Animating Quill: Creating an Emotional Experience

all right looks like it's go time here you guys all ready all right so I just want to say thanks to GDC for inviting me out here it's real honor to be up here on stage in front of you all and I really hope you enjoy what you're about to see my name is Rick Leko animation director at the new VR game studio poly arc I've been animating in the gaming industry professionally now for about 18 years primarily focused on the gameplay animation roll you may recognize my work with quill here or from many other games series such as destiny halo condemn'd many many more I've had the sincere pleasure of being a part of today I'm here to talk about animating quell the protagonists of our first game Maus but for those who may not know yet who is quill and what is Maus your time has come at last

dear reader but this story is not yours alone no it is tied to another little falls before me [Music] unblocking my son where trees meet with the sky and in the stillness for when all the pies dancing and it's child [Music]

[Music] your bond with quill grows strong and the journey you take together could change the base of both our worlds shall we begin [Music] all right the world of moss it's it's vast and it includes many stories but we decided to start with quill and the adventure she takes with you the reader luckily people seem to like the idea of adventuring with her in fact I've never seen people have such a stark emotional

reaction to a character I've worked on before and and much of this reaction happened prior to the release of the game itself people often talk to her like she's listening they'll feel protective of her that they'll rely on her for help and want to help her in return people seem to feel a genuine emotional bond with this little virtual mouse but what is it about quill that creates this emotional response now this is the question driving the purpose of this talks over the next 45 minutes or so I'm going to try and answer this question from the animators perspective so online and and even here at GDC you're going to be exposed to an incredible amount of information about the latest tools and technologies but it's really quite difficult to find information and steps you can take to draw emotion from players this is inherently a subjective topic but I'm going to try my best to make it less subjective by describing the constraints we have faced and the steps we have taken my hopes are that by hearing our story you'll be able to take home some

ideas you can use for your own characters so to help frame our constraints I need to give you a little information about poly arc the studio and poly are consists of 15 full-time employees or a handful of industry vets from the triple-a space and we're taking a risk making our first game as an indie startup and we're learning a lot about what it takes to make it on our own we've learned just how expensive it is to run our own studio that people are the largest cost of business operation so it's essential to keep the team as slim as possible that getting the attention of gamers and creating an experience that stands out is super challenging and that despite our situation I'm also be judged in comparison to games staffed by much larger and more substantial budgets to help overcome these constraints we focused our animation efforts towards these three main principles you see here building a character from honest life experiences creating compelling

characters in VR and creating a realistic workflow that enables us to accomplish our goals now let's begin by illustrating why we've chosen to go the VR route if you look at the history of our industry the competitive playing field is often reset and when new technologies and markets become viable in the retail space and a good example of this is mobile gaming from 10 years ago now in these marketplaces the competitive field is much smaller with the expectations relatively undefined now we knew it would be incredibly difficult to go toe-to-toe with well-established and well-funded console and PC developers as a as an indie startup and also the technology is really interesting to us it's super exciting to be working in VR so because it's unexplored terrain there's so much undiscovered potential so we decided to take advantage of what we saw as the next industry reset but VR introduces some unique challenges we've never really encountered before everyone's first VR experience is usually fairly

shocking and it creates a sense of realism and a feeling of immersion yet many of the early VR character performance is found in most games were anything but immersive they often rely on the tried-and-true methods of the past where characters are simple avatars who respond to basic gameplay you know inputs and basic AI routines two years ago when we started our journey we made a demo from moss where quill is just that she was a simple gameplay avatar and you can see here some of the early footage we of the world of moss and quill herself and she doesn't feel alive yet if we would have stuck with this approach and simply polished what's here is she would have been maybe proficient but not memorable it's because she's merely a game play abstraction here now decades of gaming have trained us to accept certain abstract concepts we press buttons on an input device in a character on screen in front of us performs an action as a result this is

completely normal to us gamers now if you were to show somebody this from 300 years ago what a traditional game was they they wouldn't understand it the interface and the resulting actions they're really a learned experience and this learned acceptance of the abstract makes us more forgiving when we see our game characters merely react to controller inputs instead of acting like living breathing people and creatures we've gotten used to seeing things like popping between animation samples playing a limited repetitive set of animation clips or defying physics with reckless abandon and these are all standard issues we see everyday of gamers and we've grown numb to them but VR it relies on immersion to be most effective and characters who pop and play repetitive animations and defies physics even slightly it really removes you from the experience but more importantly characters who pander to gameplay systems acting as simple

avatars devoid of personality have a substantially negative impact on Vieira's immersion you're inside a world that feels real but you're joined by creatures who cater to you as the player it reminds us that we're just playing a game now I'm not saying player avatars no longer have a place in Maus you play the role as the reader who intentionally has no predefined personality so you could define it with your own the reader is our avatar character I'm referring to characters you interact with inside of the era such as quill here in the world of Maus you and quill exist simultaneously you as the character in the game can reach into the world and interact with it in natural ways while the movements stick and the face buttons of your controller they guide quill much like a standard third-person action-adventure game where X's jump squares attack both of them are dodged so to help will be more than just a standard third-person action avatar we need her to be able to interact with the player in more natural ways more in more meaningful ways you know when you're in

VR you're not just witnessing the characters adventure you're part of it so we can and should break the fourth wall in VR but how do we do that well the most obvious way to do this is to simply have Clos make eye contact with you looking people in the eye is a powerful way to connect and is no different here the frequency of this and the times when she chooses to do it will tell you what she's thinking now if you saw your friend from across the room and and wanted to get their attention what would you do you wouldn't look at him and press the X button it's just not how people work instead you you'd wave at them right this is one of the advantages of VR waving is natural and having a character that responds to what feels natural it really begins to make them feel alive so remember that fictional gamer I mentioned from 300 years ago they'd instantly get this look at that I

am so glad I'm not the one jumping on the lever good job coach do not jump so quit will tell you when she's had enough waving to ya by having her do this we're establishing personal boundaries like real friends have from the player's perspective she's not just a representation of a gameplay system she's asserting her own personality here now being able to pet quill and have her react to it is another way to make her more than just an avatar every time we had something like this into the gameplay space we helped tell her story outside of cinematics and the traditional narrative oh and she will get sick of being pet too much smack that hand away setting those personal boundaries are really important in a friendship right so a lot of people who played our e3 demo really wanted to

high-five quill after beating some enemies or solving a puzzle but they didn't really expect her to high-five them did they I mean that's that's not what a character does and to get in a game right they do now and if you let her wait too long she'll high-five herself you know she doesn't need you she's she's a confident insecure self-reliant Mouse who doesn't need your approval to party you missed out so many games often use a hint system to help players stuck on puzzles or lost in an environment and this is often done inside the game's UI but we couldn't do this in VR there's there's no UI we had to find a way to communicate the same information but in a way that fit the platform in the game's design yes you okay you get oh you're gonna go around that way pick him up

okay come back all right cool I think I know what you mean so by helping you solve puzzles the quill becomes your partner it implies that she's capable of complex thought not just sitting in an idle animation awaiting your next input this helps to build a bond with the player now speaking of the player having an elephant charge at you in VR is absolutely terrifying unless you're not looking at it in VR the players head is the camera so they may choose to look anywhere traditional games have been struggling with this as well how do you make sure the player sees what you want them to see in VR this becomes even more challenging since looking around in 3 6 G's part of the experience so forcing the player to look where you want them to may be that's really the wrong perspective to have of the problem for thousands of years people have been entertaining others who can choose to look away so instead of approaching this problem like we would a traditional game or even a film why not

think about this as a stage show in the image above events of interest will be happening on that stage while the rest of the environment is really just there to support it by thinking about our characters like their performers on a stage it provides us with a more relevant perspective but even with the most interesting things happening where we'd like you to focus we can't guarantee the player will choose to watch so what do we do well there's the obvious things like using sound cues lighting and motion to help draw the players attention we do certainly utilize these tricks and they do help some we can also choose to accept it and simply provide players with enough opportunities to understand the game without seeing everything that hopefully players might even want to go back and see what they missed but the most effective way we've discovered is to use interactivity as a form of camera control you're not just controlling quill you're partnering with her on her adventure which which means you're

inside the world with her there may be a door too large for a mouse to open but certainly not too big for a human by giving you a role in the world we know where you'll be likely-looking and we can target quills performances around these interactive moments we can even give her multiple performances based on her position relative to the interactive object since we know you'll see it so to sum up the first section VR certainly adds new opportunities for character exploration beyond traditional gaming due to its more natural forms of interaction and immersion we're really just scratching the surface but maybe maybe if we do this right we can make a name for ourselves but VR is not enough to make will stand out there have been tons of amazing memorable characters who define gaming so I've heard somewhere that Mario's more recognizable than Mickey Mouse these days which blows my mind the competition is absolutely intimidating

so how could we possibly make will stand out make make her a character that people will remember again I'll illustrate our constraints to frame how we've chosen to answer this question now among our staff the average number of years in the industry is 11 or so and most of us came from Bungie a studio with an intense laser focus on gameplay field so we as a team we've got a lot of experience making solid gameplay but now we're faced with a challenge of creating quill a partner character which you directly control who has the emotional depth of a friend and this is uncharted territory for us so we started by defining what we didn't want to do so to make will stand out we knew we wanted to avoid common gameplay tropes so what are common gameplay tropes even asked there are the things that almost all game characters do out of necessity or even tradition they're so well known that they transcend the character using it and become iconic in their own right examples such as the the ground pound

here how many games have a ground pound in it it's insane destiny has what three of them are right the fist clench I'm gonna get you evil guy you know like it's so common giving Aquila conic actions such as these would be the best way to remove her personality another example is the over tidal loop used to keep characters looking alive while they stand still certainly a necessity of the past but I don't think it believed it it applies anymore to modern game design and no cheese like we honestly it's such a trope like we really just wanted to focus on honest actions taken from real life not not cheese okay so if we can't rely on tropes what can we do well we can observe people we can observe the world around us we could be students of the human condition become amateur psychologists should be looping ask

ourselves who was quill at her core and how would she genuinely respond to her world in her context when confronted with a floating player head how would she react our goal wasn't just to write a narrative describing the events that took place on her adventure it was to create a world that you could inhabit and build a relationship with a mouse in now let me ask this question what's the job of an animator now unfortunately I've met too many non animators that would respond with while animators create movement but as I'm sure you all know the job of an animator is actually to bring a character to life to do this you need to do a hell of a lot more than make a character move it's an animators job to help create a history for their characters to give them a perspective to communicate complex emotion merely by seeing them think to further the narrative by showing how it affects your characters belief system in actions and who quill is emotionally at the start of mosses is quite different from who she becomes but

at the end during your adventure with her we hope you'll see her think we hope you'll feel her emotions and we hope you believe that she and the world of Maus have a history so how are we doing this by giving quill the ability to speak in sign language we're implying that at some point she learned this language this creates a history and a demonstration of her intelligence level but it's also very honest and based in reality sign language exists to help overcome a disability even though quill can hear just fine she's incapable of communicating verbally with you we've chosen to keep our use of signing fairly basic so not to alienate anyone who may not know it within the proper context even people who don't speak sign like myself will likely still understand the message she's trying to convey but honest performances they don't end with sign language in this video we expect most players to be excited to dig into quills world so quill invites you in

this is one of the 40 or so response animations quill used to express her opinions in react to you by doing this her actions helped define her and reinforced the feelings we want you the player to be experiencing quill was designed to amplify your experience she to be a mirror of your own emotional arc her response animations were created with that reflection in mind so there may be the occasional moment were you're not probably where you're pressing a direction on the analog stick and and she's just not reacting and honestly isn't this like game play no no 101 right like that's a bad thing right ensure it's it's often frowned upon by players and certainly cause our design team to puff up like puffer fish when I first suggested it but it's a calculated trade-off to provide quill with perceived autonomy to help minimize any potential frustration we limit when and where she'll chew we choose to allow

