The Sound of Anthem

[Music] hi there and welcome to the sound of anthem we're very excited to be here and have an opportunity to show you some of the things that we accomplished on this game today we're gonna walk you through some of the high-level concepts of the core audio experience but then also dive deep into some of the more interesting sound challenges we faced while working on this project we at Bioware take great pride in building unique engaging audio experiences that pride is further bolstered by seeing all of you here today with an interest in what we do our studio is full of passionate and dedicated developers trying not only to make the best experience for the players but for ourselves as well I mean why else would anyone move to the Canadian north where the deserts of Texas if not to work at Bioware with that let's start off by telling you a little bit about who we are hey everyone I'm Cody I'm an audio director here at Bioware I've been working in the industry for a little over six years now I got my start

doing cinematics for Dragon Age Inquisition after graduating from Vancouver film school for sound design anthem I had a large chunk a product ownership for the game including overseeing the gameplay audio crew and managing technical system design for audio as for me I've been working in the games industry since 2006 initially I studied out of the Art Institute of Vancouver and then like Cody and many of the others at Bioware I took the audio program at the Vancouver Film School more relevant to our talk event at Bioware for the last six years where among other things I'm the director of audio up next I want to introduce to you the core audio team this is our work family this is or at least was the anthem shipping team at Bioware what Cody and I will be representing today as a result of many months and years of hard work by these fine people and we wouldn't be and we couldn't be prouder to be doing so now I'm gonna walk you through some of the core tenets and understandings about the audio vision and how we set ourselves up to execute

execute upon those and there was a project that's been in the works for a number of years of Bioware starting small but even in its infancy it had been a noticeable departure from the titles and play styles that were typically known for where previously we went into fantasy of dragon age or sci-fi at Mass Effect anthem was more of a combination of those a science fantasy reality well you can see elements of our past franchises with some events focus on character development or exploration of exotic locales it was primarily fixated from day one to bringing super heroic rewarding and engaging gameplay complex and deadly enemy encounters all while playing together with your friends so with that the audio vision we set for ourselves followed suit our drive was to produce the heroic action and technology of the Marvel Universe set in the living living foreign hostile world of avatar in addition to that we focused heavily on immersion and believability while the

overarching idea of the game and direction of our sounds had them lean into the otherworldly and fantastical our systems within the engine focused on treating them as if they had originated within the world as we know it these were systems of dynamic reverbs to be the glue that pulled all the sounds together interior versus exterior mixing based on your position or our sources and defined areas early and late reflection calculations providing dynamic feedback from the area around you to inform slap off walls etc switching content tails on area type this is for weapons and explosions having unique assets whenever they originate for wherever they originated on the map roll offs based on the speed of sound calculations for accuracy in proper delays on big explosions and thunderclaps and then in relative orientation to focus the mix on what the player is looking at the result being an immersive but out-of-this-world audio experience next I'm gonna walk you through some of the tools that we

utilized to achieve these goals to begin our tool set for those of you not aware biyo are nearly EA wide we use frostbite as our development game engine created initially by Dyson Sweden for the battlefield games Bioware was a pioneer and bringing it bringing a tool that was built for first person shooters primarily and then habits sat and serve a wider variety of titles our first game using it was Dragon Age Inquisition we're both Cody and I started with the company it's a very powerful tool its main benefit being that it's set up in a way to allow teams to access all other portions of the game in one way or another so essentially whether whatever you see in a level or in a character blueprint or on an animation timeline you can access back to its route then pull events and parameters off to inform your sounds as you can imagine that can be very empowering here just a couple screenshots of three of our audio patches to show some of the levels complexity on that back end after we've received events first you can see it

takes a modular approach to building up patches using nodes of various functions moving from inputs on the left to sample playback and manipulation in the middle to outputs on the right you can see that there are vast possible possibilities and complexity within as well this last patch here is our master patch for our player weapons the primarily yellow stack in the middle representing many elements firing together to essentially build a sound at playback rather than just playing something preconceived very similar to stacking up audio assets in a dock I'm not going to go into much more detail about the toolset itself as there are more interesting things to see in here I just wanted to give you a brief peek behind the scenes to show that to show that many of our sounds do fall same approach source building so a major foundation and tool for our work is finding unique and interesting source materials over the project cycle or internal libraries grow exponentially as we put heavy importance on gathering content as well that will help us create

