EA DICE – Ben Minto [Sound Director]

Despite having spoken to EA DICE’s Andreas Almström not too long ago, I found that there was more to learn about the inside of DICE’s audio team. So I reached out to Ben Minto, the company’s Sound Director to ask for an interview, and he obliged. You can read our chat below, where we talk about his time in the industry and his work on titles like Mirror’s Edge and Star Wars: Battlefront.

Hi Ben. Can you talk about your beginnings in audio and how you ended up at DICE?

Whilst working as a Pro Tools installation engineer in London, one of my clients and soon-to-be boss, Stephen Root, persuaded me to join him in working as a sound designer at Acclaim Entertainment’s studio in Croydon. Over time, Acclaim UK gradually shrunk in size until I was one of four people left working out of their London offices. Criterion Games was a developer who had previously worked with Acclaim on TrickStyle, and they wanted to improve the audio quality of their projects, so they offered to hire the remaining audio team from Acclaim as a ready-made solution – we all accepted! Whilst working at Criterion, I helped develop many games, including two of my favorite classics “Burnout 3” and “Black”, both of which were published by Electronic Arts. EA eventually went on to buy Criterion in 2004.

Having made racing games at Criterion for seven years, I eventually become fatigued of doing the same thing repeatedly. It was a perfect time for me to shift from car titles like “Burnout” to a first-person shooter like “Black”. Whilst we were working on the pre-production for “Black 2”, it was decided to pause the production in order see where the next generation of consoles would take us. It was suggested that I take a position at EA in the meantime, so I contacted some of the audio people I knew there, and Magnus Walterstad at DICE made me an offer, saying that he was working on a project called “Mirror’s Edge“, and that he needed someone to help him with guns and foley. I agreed, thinking I’d work here for about six months and then return to finish “Black 2”. Nine years later, I’m still here (laughs).

And you now work as the Sound Director at DICE, correct?

Yes, although each project can have a different Sound Director. For “Mirrors Edge” it was James Slavin. On the current “Battlefield 1“, it’s Ben Pajor. On “Star Wars Battlefront“, it’s me. In addition to my Director role, I also look after the entire audio department in general, along with the Audio Line Manager, Björn Hedberg.

I’ve heard that you received a degree in Computational Fluid Dynamics, which sounds advanced. Has this field of study been applicable to what you do today?

Very much so, since it has a lot to do with problem-solving and crafting elegant, yet optimal solutions. I’ve always enjoyed the certainty of mathematics, which contrasts to the open-ended nature of creative sound design. You know that 1 + 1 is always going to be 2, but you might have different ideas about how an exploding mutant chicken should sound.

What I like about my current role is that it offers me the chance to work in both the technical and creative worlds. We use a cradle-to-grave approach at DICE: if someone is working on vehicles, they’ll start by looking at the vehicle specifications and then come up with an overall sound design plan. They’ll go out to record source material, create the necessary sound assets and implement these assets within the engine via the editor. They also need to mix the vehicle components and balance the completed sound model against the already existing sounds in the game. So even in this simple example, there’s a lot of creative and technical aspects to address.

I’ve found that all sound designers at DICE are more technically proficient than the industry average. This is essential when using a cradle-to-grave approach, versus splitting the work up and giving each part to different specialist individuals. It also helps us keep a high level of consistency and quality across each work area.

What happens when there’s a conflict between you and a member of your team? How do you diffuse that?

Making mistakes is important, though it’s better to make them quickly and cheaply (laughs). Learning to diffuse and avoid conflict takes time. If I ask someone to design an ambience for the first time, I need to be patient enough to let them learn how to do it their way. I can offer mentoring and guidance, but I can’t enforce my own understanding and process on the. It’s easy to fall into the “old man” syndrome after you’ve been in the industry a while, where you resist changing your way of working. But in the game industry, we discard the current technology every five years or so as new platforms come out, so being open to new ways of thinking is key to remaining at the edge of game development. The processes we developed for the original Xbox are still slightly relevant, but aren’t as important as the innovations we need to develop for the Xbox One.

(Above: Ben Minto)

What are the challenges that DICE faces as a game developer when confronted with the specs of the new generation of consoles?

