What makes for a really good sound library, what would a game be like without audio and can a sound designer ever have enough source material? We had an interesting chat with industry veteran Ben Minto, Studio Audio Director, EA Dice.
So Ben, how did you first start out working in this industry?
After studying Applied Mathematics, and then a Masters in Computational Fluid Dynamics, at Leeds University in the UK, I followed my ‘sound’ hobby instead of getting a ‘real’ job and moved to London; first renovating and selling vintage analogue synths and modular systems, before moving on to spec’ing, installing and training users how to operate ProTools rigs. During that time, I got to try my hand/ear out at various freelance and side gigs, across the whole spectrum of the media industry – music recording, music production, post production, live sound etc. and was fortunate enough to get to know some of the games studios, and their audio teams, in and around London. After a year or so, one of the studios decided to offer me a job totally out of the blue, and at first I thought it was a joke! Thankfully they were persistent and came back 6 months later with a new (and higher!) offer to join them. I was the first ‘technical’ sound designer at the studio, coming entirely from a non-game, non-musical background, utilizing my formal education combined with my passion for ‘playing’ around with sounds. I had a lot to learn!
Through the years you’ve contributed to several major projects such as the Burnout, Battlefield and Battlefront series. What’s the role of a sound designer during a project?
The term sound designer is used as a bit of a catch all term in the games industry, and generally refers to someone working with the audio for a title, and more specifically to an individual working with the creation and implementation of sound effects (as opposed to music and VO for example). A good sound designer is a person who, as an individual or as part of a team, can consistently deliver a compelling and engaging experience which supports and enhances the title. There is a widely accepted subdivision of the role into technical sound designer and (creative) sound designer, although some people do both and almost every ‘sound designer’ does a bit of both in their role.
Broadly speaking a creative sound designer looks after one, or a few discrete areas, say for example vehicles, for a project, and can operate using a ‘cradle to grave’ model, i.e., they are responsible for every aspect of sound relating to those areas. So that can start with specifying the audio needs for the area, sourcing raw material (either by recording or obtaining new material), designing sounds, creating game assets, implementation, testing, mixing and iterating on these, for example. This can be seen as a horizontal path (intime), parallel to the game’s development and usually sees the sound designer supporting and being a member of a cross discipline team (art, gameplay, code etc.) responsible for that area.
A technical sound designer usually works more in a vertical/series approach across many areas and is more concerned with the initial setting up of the project and designing the audio subsystems, processes and conventions, at the start of the project/production and then optimizing and iterating on these designs once the game has been populated with content and assets, towards the end of production.
What would a game be like without audio? A glass of water without water?
It really depends on the game. Not all games need sound, but most will benefit from the right audio for sure. For a lot of the titles I work on, the primary purpose of sound is to encode information about what is happening around the player. Visually we are limited to what we can see on the screen in front of us, maybe limited to a viewable angle of 30 degrees either side of us. The rest of the sphere around us must be described via the power audio – ‘beyond the visually verifiable’. The more consistent and discrete that this description is, the more detail and degrees of discernible information we can encode into any given sound.
For example, we hear a gunshot. Into that one sound event we could encode real information such as distance, angle, height difference and orientation with respect to the listener, type of weapon (maybe not as detailed as a given model, but type ((eg. pistol) and calibre?)), the environment in which it was fired and the environment in which the listener is based, as well as ‘non-real’ information such as friendly or enemy, remaining bullets, threat level or narrative importance, for example.
Sound of course can support many other facets of a title, including narrative and immersion, triggering the desired emotional response or setting the mood, and aid with engagement and perception of a title’s overall quality.
Would you say that audio is used in a different way today in games than in let’s say the nineties?
If you were being bold (or brash!) you could say we have moved from representational figurative sound to ‘real’ sound; if you say we aren’t there yet, then we are definitely moving in the right ‘initially informed by physical reality’ direction. The underlying tech, processing power, memory etc. of the platforms we deliver on have all obviously increased over the years and with that the silicon limitations of the previous generations get eased or even nullified with the dawn of the next generation and associated technical advances.
