How The Outer Worlds Created Lively Environments Using Sound Particles

After following their passion for music since childhood, both Justin Bell and Dylan Hairston decided to combine that passion with their love for video games, which led to them working at Obsidian Entertainment. As Audio Director/Composer and Sound Designer respectively, and together with the rest of the audio team, they brought The Outer Worlds to life, a single-player sci-fi RPG about space exploration with lots of personality. Here's how that project was developed and how Sound Particles was an integral part of the process!

How THE OUTER WORLDS CREATED LIVELY ENVIRONMENTS USING SOUND PARTICLES

What do you do when going into a new project and how do you work together with the other teams?

JB: Initially it starts with the Game Design Documentation (GDD) and that involves a multidisciplinary team. Directors, leads and developers discussing what the goals of the project should be. Usually that starts off at a very high level where everything is a blue sky, meaning there’s no limits to our imagination, and we start just by putting that out there and get all the ideas out to the table. And then we start to refine things by figuring out what makes sense and what doesn’t, given what the vision of this game should be.

Consequently, as that starts to come together, it becomes part of our job on the audio side to see what those goals and aspirations are, and to see how audio can enhance that experience. This usually involves a lot of meetings and planning, where we’ll be talking to various stakeholders, understanding what the game goals are and collaborating with them to understand how audio can elevate it. And so it’s easiest when the process happens early on because nothing is committed yet and we’re still in the idea phase, so the team is still trying things out and prototyping, and when we have those discussions upfront then we can have more meaningful ways to integrate audio as a fundamental part of the experience and something that we rely on. For example, someone may come up with a quest where the audio is presenting as if the environment is very loud and characters can’t really hear one another, and so the writing will reflect that they can’t hear one another, and so one of them will say “I can’t hear you, hold on!” and then all the audio goes way, and so that’s some environmental storytelling where we’re using audio as a means to reflect the state of the world, and it may seem like a small thing but it is actually quite meaningful, as it makes the world feel like a believable, organic place where sound exists and is a factor on how these imaginary characters live their lives.

When do you establish the type of sounds and the direction that the overall themes of the sounds are going to be?

JB: Once there’s artwork of any kind where we are able to see what kind of mood is being established and we understand the story and the context, we can start to get a very good idea of how the environment is to be presented. Context is everything, right? So in the case of environment sounds you could have a beach, and if you went 100% realistic and you had the seagulls and the waves lapping on the coast, you might be missing some context. But maybe this is a beach that’s haunted or something, that’s kind of a cheesy example, but having that context helps to set the right tone; or maybe the point at which you arrive at the beach in the story of the game is a poignant moment, so do you really want to have intense waves and seagulls everywhere? Maybe not. You probably want it to be more understated and that will shape how you present. Dylan actually worked on the environment sounds for The Outer Worlds completely, he effectively made all those sounds and I’m sure he could speak too in his approach to that.

DH: The audio environment for Outer Worlds was pretty daunting because it’s such a big, really big game, with lots of different levels and different areas. So at the very beginning we’re looking at this 10 miles high task that just was so huge that we had to be smart about the way that we tackled it, it was sort of about figuring out how to get the biggest bang for our buck a lot of times. For instance, the intro of the game had an environment that clearly needed to be unique, it takes action in a place called Emerald Vale and it’s a pretty lush exterior area that has a lot of cool locations the player can go explore, and we knew right off the bat it needed to be something very distinct that the player could easily make a sonic association to, almost as if they could close their eyes and just listen and say “Yes, that is Emerald Vale for sure”.

On the other hand, for instance with interiors, sometimes we wanted a certain area to be sort of spooky, but then we wanted this other area to be more on the generic side, so in order to accomplish that we had to kitbash different types of sounds together. And so this was a general high level approach as to how we went about actually filling out all this massive game with environmental audio that was also fitting the context Justin was just talking about.

In The Outer Worlds what were the main goals you wanted to achieve and how did you tackle that?

JB: Obsidian makes role playing games and we want to provide a form of escapism by providing the player with an adventure that is highly immersive. And so it can be argued that the environment sounds are almost the single most important sounds in the entire game, in terms of creating a place that’s believable, and that’s why we invested so heavily in those sounds. Dylan spent months and months generating thousands and thousands of sounds and iterating on those sounds to really dial in the right balance to achieve that. These worlds are alien and so on the spectrum of highly stylized to highly realistic, so our goal was to find the right ratio between those two. And due to the visual style, if we skewed too far and too highly stylized it just wouldn’t work, because it would stand out, but if it’s too realistic there was the risk of being boring and predictable, so that was kind of a challenge throughout.

DH: I definitely agree that it’s a fine line to walk and sometimes I’d make something and it’s kind of “Well… that just sounds as if I walked outside, listening to my backyard and hearing some crickets and generic bird chirping”. And I would have that typically as a starting point: I would get these organic, real world animal sounds, but find cool ways to modify and manipulate those in order to make them sound alien. But the fact that it stems from something realistic allows players to have a sort of a conscious call back, “Right, that’s a bird!”. As if you know it’s a bird even though it doesn’t sound like a bird that actually exists, you still make the association because it has those qualities still inherent in the sound.

How did you implement Sound Particles in creating the sounds for Outer Worlds?

DH: We knew that we wanted to go very deep on environment sounds and make these rich soundscapes as immersive as possible. So after doing a lot of work on the ambience bits; the emitters and one shots; the reverbs and all this stuff, we realized that a lot of our town areas just felt really flat in terms of the ambience, as if it just wasn’t feeling right yet. And then we finally realized it needed walla, which is kind of the heartbeat of what, from an audio perspective, makes a town in a videogame feel like its populated with people.