Recently, when I was reviewing the work of another sound designer, I pulled up some of their raw samples and noticed that they had left about half a second’s worth of silence at the end. That doesn’t matter too much these days, since that amount of sample data doesn’t take up much space, and our compression formats deal with silence very effectively, but old ways of thinking about RAM as a scarce resource drew my attention to it. It also reminded me that certain things don’t matter today as much as they used to; the sound for a game could utilize 500 MB of RAM on the Playstation 4, which is like loading up a whole album of lossless music, but as a combination of music, sound effects and voice-over! So something like RAM is a manageable resource these days, whereas CPU and optimization is still an area of challenge for us.

When you look at a screenshot of a Frostbite engine patch, you’ll see that part of our CPU cost is due to running multiple instances of connections, logic, samplers, DSP and mixing within each patch. However, our sound design demands can be roughly two to three times more complex than what those patches show, so we have be clever and get the desired results in other ways. It’s nice to build behaviorally complex abd reactive systems for our sounds, but if it’s going to take two months to implement, be expensive at run-time, and have many similar versions running at the same time within one level, then it’s safe to say it’s overkill. So, spending too much of your CPU budget on overly complex systems for every sound isn’t the way to go. Some sounds can be treated simply, whilst others will require an extra level of implementation to deliver the sound experience we want. In a few more generations though, I don’t think this will be as much of an issue anymore.

Today we also have many large-data challenges: if you’re making a car game with 1000 cars, how are you going to maintain the audio for that, in terms of consistency, quality and feel? If you make one car a day, your idea of what a good car sounds like will be totally different by the 1000th day than from when you started, so you’d have to re-tackle your initial cars to bring them up to the new level of quality. To handle this at DICE, we’ve implemented hierarchical systems based on parent and child patches. For example, in a first-person shooter game, we could start with a parent class called “Weapon”, and that might split into an “Automatic” and “Non-Automatic” class. “Automatic” might lead into an “SMG”, an “Assault Rifle” and so on. If at a later stage we want to add a “Silencer” or maybe a filter to all our weapons, we could add it at the parent stage, so it propagates down through the generations of patches. Previously, we’d have to add the functionality to each child patch or create a whole new series of unique patches, which is time consuming and error-prone. When designing new core systems, we now utilize the power of inherited configurations and shared content. As an example, whilst the initial “bang” of a pistol should be made from a unique recording, the reverb tail and reflections should be shared amongst a group of similar pistols. This reduces the initial amount of work needed to be done, whilst maintaining a consistent feel and sound to the patches within that family. When it’s time to add additional content via updates or DLC, you just need to add the new elements and logic for a specific item, since the default core elements already exist and can be reused.

How do you feel about the fact that the average consumer may not appreciate the effort dedicated to game audio? A nice-sounding fire from a game in 2008 would be hard for the average gamer to distinguish from a fire made in 2016, right? 

There are two things I’d address in that question. Your description of the “average consumer” may very well be correct, but we both know that you only have to read a few Reddit posts to see that certain people dig through a new game to find out what’s behind your creation, at which point they have a similar understanding to you about the games’ technical aspects; we’ve seen online conversations that go deeper than the ones we have amongst our own staff. Sometimes we hear sounds in the game that we don’t even know we made, since the audio experience is a complex interaction of random elements combined at run-time. So the scale of player awareness, from someone who plays with the sound muted, to someone who dissects the title over many hours of gameplay, is huge.

Secondly, I look at three different groups of listeners: the public, the press and our peers. Each groups is present in my mind as I work, and I know that each one will have a different perception of the game audio. Regarding your comment on a fire sound in 2008: that fire would have probably been mono, static and short-looping in nature; you’d have heard the same thing again and again. Had there been three fires near each other, there could have been phasing between the audio. Also, the burning of the wet wood would have sounded exactly the same as the dry wood. If the wind was blowing, the fire wouldn’t have reacted at all to that. Most people might not care about that sort of thing in terms of sound, but it’s bigger than just sound; you could encode information into the fire. For example, a change in the wind might be indicated by what was burning, how long it had been burning, or if the fire suddenly flared up in a certain way.

People appreciate the small details that sound designers put into their work. It’s possible that our staff could obsess over details, but that’s where my job of keeping everyone on course comes in.

In what way do you think the audio experience contributes to the commercial success of a game? I’ve seen cases where certain soundtracks receive praise from reviews, as happened for “Shovel Knight“, but I rarely read about people appreciating the foley or VO.

Sound is hard to gauge, because it rarely exists in isolation. Very few people close their eyes while playing to just listen to the audio. People today are a lot less concerned about breaking down the audio as much as enjoying the whole game experience. If the sound is crap, people will notice. However, if the sound is what it needs to be, people won’t comment on it, since it’s playing its role of supporting the larger game experience. I mean, the average consumer of movies has no clue that most of the audio in a film is replaced in post-production. Nature programs made me realize this when I was younger; I’d always wonder how the camera-man recorded the lion eating when he’s three miles away. In reality, the sound was created by a foley person afterwards.

What game titles do you personally think are exemplary for their audio and soundtracks?

Red Dead Redemption springs to mind. I put quite a few hours into it and thoroughly enjoyed the audio and soundtrack. Games like “Journey” and “Limbo” also have beautifully constructed soundscapes that matches and supports the gameplay and artwork. Other titles like “The Last Of Us”, “Bioshock” and “Uncharted” are well-regarded champions of game audio, and rightly so. It’s hard to mention a game that people don’t know about, since the best games tend to garner a lot of attention and win awards. So people do respond to good audio.

Personally, “Left For Dead” is one of my favorites because of the audio language it used. Anytime a tank would appear, the game used one of three ways to communicate that to you: a character would shout, “Oh my god, tank! “, or you’d hear the guttural bellow of the tank, or you’d get a music cue warning of the tank’s approach. So the game was chaotic with zombies and violence, but the cues that you needed to navigate that chaos were simple and didn’t feel forced. Even though other aspects of the audio weren’t the best, like the samples or the mix, I appreciated that the audio created a very readable language that gave the players the relevant information about what was happening around them so they could react and play the game better. I use it as a reference a lot for my work today – is the information that we are encoding into our sounds readable by our players?

Do you feel like certain genres of games tend to be more consistent in delivering compelling audio experiences?

Yes, sort of. Everyone likes a good story, and when the audio is supporting the narrative, the playing experience tends to be more compelling. For “Overwatch“, the developers have talk about how players can play the game just by listening to what’s happening around them, which is similar to my experience with “Left for Dead”. The audio is not only good, but also functional in that it contains decodable information that allows a player to tell what’s happening around them, thereby giving them the ability to react accordingly.

When you’re the boss of a team of sound designers for a major title like “Star Wars: Battlefront”, what kind of responsibility rests on your shoulders? 

Some development teams have a hierarchy where the Audio Director is the boss and has minions working under him. If that’s the setup a company wants, then I think the Audio Director’s job should very well be on the line if anything goes wrong. Thankfully, DICE isn’t like that. We’re a team that works together, and as much as I’m ready to take the responsibility for potential mistakes, I focus less on ordering people around and more on enabling them. Thankfully, we’ve never shipped a bad sounding game.

I’ve been doing this for 18+ years alongside many other talented people who are veterans in the industry, so we’ve been able to tackle most things quite well. Take “Star Wars: Battlefront” for instance: this was a brand-new IP for us at DICE, and one of the most beloved IPʼs in the world, to the extent that people can tell when individual recordings aren’t the real ones from the films. So, as the DICE audio team, we knew that we had to get it right. Despite the pressure we all felt at delivering a great-sounding game, I still felt confident about taking paternity leave four months before “Battlefront” shipped. That was only possible because the team could continue working despite my absence, since our setup and lack of hierarchy allows everything to keep running even when the “boss” isn’t around.

Do you think the learning curve for aspiring sound designers today is bigger than what you needed in 90s? Wouldn’t a new sound designer need to know more about computers and audio recording than in the past?