an experience unique to us with regards to recording here it is ahem here are some pictures of a handful of sessions that our team has done in the last couple years covering everything from dropping and smashing cars in a junkyard handling antique knobs and switches did redoing a didgeridoo taking part in cross team weapon recordings we did a big walla sessions lots of ambiences electro solution imagine magnetic captures of everything plunging yogurt blacksmith shops a bumblebee in a water bottle singing singing bowls and squeaky floors and on and on and on so in addition to that modular synthesis was a big contributor for this project as you can see our previous sound effects supervisor Mike Kent had caught the modular bug pretty bad over the over the years leading to ship this condition fortunately and unfortunately spread within our team seen some people building their own racks and others experimenting more with her multiple

hardware since that we passed around the studio but from this month this Rack Mike had developed and defined many of our sound palates and iconic earth-shattering content it inspired an often function as a communal collaboration spot within the halls of AI or Edmonton but always seemed to deliver something new and amazing here are a couple clips of the team catching some modular source and some of the in-game content where the results were worked into layers [Music] [Applause] [Music] [Applause] [Music] [Applause] [Music]

[Music] moving off of our tools were saying into some game content first we'll talk about music music is often seen as the most important element in a game or any media for that matter it provides the emotion the suspense the drive to an experience that in many cases would have been void of it it's an understatement that in any of our projects when the music is implemented it changes from essentially being a tech demo to a living breathing experience it always acts as a great unifier with the team at large something for everybody to get behind with the music for anthem we got a mix of otherworldly sounds driving action filled beats and a mysterious suspense all in the right amounts to get this something we do for all the games of Bioware's develop a series of pallets for our content we do this for defining various factions of enemies suit

variants in the case of anthem weapon types areas within the world or in this case the emotional pillars of music so working with our projects creative director and lead writers we build a compilation of descriptors for the overall core emotions of the game that we then work with the composer to bring across than the score as you can see the common theme is mentioned before and the audio vision are presented here creating a cohesive language for all to follow now let's talk about our composer for anthem we selected an amazing talent and Sarah Shatner before drew joining our team she had brought music to several excellent franchises and films from Assassin's Creed to Call of Duty she's being able to show great breadth in their content and the ability and that ability was needed with the direction we were going she truly is one of the best for creating a seamless combination of classic Orchestra and choir modular synthesis vocoder and vocalizations and then often using non-traditional source

in traditional ways with this she provided a musical landscape that was not only full of emotion but one that walked the line between a classic action game and this otherworldly sound so here's a little bit of that [Music] [Applause] [Music] [Applause]

[Music] [Applause] [Music] [Music] [Music] [Music] [Music]

at this point I'm going to turn it over to Kody for a bit he's gonna run you through some of the more in-depth implementations of the core gameplay experience with some examples of each should be a great showcase of the complexity that goes into some things you might not have really thought about as you blast around the map Kody cool thanks there so since the large exosuits referred to as javelins are one of the biggest and most unique components of anthems action gameplay we're gonna do a deeper dive and how we went about constructing the audio for them as eric mentioned one of the core tenets of the initial audio evasion was this idea of having everything feel grounded in reality within the game's environments this needed to extend to the suit designed to deliver physicality and a mechanical accuracy to the technology of the world this vision is most obviously

represented by the Ranger and the Colossus suits but we retain those ideas in every possible way for the more extravagant storm and interceptor suits as well let's take a quick look at the ideas that we use for each of the four suits as well as some short videos of the full suite of actions that each suit was capable of performing so starting off at the Ranger suit this was intended to be our most accessible javelin the it's it's a basic soldier class and from an audio point of view we wanted this class to be a simple expression of the game's audio vision creating an obvious physical correlation to all the sounds used on the soup it's what we described as medium tech so it's not to go too far into a high tech or a sci-fi vibe it's also kept us away from the low tech and janky feeling it's thrusters are explosive and punchy with high-frequency detail layers for the mission triggers its locomotion uses realistic servos and detailed mechanical movements to create a really clean but grounded feel so

here's the ranger so then from a production in a technology basis building the Rangers who gave us the template for all the suits that came after it and while there are minor deviations from the design for that template pretty much held held true throughout production next up at the Colossus suit this was in every sense of the word our our heavy class it's the tank the audio team affectionately referred to it as a garbage truck we wanted it to be the most aggressive and lowest tech it needed to sound dirty and

weighty but not messy if possible and we we pushed the furthest with distortion effects in the thruster sounds to bring out that realistic rocket launch and crackle its servos were really grinding and its mechanics we're we're thick and chunky so here's what the Colossus sounds like [Music] up next is the storm suit this was our spellcaster class and this really provided us with an opportunity to break away from the very grain that

grounded in realistic aesthetic oh the classes in the Ranger suit this suit wields the power of Ember to create magical spells and propel itself around the environment so for the audio we wanted to have a consistent feel to all the things the suit was capable of doing with that specific power so you'll hear a lot of granular and glassy liquidy elements of summer as well as some Rolling Thunder to give the ability to really grand explosive feel in the environment its locomotion contains more Airy servos and and then more resonant kind of plastic sounding mechanics and then of course we included additive layers for for the cape flaps for the storm because that's an iconic part of its part part of its look so here's what the storm sounds like

[Music] [Music] next up the Interceptor suit is our melee Engadget soup its gameplay speedy and agile and it was a really great opportunity for us to explore that hi-tech vibe the concept of high-tech agility played in the sound design really nicely with a lot of tight punchy thrusters more sci-fi laser elements another interesting part of the sound design where the servos one random idea that was thrown out was to reference Bruce Lee's iconic screams as a tonal basis for the movements and Riau the the principal sound designer for the suit actually did design a sound based on that idea so remember when you hear this

little squirrely synth sound layered into the movements it was actually inspired by this okay here's the Interceptor you so as I said when we built the Ranger I provided a good template for us to base the remaining suits on so let's go over

what the shared components of that template were and then we'll dive a bit deeper into some of the specifics of how we implemented this content so for the template we had basic locomotion so this is all footsteps and movement thrusters these were any actions that the suit uses it's jet boosters for so this include all jumping sprinting evading hovering and flying the hovering and flying had more specific logic for the looping components of those states to get them to react to the character's flight velocity pitch and lean and we had swimming which was essentially a simplified version of the flight loop system we had falling and landing we had sliding on steep surfaces we had melee both on the ground and from the air and then all the additional animation coverage for things like mantling reloads emotes arrivals victory poses and I'll expand on that system a little bit more shortly so first off would

locomotion like much of our content our locomotion was built in modular layers the modular design gives us a lot of flexibility to better react to the animations by using real-time game parameters to control the sounds playback the layer is built for the javelin locomotion were a base layer which is a simple initial transient layer then there's the mech layer which is the second transient and the mechanical detail then what we do is we use the character speed to scale a delay between the triggering of the base in the mech lair and then that kind of treats them like a heel and a toe respectively for a more pronounced difference between walking and running that sounds like this and

then we have the servo lair this is for synching with our movement and filling the gaps between the steps we adjust the pitch of the servo lair based on the character speed interestingly going against our initial instinct to pitch it up when the player ran or sprinted we found that pitching it down had the effect of making the suit sound like it was under greater strain to propel itself forward next we had the material layer this is just a generic material sound with different content based on the javelin size so for this we used the current ground material softness rating

and controlled which layer of these layers held the initial transient for the step so on hard surfaces the base layer would be the the base transient the footstep but on soft surface we switched and the material layer held the initial transient and this helped prevent the stacking of multiple transients which would have caused the sound to get muddied and then finally the wait layer was just a simple low-end bump to give a consistent weight to the suit our flight sounds for the games for the game were built into three layers so we had the body which is the main thruster sound typically an engine afterburner distortion with the wind to sell the movement through the open air and then a noise layer to represent the environment the player was flying through the body and the wind layer had to blended intensities that were constantly

crossfade in between each other based on physics parameters noise would transition between interior and exterior so controlling all three of these lost these layers is something called thruster load we use the pitch in the lean of the character as well as the the user's up-down stick deflection to generate a single value representing the thruster load this controlled the intensity of the sound as well as the intensity of the thruster VFX so flying straight down and gave you more wind and less thruster and flying straight up or banking brought more intensity to the Boosters so once we had the the base flight

sounds established we thought it would be interesting to try and get a little bit more reactivity from the physical environment built into the flying sound so for this we repurposed our raycast system and this is the the same system we used to simulate early reflections in the game doing these ray checks we would know when you were flying past transient objects as well as skimming a contiguous surface and we at one shot whooshes as well as loops for skimming on the left and the right as well as below the character this actually evolved into a really cool multi Department collaboration whereby the system that the audio team had built for flybys was able to be fully integrated with the VFX and actually design do things like the cooling of the gel on thrusters once given the water

one of the challenges we had with the idea design for the suits was getting them to feel really heavy and complex without going overboard in wearing too much audio onto them this issue was especially prevalent with lands we serve the sounds for the suits landing state of course and had even conditionally swapped them based on the fall velocity but we still weren't properly really capturing the full weight and energy of the the 1-ton neck suits falling through the air impacting the ground so to demonstrate this here's what they the land sound like just by themselves in game

so our first thought our solution was to add a system for playing audio on the falling state as it would lead into the landing state so at first we we just had a simple wind loop that played based on the fall velocity of the character and that was kind of interesting but we weren't really capturing the kind of dynamics that would make it feel Germanic so we need to be able to play one shots before the landing impact so we could get a lead in whoosh that would capture the energy better luckily animation had already built a way to generate the predicted time to land while falling through the air and then using that we could hand as I'm 1 shots of a fixed-length and trigger them at the exact right moment for this we had three intensities of landing possible depending on how far you'd fallen so I'm eared that with you

incoming one shots and then we had a hard Lance - it was a specific state where you go impact the ground and have a short recovery time to get up so I made that incoming Sun much longer to build up that energy we also had a second loop sweetener that was more chaotic and then with a consistent timing I then could apply that to all of the the suits and then it didn't take too long to get everything covered and and the results were pretty good [Applause]

so for nearly all the javelin movements we knew it'd be important to give ourselves the tools to hit all the animation content the team would be making without having to do hand edited and bespoke audio for each animation so this process is pretty much identical to what we've developed on previous projects for taking cloth and movement foley here we just needed to abstract the content to fit with the javelin sounds of servos and mech movements to do this we basically identify a short list event types so that we could design a bucket of javelins specific content that could be reused for all the animations the content for the event sounds like this

[Music] so as baseline this would give us decent coverage and then anything super specific we could we could layer on top these events then you that could be called on the timelines for the animations the event calls are all generic because this timeline can play on any javelin so the specific javelin that they were played on is passed to the sound patch on the back end to do the correct asset selection so here's an example of one emote tagged entirely with generic calls playing on four different javelin types [Music] all right so now moving on from javelins

the creatures of anthem were always conceptualized to be that thing that brings forth a sense of danger and discovery to the world in order to help us deliver on this vision we partnered with bloom audio to get a couple brand-new sound libraries made featuring some really incredibly unique creature source an animal source for our games in addition to all of our own recording that Eric mentioned collaborating with boom is a huge help and delivering a lot of these aggressive visceral and foreign sounding creatures their source sounds can be heard throughout the games creature sound design so here's some examples of some of the creatures in anthem [Applause] [Applause]

[Applause] interesting tidbit is that one of the source sounds for the arachnid Queen at the end there was actually a really squeaky dishwasher that was in our old building here in Edmonton so moving away from the wild beasts of anthem we wanted to also talk about how we approach building a creature language through enemy vocal sounds it's a really

difficult task that requires building a small collaborative group working across several departments including writing our InDesign additionally at Bioware our vo is handled largely by dedicated by our dedicated vo and performance department so it's really imperative that all these stakeholders work closely together in order to deliver the highest quality and results for the SCAR specifically the direction for the palette was that these are actually swarms of insect like creatures and they'd actually just began mimicking humans in their appearance in their technology and ultimately in their speech but it wouldn't be something that they actually were speaking any intelligible language so the challenge here was to create a characteristic for this faction that preserved the relatability and emotion of a language without containing any significantly recognizable speech our first step with this was to engage with our writing department and start brainstorming ideas one of our writers wrote up a script for the scar that she loosely based off some

Hungarian that she had learned when she was young so she wrote up this cue sheet and then we recorded a scratch session locally to test out how the language worked so here are some examples of raw audio from that scratch session for your dark eyes because Oh so we were able to take that and then start working through somebody a lot of our ideas for for processing for the scars here are those so as you can hear having the language scripted was only half the battle our challenge would be providing a solid and reproducible direction para how how we wanted actors to actually perform the lines the idea being that we needed to

be able to scale up to several different actors performing these scripts to get variations but that meant we really needed to put a lot of effort into capturing our gold standard example before we were able to trigger a pipeline to expand on the content for this we hired some outside help we have a very talented voice actor named Olaf Johnson and Olaf actually specializes in creature vocalizations so he took a shot at our script and this is what he came back with lucky for us that performance did a great job of capturing the aggression and the Menace we were needed from the pallet and then after doing some extensive work on the processing changed we finally landed on the vision of a swarm of insects speaking of broken non-language through a janky radio here's Olaf's process after being