And so one day I was at a game audio meetup in Orange County and we were talking to some folks at Blizzard on their cinematic audio team, and they were telling us about this program called Sound Particles and, you know, me and one of my co-workers, we were really interested in it, it sounded awesome and he was saying all these different applications you can use it for. So, we got back to work and we downloaded it, and then at work we have these things called R&D days where we basically get to take a day to pull up a program or an audio plugin or whatever, and basically take the whole day to learn it and so, right after we downloaded Sound Particles, I was absolutely sure how I was going to spend my next R&D day because this looked really cool, and so when the time came I spent the whole day learning about it. Then, about two months after that R&D day I had walla on my plate, and I knew Sound Particles was the perfect tool for this task because I already had some sort of the particle-like image in my head of what shape I wanted and all that.

On the other hand we were consulting our audio library for what kind of sounds we were going to use for this walla, and we realized there was no stock sound in our library that we felt like “Yes, that’s perfect! Let’s just drag and drop right into the game”; our library didn’t really have that. And so I started to employ Sound Particles. I already had a template project in Sound Particles that had a torus shape with basically just a bunch of particles that would move counterclockwise around the listener from pretty far away, so that it would give a nice stereo image, and then it was only a matter of loading up a bunch of individual files for each location.

For Edgewater, for example, I would load up merchant type walla: people in a church, people in a little village, and many other different types, and I would get a huge sampling of all those different types and then just let Sound Particles do its thing, where it would batch render a lot of different takes. Then out of those renders each one would have maybe a 10 to 15 seconds sort of like golden slice, because the renders would be almost like a football waveform, where we get really quiet, then it would be at an appropriate level and then it would temper back off, and we sort of sliced out that middle section from 5 or 6 different takes and made a loop out of it and occasionally I would add a bit of reverb if it needed to be spatialized in a different way, like it was in a colosseum or something. But as far as getting things to sound appropriately dense and layered properly and have the right stereo image and everything, Sound Particles just did it for me, and I just had to basically set things up at the beginning and once that initial set up was done, it was just as simple as hitting render and batch processing, which was great.

How did you move those assets to the game?

DH: Basically I would take my golden 10 seconds slice out of each of the renders and I’d bring them into Nuendo where I would make those into a loop, and then I would export that loop into Wwise, our middleware for Outer Worlds. And because in our game we really try to encourage player choice and let the player do what they want to do, that means the player can go into a town and eliminate all the other people in the town and we had to take that into account: if the player can do that, and if they do that, we had to anticipate what would be the repercussion for walla because everyone would be absent. And so, in these cases, we needed the walla to be gone, and it was kind of a trickier problem to solve then you would think, because the player can do so many things that it’s just so unpredictable. But the way that we at least attempted to solve this issue was to ask our programming team to basically give us a number, that was essentially the number of people that were NPCs within a certain radius of the player, and we had three types of walla that we would play depending on this number: walla-near, walla-medium and walla-far.

So, for instance, let’s assume that you walk into a bar and it’s a really small room (15 by 15 foot) with 4 other people. For that we would use the walla-near emitter to basically say “Okay, there are 4 people in this room”, so the engine would play the walla at maximum volume and also play it at maximum density, but then let’s say that 2 of those 4 people left, then the walla would start playing at -6 dB and half density. And we could also extrapolate that concept to the walla-far where if you’re outside in Edgewater and you could hear the whole town with 50 NPCs all around within this huge radius, and so we would play the walla at maximum volume and maximum density, and then if the player decided to eliminate half the town we would play the walla at -6 dB and at half the density. Essentially we tried to employ things like that in order for the walla to be reactive to the player’s choice and sort of be an accurate reflection of what the game world around them was.

How is your relationship with audio programmers and how do you usually work together?

JB: So, it really varies depending on the studio, different studios have different roles for their programming staff. In Obsidian, we generally hire generalist programmers and they’re kind of broken up into 3 main categories: we have a group of tools programmers, the ones who are pushing forward on the tool side, authoring content and creating those systems for the next group of people, which is our engine team who takes those tools and integrates those tools into our game engine; and then on the gameplay side there’s a group of programmers who take the engine technology and the tools we have and create specific features for the type of game that we’re building. This is the general development form to be at Obsidian where when we develop a feature, for example, swinging a sword, that is a cross disciplinary feature, it requires all of these different groups to come together to produce the end result: the animation of swinging a sword and the sounds that you hear; the artwork that you see and the character art; the artwork of the sword and how fast the sword is being swung and how much damage it applies and what kind of damage. So all those things are required in order to consider that sword to be complete.

Basically, for each one of those features that we develop for the team, we have a cross disciplinary group that comes together to realize that end result. For the gameplay programmer who is working on that feature, they are the ones who are adding in all the necessary end data for designers and artists to produce that end result, and that includes audio. So, we don’t really have a single audio programmer at Obsidian, in reality it’s more when audio work is necessary, it falls onto the group of programmers itself to create that tech.

Thank you for sharing your process with us, Justin and Dylan. It's impressive to hear about the behind-the-scenes of such a stunning and large game, and specially gratifying to hear how Sound Particles helped you in the process! We hope you keep creating amazing games and that you'll share them with all of us.

If you liked this interview, check out this interview we did with Paul Boechler, sound designer oF EA Games FIFA, or this interview with our very own founder, Nuno Fonseca!