Well, some of the stuff I was doing fifteen years ago is irrelevant today, such as how to make looped material that obeyed proprietary compression formats – now you need to learn different and more relevant skills. To be honest, if you’re doing sound design today, then you are probably very tech-savvy already. Most game audio academic courses have a fair degree of technical studies in there, so it’s not unreasonable to expect applicants to know the basics. We’re not going to ask someone how well they know Frostbite. That wouldn’t make any sense, since it’s a proprietary system that we don’t share with anyone outside of EA. But it’s not always about hard skills. If someone comes for an interview at DICE and has a ton of knowledge, but is competing against another person with half the knowledge that comes across as a better team fit, I’d go with the latter person. You can teach people technical things as long as they’re open to learning, whereas fitting with an existing team is not something you can necessarily learn or even teach.

You’ve mentioned in past interviews that library sounds are part of your sound palette. But other sound folks stay away from library sounds, so as to not create a sound signature that’s generic. How do you avoid this if you’re open to using library stuff?

Well, just to clarify, it’s not like I’d take a general sound effects library, pull out sound effect #73 and stick it in a game; it’s not like that. I make use of samples and source material to create other sounds, whether it’s something from a library or one of my own recordings. It’s very rare that you find a sound that’s exactly right for the task at hand. You’re either going to have to add something to it or take something away. Most of the audio team here know which sounds are overused, and it’s important that you don’t use them in a way that triggers people’s memories of that sound from another game; that would break the experience.

I think the heaviest critique against sound libraries was around fifteen years ago when we only had around ten of them, like Hollywood Edge and Sound Ideas, whereas today we have more material to work with. So, it’s not as taboo as it used to be. I think today’s stigma is more connected to aspiring sound designers who show their demo reels with sound effects straight from libraries. You wouldn’t like to hear that when reviewing someone’s show-reel, as you get pulled out of the experience that you’re trying to review.

The traditional recording industry is still having the “analog” vs “digital” debate when it comes to gear, whilst sound designers don’t seem to have any aversion to going digital at all. Do you think an AAA game can be made using only analog gear?

Analog is great, but at some point, you might need to transfer a session to one of your co-workers. For example, someone might want to adjust the reverb track volume, which would normally be an audio bounce on a separate track. But with an analog plate reverb, you’d have print the effect into the original audio file, since you can’t ship the plate to your colleague. Also, if you have to process 10,000 lines of dialogue with reverb on them, you don’t want to do it in real-time with an analog unit. So analog gear is better suited for the source side of things, when you’re creating a sound. If I’m mass-processing my audio, I don’t want to have to run it to tape four times and then run it back into my workstation; I’d rather use a chain of plugins for that.

In your opinion, does the analog and digital world sound the same?

No, definitely not. But it all comes down to what works best for the task at hand. It’s like Max Lachmann said in his interview, do you add distortion at the recording stage or afterwards? If it sounds good, it is good. Analog gear is expensive, it’s hard to maintain, and you have to constantly run things back and forth. If you turn off the older gear, it won’t sound the same when you turn it on again, due to temperature changes and components wearing down. If you buy the exact same 70s analog unit as I have, the difference in the tolerances of the components means that yours could sound different than mine, even if they’re the same brand and model. Sometimes it’s more important to be consistent.

In a past interview, you said that you had your recorder with you on a trip to Dubai, where you’d record the sound of the wind blowing against grass and rocks. How did come to learn that field recordings are affected by geography in this way?

By listening and making mistakes. If you can make mistakes cheaply and quickly in your free time, then do it. If you want to make wind recordings, go and do that in your city. You might find that it’s hard to do because of the traffic and city noise, so perhaps you’ll aim for coming out in the morning, but then you’ll discover that the bird singing is a problem, and so you’ll have to be increasingly creative with your approach. Also, the wind makes no sound on its own, but rather when it interacts with other surfaces. So, if you know that, you’ll start looking for surfaces that interact with the wind, like grass for example. Then you might think, “My hearing is better than the signal to noise ratio on my handheld recorder, so it’s better to find a bigger patch of grass that blows more in the wind. And it’ll help the movement if I brush the grass with my hand as the wind blows.” This kind of thinking is how you learn your way forward.

So what’s next for you now that “Battlefront” is out and doing so well?

Mostly more Star Wars! I’m helping with the DLC for “Battlefront”, and have spent a few months on “Battlefield 1”, but now it’s onwards with the next “Star Wars” title.