quill this autonomy occasionally will allow it after a player input such as climbing or ledge or pulling a lever actions that require a button press to activate so it doesn't feel too out of place otherwise we relegate it to non-combat situations while quill standing still already active players who who often keep quill on the move will may miss out on a lot of her personality moments so we're actually still trying to figure out the proper cadence for this so I love uncharted The Last Guardian Tomb Raider the Assassin's Creed games they in my opinion represent some of the finest examples of characters who transcend the common game avatar and become real people and creatures beyond cinematic performances much of this is due to how well the personality of these characters are represented during gameplay they don't simply perform gameplay actions they show emotions they think and they react to their world and much like those games we focused on quills

identity during gameplay by maintaining the uniqueness of her personality throughout her gameplay actions we're supporting all the work we've done with her autonomous reactions and personality moments our goal is to represent who she is at all times not just give her a bucket of animation but gameplay feel is paramount remember when I mentioned how pops repetition and wonky physics can ruin a VR character well by alleviating as many of those issues as we could and there's still a lot of work to do we help craft a more believable performance that we honestly hope feels good to play and quill also needs to feel fluid moving around her environment she should obey the laws of physics whenever possible she should be instantly responsive respecting player intent 100% of the time she should have a robust set of animations describing her actions and unique in varied ways and she should do all of this without betraying her character or giving in to common gameplay tropes and she is quite

the risk-taker Jesus thank god there's no fall damage right so just because we're a VR game it doesn't mean we get a pass when it comes to the feel of combat either combat needs to be instantly responsive matching player intent it needs proper interrupt frames in a good flow to the sword swings player will need to be able to evade with short of instability windows like in the Street Fighter game so how do we do this without making quill feel like a game play avatar well we animate her actions with honesty how would such a small youthful Mouse affect her world well she'd be doing it by giving a hundred and ten percent by wearing her emotional conviction on her sleeve like any teenagers she'd approach her world assuming she'll never die so she'll act out of reckless abandon she doesn't just swing her sword she swings her entire body and throws herself at the effort without a thought to her own safety we make sure she

carries her personality across every Shaanxi takes and of course we also have a traditional narrative to tell about quills adventure and due to our limited budget and team size we actually choose to forego the the traditional cinematic approach and instead tell our story inside of a book format now all of the honest performance decisions we make for quill in gameplay we echo on a handful of animated pages and just because quill isn't animating their quills animating on a page doesn't mean you can't still interact with her we've made it possible for quill to peer out of the book and track your location this interactivity helps to further the bond between quill and the player and we're just scratching the surface with this idea and we barely make use of this in mosses is actually the only time that we've got this working because it was actually a real pain in the butt to set up but we we honestly hope to take this prototype much further in future projects so acting autonomously creating honest

personality driven performances focusing on gameplay while keeping her actions unique we believe these are the important factors in defining quill as a character and giving her a fighting chance of standing out amongst an industry of titans but without enough quality animation content to support these performances it actually just all falls apart and this final section details our process for creating enough animation to fill the world of Maus at a quality bar essential to his success now a basic understanding of animation terminology is going to be assumed beyond this point so let me just frame this for you imagine you're leaving your your secure job as lead animator at Bungie see me join your friends at their startup studio making Maus as the only animator and Ritter you've got enough animation experience but you've never worked as a professional rigger before yet you must make the animation pipeline from scratch

your wife and three kids come first and you cannot sacrifice them to the game God so a crushing crunch schedule it can't happen and every dollar you put towards animation for Maus will shorten the time you have to make the game itself so you can't afford to hire or outsource Indy startup at all right at one last thing if the game fails because of animation you've let your friends lifelong dream die a horrible death no pressure right so what do you do well looking holistically at the animation pipelines of of my past what does the animation team spend most of their time doing well number one supporting a model at the character pipeline now I certainly need a character pipeline in place but if I'm not careful this can take a tremendous amount of time to create now how can I make a powerful and robust pipeline super cheap so number two creating a

system that makes iteration difficult or costly yeah making a game is is all about chasing that elusive elusive notion of fun what does fun mean so to do this there's always obviously going to be a tremendous amount of iteration needed and it's traditionally a huge time sink one of the biggest time sinks in game development and honestly who nails fluid gameplay on the first try even the tenth try like it's crazy it's so much work so I need to make sure we can iterate quickly so animation was really never a limiting factor for a designer engineering and the last thing biggest time sink for animators is you yep you quill animating to a high quality bar can be a huge time sink how can I streamline this process without sacrificing quality well now that we've defined the constraints and understand the importance and the needs of the project we can focus on coming up with creative solutions so I just need to

define one term before going into all this when I talk about the rig what I'm referring to is the animation control rig at Paul Eric we have the character setup which includes the mesh the skeleton the sockets the point weights due to the constraints of working in a game engine it's easiest if we don't alter any part of this character setup other than adding the occasional leaf node for attachments the bone hierarchy it will not change once the characters finalized now our animation control rigor rig for short are the handles we interact with when animating inside of Maya when our rig is applied it lives outside of the character setup and drives the bones using constraints the rig is not exported into unreal and operates inside of Maya exclusively this allows us to alter our rig however we choose without altering the character setup so the best way to solve a problem is to make sure you fully understand the question and by

not assuming that the optimal answers have already been discovered so I questioned the need from Maya files and storing rigs in reference scenes now we can save one source Maya file per character so we can preserve our materials or display layers and file settings but beyond that is a really a reason to save Maya animation files but what's the problem with saving my animation files well referencing prevents hierarchy changes to the rig which is a workflow limiter and if we don't reference and choose to save Maya files with local rigs inside them we'd need a batching process to distribute updates but more importantly we use unreal which utilizes the FBX format as its source when we create new animations or make changes to animations inside of unreal it outputs to an FBX format if we save Maya files and unreal saves FBX files

won't this desync our data so my animation files come with a cost but aren't my animation files really a standard workflow practice yeah but each suit is different and we need to forge your own path while being mindful of the reasons driving other Studios decisions so we decided to make my files optional for animation we can save them but we no longer need to save them any Maya file will be considered ancillary not Derik ival format unreal and all major animation packages they use on FBX and so will we future animators can work in whatever animation package they want such as max MotionBuilder Mya today I'll use FBX any updates that Unreal does to the FBX file is now accounted for nothing's ever at a date and we never need a batch but what about our animation rigs where do our rigs live

well we can store our rigs inside of a script we can start an animation by opening the source Maya file for the character we're working on imports any existing animation if needed from an FBX and apply the rig over the animation data the trick here is making sure the rig doesn't cause any degradation of animation when applying but the benefits of working this way are simplicity and flexibility if I make a change to the rig itself I simply update the script and the next time I apply the rig I'm applying the latest and I can alter the rig as I please while animating and if the rig breaks I simply bake it down I burned the rig with fire and I start over again but what about all the the data traditionally saved inside of Maya files like are we really okay with not archiving that data anymore well yeah like really like aren't those benefits the same constraints complicating iteration and requiring a huge suite of

tools to support like tools that that I don't have the time to create and we can still use animation layers and motion paths and the time editor and all these cool things that Maya has to offer and we could even save those Maya scenes if we choose we just need to be mindful that those scenes are temporary so as animators gain experience they also gain workflow knowledge yet in some circumstances you may find brilliant scripting experts making workflow decisions for the animators now this usually happens because the person writing the tool or rig can be seen as the decision maker but the problem here is the workflow experts aren't the ones designing the workflow in that situation and when there's a lack of collaboration between the two the tools can become inefficient or even a barrier to the creative process you may have written you have a well-written tool but it may not

enable the animator so knowledge based decision making is needed to empower the content creators now in our situation we have the opposite problem I know animation workflow well but I don't know Python and I need to make rigs at baked over bones without any loss or alteration of the motion now I'm sure I could learn Python but that would take time away from the more important task which is making the game and I could potentially outsource our rigs but if they break I'd be in the dark about fixing the problem myself and that cost money too I'm also too much of a control freak or any risk like that so I know the the output window in Maya will echo every action I perform as a Mel command and thanks to my experience as an animator I can do all the options to manually build a rig over existing animation without any data loss that's actually the easy part therefore I can use the output windows Mel to teach myself rigging by simply performing the steps and copying the

output into a script now could a real rigger do better abso-fucking-lutely but I'm all poly arc has and this is what I'm capable of doing within the tight schedule and on a tight budget this is actually how all character rigs for moss were made and for future projects I sincerely hope to collaborate with a brilliant rigger who respects knowledge based decision making because I'm fully aware that I should not be making rigid scripting decisions so how do we make the animation process faster and more iterative while keeping a focus on quality well we start by reducing the complexity commonly found in traditional animation rigs and focus more on flexibility also by finding ways to automate some of the more routine steps now quick caveat I'm not gonna have time to explain how each and every trick I'm going to show you here it is made that would be an entire talk into itself but I want to highlight the high level goals that make these workflows

possible so optimal rig settings for the many different situations an animator will encounter is nearly in possible for one rig to do well for example so optimal settings between posing and tweeting can often be quite different let's say you set up your character your rig with an FK spine because posing your character is actually quite easy to do that way but tweeting with an FK spine it requires a ton of counter animating to prevent the head from wobbling and you wouldn't animate a bouncing ball by putting it at the end of a bone chain rotating those bones to make the ball bounce would you even though I did because I'm in saying so why try and tween a spine this way it just doesn't make sense to me yet setting up on a rig for optimal tweeting it has a completely different set of trade-offs that would impact how easy it is to do posing or some other steps now I can go into details on all the different approaches to rigging and that

the trade-offs they create but but the problem I always seem to run into is the same I need to alter my rig in meaningful ways as I work to be most effective now this means being able to convert the animation data of any object into any space or relative to any other object it also means having the option to alter the hierarchy without any animation loss essentially opening the animation sandbox to as many options as we can allowing for creative problem solving as we animate so how do we do this well most studios create rigs featuring numerous attributes used to drive everything from fingers to foot roles the idea behind this approach is to empower the animator to give them a fine-tuned control over the animation process but what it actually does is limit the animators sandbox forcing a manual animation process it prevents the animator from switching the relative

space of the control without altering the motion itself now this can lead to counter animating and editing curves that may not directly relate to the motion itself I'm sure many of you dealt with like foot setups that have like 20 million attributes on the foot and you're doing the foot lean foot role and then the balance and the bend and this is where these curves doing I don't know anymore it can get a little confusing so just to be clear I'm not saying that custom attributes are a bad way to animate what I'm saying is that for me personally they slow me down so a poly Ark we're only using custom attributes for the facial rig controls only the ability to switch the space of any rig component is essential to our workflow keeping all animation data in the form of translation rotation and scale maintains a simple and easy to understand interface while also making an 'nearly unlimited sandbox and this huge sandbox allows us to create workflows which automate the animation

process in many situations rather than providing a ton of manual animation options because manual animation options they take time any scripts I make are optional modules that can vary between broad and specific workflow optimizations here I'm gonna give you a few examples so this is how I work with quills spine so it's the I've got a motion loaded on there and I've applied the rig and the rig when it first applies is in FK meaning I rotate you know lower in the spine and it translates upper you know rotate their heads translating I do have everything set to worldspace rotations right now so when I rotate lower it doesn't rotate the the children object I can switch between FK and I K I tend to like doing FK posing because when I make changes the spine it moves it and then I can switch the spine into a namespace setup where I'm just grabbing those controls and moving them around so I pose an FK and I tween in am space and the

animation data is unaltered when I switch between those spaces now some of you may be asking yourself well how are you baking spline I can to your character well I'm not I'm using a hierarchy of embedded aim constraints for the up Vector as a child of the existing rotation of the individual bones to invoke but to avoid any of the gimbal issues that you normally get with aim and constraints and when I translate the rig components inside the hierarchy it alters the rotation of the bones as an offset from what they were previously doing essentially acting as spline okay but it works with your easy bake oven so the spine isn't the only place where conversion makes animation much easier here's another example have you have you ever tried to animate a tail swishing around and interact with the ground or a wall it can really be time-consuming to do this well well here you can see that you know the tails just flowing around in FK and it's going through the wall and it's going through the floor and