Gunshots, in games, for example, forty years ago may have been represented by a symbolic enveloped burst of white noise, as synthesis was more readily available than PCM WAV playback, but then once we could play back samples we could easily trigger gunshot.wav every time a weapon was fired – we could play a scientifically correct ‘real’ sound, which whilst being correct isn’t always ‘good’ sounding.
Whereas today we go beyond this and have the luxury of being able to design multilayered systems that craft the sound design at runtime, from many individual components combined with runtime DSP and mixing. Instead of sculpting noise, or playing a fixed, yet frozen in space and time sample, we can use multi-perspective recordings of a firearm (as the starting source) and blend between all these “fixed point” solutions to interpolate and derive all the possible use cases in between, thereby better representing a “real” gunshot ingame and taking into account a multitude of game parameters to inform our sound design.
It’s worth noting that for any given recording, for a given channel (microphone, preamp, recorder etc), that is the ‘sound reality’ according to that specific setup. Not all realities are the same. So it’s not just about altering real world physical characteristics of the recording setup, i.e. distance, orientation, line of sight etc from the gunshot, but also the flavours of equipment used and technique (for example “never ever clip” or “driving the limiters hard into the red!”). Your choice as a sound designer is then which of these different realities to choose, blend and manipulate so as to define your own take and reality construct. Real ‘real’ is only just the beginning, authenticity is a great reference and starting point, science should almost always be taken into account, as well as audience expectations, but a literal simulation isn’t our end goal. We are still learning and there is still a way to go ….
The sound libraries sound designers work with today, how are they compared to the sound libraries ten or twenty years ago?
Everything changes! When I started out, libraries were still stored on physical media; CDs, DATs and even some ¼” tape libs. The best you could hope for was 48/16 stereo and usually only one singular (in time) recording of each sound – i.e. not the simultaneous multi mic/multi location setups we expect today. Finding sounds was a bit hit or miss. You had literally a telephone directory that listed all the sounds for a given library in “American” English (So you would have to look for “Trunk Car Close”, not “Boot Car Close” for example). Once you found a sound you liked, or thought you might like – based on the description and whatever previous knowledge you had about the library/recordist, you had to go find the CD, hoping it was there and in the right case, extract it and listen to see if it fit your needs (you could audition it of course by playing it as a normal CD beforehand!). It could easily take 30 mins to grab the 10 or so elements that you wanted to work with. And there weren’t that many libraries to start with; with many sounds being too familiar and overused!
Of course, today all that has improved and the stats are through the roof! 32 bit float, ultrasonics up to 384kHz, 100+ channel recording sessions etc. and the sheer quantity of libraries available is at least a couple of orders of magnitude greater than it was 20 years ago! Libraries are all of course available digitally from local servers or the cloud, everything is named, tagged and instantly searchable. Is there too much? Never! You can never have enough source material, but damn there is a lot of really good stuff out there these days!
This goes hand in hand with how “easy” it is to go out and record yourself these days. That ease in most part is facilitated by the lower cost of entry, usability, quality and portability with regards to equipment, but also the sharing of the audio community and availability of tutorials on the internet. That said nothing beats skill, the right gear, the right recording opportunities and having multiple recordists available to record the same thing from many perspectives (more ‘realities’ to work with later!). So whilst there is more available, there are definitely top-notch providers and recordists that we go to time and time again.
What makes for a good sound library?
First it’s about sound quality. I don’t necessarily mean the quality of the recording, as in bit depth or signal to noise ratio, but more along the lines of how does the sound match the auralisation I have in my head versus the description of the recording. P51 Extreme Fast By, Low, Very Aggressive. Snow Swirls and Fall on Pine, Delicate. Rockfall and Tumble, Large. When I read those sound descriptions, I have a feel for how the sounds should sound. I have an expectation. A good sound library matches or even exceeds those expectations. Oh and variations, if I need 10 variations, a good library has 20. I like options, I like to be able to pick and choose, and not have to make 10 variations from only 3 recordings. Sure it can be done with mix and matching, pulling from other source material, but we are talking ideal here!
More from Pole: