in the game that didn’t work?
Andersen: The most challenging to add was the spider. I think I did three or four iterations on that. Usually, I like to start working on the most challenging things first because otherwise I get frustrated. I actually started out working with that before the entire sound of Limbo was defined. You work in different areas and as time goes by, you come up with a kind of general sound for it. I ended up having to record all the sounds myself, actually. That was a lot of work.
Gameological: What about those early iterations of the spider didn’t work?
Andersen: In the beginning, I didn’t want it to sound too natural. Also, it was a question about the scale. It was a big spider compared to the boy, but I didn’t want it to be too big. It was these kinds of questions. I started out working with a lot of debris sounds, sounds with stones. Then I would manipulate it on the tape recorder—pitch it up and down, and even warping the tapes. But it ended up sounding too abstract.
In the beginning, I didn’t want it to sound too real or too naturalistic, but at that time I hadn’t developed the idea of distorting all the sounds and making it sound bleak through old equipment. In the end it became more natural, but not natural in the true sense, more like an alternative naturalism.
I recorded all the sounds myself, like the spear, and the ground and leaves, and I would just distort them to make them fit into Limbo. It ended up that no matter what I did, when I distorted the sounds and I got the feel of the sound universe or whatever I was creating, everything would just kind of stick and just get sucked into this world of Limbo. It was a long process.