here I'm proving that it's an FK apparently I didn't trust myself and then I'm converting it into the aim space that I use for the spine so now I can just grab those points and I can translate them and the cool thing here is I can take anything that's translating below the ground grab those translation curves and zero them out and now the tail slides along the ground and let's say that I don't want it going through the wall I find out the distance of the wall I make sure the those translations never go past that wall and now I've taken the FK rotations and turned it into interacting with the wall on the ground and I can even offset it in time and the cool thing about offsetting translations through space is when your offset rotation you can cause wobble at the end of your chain but when you're offsetting translations in space and it's updating your rotations for you you never get that wobble so manually animating your overlap in your detail it can be a lot of fun for an animator that's one of the animators favorite things to do but it

also takes time and depending on the skill of the animator it may or may not look consistent so leveraging physics to supplement this area is a huge time-saver and it allows for more realism if done correctly and maybe even more consistency so this is how I work with physics on quills spine so you can see here I started out with just a core posing of her waving and I'm just gonna send things to auto tangent real quick and then I'm just gonna clean up her up and down on her her cog just to give her some flow to it and that's all I did to get to this point and now I'm going to convert her spine into that aim space and then I'm just going to I'm gonna prove that it's a namespace and then I'm just gonna run physics on the different spine components and I can run different levels of physics and I can run them multiple times to give me the proper offsets that I want and now when I press play free overlap I don't have to do this work

anymore like it's done for me and I can even offset like the middle of the spine in time because I'm in the same space set up and by offsetting it in time I'm not screwing up any of the other bones or any of the other bits of animation I'm just creating even more overlap and it's because I'm in this type of a setup right here so finally I've got one more cool trick to show you guys this is actually how I animate quills ears so I start out with just basic core posing where the ears just hit a few poses and then I press the animate this [ __ ] button okay now they're animated and that's just off of a few basic poses and what her head was already doing so I could manually animate quills ears if I chose to and there are certain high action Clips that simply require an animators hand to accomplish like when she's rolling on the ground and stuff but the point here is it's an optional

tool not my replacement as an animator I'm able to make plug and play scripts for common animation needs myself and this flexibility comes from a simple and open sandbox set of tools when we started moss I was actually animating all of this stuff by hand and I was I was only able to produce maybe three or four seconds of quality animation in a day but after we after I wrote all these little optional tools and little workflow tricks I'm actually able to produce up to 15 seconds of quality footage a day for quill and that really made the difference for trying to get this game done on time so rethinking the animation pipeline and then need for Maya files preventing animation rigs from becoming overly complex and constraining allowing for space switching which in turn enables us to do outright automation of the animation process this is how I was able to accomplish the one-man team approach and how we were able to work

iteratively without animation be ever becoming a blocking issue for designer engineering and unlike characters from film TV and books quill isn't just a character you witness with those formats you follow a character on their adventure you're a passive observer quills not just a character you control in a game with a traditional game format gameplay mechanics and story events they take center stage as the character becomes your avatar into that world quill is a living breathing character who you'll interact with in meaningful ways while you guide her on your shared adventure and that is why this girl's crying so this session is one of two talks that Paul Eric's doing this year if you missed it yesterday

Brendan Walker did a great talk on a lot of the back end and how he hooked up a lot of the information for quill and it was a good companion talk to this and he also talks a lot about how to wipe tears from from headsets so questions [Applause] hello oh this is more of a design question but with this sort of strong relationship you've established with Quillin the players for future games or places where she might show up have you thought about persistence like if I play a game later with quill will she recognize me from our experiences in the first game that's a great idea like it's it's something that we've actually discussed in the past so that

the question was are we gonna maintain persistence with quill and future products and the answer essentially is yes and we're trying to figure out what that means at this point this is amazing I'm sure we'll all understand if the answer is no but are you making any of your rigging scripts available publicly I don't think you'd want them I'm such a poor script ER and rigor that you guys would laugh your butts off if you saw my script but I do know that some people on the internet have seen my Vimeo tutorials and have made marketable scripts off of those videos Yuri Yuri Munter II I don't know he's a Spanish dude I'm not saying his name properly I'm sure but you could look him up in the physics script for the ears and the body is available through him thank you yep Wow looks like we're wrapping it up

pretty quickly here all right well I guess we'll call it so if if you have any questions that you don't want to say out loud on the mic come see me and thanks for being here everybody it's been my pleasure to be here [Applause]