performance after being processed so now of course we had this fantastic example content it was much easier to get other actors to perform that script provide consistent high-quality performances that then we could reliably batch process using our scar processing and then here so here are our final versions from other performances and that's all for me so I'm gonna throw it back over to Erik to talk about some of the

systems we use to build anthems dynamic world audio Thank You Cody so to wrap up our more detailed looks into the audio I wanted to quickly walk you through a couple of our processes for getting the world of anthem to stand up to a benchmark of realism and its function but fantastical in its presentation the approach to anthems world audio was essentially having elements having all elements work together to inform the whole over the previous two projects we started to build a core set of level audio tools that have enabled us to cover our vast worlds with great accuracy so we achieve we're achieving this by systems that deal with automatic mixing and switching of content interior versus exterior prioritization systems dynamic day/night cycles with advanced - determining density and frequency of wildlife to be heard throughout the days supportive weather sitting in top dynamic real world based weather systems and then a set of tools that lean into procedural placement and data collection

- that last point essentially we have a way to tag up all the objects in our worlds giving them unique identifiers to then feed back to a centralized collation point those tags are then weighed against each other prioritized and eventually serve as drivers for which quorum Ian says to play as placements for points or sounds or as locations for contextual audio content to feedback from so you can see in this little animation the player essentially acts as a polling point for the data around you at any given moment that data then drives the underlying systems to do the things mentioned just before you can imagine setting all this up as quite a task but the result was a more or less automatic coverage of our large regions this then allowed our teams to focus on custom sounds and situations to override that base layer so let's take a quick look at that in action just a little bit of context before we start here what you're going to see as we get close as the debug start appearing on the screen

these points that will initialize this blue then they'll turn green as they're active and transmitting back to the system the type and size of content that they represent so because in this example these are mainly foliage and it is daytime the core ambience will switch to our day jungle palette with the corresponding reverb area type settings as well as that day time point source content [Music]

next peripheral to that but still a core level system that we leaned on heavily is what we call big world sounds while we cover the base of the ambience with a quad representative error wins in some cases through that system we just spoke about we use the big world system to populate the world with all types of dynamic environmental content the system in his most simplistic form essentially picks sounds from a set of defined buckets and throws in randomly 360 into the world around the player nothing new there this being a definable distances and timings between is triggering these could be anything from the obvious like dissing construction where a creature calls and all the way to the more area defining a mood setting swells Moo setting ambience walls that we lean into pretty heavily essentially these are sounds of unseen source and it's more complex form we can then associate its placements to not just pick randomly but to look for the points that are demonstrated in the previous section this then allows us to do things like having birds only trigger from placements and tree canopies or random

column conversations playing off of actual npc groupings etc etc so the end you have a level of density and accuracy and the soundscape that is again more or less built procedurally as you move within it so let's take a little peek of what I'm talking about the first part of the video we'll be showing debug of queueing sounds looking to play back in the big world system at that given time and location what follows is a mix of level ambiences with one shot big world's triggering from across the game [Music] [Music] [Music]

the mix was a time where our adherence to organization and pre-planning really paid off this was compounded when put in combination when the games were procedurally driven mixing attributes the things like making sure our audio patches were as lean and approachable as possible so the team could focus on making great sounds versus struggling with implementation we had RMS driven e cues to give us control of different bus interactions on the fly setting up our sound content to be prioritized properly against the many thousand other sounds trying to find a spot having specific

mix States and dynamically driven mixers set for pulling out elements when we want to focus your ear HD our mixing based on loudness values we associate two given sounds or groups of sounds we centralized or reverb controls to have Universal uniformity across all levels and then having implementation asset prep and design standards set up for the team for content going into the game these and many other things allowed us to get a handle on what is a very busy mix environment additionally and lucky for us this past year in live service has given us the ability to continue to evolve our mix in ways we haven't been able to in past projects so let's take a little look at what I'm talking about with some action and some gameplay Oh shucks lucky I guess

[Music] [Music]

[Music] and that's it on behalf of both of us the core audio team back in Edmonton in Austin and all the others that you can see scrolling by that have helped us and our team deliver another great sounding experience at Bioware we thank you so much thanks for having us to here today